Keywords: Cardiovascular and respiratory signal processing - Blood pressure measurement, Vascular mechanics and hemodynamics - Pulse wave velocity, Cardiovascular, respiratory, and sleep devices - Monitors
Abstract: Two mini-symposia on the popular cuff-less blood pressure measurement field are proposed. This mini-symposium represents part one and focuses on PPG waveform analysis methods presented by three industrial and two academic speakers.

Keywords: Cardiovascular and respiratory signal processing - Blood pressure measurement, Vascular mechanics and hemodynamics - Pulse wave velocity, Cardiovascular, respiratory, and sleep devices - Monitors
Abstract: Two mini-symposia on the popular cuff-less blood pressure measurement field are proposed. This mini-symposium represents part two and covers innovative sensors and pulse transit time methods presented by two industrial and three academic speakers.

Keywords: BME and global health, Novel approaches to BME education, Biomedical engineering curricula
Abstract: In this special session we will present and discuss the non-technical part of different biomedical engineering related curricula. Non-technical, because the key 21st century skills will likely be in this area as essential differentiators and core elements for successful programs that are geared towards innovation generation. Novel online learning methodologies are needed that teach global perspectives and enable students to create future perspectives converging technology developments and their effect on business models with health stakeholder empathy. After presentations from four different institutions on three continents we will discuss a future curriculum by listing current issues, foreseeable needs, and skills that are relevant.

Keywords: Therapeutic robotics in rehabilitation
Abstract: In the past decade there have been several important advances in Robotics and an expansion of the Robotics -based rehabilitation. This symposium will focus on Innovative rehabilitation robotic technologies and its impact on recovery of upper and lower extremity function in people with neurological deficits (stroke, traumatic brain injury, and spinal cord injury). The symposium will present an overview of four areas of rehabilitation technology: Development of rehabilitative and assistive robots, clinical use of robots, functional and biomechanical changes due to robotic therapy, and the corresponding adaptive induced neuroplasticity. An expert will present an overview of each area with an emphasis on patient populations that are best served by each family of technologies, progress that has been made beyond the original research, and the current directions of research in each of these fields. . Investigators will examine immediate and long term training responses for biomechanical (joint kinematics, limits of stability, 3D loading forces, center of pressure, joint torques, and metabolic), electrophysiological (EEG, fNIRs, MRI) and peripheral neural (EMG) variables in order to understand the changes in peripheral, spinal, supraspinal/cortical control of gait. The overall focus of the symposium is to illustrate efficacy of using different powered robotics in upper and lower extremity rehabilitation for a cross section of different neurologically impaired populations.

Keywords: Brain-computer/machine interface, Neurorehabilitation, Brain functional imaging - Mapping
Abstract: Brain-computer interfaces are used as a treatment for the motor rehabilitation in stroke survivors. The motor impairment that affects walking is one of the most frequent limitations, which causes a high functional deficit in the patient's autonomy. The objective of this work is to analyze the effectiveness of the BCI technology combined with Functional Electrical Stimulation for the motor rehabilitation of the lower extremity.

Recent studies have demonstrated the possibility of decoding various speech representations such as phonemes, words, and phrases directly from intracranial recordings of brain activity. With the aim of progressing toward a transparent speech neuroprosthetic, the present work presents approaches for directly synthesizing speech using intracranial brain activity from speech-production areas.

Dementia, especially the age-related memory decline, is one of the most significant global challenges in the 21st century’s mental well-being and social welfare. The presented methodology showcases the possible social benefits of artificial intelligence (AI) applications for the elderly and establishes a step forward to develop AI techniques for applying simple wearable devices. We present a behavioral and brainwave (EEG or fNIRS) data collection concept for subsequent AI-based employment together with a range of machine learning encouraging results of early dementia onset elucidation.

Brain-machine interfaces are used as communication and control channels for people with neurological disorders such as amyotrophic lateral sclerosis (ALS). This work presents the design, implementation, and evaluation of a P300-based brain-machine interface (BMI) coupled with a robotic hand-orthosis aiming to assist ALS patients to freely open and close their fingers and hands.

Keywords: Sleep - Cardiovascular & Metabolic consequences of sleep disorders, Cardiovascular and respiratory system modeling - Sleep-cardiorespiratory Interactions, Cardiovascular and respiratory signal processing - Cardiovascular signal processing
Abstract: Purpose of the work is to identify the directional coupling between the structures of the brain and the autonomic control of the heart rate variability, to analyze the changes in these coupling in sleep and in wakefulness. Infra-slow oscillations of the electroencephalograms potential and low-frequency components (0.04-0.15 Hz) of the interbeat intervals signal where analyzed using a sensitive method for identifying the directional coupling. The technique, based on modeling the dynamics of instantaneous phases of oscillations, made it possible to reveal the presence and quantify the directional couplings between the structures of the brain and the autonomic control of the heart rate variability. It was shown that the coupling coefficients in the frequency band of 0.04-0.15 Hz (associated mainly with sympathetic control of blood circulation), on average, decrease with falling asleep. We have also shown the asymmetry of coupling. At the same time, stronger connections were revealed in the direction from the autonomic control of the heart rate variability to the brain structures than in the opposite direction. It has been shown that the strength of such couplings decreases with increasing of sleep depth.

Keywords: Cardiovascular and respiratory system modeling - Cardiovascular-Respiratory Interactions, Cardiovascular and respiratory system modeling - Cerebrovascular models, Cardiovascular regulation - Autonomic nervous system
Abstract: This study tested the hypothesis that respiration (RESP) is a confounder or suppressor of the closed loop relationship responsible for the cerebrovascular dynamical interactions as assessed from spontaneous variability of mean arterial pressure (MAP) and mean cerebral blood flow (MCBF). The evaluation was carried out in the information domain via transfer entropy (TE) estimated through a linear model-based approach comparing TE markers computed solely over MAP and MCBF series with TE indexes accounting for the eventual action of RESP over MAP and MCBF. We considered 11 patients (age: 76±5 yrs, 7 males) undergoing surgical aortic valve replacement (SAVR) at supine resting (REST) and during active standing (STAND) before and after SAVR surgery. The decrease of the predictive ability of MCBF to MAP when accounting for RESP compared to the one assessed when disregarding RESP suggested that RESP is a confounder of the link from MCBF to MAP along the Cushing reflex instead of being a suppressor. This result was more evident in POST when autonomic control was dramatically depressed and in an unchallenged condition such as REST. RESP did not affect significantly the link from MAP to MCBF along the pressure-to-flow relationship. Clarification of the type of RESP influence on the MAP-MCBF closed loop relationship could favor a deeper characterization of cerebrovascular interactions and the comprehension of cerebral autoregulation mechanisms.

Keywords: Cardiac mechanics, structure & function - Artificial heart and valves, Cardiac mechanics, structure & function - Ventricular assist devices, Cardiovascular, respiratory, and sleep devices - Implantables
Abstract: This paper describes the application of a proposed spiral coil to the transformer of a transcutaneous energy transfer system for a totally implantable artificial heart. To reduce the number of rectifier components in the power receiving circuit, the shape of the power receiving transformer was reviewed. The results indicated that the power transmission efficiency between the transformers was almost the same as that of the receiving transformer with the same shape. In addition, the calculations indicated that the power transmission efficiency including that of the power receiving circuit was increased, and the number of components in the power receiving circuit was reduced.

Keywords: Cardiovascular regulation - Heart rate variability, Cardiovascular regulation - Autonomic nervous system, Cardiovascular and respiratory signal processing - Cardiovascular signal processing
Abstract: Traditional frequency domain analysis of heart period (HP) variability allows the estimation of the parasympathetic modulation directed to the heart but the sympathetic one remains largely unknown. Recently, sympathetic and parasympathetic activity indexes (SAI and PAI) have been proposed to address this issue. SAI and PAI were derived from HP variability via the application of an orthonormal Laguerre expansion allowing the separation of HP variations driven by sympathetic and parasympathetic outflows. In this study, SAI and PAI were validated against tonic and variability measures of muscle sympathetic nerve activity (MSNA) and more traditional markers derived from HP variability. Indexes were calculated in 12 healthy subjects (9 females, age from 20 to 36 years, median 22.5 years) undergoing incremental head-up tilt. Results showed that traditional HP and MSNA variability markers as well as SAI and PAI were modified in proportion to the magnitude of the postural challenge. However, SAI was not correlated with any MSNA markers and PAI was not linked to respiratory sinus arrhythmia. SAI and PAI can capture modifications of cardiac control induced by the orthostatic challenge but they might be weak surrogates of vagal and sympathetic activities and/or modulations.

Clinical Relevance— SAI and PAI markers are useful to characterize cardiac control but poorly linked with autonomic nervous system state.

Keywords: Cardiovascular and respiratory signal processing - Heart Rate and Blood Pressure Variability, Cardiovascular and respiratory signal processing - Cardiovascular signal processing
Abstract: This study investigates the relationship between respiration and autonomic nervous system (ANS) activity and proposes a parallel detection method that can simultaneously extract the heart rate (HR) and respiration rate (RR) from different pulse waves measured using a novel biodegradable piezoelectric sensor. The synchronous changes in heart rate variability and respiration reveal the interaction between respiration and the cardiovascular system and their interconnection with ANS activity. Following this principle, respiration was extracted from the HR calculated beat-by-beat from pulse waves. Pulse waves were measured using multiple biodegradable piezoelectric sensors each attached to the human body surface. The Valsalva maneuver experiment was conducted on seven healthy young adults, and the extracted respiratory wave was compared with a reference respiratory wave measured simultaneously. The experimental results are consistent with the observations from reference waves, where R²=0.9506, p<0.001 for the extracted RR and the reference RR, thus demonstrating the detection capability under different respiratory statuses.

Keywords: Cardiac mechanics, structure & function - Ventricular assist devices, Cardiac mechanics, structure & function - Heart failure, Vascular mechanics and hemodynamics - Vascular Hemodynamics
Abstract: High and low shear regions in rotary blood pumps require sufficient washout flow to minimize blood residence time, thereby preventing hemolysis or regions of stasis that can lead to pump thrombosis. Spiral groove bearings (SGBs) both enhance pump washout and reduce erythrocyte exposure to high shear. Nar-row groove theory (NGT) has been used as an analytical tool to estimate the flow performance of a flat SGB during the design stage. However, NGT cannot accurately predict the performance of a conical SGB. In this study, we formulated an analytical model from the established NGT by adding an inertia correction term to incorporate variations in centrifugal force, which improved washout prediction in a conical SGB. The modified NGT model was then validated by comparison with experimental results. The results show that the modified NFT analytical model can reasonably predict washout rate when the spiral groove geometry favors creep flow conditions. The conical half angle of the SGB had the most significant impact on washout, with a decrease in half angle leading to large increases in washout flow. Small half angles also maintained viscous pumping at larger Reynolds numbers. In summary, the modified NGT can be a useful tool for designing conical SGBs for rotary blood pump washout within the creep flow regime.

Keywords: Cardiovascular and respiratory system modeling - Cardiac models, Cardiac electrophysiology - Patient-specific approaches to treatment of heat disease
Abstract: State-of-the-art solvers for in silico cardiac electrophysiology employ the Finite Element Method to solve complex anatomical models. While this is a robust and accurate technique, it requires a high-quality mesh to prevent its accuracy from being severely deteriorated. The generation of a good quality mesh for realistic anatomical models can be very time-consuming, making the translation to the clinics challenging, especially if we try to use patient-specific geometries.

Aiming to tackle this challenge, we propose an image-based model generation approach based on the meshfree Mixed Collocation Method. The flexibility provided by this method during model generation allows building meshfree models directly from the image data in an automatic procedure. Furthermore, this approach allows interpreting the simulation results directly in the voxel coordinates system of the image.

We simulate electrical propagation in a porcine biventricular model with the proposed method and we compare the results with those obtained using the Finite Element Method. We conclude that the proposed method can generate results that are in good agreement with the Finite Element Method solution, alleviating the requirement of a mesh and user-input during modeling with only minimum efficiency overhead.

Keywords: Cardiovascular and respiratory signal processing - Pulse transit time, Cardiovascular and respiratory signal processing - Blood pressure measurement, Cardiovascular and respiratory signal processing - Cardiovascular signal processing
Abstract: Blood Pressure (BP) is a critical biomarker for cardiorespiratory health. Conventional non-invasive BP measurement devices are mostly built on the principle of auscultation, oscillometry, or tonometry. The strong correlation between the Pulse Arrival Time (PAT) and BP has enabled unconstrained cuff-less BP monitoring. In this paper, we exploited that relationship for estimating Systolic Blood Pressure (SBP), Diastolic Blood Pressure (DBP), and Mean Arterial pressure (MAP) values. The proposed model involves extraction of PAT values by denoising the signals using advanced filtering techniques and finally employing machine learning algorithms to estimate cuff-less BP. The results are validated against Advancement of Medical Instrumentation (AAMI) standards and British Hypertension Society (BHS) protocols. Proposed method meets the AAMI standards in the context of estimating DBP and MAP values. The model’s accuracy achieved Grade A for both MAP and DBP values using the CatBoost algorithm, whereas it achieved grade A for MAP and Grade B for DBP using the XGBoost algorithm based on the BHS standards.

Keywords: Vascular mechanics and hemodynamics - Vascular Disease, Cardiovascular and respiratory system modeling - Cardiovascular Disease, Coronary artery disease
Abstract: Atherosclerosis is a chronic inflammatory disease associated with heart attack and stroke. It causes the growth of atherosclerotic plaques inside the arterial vessels, which in turn results to the reduction of the blood flow to the different organs. Drug-Eluting Stents (DES) are mesh-like wires, carrying pharmaceutical coating, designed to dilate and support the arterial vessel, restore blood flow and through the controlled local drug delivery inhibit neo-intimal thickening. In silico modeling is an efficient method of accurately predicting and assessing the performance of the stenting procedure. The present in silico study investigates the performance of two different stents (Bare Metal Stent, Drug-Eluting Stent) in a patient-specific coronary artery and assesses the effect of stent coating, considering that the same procedural approach is followed by the interventional cardiologist. The results demonstrate that even if small differences are obtained in the two models, the incorporation of the stent coatings (in DES) does not significantly affect the outcomes of the stent deployment, the stresses and strains in the scaffold and the arterial tissue. Nevertheless, it is suggested that regarding the DES expansion, higher pressure should be applied at the inner surface of the stent.

Keywords: Cardiovascular and respiratory signal processing - Blood pressure measurement, Cardiovascular and respiratory signal processing - Pulse transit time
Abstract: Objective: We investigate the effect of selective single parameter personalization on the performance of multi-parameter models for pulse arrival time (PAT) and pulse wave velocity (PWV) based blood pressure (BP) surrogates. Methods: Our data set stems from 15 surgery patients, and we selected from each patient 5 segments of 30 min length each. We evaluate the root mean squared BP tracking error of the two models with and without single parameter personalization. We further compare the BP tracking performance to a surrogate-free sample-and-hold approach, e.g., as afforded by conventional non-invasive blood pressure (NIBP) oscillometry. Results: Parameter personalization is key to realizing a tracking performance benefit of PAT-based or PWV-based BP surrogates. The highest tracking error reduction of about 3.7 mmHg with respect to a sample-and-hold approach was reached with a personalized PWV-to-BP model, which achieves an estimation error of 7.8 mmHg with respect to a continuously measured invasive reference.

Keywords: Cardiovascular, respiratory, and sleep devices - Sensors, Cardiovascular, respiratory, and sleep devices - Implantables, Cardiovascular, respiratory, and sleep devices - Diagnostics
Abstract: Central airway obstruction (CAO) is a respiratory disorder characterized by the blockage of the trachea and/or the main bronchi that can be life-threatening. Airway stenting is a palliative procedure for CAO commonly used given its efficacy. However, mucus impaction, secretion retention, and granulation tissue growth are known complications that can counteract the stent’s benefits. To prevent these situations, patients are routinely brought into the hospital to check stent patency, incurring a burden for the patient and the health care system, unnecessarily when no problems are found. In this paper, we introduce a capacitive sensor embedded in a stent that can detect solid and colloidal obstructions in the stent, as such obstructions alter the capacitor’s dielectric relative permittivity. In the case of colloidal obstructions (e.g., mucus), volumes as low as 0.1 ml can be detected. Given the small form factor of the sensor, it could be adapted to a variety of stent types without changing the standard bronchoscopy insertion method. The proposed system is a step forward in the development of smart airway stents that overcome the limitations of current stenting technology.

Keywords: Cardiovascular and respiratory signal processing - Heart Rate and Blood Pressure Variability, Cardiovascular regulation - Baroreflex, Cardiovascular regulation - Blood pressure variability
Abstract: Abstract— Idiopathic pulmonary fibrosis (IPF) is a progressive interstitial lung disease that can lead to chronic arterial hypoxemia, hypercapnia, and dyspnea. To improve clinical symptoms in IPF patients, supplemental oxygen (SupplO2) has been prescribed with the aim to maintain SpO2 level, and consequently to relieve dyspnea, increase physical activity and improve quality of life. In this study, we investigated the effect of disease and short-term SupplO2 on cardiovascular and respiratory autonomic regulation. Linear and nonlinear indices were extracted from the beat-to-beat variability of heart rate (HR), systolic (SYS) blood pressure and respiration (RESP) in IPF patients and healthy subjects spontaneously breathing ambient air (AA) and during SupplO2 at 3 L/min. It was found that the effects on autonomic nervous systems (ANS) regulation were better demonstrated by the Granger causality (GC) method. GC was significantly higher (p<0.01) in patients compared to controls for the interactions RESP→SYS and BBI→SYS.

Clinical Relevance—Short-term SupplO2 in IPF could adversely affect systolic blood pressure variability in particular. This study may help in the management of SupplO2 administration.

Keywords: Cardiovascular and respiratory system modeling - Cardiac models, Cardiac mechanics, structure & function - Cardiac structure from imaging, Vascular mechanics and hemodynamics - Vascular Hemodynamics
Abstract: In this paper, we present a cardiac computational framework aimed at simulating the effects of ischemia on cardiac potentials and hemodynamics. Proposed cardiac model uses an image based pipeline for modeling and analysis of the ischemic condition in-silico. We compute epicardial potential as well as body surface potential (BSP) for acute ischemic conditions based on data from animal model while varying both local coronary supply and global metabolic demand. Single lead ECG equivalent signal processed from computed BSP is used to drive a lumped hemodynamic model and derive left ventricular dynamics. Computational framework combining 3d structural information from image data and integrating electrophysiology and hemodynamics functionality is aimed to evaluate additional cardiac markers along with conventional electrical markers visible during acute ischemia and give a broader understanding of ischemic manifestation leading to pathophysiological changes. Simulation of epicardial to body surface potential followed by estimation of hemodynamic parameters like ejection fraction, contractility, blood pressure, etc, would help to infer subtle changes detectable beyond conventional ST segment changes.

Keywords: Cardiovascular and respiratory signal processing - Heart Rate and Blood Pressure Variability, Cardiovascular and respiratory signal processing - Blood pressure measurement, Cardiovascular, respiratory, and sleep devices - Monitors
Abstract: Blood pressure (BP) variability (BPV) is one of the important risk factors of cardiovascular (CV) disease. Particularly, nocturnal short-term BPV, characterized as acute transient BP elevation over several tens of seconds (BP surge), can trigger CV events. To accurately detect BP surge, it is necessary to monitor BP at each heartbeat. Although continuous BP monitors have been developed and validated, they are too large to measure beat-by-beat (BbB) BP at home. Therefore, we developed a small and lightweight BbB BP monitoring device (BbB device) based on tonometry. In this study, the BbB device was evaluated in terms of size, weight, and performance compared with a validated conventional continuous BP monitoring device based on tonometry (conventional device). The performance was evaluated using the correlation coefficient of pulse wave signals and the difference in BbB BP values between the two devices. Measurement data obtained from 30 subjects with a total of 81 sets, including short-term BPV by the Valsalva maneuver, was used for the evaluation of the performance. The results showed that the conventional device consists of two units (a control unit and a sensor unit), while the BbB device has integrated them into a single unit, with a weight of 150 g (approximately 1/45th of the conventional device). The BbB device was significantly smaller and more lightweight than the conventional device. The correlation coefficient of pulse wave signals between the two devices was 0.98 ± 0.02. The BbB systolic BP and diastolic BP differences were −0.3 ± 4.7 mmHg and 0.7 ± 3.4 mmHg, respectively. The developed BbB device was demonstrated to have an almost equivalent performance as the validated conventional device. In conclusion, we realized a small and lightweight continuous BP monitor that can evaluate the BP for each heartbeat using the BbB device without limitations regarding measurement location.

Keywords: Cardiovascular and respiratory system modeling - Cardiac models, Cardiovascular and respiratory system modeling - Gas exchange models, Cardiovascular and respiratory system modeling - Cardiovascular-Respiratory Interactions
Abstract: Abstract— In this study, we used a high-fidelity integrated computational model of the respiratory and cardiovascular systems to investigate cardiopulmonary resuscitation (CPR) after cardiac arrest in a virtual healthy subject. For the purpose of this work, a newly developed thoracic model has been integrated to the current model, to study the influence of external chest compressions upon the arrested circulation during CPR. We evaluated the chest compression (CC) parameters, namely, end compression force, compression rate, and duty cycle to optimize the coronary perfusion pressure and the systolic blood pressure, using a genetic algorithm. While the sternal displacement associated with the CC force agreed with the ERC guidelines, the CC rate and duty cycle were respectively higher and lower than the ones recommended by the ERC guidelines. The effect of these CC parameters on cardiac output (CO) were also assessed. The end compression force was the parameter with the largest impact on CO, while the compression rate and duty cycle scarcely influence it.

Relevance— Our results may aid in understanding the underlying pathophysiology of cardiac arrest and help guide research into the refinement of CPR strategies, without sacrificing animals or conducting clinical trials, which are difficult to undertake in crisis scenarios.

Keywords: Pulmonary and critical care – Multimodality monitoring in intensive care, Pulmonary and critical care - Bioengineering applications in Intensive care, Sleep - Periodic breathing & central apnea
Abstract: In neonatal intensive care units, respiratory traces of premature infants developing late onset sepsis (LOS) may also show episodes of apneas. However, since clinical patient monitors often underdetect apneas, clinical experts are required to investigate patients’ traces looking for these events. In this work we present a method to optimize an existing algorithm for central apnea (CA) detection and how we used it together with human annotations to investigate the occurrence of CAs preceding LOS. The algorithm was optimized by using a previously-annotated dataset consisting of 90 hours, extracted from 10 premature infants. This allowed to double precision (19.7% vs 9.3%, median values per patient) without affecting recall (90.5% vs 94.5%) compared to the original algorithm. This choice caused the missed identification of just 1 additional CA (4 vs 3) in the whole dataset. The optimized algorithm was then used to annotate a second dataset consisting of 480 hours, extracted from 10 premature infants diagnosed with LOS. Annotations were corrected by two clinical experts. A significantly higher number of CA annotations was found in the 6 hours prior to sepsis onset (p-value < 0.05). The use of the optimized algorithm followed by human annotations proved to be a suitable, time-efficient method to annotate CAs before sepsis in premature infants, enabling future use in large datasets.

Keywords: Cardiovascular and respiratory signal processing - Complexity in cardiovascular or respiratory signals, Cardiovascular regulation - Heart rate variability, Cardiovascular and respiratory signal processing - Cardiovascular signal processing
Abstract: In the last decades, a considerable effort has been devoted to quantify complexity in physiological time series, with a particular focus on heart rate variability (HRV). To this end, exemplary quantifiers including Approximate Entropy and Sample Entropy have successfully been applied by leveraging on statistical approximation and further parametrization through the definition of tolerance and embedding dimension, among others. In this study, we investigate the use of the Algorithmic Information Content, which is estimated through an effective compression algorithm, to quantify partition-based Kolmogorov-Sinai (K-S) entropy on HRV series. We test such a K-S estimate on real data gathered from the Fantasia database, aiming to discern young vs. elderly complex dynamics. Experimental results show that elderly people are associated with a lower HRV complexity and a more predictable behavior, with significantly lower partition-based K-S entropy than the young adults. We conclude that partition-based K-S entropy may effectively be used to investigate pathological conditions in the cardiovascular system, complementing state-of-the-art methods for complexity assessment.

Keywords: Cardiovascular and respiratory signal processing - Non-linear cardiovascular or cardiorespiratory relations, Cardiovascular and respiratory signal processing - Heart Rate and Blood Pressure Variability
Abstract: The present study investigates the differences in autonomic nervous system (ANS) function and stress response between patients with major depressive disorder (MDD) and healthy subjects by measuring changes in ANS biomarkers. ANS-related parameters are derived from various biosignals during a mental stress protocol consisting of a basal, stress, and recovery phase. The feature set consists of ANS biomarkers such as the heart rate (HR) derived from the electrocardiogram, the respiratory rate derived from the respiration signal, vascular parameters obtained from a model-based photoplethysmographic pulse waveform analysis, and cardiorespiratory coupling indices derived from the joint analysis of the heart rate variability (HRV) and respiratory signals. In particular, linear cardiorespiratory interactions are quantified by means of time-frequency coherence, while interactions of quadratic nonlinear nature between HRV and respiration are quantified by means of real wavelet biphase. The intra-subject difference of a feature value between two phases of the protocol, the so-called autonomic reactivity, is considered as a ANS biomarker as well. The performance of ANS biomarkers on discriminating MDD patients is evaluated using a classification pipeline. The results show that the most discriminative ANS biomarkers are related with differences in HR and autonomic reactivity of both vascular and nonlinear cardiorespiratory coupling indices. Differences in autonomic reactivity imply that MDD and healthy subjects differ in their ability to cope with stress. Considering only HR and vascular characteristics a linear support-vector machine classifier yields to accuracy 72.5% and F1-score 73.2%. However, taking into account the nonlinear cardiorespiratory coupling indices, the classification performance improves, yielding to accuracy 77.5% and F1-score 78.0%. Thus, changes in the nonlinear properties of the cardiorespiratory system during stress may yield additional information.

Keywords: Cardiovascular and respiratory system modeling - Cardiac models, Cardiovascular regulation - Heart rate variability, Cardiovascular regulation - Blood pressure variability
Abstract: Vagus nerve stimulation (VNS) is an emerging ther- apeutic strategy for pathological conditions in a variety of diseases; however, several challenges arise for applying this stimulation paradigm in automated closed-loop control. In this work, we propose a data driven approach for predicting the impact of VNS on physiological variables. We apply this approach on a synthetic dataset created with a physiological model of a rat heart. Through training several neural network models, we found that a long short term memory (LSTM) architecture gave the best performance on a test set. Further, we found the neural network model was capable of mapping a set of VNS parameters to the correct response in the heart rate and the mean arterial blood pressure. In closed-loop control of biological systems, a model of the physiological system is often required and we demonstrate using a data driven approach to meet this requirement in the cardiac system.

Keywords: Cardiovascular and respiratory signal processing - Cardiovascular signal processing
Abstract: This paper proposes a new generative probabilistic model for phonocardiograms (PCGs) that can simultaneously capture oscillatory factors and state transitions in cardiac cycles. Conventionally, PCGs have been modeled in two main aspects. One is a state space model that represents recurrent and frequently appearing state transitions. Another is a factor model that expresses the PCG as a non-stationary signal consisting of multiple oscillations. To model these perspectives in a unified framework, we combine an oscillation decomposition with a state space model. The proposed model can decompose the PCG into cardiac state dependent oscillations by reflecting the mechanism of cardiac sounds generation in an unsupervised manner. In the experiments, our model achieved better accuracy in the state estimation task compared to the empirical mode decomposition method. In addition, our model detected S2 onsets more accurately than the supervised segmentation method when distributions among PCG signals were different.

Keywords: Cardiac electrophysiology - Defibrillation, ablation, and cardioversion, Cardiac electrophysiology - Simulation for cardiac arrhythmia
Abstract: The main treatment option for Ventricular Fibrillation (VF), especially in out-of-hospital cardiac arrests (OHCA) is defibrillation. Typically, the survival-to-discharge rates are very poor for OHCA. Existing studies have shown that rotors may be the sources of arrhythmia and ablating them could modulate or terminate VF. However, tracking rotors and ablating them is not a feasible solution in a OHCA scenario. Hence, if the sources (or rotors) can be regionally localized non-invasively and this information can be used to direct the orientation of the shock vectors, it may aid the termination of rotors and defibrillation success. In this work, using computational modeling, we present our initial results on testing the effect of shock vector orientation on modulating (or) terminating rotors. A combination of Sovlij’s and Aliev Panfilov’s monodomain cardiac models were used in inducing rotors and testing the effect of shock vector magnitude and direction. Based on our simulation results on an average with four experimental trials, a shock vector directed in the perpendicular direction along the axis of the rotor terminated the rotor with 16% lesser magnitude than parallel direction and 38% lesser magnitude than in oblique direction.

Keywords: Vascular mechanics and hemodynamics - Vascular mechanics, Vascular mechanics and hemodynamics - Vascular Hemodynamics, Cardiovascular and respiratory system modeling - Vascular mechanics and hemodynamics
Abstract: INTRODUCTION: Allometry describes the disproportionate changes in shape, size or function that are observed when comparing separate isolated features in animals spanning a range of body sizes. Scaling of the energy dissipation has been also observed in warm blooded animals, essentially varying as mammal’s body mass (BM). Part of the energy stored in the arterial wall during elastic distension corresponding to the viscous deformation is dissipated within the arterial wall. OBJECTIVE: To elucidate the allometric existing relationship between BM and arterial wall viscosity, as a measure of energy dissipation. MATERIAL AND METHODS: Arterial viscous dissipation (WVD) was assessed in dogs, sheep, and humans in terms of BM and heart rate (HR) variations. RESULTS: An allometric law was found between WVD and BM, jointly with the assessment of WVD in terms of HR. CONCLUSION. The existence of a power-law link for viscous dissipation and BM that involve different mammals was demonstrated.

Keywords: Cardiovascular, respiratory, and sleep devices - Diagnostics, Sleep - Obstructive sleep apnea, Sleep - Cardiovascular & Metabolic consequences of sleep disorders
Abstract: A new method for calculation of an overnight oximetry signal metric which is predictive of cardiovascular disease (CVD) outcomes in individuals undergoing an overnight sleep test is presented. The metric – the respiratory event desaturation transient area (REDTA) - quantifies the desaturation associated with respiratory events. Data from the Sleep Heart Health Study, which includes overnight oximetry signals and long-term CVD outcomes, was used to develop and test the parameter. Performance of the REDTA parameter was assessed using Cox proportional hazard ratios and compared to established metrics of hypoxia. Results show that hazard ratios in adjusted Cox analysis for predicting cardiovascular death using REDTA are up to 1.90 (95%CI: 1.22-2.96) which compares with the best of the established metrics. A big advantage of our metric compared to other high performing metrics is its ease of computation.

Keywords: Cardiovascular and respiratory signal processing - Cardiovascular signal processing, Cardiovascular regulation - Heart rate variability, Cardiovascular and respiratory signal processing - Heart Rate and Blood Pressure Variability
Abstract: Due to the widespread use and simplicity of photoplethysmography (PPG) signals, and because this signal contains information related to pulse rate, several studies have started to propose the use of Pulse Rate Variability (PRV) for the assessment of cardiovascular autonomic nervous activity, instead of using Heart Rate Variability (HRV) obtained with the electrocardiogram (ECG). However, there is a lack of standardisation and guidelines for the measurement of PRV from PPG signals, which might hinder comparability among studies and validation of results. The aim of this study was to evaluate different digital filters on PPG signals and their effects on PRV information, compared to HRV obtained from ECG. PPG and ECG signals obtained from healthy volunteers were used to measure HRV and PRV. PPG signals were filtered using different FIR and IIR digital filters, with several cut-off frequencies. The results indicate that filtering PPG signals using IIR filters and lower low-cut-off frequencies allow for the acquisition of more reliable PRV information, with lower Bland-Altman ratios and higher cross-correlations when compared to HRV. This is a first step in establishing guidelines and standards for the analysis of PRV information using PPG signals.

Keywords: Cardiovascular regulation - Heart rate variability, Cardiovascular regulation - Autonomic nervous system
Abstract: Prenatal uptake of valproic acid (VPA) was associated with increased risk of fetal cardiac anomalies and autism spectrum disorder (ASD), but uptake of VPA is considered the only effective treatment for epilepsy and other neurological disorders. Up until now, little is known about the effect of VPA on maternal – fetal heart rate (HR) coupling patterns; therefore, this study aims at studying such patterns in mice on embryonic day 15.5 (E15.5). At E12.5, 8 mothers were injected with VPA (VPA group) and another 8 mothers were injected with saline (control group). At E15.5, electrocardiogram (ECG) records of 15 minutes were collected from the 16 mothers and 25 fetuses. A maximum of 5-minutes and a minimum of 1-minute were selected from the ECG data for analysis. Mean RR intervals and coupling ratios and their occurrence percentages were calculated per 1-minute. 1-minute analysis was done for periods with no arrhythmia and clear R peaks. The total number of 1-minute segments that were analyzed was 56 for the saline group and 54 for the VPA group. The correlation analysis between the 1:3 and 2:6 coupling ratios and RR intervals revealed that the ratios were significantly correlated in the saline group, whereas no significant correlations were observed in the VPA group. The results further revealed that fetal RR intervals are strongly correlated with maternal RR intervals in the saline group, but the same correlation is different in the VPA group. The presented results imply that maintaining certain coupling patterns are important for proper fetal cardiac development and maternal uptake of VPA may affect maternal-fetal HRs interactions.

Keywords: Cardiovascular, respiratory, and sleep devices - Wearables, Cardiovascular, respiratory, and sleep devices - Monitors, Cardiovascular, respiratory, and sleep devices - Diagnostics
Abstract: Many studies have focused on novel noninvasive techniques to monitor respiratory rate such as bioimpedance. We propose an algorithm to detect respiratory phases using wearable bioimpedance to compute time parameters like respiratory rate, inspiratory and expiratory times, and duty cycle. The proposed algorithm was compared with two other algorithms from literature designed to estimate the respiratory rate using physiological signals like bioimpedance. We acquired bioimpedance and airflow from 50 chronic obstructive pulmonary disease (COPD) patients during an inspiratory loading protocol. We compared performance of the algorithms by computing accuracy and mean average percentage error (MAPE) between the bioimpedance parameters and the reference parameters from airflow. We found similar performance for the three algorithms in terms of accuracy (>0.96) and respiratory time and rate errors (<3.42 %). However, the proposed algorithm showed lower MAPE in duty cycle (10.18 %), inspiratory time (10.65 %) and expiratory time (8.61 %). Furthermore, only the proposed algorithm kept the statistical differences in duty cycle between COPD severity levels that were observed using airflow. Accordingly, we suggest bioimpedance to monitor breathing pattern parameters in home situations.

Keywords: Vascular mechanics and hemodynamics - Vascular Hemodynamics, Cardiovascular and respiratory signal processing - Cardiovascular signal processing, Cardiovascular and respiratory signal processing - Blood pressure measurement
Abstract: Central blood pressure is a vital signal that provides relevant physiological knowledge about cardiovascular diseases risk factors. Generally, the standard invasive clinical protocols for measuring central blood pressure are challenging. On the other hand, noninvasive methods are more convenient but are not very accurate as they are commonly based on estimation approaches to approximate the central waveform from peripheral ones. In this paper, we propose a novel data-driven approach that combines machine learning tools and the cross-relation-based blind estimation methods to evaluate the aortic blood pressure waves using peripheral signals. Due to the lack of large real datasets to train the machine learning models, in this study, we utilize virtual pulse waves in-silico databases that are useful resources to evaluate the pre-clinical assessment of the hemodynamic analysis and algorithms. The estimation's performance of the proposed approach was compared with a pure machine learning-based model and the cross-relation-based blind estimation approach. In both cases, the hybrid approach shows promising results as the root-mean-squared error has been decreased by 25% with regards to the pure machine learning method and by 40% with respect to the cross-relation approach.

Keywords: Cardiovascular, respiratory, and sleep devices - Smart systems, Cardiovascular, respiratory, and sleep devices - Diagnostics, Cardiovascular and respiratory signal processing - Cardiovascular signal processing
Abstract: This paper presents a generic method to enhance performance and incorporate temporal information for cardiorespiratory-based sleep stage classification with a limited feature set and limited data. The classification algorithm relies on random forests and a feature set extracted from long-time home monitoring for sleep analysis. Employing temporal feature stacking, the system could be significantly improved in terms of Cohen’s κ and accuracy. The detection performance could be improved for three classes of sleep stages (Wake, REM, Non-REM sleep), four classes (Wake, Non-REM-Light sleep, Non-REM Deep sleep, REM sleep), and five classes (Wake, N1, N2, N3/4, REM sleep) from a κ of 0.44 to 0.58, 0.33 to 0.51, and 0.28 to 0.44 respectively by stacking features before and after the epoch to be classified. Further analysis was done for the optimal length and combination method for this stacking approach. Overall, three methods and a variable duration between 30 s and 30 min have been analyzed. Overnight recordings of 36 healthy subjects from the Interdisciplinary Center for Sleep Medicine at Charité-Universitätsmedizin Berlin and Leave-One-Out-Cross-Validation on a patient-level have been used to validate the method.

Keywords: Cardiovascular and respiratory system modeling - Cardiac models, Cardiovascular and respiratory system modeling - Compartmental modeling
Abstract: This paper investigates a subject-specific lumped parameter cardiovascular model for estimating Cardiac Output (CO) using the radial Arterial Blood Pressure (ABP) waveform. The model integrates a simplified model of the left ventricle along with a linear third order model of the arterial tree and generates reasonably accurate ABP waveforms along with the Dicrotic Notch (DN). Non-linear least square optimization technique is used to obtain uncalibrated estimates of cardiovascular parameters. Thermodilution CO measurements have been used to evaluate the CO estimation accuracy. The model achieves less than 15% normalized error across 10 subjects with different shapes of ABP waveform.

Keywords: Vascular mechanics and hemodynamics - Pulse wave velocity, Vascular mechanics and hemodynamics - Arterial pressure in cardiovascular disease, Cardiovascular regulation - Blood pressure variability
Abstract: Cardiomyopathies diseases affects a great number of the elderly population. An adequate identification of the etiology of a cardiomyopathy patient is still a challenge. The aim of this study was to classify patients by their etiology in function of indexes extracted from the characterization of the pulse transit time (PTT). This time series represents the time taken by the pulse pressure to propagate through the length of the arterial tree and corresponding to the time between R peak of ECG and the mid-point of the diastolic to systolic slope in the blood pressure signal. For each patient, the PTT time series was extracted. Thirty cardiomyopathy patients (CMP) classified as ischemic (ICM – 15 patients) and dilated (DCM – 15 patients) were analyzed. Forty-three healthy subjects (CON) were used as a reference. The PTT time series was characterized through statistical descriptive indices and the joint symbolic dynamics method. The best indices were used to build support vector machine models. The optimal model to classify ICM versus DCM patients achieved 89.6% accuracy, 78.5% sensitivity, and 100% specificity. When comparing CMP patients and CON subjects, the best model achieved 91.3% accuracy, 91.3% sensitivity, and 88.3% specificity. Our results suggests a significantly lower pulse transit time in ischemic patients.

Keywords: Cardiac mechanics, structure & function - Cardiac muscle mechanics, Cardiac mechanics, structure & function - Ventricular mechanics
Abstract: INTRODUCTION: Athletes training is often associated with morphological changes in the heart. In this sense, the ventricular pressure-volume (PV) relation provides a complete characterization of cardiac pump performance. Regarding the arterial system (AS), arterial wall viscosity is a source of energy dissipation, that takes place during mechanical transduction. Left ventricular stroke work (SW) constitutes the useful fraction of ventricular energy that is delivered to the AS. OBJECTIVE: Left ventricular PV-loops were evaluated in terms of AS viscous property, by means of the interaction of two SW components (Stroke Work Damping Ratio, SWDR), both in untrained and trained subjects. MATERIAL AND METHODS: Fourteen healthy individuals (seven trained) were noninvasively evaluated in terms of echocardiographic and aortic pressure measurements. RESULTS: SWDR was observed to be increased in trained subjects. CONCLUSION: SWDR was evaluated in trained individuals, being increased in comparison with the non-trained group. This effect is a consequence of a significant increase of SWD, which could be related with the viscous mechanical property of AS.

Keywords: Cardiovascular and respiratory signal processing - Heart Rate and Blood Pressure Variability, Sleep - Obstructive sleep apnea, Cardiovascular and respiratory signal processing - Cardiovascular signal processing
Abstract: Patients suffering from obstructive sleep apnea (OSA) usually present an increased sympathetic activity caused by the intermittent hypoxia effect on autonomic control. This study evaluated the relationship between sleep stages and the apnea duration, frequency, and type, as well as their impact on HRV markers in different groups of disease severity. The hypnogram and R-R interval signals were extracted in 81 OSA patients from night polysomnographic (PSG) recordings. The apnea-hypopnea index (AHI) defined patient classification as mild-moderate (AHI<=30, n=44) or severe (AHI>30, n=37). The normalized power in VLH, LF, and HF bands of RR series were estimated by a time-frequency approach and averaged in 1-min epochs of normal and apnea segments. The autonomic response and the impact of sleep stages were assessed in both segments to compare patient groups. Deeper sleep stages (particularly S2) concentrated the shorter and mild apnea episodes (from 10 to 40 s) compared to light (SWS) and REM sleep. Longer episodes (>50 s) although less frequent, were of similar incidence in all stages. This pattern was more pronounced for the group of severe patients. Moreover, during apnea segments, LFnu was higher (p=0.044) for the severe group, since V LFnu and HFnu presented the greatest changes when compared to normal segments. The non-REM sleep seems to better differentiate OSA patients groups, particularly through V LFnu and HFnu (p<0.001). A significant difference in both sympathetic and vagal modulation between REM and non-REM sleep was only found within the severe group. These results confirm the importance of considering sleep stages for HRV analysis to further assess OSA disease severity, beyond the traditional and clinically limited AHI values. Clinical relevance: Accounting for sleep stages during HRV analysis could better assess disease severity in OSA patients.

Keywords: Cardiac mechanics, structure & function - Heart failure, Cardiac mechanics, structure & function - Ventricular assist devices, Cardiovascular and respiratory system modeling - Blood flow models
Abstract: Left ventricular assist devices (LVADs) have long been used to treat adults with heart failure, but LVAD options for pediatric patients with heart failure are lacking. Despite the urgent need for long-term, implantable pediatric LVADs, design challenges such as hemolysis, pump thrombosis, and bleeding persist. We have developed a Hemocompatibility Assessment Platform (HAP) to identify blood trauma from individual LVAD components. A HAP would aid in refining pump components before in vivo testing, thereby preventing unnecessary animal sacrifice and reducing development time and cost. So that the HAP does not confound hemolysis data, the HAP drive system consists of an enlarged air-gap motor coupled to a magnetic levitation system. Although it is known that an enlarged air gap motor will have diminished performance, while the larger gap in the motor will cause less blood damage, the trade-offs are not fully characterized. Therefore, in this study we evaluated these trade-offs to determine an optimal rotor diameter for the HAP drive motor. The motor performance was characterized with an experimental method by determining the torque constant for the HAP drive motor with varied rotor diameters. The torque threshold was set as 10 mNm to achieve a nominal current of 3.5A. Hemolysis in the HAP drive motor gap was estimated by calculating scalar shear stress generated in the HAP motor gap analytically and numerically. A design criterion of 30 Pa was selected for scalar shear stress to achieve minimal hemolysis and platelet activation in the HAP drive system.

Keywords: Cardiac electrophysiology - Ventricular arrhythmia mechanisms, Cardiac electrophysiology - Simulation for cardiac arrhythmia
Abstract: Brugada Syndrome is a rare arrhythmia, hereditary in nature. It is caused due to mutation in genes that encodes sodium ion channels and it results sudden cardiac death in young adults. This paper aims to model a two dimensional SCN5A L812Q mutated endocardial tissue by modifying the model equations for sodium ion channel in the Ten Tusscher model for human ventricular tissue. Results show that the propagation of electrical activity in the mutated cells is slower when compared to the normal cells of the endocardial tissue. From this it is concluded that there is a large reduction of sodium current in the mutated region of the endocardial tissue. This leads to reduction in the total ionic current as well and further reduces the membrane potential. It also leads to the slower propagation of action potential in the mutated region when compared to the normal endocardial tissue.

Keywords: Cardiovascular and respiratory signal processing - Cardiovascular signal processing, Vascular mechanics and hemodynamics - Vascular Hemodynamics, Cardiovascular and respiratory system modeling - Cardiac models
Abstract: The arterial pulse waveform has an immense wealth of information in its morphology yet to be explored and translated to clinical practice. Wave separation analysis involves decomposing a pulse wave (pressure or diameter waveform) into a forward wave and a backward wave. The backward wave accumulates reflections due to arterial stiffness gradient, branching and geometric tapering of blood vessels across the arterial tree. The state-of-the-art wave separation analysis is based on estimating the input impedance of the target artery in the frequency/time domain, which requires simultaneously measured or modelled flow velocity and pressure waveform. We are proposing a new method of wave separation analysis using a multi-gaussian decomposition. The novelty of this approach is that it requires only a single pulse waveform at the target artery. Our method was compared against the triangular waveform-based impedance method. We successfully separated forward and backward waveform from the pressure waveform with maximum RMSE less than 5 mmHg and mean RMSE of 1.31 mmHg when compared against the triangular flow/impedance method. Results demonstrated a statistically significant correlation (r>0.66, p<0.0001) for Reflection Magnitude (RM) and Reflection Index (RI) for the multi-gaussian approach against the triangular flow method for 105 virtual subjects. The range of RM was from 0.35 to 0.97 (RI: 27.53% to 49.29%). This method proves to be a technique for evaluating reflection parameters if only a single pulse measurement is available from any artery.

Keywords: Cardiovascular and respiratory signal processing - Cardiovascular signal processing, Vascular mechanics and hemodynamics - Vascular Hemodynamics, Cardiovascular and respiratory system modeling - Cardiac models
Abstract: Conventional methods to calculate reflection transit time (RTT) is based on pulse counter analysis. An alternative to this approach is separating forward and backward components from a pulse waveform to calculate the RTT. State-of-the-art in wave separation requires simultaneously measured pressure and flow velocity waveforms. Practically, getting a simultaneous measurement from a single arterial site has its limitations, and this has made the translation of wave separation methods to clinical practice difficult. We propose a new method of wave separation analysis that requires only a single pulse waveform measurement using a multi-Gaussian decomposition approach. The novelty of the method is that it does not require any measured or modelled flow velocity waveform. In this method, the pulse waveform is decomposed into the sum of Gaussians and reconstructed based on model criteria. RTT is calculated as the time difference between normalized forward and backward waveform. The method’s feasibility in using RTT as a potential surrogate is demonstrated on 105 diverse selections of virtual subjects. The results were statistically significant and had a strong correlation (r>79, p<0.0001) against clinically approved artery stiffness markers such as Peterson’s elastic modulus (Ep), pulse wave velocity (PWV), specific stiffness index (β), and arterial compliance (AC). Out of all the elasticity markers, a better correlation was found against AC.

Keywords: Cardiovascular and respiratory system modeling - Cardiovascular control models
Abstract: Left ventricular assist device (LVAD) is a therapeutic option for advanced heart failure (HF) patients. This mechanical device assists a failing heart to circulate blood in the human body by adjusting its pump speed according to cardiac output. However, to use an LVAD for bridge-to-recovery, other criteria (e.g., aortic valve function) should be also considered to reduce complications of the LVAD implantation. In this work, we present an optimization-based control approach to meet the circulatory demand of blood, while maintaining the aortic valve to open and close repeatedly in a cardiac cycle. To validate the performance of the control method, several case studies were investigated, which incorporate different levels of HF severity and physical activity. The results show that the optimization-based control algorithm can quantify the trade-off between the aortic valve function and the blood flow, which will meet clinicians’ long quest to improve the myocardial functions for the use of an LVAD as bridge-to-recovery.

Keywords: Cardiovascular and respiratory system modeling - Compartmental modeling, Cardiovascular and respiratory system modeling - Vascular mechanics and hemodynamics, Cardiovascular and respiratory system modeling - Cardiac models
Abstract: Arterial compliance is a vital determinant of the ventriculo-arterial coupling dynamic. Its variation is detrimental to cardiovascular functions and associated with heart diseases. Accordingly, assessment and measurement of arterial compliance are essential in the diagnosis and treatment of chronic arterial insufficiency. Recently, experimental and theoretical studies have recognized the power of fractional calculus to perceive viscoelastic blood vessel structure and biomechanical properties. This paper presents five fractional-order model representations to describe the dynamic relationship between the aortic blood pressure input and blood volume. Each configuration incorporates a fractional-order capacitor element (FOC) to lump the apparent arterial compliance's complex and frequency dependence properties. FOC combines both resistive and capacitive attributes within a unified component, which can be controlled through the fractional differentiation order factor, alpha. Besides, the equivalent capacitance of FOC is by its very nature frequency-dependent, compassing the complex properties using only a few numbers of parameters. The proposed representations have been compared with generalized integer-order models of arterial compliance. Both models have been applied and validated using different aortic pressure and flow rate data acquired from various species such as humans, pigs, and dogs. The results have shown that the fractional-order framework is able to accurately reconstruct the dynamic of the complex and frequency-dependent apparent compliance dynamic and reduce the complexity. It seems that this new paradigm confers a prominent potential to be adopted in clinical practice and basic cardiovascular mechanics research.

Keywords: Respiratory transport, mechanics and control - Work of breathing, Respiratory transport, mechanics and control - Periodic breathing, Respiratory transport, mechanics and control - Pulmonary mechanics in disease
Abstract: This work details a methodology of design and test of a new prototype emergency mechanical ventilator called Fenix for the COVID-19 crisis in Peru. This equipment was manufactured with industrial equipment for the embedded and pneumatic systems, such as a Programmable Logic Controller (PLC), proportional flow valves, sensors, uninterruptible power supply (UPS), industrial panel HMI 15" and other electrical and pneumatic parts from Festo and Schneider Electric. This selection was in accordance with safety requirements based on ISO 80601-2-12: 2020-02. This study included two ventilatory modes, pressure- controlled in continuous mandatory ventilation (PC-CMV) and volume-controlled in continuous mandatory ventilation (VC-CMV), these control algorithms were evaluated analytically and experimentally in a FLUKE VT-650 Gas Flow Analyzer and an Acculung Fluke connected with a computer for comparing 9 ventilatory parameters in 4 different states as μ, simulation of the variation of the pressure control in a patient, and ϴ, simulation of alveolar recruitment in an intensive care patient, both states to PC-CMV, and also 𝛽, simulation of the variation of the flow control in a patient, and 𝛼, simulation of alveolar recruitment in an intensive care patient, both last states to VC-CMV. Additionally, we study the pressure, volume, and flow graphs in the Fenix user interface for comparison with data recovered from Fluke Medical VT650 Gas Flow Analyzer. The results demonstrate an error in the flow measurement for the 4 states due to the peaks that are not detected by the low-pass filter of the sensor, however, a similar trend is seen in the control ventilatory graphs of the calibrator. Finally, the ventilator prototype provides ventilatory support, with a maximum tidal volume error of 12.93 % and inspiratory pressure of -20.15 % with respect to the set value; and it allows to monitor the main ventilation parameters with a calculation error between -6 to 25%

Keywords: Cardiovascular, respiratory, and sleep devices - Diagnostics, Cardiovascular and respiratory system modeling - Cardiovascular Disease, Cardiac mechanics, structure & function - Cardiac structure from imaging
Abstract: Cardiac cine-MRI is one of the most important diagnostic tools for characterizing heart-related pathologies. This imaging technique allows clinicians to assess the mor-pho-logy and physiology of the heart during the cardiac cycle. Nonetheless, the analysis on cardiac cine-MRI is highly dependent on the observer expertise and a high inter-reader variability is frequently observed. Alternatively, the ejection fraction, a quantitative heart dynamic measure, is used to identify potential cardiac diseases. Unfortunately, this type of measurement is insufficient to distinguish among different cardiac pathologies. This quantification does not exploit all the heart functional information conveyed by cine-MRI sequences. Automatic image analysis might help to identify visual patterns associated with cardiac diseases in the cine-MRI sequences and highlight potential biomarkers. This paper introduces a conditional generative adversarial network that learns a mapping between the latent space and a generated cine-MRI data distribution involving information from five different cardiac pathologies. This net is guided from the left ventricle segmentation and the velocity field that is computed as prior information to focus on the deep representation of salient cardiac patterns. Once the deep neural networks are trained, a set of validation cine-MRI slices is represented in the embedding space. The associated embedding descriptor, in the latent space, is found by minimizing a reconstruction error in the generator output. We evaluated the obtained embedded representation as a disease marker by using different classification models in 16000 pathological cine-MRI slices. The representation retrieved by using the best conditional generative model configuration was used on the classifier models yielding an average accuracy of 90.04% and an average F1-score of 89.97% in the classification task.

Keywords: Sleep - Obstructive sleep apnea, Sleep - Periodic breathing & central apnea, Cardiovascular, respiratory, and sleep devices - Diagnostics
Abstract: Obstructive sleep apnea (OSA) is a sleep disorder in which repetitive upper airway obstructive events occur during sleep. These events can induce hypoxia, which is a risk factor for multiple cardiovascular and cerebrovascular diseases. Chronic obstructive pulmonary disease (COPD) is a disorder which induces a persistent inflammation of the lungs. This condition produces hypoventilation, affecting the blood oxygenation, and leads to an increased risk of developing lung cancer and heart disease. In this study, we evaluated how COPD affects the severity and characteristics of OSA in a multivariate demographic database including polysomnographic signals. Results showed SpO2 subtle variations, such as more non-recovered desaturations and increased time below a 90% SpO2 level, which, in the long term, could worsen the risk to suffer cardiovascular and cerebrovascular diseases.

Keywords: Cardiovascular and respiratory signal processing - Cardiovascular signal processing, Cardiac mechanics, structure & function - Atrial Fibrillation
Abstract: Atrial Fibrillation (AF) is the most common cardiac arrhythmia, and its progressive nature is associated with gradual atrial remodeling. The P-wave in the surface Electrocardiogram (ECG) reflects the atrial activation, while the modification of the atrial pathophysiological properties leads to P-wave morphology (PWM) alternations. In paroxysmal AF (pAF), the modifications of the PWM may have a spontaneous rather than permanent presence in the ECG signal. The analysis of the P-waves, during sinus rhythm, on a beat-to-beat basis, has revealed the existence of at least two PWM. In addition, the wavelet characteristics of the P-wave matching the main morphology can accurately distinguish the patients with pAF from healthy volunteers. In this work, we examine the hypothesis that there is an effect of the anti-arrhythmic medication on beat-to-beat PWM alternations of pAF patients. ECG signals of high frequency (1000Hz), in the three orthogonal leads, were collected for 81 pAF patients of minimal and mild AF burden, 47 of which receiving antiarrhythmic medication treatment, and from 56 healthy volunteers. Kruskal-Wallis test was performed, and the preliminary results denote the existence of statistically significant differences between the groups. A 3-class Random Forest classifier was trained, using the forward wrapper approach, resulting in a high overall classification performance (AUC = 85.75%). This analysis is a step towards improving understanding of medication effect on the variability of P-wave.

Keywords: Cardiovascular and respiratory signal processing - Cardiovascular signal processing
Abstract: Abstract— Fixed sample entropy (fSampEn) is a promising technique for the analysis of respiratory electromyographic (EMG) signals. Its use has shown outperformance of amplitude-based estimators such as the root mean square (RMS) in the evaluation of respiratory EMG signals with cardiac noise and a high correlation with respiratory signals, allowing changes in respiratory muscle activity to be tracked. However, the relationship between the fSampEn response to a given muscle activation has not been investigated. The aim of this study was to analyze the nature of the fSampEn measurements that are produced as the EMG activity increases linearly. Simulated EMG signals were generated and increased linearly. The effect of the parameters r and the size of the moving window N of the fSampEn were evaluated and compared with those obtained using the RMS. The RMS showed a linear trend throughout the study. A non-linear, sigmoidal-like behavior was found when analyzing the EMG signals using the fSampEn. The lower the values of r, the higher the non-linearity observed in the fSampEn results. Greater moving windows reduced the variation produced by too small values of r.

Clinical Relevance— Understanding the inherent non-linear relationship produced when using the fSampEn in EMG recordings will contribute to the improvement of the respiratory muscle activation assessment at different levels of respiratory effort in patients with respiratory conditions, particularly during the inspiratory phase.

Keywords: Cardiovascular, respiratory, and sleep devices - Monitors, Cardiovascular, respiratory, and sleep devices - Sensors
Abstract: This study proposes a novel respiratory signal detection system for 4D-CT in radiotherapy by measuring back pressure changes at multiple positions on CT couch. The 12-channel pressure sensor is fixed on CT couch to obtain patient’s back pressure signal. The 12-channel signal is transmitted to a PC at a sampling rate of 50 Hz after a signal conditioning circuit and an analog-digital converter. The amplitude of pressure changes is characterized to select the optimal channel. This system is validated by comparing with the respiratory signal collected synchronously with a real-time position management (RPM) system on 10 healthy volunteers. The correlation coefficient between the signals is 0.82 ± 0.09 (standard deviation) and the time shift is 0.32 ± 0.15 second. We conclude that the back pressure signal acquired by the proposed system has the potential to replace the clinical RPM system for respiratory signal detection in 4D-CT data acquisition.

Keywords: Cardiac mechanics, structure & function - Cardiac muscle mechanics, Cardiac mechanics, structure & function - Ventricular mechanics
Abstract: INTRODUCTION: Left ventricular (LV) interaction with the arterial system (arterial-ventricular coupling, AVC) is a central determinant of cardiovascular performance and cardiac energetics. Stress Echocardiography (SE) constitutes a valuable clinical tool in both diagnosis and risk stratification of patients with suspected and established coronary artery disease. Cluster Analysis (CA), an unsupervised Machine Learning technique, defines an exploratory statistical method which can be used to uncover natural groups within data. OBJECTIVE: To evaluate the capacity of CA to identify uncoupled groups with ischemic condition based on SE baseline information. MATERIAL AND METHODS: CA was applied to SE data acquired at baseline and peak exercise (PE) conditions. Obtained clusters were evaluated in terms of coupling conditions and LV wall motility alterations. RESULTS: Inter cluster significant AVC differences were obtained in terms of baseline data and wall motility alterations, confirmed by CA applied to PE data. CONCLUSION: AVC impairment was evidenced in both normal and ischemic subjects by applying CA.

Keywords: Vascular mechanics and hemodynamics - Vascular Disease, Vascular mechanics and hemodynamics - Pulse wave velocity, Vascular mechanics and hemodynamics - Arterial pressure in cardiovascular disease
Abstract: Abstract— A series of physiological measures can be assessed from the arterial pulse waveform, which is beneficial for cardiovascular health diagnosis, monitoring, and decision making. In this work, we have investigated the variations in regional pulse wave velocity (PWVR) and other pulse waveform indexes such as reflected wave transit time (RWTT), augmentation index (Alx), ejection duration index (ED), and subendocardial viability ratio (SEVR) with blood pressure (BP) parameters and heartrate on a vasoconstrictor drug-induced porcine model. Two healthy female (nulliparous and non-pregnant) Sus scrofa swine (~ 80 kg) was used for the experimental study. The measurement system consists of a catheter-based system with two highly accurate pressure catheters placed via the sheath at the femoral and carotid artery for acquiring and recording the pressure waveforms. The pulse waveform indexes were extracted from these recorded waveforms. Results from the pulse contour analysis of these waveforms demonstrated that Phenylephrine, as a post-synaptic alpha-adrenergic receptor agonist that causes vasoconstriction, produced a significant increment in the carotid BP parameters and heartrate. Due to the drug's effect, the PWVR and SEVR were significantly increased, whereas the RWTT, AIx index and ED index significantly decreased.

Clinical Relevance— This experimental study provides the usefulness of the pulse contour analysis and estimation of various pulse waveform indexes for cardiovascular health screening and diagnosis.

Keywords: Vascular mechanics and hemodynamics - Vascular mechanics, Vascular mechanics and hemodynamics - Vascular Hemodynamics, Cardiovascular and respiratory signal processing - Cardiovascular signal processing
Abstract: Measurement of arterial wall thickness is an integral component of vascular properties and health assessment. State-of-the-art automated or semi-automated techniques are majorly applicable to B-mode images and are not available for entry-level in-expensive devices. Considering this, we have earlier developed and validated an image-free (A-mode) ultrasound device, ARTSENS® for the evaluation of vascular properties. In this work, we present a novel gaussian-mixture modeling-based method to measure arterial wall thickness from A-mode frames, which is readily deployable to the existing technology. The method’s performance was assessed based on systematic simulations and controlled phantom experiments. Simulations revealed that the method could be confidently applied to A-mode frames with above-moderate SNR (>15 dB). When applied to A-mode frames acquired from the flow-phantom setup (SNR > 25 dB), the mean error was limited to (2 ± 1%), and RMSE was 19 μm, on comparison with B-mode measurements. The measured and reference wall thickness strongly agreed with each other (r = 0.88, insignificant mean bias = 7 μm, p = 0.16). The proposed method was capable of performing real-time measurements.

Keywords: Vascular mechanics and hemodynamics - Vascular mechanics, Vascular mechanics and hemodynamics - Vascular Hemodynamics, Vascular mechanics and hemodynamics - Pulse wave velocity
Abstract: Capturing vascular dynamics using ultrasound at a high framerate provided a unique way to track time-dependent and transient physiologic events non-invasively. In this work, we present an A-model high-framerate (500 frames per second) image-free ultrasound system for monitoring vascular structural and material properties. It was developed based on our clinically validated ARTSENS® technology. Following in-vitro verification on arterial flow phantoms, its measurement accuracy and high-framerate data acquisition and processing were verified in-vivo on 2 anesthetized Sus scrofa swine. Measurements of the carotid artery (the luminal diameter, distension, and wall thickness) obtained using the high-framerate system were comparable to those provided by a clinical-grade reference ultrasound imaging device (absolute error < 4%, < 6.3%, and < 6.6%, respectively). Notably, the morphology of the arterial distension waveforms obtained at high-framerate depicted vital physiological fiduciary points compared to the low-framerate reference waveform. The compression-decompression pattern of the arterial wall was also captured with the high-framerate system, which is challenging with low-framerate ultrasound. Potential applications of these high temporal structural waveforms have also been discussed.

Keywords: Cardiac mechanics, structure & function - Ventricular assist devices, Cardiovascular and respiratory system modeling - Cardiovascular control models, Cardiovascular and respiratory system modeling - Blood flow models
Abstract: This paper deals with designing a physiological adaptive control law for a turbodynamic ventricular assist device (TVAD) using a lumped parameter time-varying model that describes the cardiovascular system. The TVAD is a rotary blood pump driven by an electrical motor. The system simulation also includes the adaptive feedback controller, which provides a physiologically correct cardiac output under different preload and afterload conditions. The cardiac output is estimated at each heartbeat, and the control objective is achieved by dynamically changing the motor speed controller's reference based on the systolic pressure error. TVADs provide support for blood circulation in patients with heart failure. To improve the performance of these devices, several control strategies have been developed over the years, with an emphasis on the physiological strategies that adapt their parameters to improve the patient's condition. In this paper, a new strategy is proposed using a variable gain physiological controller to keep the cardiac output in a reference value under changes in both preload and afterload. Computational models are used to evaluate the performance of this control technique, which has shown better results of adaptability than constant speed controllers and constant gain controllers.

Keywords: Vascular mechanics and hemodynamics - Pulse wave velocity, Cardiovascular and respiratory signal processing - Pulse transit time, Vascular mechanics and hemodynamics - Vascular mechanics
Abstract: Cardiovascular community has started clinically adopting the assessment of local stiffness, contrary to the traditionally measured carotid-femoral pulse wave velocity (PWV). Though they offer higher reliability, ultrasound methods require advanced hardware and processing methods to perform real-time measurement of local PWV. This work presents a system and method to perform online PWV measurement in an automated manner. It is a fast image-free ultrasound technology that meets the methodological requirements necessary to measure small orders of local pulse transit, from which PWV is measured. The measurement accuracy and repeatability were assessed via phantom experiments, where the measured transit time-based PWV (PWVTT) was compared against the theoretically calculated PWV from Bramwell-Hill equation (PWVBH). The beat-to-beat variability in the measured PWVTT was within 3%. PWVTT values strongly correlated (r=0.98) with PWVBH, yielding a negligible bias of -0.01 m/s, mean error of 3%, and RMSE of 0.27 m/s. These study results demonstrated the presented system’s reliability in yielding online local PWV measurements.

Keywords: Cardiac electrophysiology - Simulation for cardiac arrhythmia, Cardiovascular and respiratory signal processing - Cardiovascular signal processing, Cardiovascular and respiratory signal processing - Heart Rate and Blood Pressure Variability
Abstract: A lumped parameter synergistic model of the human cardiovascular system (CVS) is proposed to integrate the heart's electrical activity with its mechanical activity. This model can represent the physiological condition of a patient in an effective way, whether it is considered normal or with some cardiac disorders. The electrical activity is coupled to the CVS model through electrocardiogram (ECG) signals in the suggested model. The variations in ECG morphology are detected by appropriate algorithms and changes parameters of the CVS model, such as systemic resistance and end-systolic pressure-volume relationship.

Keywords: Cardiovascular and respiratory signal processing - Cardiovascular signal processing, Cardiovascular, respiratory, and sleep devices - Wearables
Abstract: This work presents ReBeatICG, a real-time, low-complexity beat-to-beat impedance cardiography (ICG) delineation algorithm that allows hemodynamic parameters monitoring. The proposed procedure relies only on the ICG signal compared to most algorithms found in the literature that rely on synchronous electrocardiogram signal (ECG) recordings. ReBeatICG was designed with implementation on an ultra-low-power microcontroller (MCU) in mind. The detection accuracy of the developed algorithm is tested against points manually labeled by cardiologists. It achieves a detection Gmean accuracy of 94.9%, 98.6%, 90.3%, and 84.3% for the B, C, X, and O characteristic points, respectively. Furthermore, several hemodynamic parameters were calculated based on annotated characteristic points and compared with values generated from the cardiologists' annotations. ReBeatICG achieved mean error rates of 0.11 ms, 9.72 ms, 8.32 ms, and 3.97% for HR, LVET, IVRT, and relative C-point amplitude, respectively.

Keywords: Cardiovascular and respiratory signal processing - Cardiovascular signal processing, Cardiovascular and respiratory signal processing - Time-frequency, time-scale analysis of cardiorespiratory variability
Abstract: Photoplethysmography (PPG) is a non-invasive and economical technique to extract vital signs of the human body. Although it has been widely used in consumer and research grade wrist devices to track a user's physiology, the PPG signal is very sensitive to motion which can corrupt the signal's quality. Existing Motion Artifact (MA) reduction techniques have been developed and evaluated using either synthetic noisy signals or signals collected during high-intensity activities - both of which are difficult to generalize for real-life scenarios. Therefore, it is valuable to collect realistic PPG signals while performing Activities of Daily Living (ADL) to develop practical signal denoising and analysis methods. In this work, we propose an automatic pseudo clean PPG generation process for reliable PPG signal selection. For each noisy PPG segment, the corresponding pseudo clean PPG reduces the MAs and contains rich temporal details depicting cardiac features. Our experimental results show that 71% of the pseudo clean PPG collected from ADL can be considered as high quality segment where the derived MAE of heart rate and respiration rate are 1.46 BPM and 3.93 BrPM, respectively. Therefore, our proposed method can determine the reliability of the raw noisy PPG by considering quality of the corresponding pseudo clean PPG signal.

Keywords: Cardiovascular and respiratory signal processing - Lung Sounds, Cardiovascular, respiratory, and sleep devices - Smart systems, Cardiovascular, respiratory, and sleep devices - Wearables
Abstract: Mobile and wearable devices are being increasingly used for developing audio based machine learning models to infer pulmonary health, exacerbation and activity. A major challenge to widespread usage and deployment of such pulmonary health monitoring audio models is to maintain accuracy and robustness across a variety of commodity devices, due to the effect of device heterogeneity. Because of this phenomenon, pulmonary audio models developed with data from one type of device perform poorly when deployed on another type of device. In this work, we propose a framework incorporating feature normalization across individual frequency bins and combining task specific deep neural networks for model invariance across devices for pulmonary event detection. Our empirical and extensive experiments with data from 131 real pulmonary patients and healthy controls show that our framework can recover up to 163.6% of the accuracy lost due to device heterogeneity for four different pulmonary classification tasks across two broad classification scenarios with two common mobile and wearable devices: smartphone and smartwatch.

Keywords: Cardiovascular and respiratory signal processing - Complexity in cardiovascular or respiratory signals, Cardiovascular and respiratory signal processing - Non-linear cardiovascular or cardiorespiratory relations
Abstract: Preeclampsia (PE) is one of the leading causes of maternal mortality worldwide. Although clinical strategies to prevent the early onset of PE have been proposed, the ultimate solution is to end the pregnancy. Therefore, patients’ identification with major PE risk is important towards the prevention and better management of a severe manifestation of the illness. This study aims to analyze the systolic blood pressure (SBP) and diastolic blood pressure (DBP) time series through a nonlinear perspective using symbolic dynamics and to incorporate a multi-scale assessment in the first trimester of pregnancy, previous to the clinical manifestation of PE. The study group of normotensive women who developed and were diagnosed with PE included 14 pregnant women, a normotensive throughout pregnancy control group (N) consisting of 14 participants, and a group of 14 normotensive women during pregnancy without comorbidities (S) were matched with PE by age, body mass index, gestational age and comorbidities. The preliminary results of this study showed a decreased complexity of SBP, assessed by multiscale symbolic entropy in the first trimester in PE patients, in comparison with normotensive pregnant women.

Keywords: Cardiovascular and respiratory system modeling - Cardiac models, Vascular mechanics and hemodynamics - Vascular Hemodynamics, Cardiac electrophysiology - Inverse problems
Abstract: Left ventricular assist devices (LVADs) are mechanical pumps that help patients with chronic heart failure waiting for a heart transplant. Mathematical models of these devices can be used along cardiovascular system (CVS) models to evaluate the assistance performance under different operating modes. The estimation of the CVS model parameters for a particular patient and numerical simulations allow the implementation of adequate LVAD operation mode. This work presents a method to estimate the parameters of a CVS model using only one hemodynamic variable: the systemic arterial pressure (P_s). Synthetic signals of P_s are used to solve this ill-posed inverse problem partially, and the results show the high accuracy of the proposed method, which achieves 0.5%.

Keywords: Respiratory transport, mechanics and control - Work of breathing, Respiratory transport, mechanics and control - Respiratory variability, Respiratory transport, mechanics and control - Pulmonary mechanics in disease
Abstract: In clinical practice, when a patient is undergoing mechanical ventilation, it is important to identify the optimal moment for extubation, minimizing the risk of failure. However, this prediction remains a challenge in the clinical process. In this work, we propose a new protocol to study the extubation process, including the electromyographic diaphragm signal (diaEMG) recorded through 5–channels with surface electrodes around the diaphragm muscle. First channel corresponds to the electrode on the right. A total of 40 patients in process of withdrawal of mechanical ventilation, undergoing spontaneous breathing tests (SBT), were studied. According to the outcome of the SBT, the patients were classified into two groups: successful (SG: 19 patients) and failure (FG: 21 patients) groups. Parameters extracted from the envelope of each channel of diaEMG signal in time and frequency domain were studied. After analyzing all channels, the second presented maximum differences when comparing the two groups of patients, with parameters related to root mean square (p = 0.005), moving average (p = 0.001), and upward slope (p = 0.017). The third channel also presented maximum differences in parameters as the time between maximum peak (p = 0.004), and the skewness (p = 0.027). These results suggest that diaphragm EMG signal could contribute to increase the knowledge of the behaviour of respiratory system in these patients and improve the extubation process.

Keywords: Cardiovascular and respiratory signal processing - Pulse transit time, Cardiovascular and respiratory signal processing - Blood pressure measurement, Vascular mechanics and hemodynamics - Pulse wave velocity
Abstract: Background: Non-contact measurement of physiological vital signs, such as blood pressure (BP), by video-based photoplethysmography (vPPG) is a potential means for remote health monitoring. However, the signal-to-noise ratio of cardiovascular signals within the vPPG is very low. Objective: This study investigates the potential of BP estimation from vPPG. Methods: In 10 healthy volunteers (4 females, 28 ± 7 years), continuous electrocardiogram, finger BP and video of the face and palm of the hand were recorded. BP was varied by isometric hand grip exercise and leg ischemia. Four vPPG methods were compared: (i) averages of the green (GREEN) color intensity; (ii) the best linear combination of color channels using independent component analysis (ICA); (iii) a linear combination of chrominance-based (CHROM) signal by standardizing the skin color profile; (iv) plane orthogonal to the skin tone (POS) as vPPG signal. These were applied to 14 regions of interest (ROIs) on the face and 5 ROIs on the palm. Pulse transit time (PTT) between ROIs, for all permutations, were calculated and the correlation with BP quantified. Results: A significant, negative PTT-BP correlation was defined as success. A maximum success rate of 80% was achieved, occurring for the GREEN, POS and ICA methods only for specific ROIs within the face, but not for any permutation using the hand. Conclusions: These results indicate that the use of vPPG for estimation of BP will be challenging. A combination of different vPPG methods and within-face ROIs may yield useful information.

Keywords: Cardiovascular and respiratory signal processing - Blood pressure measurement
Abstract: In this paper, we introduce PulseLab, a comprehensive MATLAB toolbox that enables estimating the blood pressure (BP) from electrocardiogram (ECG) and photoplethysmogram (PPG) signals using pulse wave velocity (PWV)-based models. This universal framework consists of 6 sequential modules, covering end-to-end procedures that are needed for estimating BP from raw PPG/ECG data. These modules are "dataset formation", "signal pre-processing", "segmentation", "characteristic-points detection", "pulse transit time (PTT)/ pulse arrival time (PAT) calculation", and "model validation". The toolbox is expandable and its application programming interface (API) is built such that newly-derived PWV-BP models can be easily included. The toolbox also includes a user-friendly graphical user interface (GUI) offering visualization for step-by-step processing of physiological signals, position of characteristic points, PAT/PTT values, and the BP regression results. To the best of our knowledge, PulseLab is the first comprehensive toolbox that enables users to optimize their model by considering several factors along the process for obtaining the most accurate model for cuff-less BP estimation.

Keywords: Cardiovascular and respiratory signal processing - Blood pressure measurement, Cardiovascular and respiratory signal processing - Cardiovascular signal processing, Vascular mechanics and hemodynamics - Arterial pressure in cardiovascular disease
Abstract: Blood Pressure (BP) is one of the four primary vital signs indicating the status of the body's vital (life-sustaining) functions. BP is difficult to continuously monitor using a sphygmomanometer (i.e. a blood pressure cuff), especially in everyday-setting. However, other health signals which can be easily and continuously acquired, such as photoplethysmography (PPG), show some similarities with the Aortic Pressure waveform. Based on these similarities, in recent years several methods were proposed to predict BP from the PPG signal. Building on these results, we propose an advanced personalized data-driven approach that uses a three-layer deep neural network to estimate BP based on PPG signals. Different from previous work, the proposed model analyzes the PPG signal in time-domain and automatically extracts the most critical features for this specific application, then uses a variation of recurrent neural networks (RNN) called Long-Short-Term-Memory (LSTM) to map the extracted features to the BP value associated with that time window. Experimental results on two separate standard hospital datasets, yielded absolute errors mean and absolute error standard deviation for systolic and diastolic BP values outperforming prior works.

Keywords: Vascular mechanics and hemodynamics - Pulse wave velocity, Cardiovascular and respiratory signal processing - Cardiovascular signal processing, Cardiovascular and respiratory signal processing - Blood pressure measurement
Abstract: Wave intensity analysis (WIA) as a framework to assess cardiovascular hemodynamics has been successfully used in many clinical applications. Typically, wave intensity calculations require the simultaneous acquisition of blood velocity and blood pressure at the same vascular site. Unfortunately, many hemodynamic parameters that are used to monitor pre-operative patient hemodynamic state use both invasively acquired blood pressure measurements in catheterization laboratory and non-invasively acquired blood velocity measurements. To utilize wave intensity analysis to assess patients undergoing cardiac interventional procedures, we have developed a graphical user interface (GUI) that uses standard clinical measurements which include invasive blood pressure waveforms and Doppler echocardiography images to calculate wave intensity parameters. The GUI consists of three main subroutines that allow clinicians to import raw data and extract and analyze the blood pressure and blood velocity signals separately. Using the electrocardiogram signals as an alignment marker, the re-formatted signals are aligned, and wave intensity is calculated. Wave intensity features such as forward compression wave (FCW), forward expansion wave (FEW) and wave speed are calculated and output in a table for statistical analysis. The GUI represents the first attempt to create a program that encourages clinicians to use WIA for hemodynamic assessment in patients undergoing cardiac catheterization procedures with the data they have already procured.

Keywords: Cardiovascular and respiratory signal processing - Complexity in cardiovascular or respiratory signals, Cardiovascular and respiratory signal processing - Lung Sounds, Cardiovascular and respiratory signal processing - Cardiovascular signal processing
Abstract: Obtaining high quality heart and lung sounds enables clinicians to accurately assess a newborns cardio-respiratory health and provide timely care. However, noisy chest sound recordings are common, hindering timely and accurate assessment. A new Non-negative Matrix Co-Factorisation based approach is proposed to separate noisy chest sound recordings into heart, lung and noise components to address this problem. This method is achieved through training with 20 high quality heart and lung sounds, in parallel with separating the sounds of the noisy recording. The method was tested on 68 10-second noisy recordings containing both heart and lung sounds and compared to the current state of the art Non-negative Matrix Factorisation methods. Results show significant improvements in heart and lung sound quality scores respectively, and improved accuracy of 3.6bpm and 1.2bpm in heart and breathing rate estimation respectively, when compared to existing methods.

Keywords: Cardiovascular and respiratory signal processing - Cardiovascular signal processing, Cardiovascular and respiratory signal processing - Blood pressure measurement
Abstract: Arterial blood pressure (ABP) waveform is a common physiological signal that contains a wealth of cardiovascular information. According to the cardiac cycle, the ABP waveform is divided into rapid ejection, systolic and diastolic phases. Therefore, the characteristic points of the arterial blood pressure waveform, i.e. their onsets, systolic peaks, represent the timing of the minimum and maximum pressures. It is important to detect these characteristic points accurately. Recently, many researchers have introduced some feature points detection methods, but the accuracy is not particularly high. In this paper, a deep learning method is proposed to achieve periodic segmentation and feature points detection of ABP signals using a one-dimensional U-Net network. The network can split the ABP signal into two parts and accurately detect the feature points. The method is validated on an ABP dataset of 126 people, 500 people each. Performances are good at different tolerance thresholds, with an average time difference of less than 1.5 ms. Finally, the method performs with 99.79% and 99.79% sensitivity, 99.99% and 99.94% positive predictivity, and 0.23% and 0.27% error rates for both onsets and systolic peaks at a tolerance threshold of 30 ms. To our knowledge, this is the first paper to use deep learning methods for the onsets and systolic peaks detections of ABP signals.

Keywords: Smart neural implants, Neurological disorders - Sleep, Neural interfaces - Bioelectric sensors
Abstract: Stimulation of target neuronal populations using optogenetic techniques during specific sleep stages has begun to elucidate the mechanisms and effects of sleep. To conduct closed-loop optogenetic sleep studies in untethered animals, we designed a fully integrated, low-power system-on-chip (SoC) for real-time sleep stage classification and stage-specific optical stimulation. The SoC consists of a 4-channel analog front-end for recording polysomnography signals, a mixed-signal machine-learning (ML) core, and a 16-channel optical stimulation back-end. A novel ML algorithm and innovative circuit design techniques improved the online classification performance while minimizing power consumption. The SoC was designed and simulated in 180 nm CMOS technology. In an evaluation using an expert labeled sleep database with 20 subjects, the SoC achieves a high sensitivity of 0.806 and a specificity of 0.947 in discriminating 5 sleep stages. Overall power consumption in continuous operation is 97 uW.

Keywords: Human performance - Cognition, Human performance - Ergonomics and human factors
Abstract: To explore the effectiveness of using EEG spectral power and multiscale sample entropy for accessing mental workload in different tasks, working memory tasks with different information types and various mental loads were designed based on the N-Back paradigm. EEG signals from 18 normal adults were acquired when tasks were being performed. Linear (relative power in Theta and Alpha band, etc.) and nonlinear (multiscale sample entropy) features of EEGs were then extracted. Indices that can effectively reflect mental workload levels were selected by using multivariate analysis of variance statistical approach. Results showed that with the increment of task load, power of frontal Theta, Theta/Alpha ratio, and sample entropies (scales>10) in parietal regions increased significantly first and decreased slightly then, while the power of central-parietal Alpha decreased significantly first and increased slightly then. Considering the difference among task types, no difference in power of frontal Theta, central-parietal Alpha, and sample entropies (scales>10) of parietal regions were found between verbal and object tasks, as well as between two spatial tasks. No difference of frontal Theta/Alpha ratio was found in all the four tasks. The results can provide evidence for the mental workload evaluation in tasks with different information types.

Keywords: Brain-computer/machine interface, Neurological disorders, Neural signals - Machine learning & Classification
Abstract: Aim: Brain-Computer Interfaces (BCIs) hold promise to provide people with partial or complete paralysis, the ability to control assistive technology. This study reports offline classification of imagined and executed movements of the upper and lower limb in one participant with multiple sclerosis and people with no limb function deficits. Methods: We collected neural signals using electroencephalography (EEG) while participants performed executed and imagined motor tasks as directed by prompts shown on a screen. Results: Participants with no limb function attained >70% decoding accuracy on their best-imagined task compared to rest and on at-least one task comparison. The participant with multiple sclerosis also achieved accuracies within the range of participants with no limb function loss. Conclusion: While only one case study is provided it was promising that the participant with MS was able to achieve comparable classification to that of the seven healthy controls. Further studies are needed to assess whether people suffering from MS may be able to use a BCI to improve their quality of life.

Keywords: Neural stimulation, Neurorehabilitation
Abstract: Our recent study showed that low-intensity focused ultrasound stimulation (FUS) of the vagus nerve is capable of lowering blood pressure (BP). However, it remains unknown that what is the underlying mechanisms of BP modulation with FUS. In our preliminary experiments, we noticed that there was temperature elevation accompanied the FUS. Thus, to verify whether the thermal effect of ultrasound contributes in the BP lowering effect, this study compared the BP response under the FUS (with thermal effect and mechanical effect) and the alternative heating source treatment (AHST) (with thermal effect only) of left vagus nerve. Six Sprague Dawley rats were randomly divided into two groups (FUS, n=3 and AHST, n=3). In vivo temperature measurements were conducted to evaluate the heating performance of the FUS and the AHST. Blood pressure (BP) waveform was continuously recorded from the right common artery and was used for analyzing systolic BP (SBP), diastolic BP (DBP), mean BP (MBP), and heart rate (HR). The results showed that the SBP, DBP, MBP and HR decreased during the 15-min FUS. However, most of the SBP, DBP, MBP and HR increased during the 15-min AHST, which had the approximate temperature elevation of the FUS. Thus, the thermal effect of ultrasound probably does not contribute in the BP-lowering effect induced by low-intensity FUS of the vagus nerve.

Keywords: Brain-computer/machine interface, Brain functional imaging - Evoked potentials, Human performance - Attention and vigilance
Abstract: This paper investigates for the first time the use of single-frequency phase-coded stimuli to detect covert visuo-spatial attention (CVSA) with steady-state visual evoked potentials (SSVEP). Two 15Hz pattern-onset stimulations were encoded with opposite phases and simultaneously presented on a LCD monitor. The effects of attending each stimulus on the amplitudes and phases of the evoked SSVEPs across the visual cortex are explored. A real-time CVSA classification experiment was simulated offline with 9 BCI-naive subjects, achieving an average classification accuracy of 88.4 ± 8% SE. Our results are, to our knowledge, the first report that CVSA can be decoded with SSVEP using single-frequency phase-coded stimuli. This opens opportunities for attention-tracking applications with largely increased number of targets.

Keywords: Human performance - Gait, Human performance - Modelling and prediction, Human performance - Activities of daily living
Abstract: Abstract— Older adults with dementia have a high risk of developing drug-induced parkinsonism; however, formal clinical gait assessments are too infrequent to capture fluctuations in their gait. Camera-based human pose estimation and tracking provides a means to frequently monitor gait in non-clinical settings. In this study, 2160 walking bouts from 49 participants were recorded using a ceiling-mounted camera. Recorded color videos were processed using AlphaPose to obtain 2D joint trajectories of the participant as they were walking down a hallway of the unit. A subset of 324 walking bouts from 14 participants were annotated with clinical scores of parkinsonism on the Unified Parkinson’s Disease Rating Scale (UPDRS)-gait scale. Linear, random forest, and ordinal logistic regression models were evaluated for regression to UPDRS-gait scores using engineered 2D gait features calculated from the AlphaPose joint trajectories. Additionally, spatial temporal graph convolutional networks (ST-GCNs) were trained to predict UPDRS-gait scores from joint trajectories and gait features using a two-stage training scheme (self-supervised pretraining stage on all walks followed by a finetuning stage on labelled walks). All models were trained using leave-one-subject-out cross-validation to simulate testing on previously unseen participants. The macro-averaged F1-score was 0.333 for the best model operating on only gait features and 0.372 for the top ST-GCN model that used both joint trajectories and gait features as input. When accepting predicted scores that were only off by at most 1 point on the UPDRS-gait scale, the accuracy of the model that only used gait features was 82.8%, while the model that also used joint trajectories had an accuracy of 94.2%. Clinical Relevance— The combination of gait features and joint trajectories capture parkinsonian qualities in gait better than either group of data individually.

Keywords: Neural stimulation, Sensory neuroprostheses - Visual, Neural signal processing
Abstract: Retinal prostheses can restore the basic visual function of patients with retinal degeneration, which relies on effective electrical stimulation to evoke the physiological activities of retinal ganglion cells (RGCs). Current electrical stimulation strategies suffer from unstable effects and insufficient stimulation positions. Therefore, it is crucial to determine the optimal parameters for precise and safe electrical stimulation. Biphasic voltages (cathode-first) with a pulse width of 25 ms and different amplitudes were used to ex vivo stimulate RGCs of three wild-type (WT) mice using a commercial microelectrode array (MEA) recording system. Based on a facile and efficient spike sorting method, comprehensive statistics of RGCs response types were performed, and the influence of electrical stimulation on RGCs response status was analyzed. There were three types of RGCs response measured from the retinas of three WT mice, and the proportions were calculated to be 91.5%, 3.11% and 5.39%, respectively. This work can provide an in-depth understanding of the internal effects of electrical stimulation and RGCs response, with the potential as a useful guidance for optimizing parameters of electrical stimulation strategies in retinal prostheses.

Keywords: Brain functional imaging - EEG, Human performance - Activities of daily living, Human performance - Modelling and prediction
Abstract: Companion robots play an important role to accompany humans and provide emotional support, such as reducing human social isolation and loneliness. Based on recognizing human partner’s mental states, a companion robot is able to dynamically adjust its behaviors, and make human-robot interaction smoother and natural. Human emotion has been recognized by many modalities like facial expression and voice. Neurophysiological signals have shown promising results in emotion recognition, since it is an innate signal of human brain which cannot be faked. In this paper, emotional state recognition using a neurophysiology method is studied to guide and modulate companion-robot navigation to enhance its social capabilities. Electroencephalogram (EEG), a type of neurophysiological signals, is used to recognize human emotional state, and then feed into a navigation path planning algorithm for controlling a companion robot’s routes. Simulation results show that mobile robot presents navigation behaviors modulated by dynamic human emotional states.

Keywords: Sensory neuroprostheses - Auditory, Neural signals - Machine learning & Classification, Smart neural implants - Cochlear
Abstract: Despite being able to restore speech perception with 99% success rate, cochlear implants cannot successfully restore pitch perception or music appreciation. Studies suggest that if auditory neurons were activated with fine timing closer to that of natural responses pitch would be restored. Predicting the timing of cochlear responses requires detailed biophysical models of sound transmission, inner hair cell responses, and outer hair cell responses. Performing these calculations is computationally costly for real time cochlear implant stimulation. Instead, implants typically modulate pulse amplitude of fixed pulse rate stimulation with the band-limited envelopes of incoming sound. This method is known to produce unrealistic responses, even to simple step inputs. Here we investigate using a machine learning algorithm to optimize the prediction of the desired firing patterns of the auditory afferents in response to sinusoidal and step modulation of pure tones. We conclude that a trained network that consists of 25 GRU nodes can reproduce fine timing with 4.4 percent error on a test set of sines and steps. This trained network can also transfer learn and capture features of natural sounds that are not captured by standard CI algorithms. Additionally, for 0.5 second test inputs, the ML algorithm completed the sound to spike rate conversion in 300x less time than the phenomenological model. This calculation occurs at a real-time compatible rate of 1 ms for 1 second of spike timing prediction on an i9 microprocessor. This suggests that this is a feasible approach to pursue for real-time CI implementation.

Keywords: Motor neuroprostheses - Neuromuscular stimulation, Neuromuscular systems - Postural and balance, Human performance - Activities of daily living
Abstract: Seated stability is a major concern of individuals with trunk paralysis. Trunk paralysis is commonly caused by spinal cord injuries (SCI) at or above the thoracic spine. Current methods to improve stability restrict the movement of the user by constraining their trunk to an upright position. Feedback control of functional neuromuscular stimulation (FNS) can help maintain seated stability while still allowing the user to perform movements to accomplish functional tasks. In this study, an individual with a SCI (C7, AIS B) and an implanted stimulator capable of recruiting trunk and hip musculature unilaterally moved a weighted jar on a countertop to and from three prescribed stations directly in front, laterally, and across midline. For comparison, the tasks were performed with constant baseline stimulation and with feedback modulated stimulation based on the tilt of the trunk obtained from an external accelerometer fed into two PID controllers; one for forward trunk pitch and the other for lateral roll. The trunk pitch and roll angles were obtained through motion capture cameras and various measures of postural sway (95% fitted ellipse area, root mean squared (RMS), path length) and the repeatability (coefficient of variation (CoV), variance ratio (VR)) were calculated. Feedback control significantly increased RMS of trunk movement along the major axis of the fitted ellipse, but decreased RMS values during bending along the minor axis of motion. As a result, the fitted ellipse area decreased when deploying the jar to one of the stations and increased with the other two. The CoV indicated reduced variation in the presence of feedback controlled stimulation for all stations, and VR showed higher repeatability in trunk pitch. Plots of the trunk pitch and roll revealed a faster return to upright motion due to feedback stimulation.

Keywords: Brain-computer/machine interface, Neural signals - Machine learning & Classification, Brain functional imaging - EEG
Abstract: The accurate emotional assessment of humans can prove beneficial in health care, security investigations and human interaction. In contrast to emotion recognition from facial expressions which can prove to be inaccurate, analysis of electroencephalogram (EEG) activity is a more accurate representation of one’s state of mind. With advancements in deep learning, various methods are being employed for this task. In this research, importance of attention mechanism in EEG signals is introduced through two vision transformer based methods for the classification of EEG signals on the basis of emotions. The first method utilizes 2-D images generated through continuous wavelet transform (CWT) of the raw EEG signals and the second method directly operates on the raw signal. The publicly available and widely accepted DEAP dataset has been utilized in this research for validating the proposed approaches. The proposed approaches report very high accuracies of 97% and 95.75% using CWT and 99.4% and 99.1% using raw signal for valence and arousal classifications respectively, which clearly highlights the significance of attention mechanism for EEG signals. The proposed methodology also ensures faster training and testing time which suits the clinical purposes.

Keywords: Neural signals - Machine learning & Classification, Neural signal processing, Brain-computer/machine interface
Abstract: The goal of this research is to evaluate the usability of new features to classify EEG data from several completely locked-in patients (CLIS), and eventually build a more reliable communication system for them. Patients in such state are completely paralyzed, preventing them to be able to talk, but they retain their cognitive abilities.

The data were obtained from four CLIS patients and recorded during an auditory paradigm task during which they were asked yes/no questions. Spectral measures such as the relative power of delta, theta, alpha, beta and gamma frequency bands, spectral edge frequencies (SEF50 and SEF95), complexity measure obtained from Poincaré plots and connectivity measures such as the imaginary part of coherency and the weighted Symbolic Mutual Information (wSMI) were used as features. The data was classified using Random Forest and Support Vector Machine, two methods successfully used to classify mental states in both healthy subjects and patients. Additionally, two cases were studied. The first case uses data recorded when the patient is answering questions, while in the second case it also includes data recorded when the experimenter is asking the questions.

The classification accuracy during training varies between 51.73 to 67.72% in the first case, and from 50.41 to 67.94% for the second case. Overall, wSMI with a time lag of 64 ms gave the best classification accuracy and in general, Random Forest appears to be the best classification method.

Keywords: Sensory neuroprostheses - Visual, Neural stimulation, Brain-computer/machine interface
Abstract: Cortical vision prostheses are being developed to restore sight in blind patients. Existing electrode arrays that electrically stimulate cortical tissue to artificially induce neural activity are difficult to position directly next to each other. Leaving space between implants creates gaps in the visual field where no visual percepts can be created. Here, we propose current steering as a solution to elicit a neural response between physical electrode locations. We assessed the centroid of neural activity produced by dual-electrode stimulation in the visual cortex of Sprague-Dawley rats. We determined that this centroid could be shifted between physical electrodes by altering the ratio of charge delivered to each electrode. This centroidal shift could enable better environmental perception for cortical implant patients by creating a complete visual field representation while maintaining safe array spacing.

Keywords: Neuromuscular systems - Postural and balance, Neurorehabilitation, Human performance
Abstract: Human balance control is a critical prerequisite to nearly all activities, and human falls are a major health concern. The most robust way to assess reactive balance is to apply external perturbations. Perturbations are typically delivered with destabilizing motorized surfaces, external forces, visual motion, or neural stimulation. However, most devices that perturb walking in research settings are not likely to see wide clinical use due to cost, space, and time constraints. In contrast, there are low-cost destabilizing clinical tests that might require similar neural control mechanisms as walking. The present study examines and compares frontal plane balance responses with a research-based surface perturbation walking device to balance responses in a clinical standing balance assessment. We found that correlations between these walking and standing tests varied widely depending on the conditions compared. Correlations between standing and walking balance were highest when 1) a perturbation was present in walking tests, 2) subjects walked slowly, and 3) the standing tests were on foam as opposed to firm surface.

Keywords: Neural interfaces - Implantable systems, Neural stimulation, Neurological disorders
Abstract: Optogenetics has the potential to transform the study of organ functions in the peripheral nervous system via relatively easy access to the nerves and a direct link between the brain and organ systems. Implementation typically requires a static skeletal feature for the securement of a fiber. Unfortunately, the soft nature of peripheral nervous systems makes the wired fiber-optic approach less ideal for the study of the peripheral nervous system. Existing wireless approaches could bypass some constraints associated with optical fibers and thereby offer organ specificity. However, they suffer from durability loss due to considerable biological strains and unable to perform longitudinal experiments. Here, we propose a new class of wireless gastric optogenetic implant for identifying signaling pathways, in particular viscerosensory pathways, that can regulate food intake to treat obesity. Robust, wireless gastric optogenetic implants with a tubing-assisted U-shaped tether directly interface with nerve endings in the stomach with chronic stability in operation (> 100 kilocycles) and allows for optogenetic stimulations of vagus nerves in a freely behaving animal. We demonstrated utilities of the proposed wireless device in in vivo experiments. Results suggest the potential for identifying interventions for the treatment of obesity.

Keywords: Neurological disorders - Diagnostic and evaluation techniques, Human performance - Cognition, Neurological disorders - Stroke
Abstract: Cerebellar ataxias are a large family of movement disorders that generally follow a stroke. The clinical picture is very complicated and normal activities become difficult for ataxic patients. For instance, dynamic ataxia involves both walk and upper-limbs movement, thus affecting the possibility to fulfill daily life tasks. Rehabilitation treatments and strategies for cerebellar ataxia are nowadays controversial, since different opinions on the several approaches are spread among the clinical community. The purpose of the present work is not to shed some light on such disagreements. Indeed, here a solution for delivering rehabilitation activities in the form of videogame is presented. Data related to patient's performance are collected and analyzed in order to provide the clinical staff with objective indicators that properly describe the activity. Such information can also be used to discuss the effectiveness and the incidence of some strategy adopted for fulfilling some task. The experimental phase is conducted on two case-studies with regards to the upper-limb rehabilitation. The adoption of the strategy of weighting the limb when performing the movement is discussed. The indicators computed in both sessions with and without strategy are compared, also referring to the practice of some healthy subjects. The present work introduces the preliminary phase of a wider study and foretells its future development conducted on a larger population.

Keywords: Brain functional imaging - fMRI, Human performance - Cognition
Abstract: Subjective cognitive decline (SCD) is a preclinical stage before cognitive impairment, which has a high conversion risk into Alzheimer's disease. However, it is still unknown on the brain functional differences between SCD and healthy controls (HC) subjects. This study therefore proposed a complex brain network analysis based on graph theory. In this study, we selected functional magnetic resonance imaging (fMRI) scans from Xuanwu Hospital of Capital Medical University, including 27 SCD and 42 HC subjects. First, we constructed brain functional connectivity network to obtain brain network topology parameters, including clustering parameters, shortest path length, global efficiency, local efficiency, small world attributes, and modularity. Then, we compared differences on the parameters between two groups. As a result, both SCD and HC groups showed the characteristics of small world. As for the shortest path length, there was no significant difference between the SCD group and HC group. Both global efficiency and local efficiency of HC groups were higher than those of the SCD group. In addition, we found that the global modularity of the SCD group (6 modules) was higher than the HC group (7 modules). Our findings indicated that there were differences in brain functional networks between SCD and HC groups. Graph theory analysis may be useful and helpful to discriminate SCD and HC subjects.

Keywords: Human performance - Activities of daily living, Human performance - Cognition, Neuromuscular systems - Postural and balance
Abstract: Recently, emerging technologies are being used to solve state of the art problems in rehabilitation and physiotherapy. The increasing power of portable sensors is making a great choice for analysis of movements during daily activities. We previously developed a method to personalize the measure of balance only using kinematic data from Kinect. This paper presents the results of simultaneous quantification for the postural balance, motion classification and its quality with Synergy Probe. Previously, it was not possible to verify what happens when the motion balance is unstable. With motion quality index along with the stability, we can quantitatively evaluate the balance stability considering the motion class and its intensity during whole-body exercise.

Keywords: Neural interfaces - Implantable systems, Neural stimulation, Neuromuscular systems - EMG processing and applications
Abstract: Spinal cord stimulation (SCS) is a widely accepted effective treatment for managing chronic pain. SCS outcomes depend highly on accurate placement of SCS electrodes at the appropriate spine level for a desired pain relief. Intraoperative neurophysiological monitoring (IONM) under general anesthesia provides an objective real-time mapping of the dorsal columns, and has been shown to be a safe and effective tool. IONM applies stimulation to multiple electrode contacts at various intensities and monitors the triggered electromyography (EMG) responses in several muscle groups simultaneously. Therefore, it requires dynamic communication between neurosurgeon and neurophysiologist and continuous real-time annotations of the responses, which makes the procedure complex and experience-based. Here, we describe an automated data visualization tool that generates patient specific activity maps using intraoperatively collected signals. Responses were collected using a High-resolution (HR)-SCS lead with 8 columns of electrodes spanning the dorsal columns. Our JavaScript/Python based graphical user interface (GUI) provides a fast and robust visualization of EMG activity via denoising, feature extraction, normalization, and overlaying of the activity maps on body images in selected colormaps. In contrast to reviewing series of EMG signals, our user-friendly tool provides a rapid and robust analysis of stimulation effects on various muscle groups and direct comparison across subjects and/or stimulation settings. Future work includes expanding analytics capabilities and operating room implementation as a real-time processing tool that can be used in conjunction with the current IONM techniques.

Keywords: Human performance - Gait, Human performance - Activities of daily living, Neurological disorders - Stroke
Abstract: We developed a virtual reality (VR)-based gait training system, which could be used by inpatients to train their gait function in a simulated home environment, to reduce the risk of falling after discharge. The proposed system simulates the home environment on a head-mounted display, in which a user can walk around freely. The system provides visual feedback in the event of a collision with an indoor object such as a wall or furniture, prompting the user to modify his or her gait pattern. We first applied the system to healthy young adults and confirmed the usefulness of visual feedback in reducing the walking time and the number of collisions in the simulated room environment. Further, we applied the system to an inpatient with stroke and lower limb paralysis. The patient performed gait training based on a scenario of daily activity using the VR environment that mimicked his house. Five days of training significantly improved the gait and balance functions of the patient. These results suggest that the proposed system foster attention to the surrounding environment and improve gait function in both healthy participants and patients with stroke.

Keywords: Human performance - Fatigue, Human performance - Ergonomics and human factors, Human performance
Abstract: The “screening” trend of modern society has been a progressively increasing burden on human visual system, and visual fatigue problems are attracting growing attentions. Nowadays, subjective testing is the most widely used measure for visual fatigue; however, the low accuracy of subjective testing has been hindering its further improvement. Motivated by the idea of weighted scoring, this study investigated the effects of two weighted scales for measuring visual fatigue in screening tasks. Specifically, a questionnaire with 10 items collected from the classic scales was performed with an eye-tracking testing in two typical screen visual fatigue experiments, i.e., searching and watching. Then the subjective scores were factor-analyzed into three subscales before attempting linear regression analyses, which set the dependents to two previously validated eye-tracking parameters, i.e., fixation frequency and saccade amplitude. Finally two weighted scales were obtained in assessing visual fatigue of varying levels, which demonstrated the potential to improve testing accuracy of visual fatigue with the calibration of objective measurement.

Keywords: Neurological disorders, Brain functional imaging
Abstract: Alzheimer’s disease (AD) is the most prevalent neurodegenerative disorder and considerably determined by genetic factors. Fluorodeoxyglucose positron emission tomography (FDG-PET) can reflect the functional state of glucose metabolism in the brain, and radiomic features of FDG-PET were considered as important imaging markers in AD. However, radiomic features are not highly interpretable, especially lack of explanation of underlying biological and molecular mechanisms. Therefore, this study used radiogenomics analysis to explore prognostic metabolic imaging markers by associating radiomics features and genetic data. In the study, we used the FDG-PET images and genotype data of 389 subjects (Cohort B） enrolled in the ADNI, including 109 AD, 134 healthy controls (HCs), 72 MCI non-converters (MCI-nc) and 74 MCI converters (MCI-c). Firstly, we performed a Genome-wide association study (GWAS) on the genotype data of 998 subjects (Cohort A), including 632 AD and 366 HCs after quality control (QC) steps to identify susceptibility loci as the gene features. Secondly, radiomics features were extracted from the preprocessed PET images. Thirdly, two-sample t-test, rank sum test and F-score were regarded as the feature selection step to select effective radiomic features. Fourthly, a support vector machine (SVM) was used to test the ability of the radiomic features to classify HCs, MCI and AD patients. Finally, we performed the Spearman correlation analysis on the genetic data and radiomic features. As a result, we identified rs429358 and rs2075650 as genome-wide significant signals. The radiomic approach achieved good classification abilities. Two prognostic FDG-PET radiomic features in the amygdala were proven to be correlated with the genetic data.

Keywords: Brain physiology and modeling - Cognition, memory, perception, Brain functional imaging - fMRI, Human performance - Modelling and prediction
Abstract: Brain decoding is able to make human interact with an external machine or robot for assisting patient’s rehabilitation. Brain generic object recognition ability can be decoded through multiple neuroimaging modalities like functional magnetic resonance imaging (fMRI). On the other hand, external machine may wrongly recognize objects due to distorted noisy or blurring images caused by many factors, and therefore deteriorate performance of brain-machine interaction. In order to create better machine, generalization capability of human brain is transferred to classifier for enhancing classification accuracy of distorted images. Since homology existing between human and machine vision has been demonstrated, through decoding neural activity features of fMRI signals into feature units of convolutional neural network layers, an enhanced object recognition method is proposed to integrate brain activity into classifier for increasing classification accuracy. Experimental results show that the proposed method is able to enhance generalization capability of distorted object recognition.

Keywords: Neural interfaces - Implantable systems, Brain-computer/machine interface, Smart neural implants
Abstract: This paper presents an ultra-low power mixed-signal neural data acquisition (MSN-DAQ) system that enables a novel low-power hybrid-domain neural decoding architecture for implantable brain-machine interfaces with high channel count. Implemented in 180nm CMOS technology, the 32-channel custom chip operates at 1V supply voltage and achieves excellent performance including 1.07μW/channel, 2.37/5.62 NEF/PEF and 88dB common-mode rejection ratio (CMRR) with significant back-end power-saving advantage compared to prior works. The fabricated prototype was further evaluated with in vivo human tests at bedside, and its performance closely follows that of a commercial recording system.

Keywords: Neural stimulation, Neurorehabilitation, Neural interfaces - Tissue-electrode interface
Abstract: Transcorneal electrical stimulation (TES) is a non-invasive approach for activating the retina and its downstream components through the application of electric current on the cornea. Although previous studies have demonstrated the clinical relevance of TES for modulating neurons with improvements in visual evoked potentials (VEPs) and electroretinograms (ERGs), there are still huge gaps in knowledge of its effect on the brain structures. To determine the short-term impact as well as the after-effects of TES on neural oscillatory power in retinal degeneration mice, we performed electrocorticography (ECoG) recording in the prefrontal and primary visual cortices at different stages of prolonged TES [transient stage, following prolonged stimulation (post-stimulation stage 1) and long after the end of the retinal stimulation (post-stimulation stage 2)]) under varying stimulation current amplitudes (400 µA, 500 µA and 600 µA). The results revealed asymmetric differences between short-term and long- pTES under different stimulation current amplitudes. Specifically, in post-stimulation stage 1 we observed significant increase in ECoG power of theta, alpha and beta oscillations respectively compared with baseline pre-stimulation results. These effects were dependent on the stimulation current amplitude and stimulation stage. Transient TES was not sufficient to cause significant changes in the ECoG power of all accessed oscillations except in medium, high and ultra-gamma oscillations which significantly decreased in 400 µA and 500 µA stimulation groups respectively compared with pre-stimulation results. Regarding long-term stimulation, the increase in ECoG power of theta, alpha and beta oscillations observed in post-stimulation stage 1 was significantly maintained in post-stimulation stage 2.

Keywords: Neurological disorders, Human performance - Cognition
Abstract: Alzheimer’s disease (AD) is the most prevalent neurodegenerative disorder and the most common form of dementia in the elderly. Because gene is an important clinical risk factor resulting in AD, genomic studies, such as genome-wide association studies (GWAS), have widely been applied into AD studies. However, main shortcomings of GWAS method were that hereditary deletions were evident in the GWAS studies, which resulted in low classification or prediction abilities by using GWAS analysis. Therefore, this paper proposed a novel deep learning genomics approach and applied it to discriminate AD patients and healthy control (HC) subjects. In this study, we selected genotype data of 988 subjects enrolled in the ADNI, including 622 AD patients and 366 HC subjects. The proposed deep learning genomics (DLG) approach was composed of three steps: quality control, SNP genotype coding, and classification. The Resnet framework was used as the DLG model in this study. In the comparative GWAS analysis, APOE ε4 status and the normalized theta-value of the significant SNP loci were seen as predictors to classify genetically using Support Vector Machine (SVM). All data were divided into one training & validation group and one test group. 5-fold cross-validation was used in 500 times. Finally, we compared the classification results between DLG model and traditional GWAS analysis. As a result, the accuracy, sensitivity, and specificity of classification for traditional GWAS analysis was 71.38%±0.63%, 63.13%±2.87% and 85.59%±6.66% in the test group; while the accuracy, sensitivity, and specificity of classification for DLG model was 92.65%±4.80%, 85.00%±16.25% and 97.10%±4.38% in the test group. Hence, the DLG model can achieve higher accuracy and sensitivity when applied to AD. More importantly, we discovered several novel genetic biomarkers of AD, including rs6311 and rs6313 in HTR2A, and rs690705 in RFC3. The roles of these novel loci in AD should be explored future.

Keywords: Human performance - Attention and vigilance, Brain-computer/machine interface, Neural signals - Machine learning & Classification
Abstract: Attention, a multi-faceted cognitive process, is essential in our daily lives. We can measure visual attention using an EEG Brain-Computer Interface for detecting different levels of attention in gaming, performance training, and clinical applications. In attention calibration, we use Flanker task to capture EEG data for attentive class. For EEG data belonging to inattentive class calibration, we instruct subject not focusing on a specific position on screen. We then classify attention levels using binary classifier trained with these surrogate ground-truth classes. However, subjects may not be in desirable attention conditions when performing repetitive boring activities over a long experiment duration. We propose attention calibration protocols in this paper that use simultaneous visual search with an audio directional change paradigm and static white noise as ‘attentive’ and ‘inattentive’ conditions, respectively. To compare the performance of proposed calibrations against baselines, we collected data from sixteen healthy subjects. For a fair comparison of classification performance; we used six basic EEG band-power features with a standard binary classifier. With the new calibration protocol, we achieved 74.37±6.56% mean subject accuracy, which is about 3.73±2.49% higher than the baseline, but there were no statistically significant differences. According to post-experiment survey results, new calibrations are more effective in inducing desired perceived attention levels. We will improve calibration protocols with reliable attention classifier modeling to enable better attention recognition based on these promising results.

Keywords: Brain-computer/machine interface, Neural signals - Information theory, Neural signals - Machine learning & Classification
Abstract: Stimulus-driven brain-computer interfaces (BCIs), such as the P300 speller, rely on using sensory stimuli to elicit specific neural signal components called event-related potentials (ERPs) to control external devices. However, psychophysical factors, such as refractory effects and adjacency distractions, may negatively impact ERP elicitation and BCI performance. Although conventional BCI stimulus presentation paradigms usually design stimulus presentation schedules in a pseudo-random manner, recent studies have shown that controlling the stimulus selection process can enhance ERP elicitation. In prior work, we developed an algorithm to adaptively select BCI stimuli using an objective criterion that maximizes the amount of information about the user's intent that can be elicited with the presented stimuli given current data conditions. Here, we enhance this adaptive BCI stimulus selection algorithm to mitigate adjacency distractions and refractory effects by modeling temporal dependencies of ERP elicitation in the objective function and imposing spatial restrictions in the stimulus search space. Results from simulations using synthetic data and human data from a BCI study show that the enhanced adaptive stimulus selection algorithm can improve spelling speeds relative to conventional BCI stimulus presentation paradigms.

Clinical relevance— Increased communication rates with our enhanced adaptive stimulus selection algorithm can potentially facilitate the translation of BCIs as viable communication alternatives for individuals with severe neuromuscular limitations.

Keywords: Brain-computer/machine interface, Neural signals - Machine learning & Classification, Brain functional imaging - EEG
Abstract: Recently, transfer learning and deep learning have been introduced to solve intra- and inter-subject variability problems in Brain-Computer Interfaces. However, the generalization ability of these BCIs is still to be further verified in a cross-dataset scenario. This study compared the transfer performance of manifold embedded knowledge transfer and pre-trained EEGNet with three preprocessing strategies. This study also introduced AdaBN for target domain adaptation. The results showed that EEGNet with Riemannian alignment and AdaBN could achieve the best transfer accuracy about 65.6% on the target dataset. This study may provide new insights into the design of transfer neural networks for BCIs by separating source and target batch normalization layers in the domain adaptation process.

Keywords: Brain-computer/machine interface, Neural signal processing, Sensory neuroprostheses - Auditory
Abstract: Auditory attention detection (AAD) seeks to detect the attended speech from EEG signals in a multi-talker scenario, i.e. cocktail party. As the EEG channels reflect the activities of different brain areas, a task-oriented channel selection technique improves the performance of brain-computer interface applications. In this study, we propose a soft channel attention mechanism, instead of hard channel selection, that derives an EEG channel mask by optimizing the auditory attention detection task. The neural AAD system consists of a neural channel attention mechanism and a convolutional neural network (CNN) classifier. We evaluate the proposed framework on a publicly available database. We achieve 88.3% and 77.2% for 2-second and 0.1-second decision windows with 64-channel EEG; and 86.1% and 83.9% for 2-second decision windows with 32-channel and 16-channel EEG, respectively. The proposed framework outperforms other competitive models by a large margin across all test cases.

Keywords: Brain-computer/machine interface, Neural signals - Machine learning & Classification, Neural signal processing
Abstract: The commonly used fixed discrete Kalman filters (DKF) in neural decoders do not generalize well to the actual relationship between neuronal firing rates and movement intention. This is due to the underlying assumption that the neural activity is linearly related to the output state. They also face the issues of requiring large amount of training datasets to achieve a robust model and a degradation of decoding performance over time. In this paper, an adaptive adjustment is made to the conventional unscented Kalman filter (UKF) via intention estimation. This is done by incorporating a history of newly collected state parameters to develop a new set of model parameters. At each time point, a comparative weighted sum of old and new model parameters using matrix squared sums is used to update the neural decoding model parameters. The effectiveness of the resulting adaptive unscented Kalman filter (AUKF) is compared against the discrete Kalman filter and unscented Kalman filter-based algorithms. The results show that the proposed new algorithm provides higher decoding accuracy and stability while requiring less training data.

Keywords: Brain-computer/machine interface, Sensory neuroprostheses - Auditory, Neural signals - Machine learning & Classification
Abstract: Detecting auditory attention based on brain signals enables many everyday applications, and serves as part of the solution to the cocktail party effect in speech processing. Several studies leverage the correlation between brain signals and auditory stimuli to detect the auditory attention of listeners. Recently, studies show that the alpha band (8-13 Hz) EEG signals enable the localization of auditory stimuli. We believe that it is possible to detect auditory spatial attention without the need of auditory stimuli as references. In this work, we firstly propose a spectro-spatial feature extraction technique to detect auditory spatial attention (left/right) based on the topographic specificity of alpha power. Experiments show that the proposed neural approach achieves 81.7% and 94.6% accuracy for 1-second and 10-second decision windows, respectively. Our comparative results show that this neural approach outperforms other competitive models by a large margin in all test cases.

Keywords: Neurological disorders - Epilepsy, Neural stimulation, Neurological disorders - Treatment methodologies
Abstract: Vagal Nerve Stimulation (VNS) is used to treat patients with pharmacoresistant epilepsy. However, generally accepted tools to predict VNS response do not exist. Here we examined two heart activity measures – mean RR and pNN50 and their complex behavior during activation in pre-implant measurements. The ECG recordings of 73 patients (38 responders, 36 non-responders) were examined in a 30-sec floating window before (120 sec), during (2x120 sec), and after (120 sec) the hyperventilation by nose and mouth. The VNS response differentiation by pNN50 was significant (min p=0.01) in the hyperventilation by a nose with a noticeable descendant trend in nominal values. The mean RR was significant (p=0.01) in the rest after the hyperventilation by mouth but after an approximately 40-sec delay. Clinical Relevance: Our study shows that pNN50 and mean RR can be used to distinguish between VNS responders and non-responders. However, details of dynamic behavior showed how this ability varies in tested measurement segments.

Keywords: Human performance - Ergonomics and human factors, Human performance - Fatigue, Brain-computer/machine interface
Abstract: Estimation of human attentional states using an electroencephalogram (EEG) has been demonstrated to help prevent human errors associated with the degradation. Since the use of the lambda response --one of eye-fixation-related potentials time-locked to the saccade offset-- enables such estimation without external triggers, the measurements are compatible for an application in a real-world environment. With aiming to apply the lambda response as an index of human errors during the visual inspection, the current research elucidated whether the mean amplitude of the lambda response was a predictor of the number of inspection errors. EEGs were measured from 50 participants while inspecting the differences between two images of the circuit board. Twenty percent of the total number of image pairs included differences. The lambda response was obtained relative to a saccade offset starting a fixation of the inspection image. Participants conducted four sessions over two days (625 trials/ session, 2 sessions/ day). A Poisson regression of the number of inspection errors using a generalized linear mixed model showed that a coefficient of the mean amplitude of the lambda response was significant (β^= 0.24, p < 0.01), suggesting that the response has a role in the prediction of the number of human error occurrences in the visual inspection.

Keywords: Human performance - Gait, Human performance - Ergonomics and human factors
Abstract: Introduction: The gait while using an intravenous (IV) pole is close to the gait of the elderly and fallers. Additionally, one survey has reported that the diagonal position is optimal for transporting an IV pole with a light load. However, in clinical practice, carrying a heavier load may be possible. Therefore, this study clarifies the optimum operation position using an IV pole with a weight closer to that in actual clinical practice. Method: Using image analysis software, we investigated several variables indicating gait, such as stride length. Participants walk with an IV pole in three ways: sideways, in front, and diagonally. We investigated two types of IV pole loads, which are 0.5 kg and 5.0 kg. Results and Discussion: In 0.5-kg settings, the sideways position is a way to suppress the narrowing of the heel–floor angle. No significant difference in the subjective appraisals was observed between the sideways and diagonal positions. In addition, the sideways position is as optimum as the diagonal position. In 5.0-kg settings, only the sideways position suppressed the narrowing of the step length. Therefore, the sideways position is optimal. However, the participants’ impressions suggested that arm strength is required for the sideways position. If a patient has weak arms and cannot maintain the sideways position, the patient may choose the diagonal position. Moreover, the front position is the way to hold the trunk most forward. However, there is a possibility that it is easy for a specific person, such as a rollator user, to choose. Therefore, further investigate of the optimum operation position depending on the walking abilities is needed. Conclusion: It was suggested that the sideways position is optimal for walking with an IV pole when transporting with a total load of approximately 5.0 kg.

Keywords: Neurorehabilitation, Brain functional imaging - EEG, Neural signal processing
Abstract: Post-stroke neuronal plasticity was always viewed as a localized gain-of-functionality. The reorganization of neurons neighboring the lesioned brain tissues is able to compensate for the function of damaged neurons. However, it was also proposed that distant interconnected brain regions could be affected by stroke. Changes in functional connections across the brain were found associated with motor deficiency and recovery. Parietal-frontocentral functional connectivity was found related to the performance of motor imagery. This study aims to evaluate the EEG-based parietal-frontocentral functional connectivity in post-stroke patients, and to investigate the immediate effect of rehabilitation training toward these connections. Pairwise functional connectivity was extracted from healthy subjects and post-stroke patients during standing and walking. Significant reductions in P3-FC4 and P3-C4 connectivity strengths were found in post-stroke patients during both standing and walking conditions. Immediate improvement in the reduced connections was observed with the intervention of a previously proposed, motivation-based rehabilitation system, which was known as the mixed‐reality music rehabilitation (MR2) system. This indicates the relationship between left parietal functional connectivity and stroke-related motor performance. These findings suggest the feasibility to evaluate the immediate plasticity of functional connectivity during post-stroke rehabilitation.

Keywords: Neurological disorders - Epilepsy, Brain functional imaging - EEG, Neural signals - Machine learning & Classification
Abstract: In this paper, we propose a time-series stochastic model based on a scale mixture distribution with Markov transitions to detect epileptic seizures in electroencephalography (EEG). In the proposed model, an EEG signal at each time point is assumed to be a random variable following a Gaussian distribution. The covariance matrix of the Gaussian distribution is weighted with a latent scale parameter, which is also a random variable, resulting in the stochastic fluctuations of covariances. By introducing a latent state variable with a Markov chain in the background of this stochastic relationship, time-series changes in the distribution of latent scale parameters can be represented according to the state of epileptic seizures. In an experiment, we evaluated the performance of the proposed model for seizure detection using EEGs with multiple frequency bands decomposed from a clinical dataset. The results demonstrated that the proposed model can detect seizures with high sensitivity and outperformed several baselines.

Keywords: Brain-computer/machine interface, Neural signal processing, Neural signals - Machine learning & Classification
Abstract: Motor imagery-based brain computer interface (MI-BCI) is a representative active BCI paradigm which is widely employed in the rehabilitation field. In MI-BCI, a classification model is built to identify the target limb from MI-based EEG signals, but the performance of models cannot meet the demand for practical use. Lightweight neural networks in deep learning methods are used to build high performance models in MI-BCI. Small sample sizes and the lack of multi-scale information extraction in frequency domain limit the performance improvement of lightweight neural networks. To solve these problems, the Filter Bank Sinc-ShallowNet (FB-Sinc-ShallowNet) algorithm combined with the mixed noise adding method based on empirical mode decomposition (EMD) was proposed. The FB-Sinc-ShallowNet algorithm improves a lightweight neural network Sinc-ShallowNet with a filter bank structure corresponding to four sensory motor rhythms. The mixed noise adding method employs the EMD method to improve the quality of generated data. The proposed method was evaluated on the BCI competition IV IIa dataset and can achieve highest average accuracy of 77.2%, about 6.34% higher than state-of-the-art method Sinc-ShallowNet. This work implies the effectiveness of filter bank structure in lightweight neural networks and provides a novel option for data augmentation and classification of MI-based EEG signals, which can be applied in the rehabilitation field for decoding MI-EEG with few samples.

Keywords: Brain physiology and modeling - Neural dynamics and computation, Brain physiology and modeling - Neural circuits, Neurological disorders - Diagnostic and evaluation techniques
Abstract: Parkinson’s disease (PD) is a neuropathy characterized by motor disorders, but it has also been associated with the presence of autonomic alterations as a result of degradation of the dopaminergic system. Studying the relation between Band Power time series (BPts) and Heart Rate Variability (HRV), has been proposed as a tool to explore the bidirectional communication pathways between cortex and autonomic control. This work presents a primer analysis on study brain ↔ heart interaction on a databse of PD patients under two conditions: without and after levadopa (L-dopa) intake. Additionally a healthy control population was also analyzed, and used as comparison level between both conditions. Results show PD affects pathways by reducing the number of connections, specially association of beta and power and the second faster component of HRV seems to be more sensitive to L-dopa administration.

Keywords: Neural signals - Machine learning & Classification, Neurological disorders - Diagnostic and evaluation techniques, Brain physiology and modeling - Neural dynamics and computation
Abstract: Parkinson’s Disease is a neuropathy that produces changes in several biomarkers, these changes could be used to evaluate even sub–clinical conditions. This paper presents an evaluation of indices extracted from electroencephalography and Heart Rate Variability (HRV), when used to classify a sample of subjects from three groups: control (healthy), medicated and non medicated subjects diagnosed with Parkinson’s Disease. Classification performance was measured using accuracy over these classes and a cross validation scheme was used to assess repeatability for the classification process. Results tend to prove that inclusion of an autonomic index derived from HRV analysis enhances classification, suggesting that Parkinson’s Disease could be related with unperceptible to mild alterations of the Autonomic Nervous System.

Keywords: Brain-computer/machine interface, Neural signals - Machine learning & Classification, Motor neuroprostheses
Abstract: Abstract— Brain-Computer Interface systems can contribute to a vast set of applications such as overcoming physical disabilities in people with neural injuries or hands-free control of devices in healthy individuals. However, having systems that can accurately interpret intention online remains a challenge in this field. Robust and data-efficient decoding---despite the dynamical nature of cortical activity and causality requirements for physical function---is among the most important challenges that limit the widespread use of these devices for real-world applications. Here, we present a causal, data-efficient neural decoding pipeline that predicts intention by first classifying recordings in short sliding windows. Next, it performs weighted voting over initial predictions up to the current point in time to report a refined final prediction. We demonstrate its utility by classifying spiking neural activity collected from the human posterior parietal cortex for a cue, delay, imaginary motor task. This pipeline provides higher classification accuracy than state-of-the-art time windowed spiking activity based causal methods, and is robust to the choice of hyper-parameters.

Clinical relevance— We have tested our decoder during delayed imaginary grasp tasks on data from the human posterior parietal cortex---a relatively understudied region of the brain thought to contribute to motor intention. Our results provide new insight into the underlying neural dynamics of this region. In fact, the most discriminating information---and the greatest utility of voting---appear to occur during the early phases of the task. This makes our approach most useful to short-latency control of brain-computer systems such as neuroprosthetics.

Keywords: Neuromuscular systems - Computational modeling, Neural interfaces - Neuromorphic engineering, Motor neuroprostheses - Prostheses
Abstract: Model-based biomimetic control with neuro-muscular reflex requires accurate representation of muscle fascicle length, which affects both force generation capability of muscle and dynamics of muscle spindle. However, physiological data are insufficient to guide the selection of range of fascicle length for task control. Here a reverse engineering approach was used to investigate the effects of different fascicle length range on controller’s force control ability, so as to justify the selection of operating range of muscle length for a grasp force task. We compared 3 different ranges of fascicle length for their effects on force generation, i.e. R1: 0.5 – 1.0 L0, R2: 0.5 – 1.3 L0 and R3: 0.5 – 1.6 L0. The rationale to test these range selections was based on both physiological realism and engineering considerations. The steady state force output and transient force responses were evaluated with a range of step inputs as controller input. Results show that the prosthetic finger can produce a linear steady state force response with all 3 ranges of fascicle length. Peak force was the largest with R3. Fascicle length range had no significant effect on the rise time in force generation tasks. Results suggest that a wider range of fascicle length may be more favorable for force capacity, since the contact point of force control may well fall near the optimal length (Lo) region.

Keywords: Brain functional imaging - Spatial-temporal dynamics
Abstract: Calcium imaging has great potential to be applied to online brain-machine interfaces (BMIs). As opposed to two-photon imaging settings, a one-photon microendoscopic imaging device can be chronically implanted and is subject to little motion artifacts. Traditionally, one-photon microendoscopic calcium imaging data are processed using the constrained nonnegative matrix factorization (CNMFe) algorithm, but this batched processing algorithm cannot be applied in real-time. An online analysis of calcium imaging data algorithm (or OnACIDe) has been proposed, but OnACIDe updates the neural components by repeatedly performing neuron identification frame-by-frame, which may decelerate the update speed if applying to online BMIs. For BMI applications, the ability to track a stable population of neurons in real-time has a higher priority over accurately identifying all the neurons in the field of view. By leveraging the fact that 1) microendoscopic recordings are rather stable with little motion artifacts and 2) the number of neurons identified in a short training period is sufficient for potential online BMI tasks such as cursor movements, we proposed the short-training CNMFe algorithm (stCNMFe) that skips motion correction and neuron identification processes to enable a more efficient BMI training program in a one-photon microendoscopic setting.

Keywords: Brain-computer/machine interface, Human performance - Activities of daily living, Neurological disorders
Abstract: Children with severe physical disabilities are often unable to independently explore their environments, further contributing to complex developmental delays. Brain-computer interfaces (BCIs) could be a novel access method to power mobility for children who struggle to use existing alternate access technologies, allowing them to reap the developmental, social, and psychological benefits of exploring their environment. In this study we demonstrate that children with quadriplegic cerebral palsy can use a simple BCI system to explore movement with a power mobility device. Four children were able to use the BCI to drive forward at least 7m, although more practice is needed to achieve more efficient driving skills through sustained BCI activations. Clinical relevance – This paper highlights the potential of a novel access technology to achieve patient-centered goals in power mobility for children with severe physical impairments who are otherwise neglected as candidates for powered wheelchairs. This paper also demonstrates how a power mobility device can be adapted to be operated by a readily available commercial-grade BCI system.

Keywords: Human performance - Cognition, Brain physiology and modeling - Cognition, memory, perception, Sensory neuroprostheses
Abstract: Sensory substitution devices (SSDs) such as the ‘vOICe’ preserve visual information in sound by turning visual height, brightness, and laterality into auditory pitch, volume, and panning/time respectively. However, users have difficulty identifying or tracking multiple simultaneously presented tones – a skill necessary to discriminate the upper and lower edges of object shapes. We explore how these deficits can be addressed by using image-sonifications inspired by auditory scene analysis (ASA). Here, sighted subjects (N=25) of varying musical experience listened to, and then reconstructed, complex shapes consisting of simultaneously presented upper and lower lines. Complex shapes were sonified using the vOICe, with either the upper and lower lines varying only in pitch (i.e. the vOICe’s ‘unaltered’ default settings), or with one line degraded to alter its auditory timbre or volume. Results indicate that overall performance increased with subjects’ years of prior musical experience. ANOVAs revealed that both sonification style and musical experience significantly affected performance, but with no interaction effect between them. Compared to the vOICe’s ‘unaltered’ pitch-height mapping, subjects had significantly better image-reconstruction abilities when the lower line was altered via timbre or volume-modulation. By contrast, altering the upper line only helped users identify the unaltered lower line. In conclusion, adding ASA principles to vision-to-audio SSDs boosts subjects’ image-reconstruction abilities, even if this also reduces total task-relevant information. Future SSDs should seek to exploit these findings to enhance both novice user abilities and the use of SSDs as visual rehabilitation tools.

Keywords: Neural interfaces - Biomaterials, Neural interfaces - Microelectrode technology, Neural interfaces - Implantable systems
Abstract: Soft, flexible polymer-based bioelectronics are a promising approach to minimize the chronic inflammatory reactions associated with metallic devices, impairing long-term device reliability and functionality. This work demonstrates the fabrication of conductive elastomers (CEs) consisting of chemically synthesized poly(3,4-ethylenedioxythiophene) (PEDOT) nanowires embedded within a polyurethane (PU) elastomeric matrix, resulting in soft and flexible, fully polymeric electrode materials. Increasing PEDOT nanowire loadings resulted in an improvement in electrochemical properties and conductivity, an increased Young’s modulus and reduced strain at failure. Nanowire CEs were also found to have significantly improved electrochemical performance compared to one of the standard electrode materials, platinum (Pt). Indirect in vitro cytocompatibility test was carried out to investigate the effect of leachable substances from the CE on primary rodent cells. Nanowire CEs provide a promising alternative to metals for the fabrication of soft bioelectronics.

Keywords: Brain-computer/machine interface, Brain physiology and modeling - Sensory-motor, Human performance - Sensory-motor
Abstract: In the recent years, brain computer interfaces (BCI) using motor imagery have shown some limitations regarding the quality of control. In an effort to improve this promising technology, some studies intended to develop hybrid BCI with other technologies such as eye tracking which shows more reliability. However, the use of an eye tracker in the control of a robot might affect by itself the sense of agency (SoA) and the brain activity in the regions used for motor imagery (MI). Here, we explore the link between the sense of agency and the activity of the motor cortex. For this purpose, we used of a virtual arm projected on a surface which is either controlled by motion capture or controlled by gaze using an eye tracker. We found out that there is an activity in the motor cortex during the task of control by gaze and that having control over a projected robotic arm presents significant differences.

Keywords: Human performance - Sensory-motor, Human performance, Human performance - Driving
Abstract: Vision seems essential for cross-modal calibration of auditory cues in spatial perception. Previous findings showed that, in some specific tasks such as sound localization, blind individuals have enhanced skills, suggesting that the audiomotor loop might partially compensate for early visual loss in the calibration of auditory space; however, direct evidence is still lacking. Here, we proposed a method based on the alteration of the audiomotor loop. Acoustic virtual reality was used to measure the audiomotor loop's influence on the space perception of blind individuals. We developed a VR steering task by head or trunk pointing to auditory sources, where the audiomotor conflict is induced by letting trunk rotations change the auditory scene together with head rotations. Early blind, late blind, and sighted participants were tested to assess their sensitivity to the induced audiomotor conflict. The platform demonstrated its effectiveness in exposing participants' sensitivity to the audiomotor loop alteration. The early blind group was significantly more affected than the sighted group, while the late blind group did not significantly differ from any of the other groups. Our results confirm the increased role of the audiomotor loop for audiospatial information processing in blindness and advocate for the development of new spatial orientation training for blind people based on exploiting the audiomotor loop itself.

Keywords: Human performance, Brain physiology and modeling - Cognition, memory, perception, Neurorehabilitation
Abstract: With this work, we introduce a novel Android app designed to monitor and enhance auditory and tactile temporal sensitivity. To assess the app’s reliability, we tested its technical performance evaluating stimuli production’s accuracy (i.e., onset, offset, and duration of stimulation). To validate the app with participants we generated temporal intervals, using either sounds or vibratory stimuli, by implementing two versions of a Two-Alternative Forced-Choice (2AFC) task. Auditory and tactile temporal sensitivity of 12 participants was evaluated using this procedure. To investigate whether temporal abilities could be enhanced using the app, participants were then divided into two groups: one group was trained for four days on the auditory temporal task, while the other was trained for four days on the tactile temporal task. Results suggest that the app can i) effectively measure auditory and tactile temporal thresholds and ii) be used to enhance temporal abilities through perceptual learning. The accessibility of the experimental protocols, combined with our findings, fosters the app’s involvement in rehabilitation programs, for example, with a specific focus on sensory disabilities that are associated with temporal deficits (e.g., deafness and Parkinson).

Keywords: Neural stimulation - Deep brain
Abstract: Deep brain stimulation (DBS) promises to treat an increasing number of neurological and psychiatric disorders. DBS outcome is directly a factor of optimal targeting of the relevant brain structures. Computational models can help to interpret a patient's outcome by predicting the volume of tissue activated (VTA) around DBS electrode contacts. Here we report results of a preliminary study of DBS in two patients with obsessive-compulsive disorder and show that VTA predictions, which are based on patient-specific volume conductor models, correlate with clinical outcome. Our results suggest that patient-specific VTA calculation can help inform device programing to maximize therapeutic effects and minimize side effects.

Keywords: Brain-computer/machine interface, Neural signals - Machine learning & Classification
Abstract: P300 Speller is a brain-computer interface (BCI) speller system, used for enabling human with different paralyzing disorders, such as amyotrophic lateral sclerosis (ALS), to communicate with the outer world by processing electroencephalography (EEG) signals. Different people have different latency and amplitude of the P300 event-related potential (ERP) component, which is used as the main feature for detecting the target character. In order to achieve robust results for different subjects using generic training (GT), the ensemble learning classifiers are proposed based on linear discriminant analysis (LDA), support vector machine (SVM), k-nearest neighbors (kNN), and convolutional neural network (CNN). The proposed models are trained using data from healthy subjects and tested on both healthy subjects and ALS patients. The results show that the fusion of LDA, kNN and SVM provides the most accurate results, achieving the accuracy of 99% for healthy subjects and about 85% for ALS patients.

Keywords: Neurorehabilitation, Motor neuroprostheses - Neuromuscular stimulation, Neurological disorders - Treatment methodologies
Abstract: Neuromuscular electrical stimulation (NMES) targeting the muscle belly is commonly used to restore muscle strength in individuals with neurological disorders. However, early onset of muscle fatigue is a major limiting factor. Transcutaneous nerve stimulation (TNS) can delay muscle fatigue compared with traditional NMES techniques. However, the recruitment of Ia afferent fibers has not be specifically targeted to maximize muscle activation through the reflex pathway, which can lead to more orderly recruitment of motor units, further delaying fatigue. This preliminary study assessed the distribution of M-wave and H-reflex of intrinsic and extrinsic finger muscles. TNS was delivered using an electrode array placed along the medial side of the upper arm. Selective electrode pairs targeted the median and ulnar nerves innervating the finger flexors. High-density electromyography (HD EMG) was utilized to quantify the spatial distribution of the elicited activation of finger intrinsic and extrinsic muscles along the hand and forearm. The spatial patterns were characterized through isolation of the M-wave and H-reflex across various stimulation levels and EMG channels. Our preliminary results showed that, by altering the stimulation amplitude, distinct M-wave and H-reflex responses were evoked across EMG channels. In addition, distinct stimulation locations appeared to result in varied levels of reflex recruitment. Our findings indicate that it is possible to adjust stimulation parameters to maximize reflex activation, which can potentially facilitate physiological recruitment order of motoneurons.

Keywords: Neurological disorders - Diagnostic and evaluation techniques, Brain functional imaging - EEG, Neural signal processing
Abstract: In the human visual system, cerebral cortex combines left- and right-eye retinal inputs, enabling single, comfortable binocular vision. In visual cortex, the signals from each eye inhibit one another (interocular suppression). While this mechanism may be disrupted by e.g. traumatic brain injury, clinical assessments of interocular suppression are subjective, qualitative, and lack reliability. EEG is a potentially useful clinical tool for objective, quantitative assessment of binocular vision. In a cohort of normal participants, we measured occipital, visual evoked potentials (VEPs) in response to dichoptically-presented vertical and/or horizontal sine-wave gratings. Response amplitudes to orthogonal gratings were greater than that of parallel gratings, which were in turn greater than that of monocular gratings. Our results indicate that interocular suppression is (normally) balanced, orientation-tuned, and that suppression per se is reduced for orthogonal gratings. This objective measure of suppression may have application in clinical settings.

Keywords: Neural signals - Machine learning & Classification, Brain functional imaging - EEG, Brain-computer/machine interface
Abstract: The error-related potential (ErrP) is an event-related potential (ERP) evoked by an experimental participant's recognition of an error during task performance. ErrPs, originally described by cognitive psychologists, have been adopted for use in brain-computer interfaces (BCIs) for the detection and correction of errors, and the online refinement of decoding algorithms. Riemannian geometry-based feature extraction and classification is a new approach to BCI which shows good performance in a range of experimental paradigms, but has yet to be applied to the classification of ErrPs. Here, we describe an experiment that elicited ErrPs in seven normal participants performing a visual discrimination task. Audio feedback was provided on each trial. We used multi-channel electroencephalogram (EEG) recordings to classify ErrPs (success/failure), comparing a Riemannian geometry-based method to a traditional approach that computes time-point features. Overall, the Riemannian approach outperformed the traditional approach (78.2% versus 75.9% accuracy, p < 0.05); this difference was statistically significant (p < 0.05) in three of seven participants. These results indicate that the Riemannian approach better captured the features from feedback-elicited ErrPs, and may have application in BCI for error detection and correction.

Keywords: Neural signals - Machine learning & Classification, Brain functional imaging, Brain-computer/machine interface
Abstract: Riemannian tangent space methods offer state-of-the-art performance in magnetoencephalography (MEG) and electroencephalography (EEG) based applications such as brain-computer interfaces and biomarker development. One limitation, particularly relevant for biomarker development, is limited model interpretability compared to established component-based methods. Here, we propose a method to transform the parameters of linear tangent space models into interpretable patterns. Using typical assumptions, we show that this approach identifies the true patterns of latent sources, encoding a target signal. In simulations and two real MEG and EEG datasets, we demonstrate the validity of the proposed approach and investigate its behavior when the model assumptions are violated. Our results confirm that Riemannian tangent space methods are robust to differences in the source patterns across observations. We found that this robustness property also transfers to the associated patterns.

Keywords: Neural signal processing
Abstract: Measuring electrical potentials in the extracellular space of the brain is a popular technique because it can detect action potentials from putative individual neurons. Electrophysiology is undergoing a transformation where the number of recording channels, and thus number of neurons detected, is growing at a dramatic rate. This rapid scaling is paving the way for both new discoveries and commercial applications; however, as the number of channels increases there will be an increasing need to make these systems more power efficient. One area ripe for optimization are the signal acquisition specifications needed to detect and sort action potentials (i.e., “spikes”) to putative single neuron sources. In this work, we take existing recordings collected using Intan hardware and modify them in a way that corresponds to reduced recording performance. The accuracy of these degraded recordings to spike sort using MountainSort4 is evaluated by comparing against expert labels. We show that despite reducing signal specifications by a factor of 2 or more, spike sorting accuracy does not change substantially. Specifically, reducing both sample rate and bit depth from 30 kHz and 16 bits to 12 kHz and 12 bits resulted in a 3% drop in spike sorting accuracy. Our results suggest that current neural acquisition systems are over-specified. These results may inform the design of next generation neural acquisition systems enabling higher channel count systems.

Keywords: Human performance - Driving, Human performance - Modelling and prediction, Human performance - Attention and vigilance
Abstract: Drug recognition expert (DRE) officers employ a set of tests to investigate drivers who are under impairment and to determine the type of drug that they have used. Horizontal Gaze Nystagmus (HGN), Walk and Turn (WAT), and One Leg Stand (OLS) are the main three tests included in the Standardized Field Sobriety Tests (SFSTs), which lead the officers to evaluate the sobriety of drivers. Performing these tests requires trained officers, but the final decision may still be subjective. These tests along with a suite of comprehensive (yet manual) at-station testing are the basis of police decision making and are subjected to scrutiny by courts. Therefore, designing an automated system to detect impairment not only will help officers in making accurate decisions, but also will remove the subjectivity and can potentially serve as a court-admissible evidence. In this paper, a new method for automated impairment detection is introduced and implemented using data analysis and machine learning algorithms based on a comprehensive suite of tests performed on 34 participants.

Keywords: Brain-computer/machine interface, Brain functional imaging - EEG, Neural signal processing
Abstract: The steady-state visual evoked potential (SSVEP) is one of the most widely used modalities in brain-computer interfaces (BCIs) due to its many advantages. However, the existence of harmonics and the limited range of responsive frequencies in SSVEP make it challenging to further expand the number of targets without sacrificing other aspects of the interface or putting additional constraints on the system. This paper introduces a novel multi-frequency stimulation method for SSVEP and investigates its potential to effectively and efficiently increase the number of targets presented. The proposed stimulation method, obtained by the superposition of the stimulation signals at different frequencies, is size-efficient, allows single-step target identification, puts no strict constraints on the usable frequency range, can be suited to self-paced BCIs, and does not require specific light sources. In addition to the stimulus frequencies and their harmonics, the evoked SSVEP waveforms include frequencies that are integer linear combinations of the stimulus frequencies. Results of decoding SSVEPs collected from nine subjects using canonical correlation analysis (CCA) with only the frequencies and harmonics as reference, also demonstrate the potential of using such a stimulation paradigm in SSVEP-based BCIs.

Keywords: Neural signal processing, Brain physiology and modeling - Sleep, Human performance - Sleep
Abstract: Sleep screening based on the construction of sleep stages is one of the major tool for the assessment of sleep quality and early detection of sleep-related disorders. Due to the inherent variability such as inter-users anatomical variability and the inter-systems differences, representation learning of sleep stages in order to obtain the stable and reliable characteristics is runoff for downstream tasks in sleep science. In this paper, we investigated the feasibility of the EEG-based symbolic representation for sleep stages. By combining the Latent Dirichlet Allocation topic model and comparing with different feature extraction methods, the work proved the feasibility of multi-topics representation for sleep stages and physiological signals.

Keywords: Brain functional imaging - EEG, Neural signals - Machine learning & Classification, Human performance - Attention and vigilance
Abstract: The study of human reaction time (RT) is invaluable not only to understand the sensory-motor functions but also to translate brain signals into machine comprehensible commands that can facilitate augmentative and alternative communication using brain-computer interfaces (BCI). Recent developments in sensor technologies, hardware computational capabilities, and neural network models have significantly helped advance biomedical signal processing research. This study is an attempt to utilize state-of-the-art resources to explore the relationship between human behavioral responses during perceptual decision-making and corresponding brain signals in the form of electroencephalograms (EEG). In this paper, a generalized 3D convolutional neural network (CNN) architecture is introduced to estimate RT for a simple visual task using single-trial multi-channel EEG. Earlier comparable studies have also employed a number of machine learning and deep learning-based models, but none of them considered inter-channel relationships while estimating RT. On the contrary, the use of 3D convolutional layers enabled us to consider the spatial relationship among adjacent channels while simultaneously utilizing spectral information from individual channels. Our model can predict RT with a root mean square error of 91.5 ms and a correlation coefficient of 0.83. These results surpass all the previous results attained from different studies.

Keywords: Human performance - Modelling and prediction, Neurological disorders - Diagnostic and evaluation techniques, Neurorehabilitation
Abstract: General motor and executive functions are integral for tasks of daily living and are typically assessed when quantifying impairment of an individual. Robotic tasks offer highly repeatable and objective measures of motor and cognitive function. Additionally, robotic tasks and measures have been used successfully to quantify impairment of children with cerebral palsy (CP). Many robotic tasks include multiple performance parameters, so interpretation of results and identification of impairment can be difficult, especially when multiple tasks are completed. This study used exploratory factor analysis to investigate a potential set of quantitative models of motor and cognitive function in children, and compare performance of participants with CP to these models. The three calculated factors achieved strong differentiation between participants with mild CP and the typically developing population. This demonstrates the feasibility of these factors to quantify impairment and track improvements related to therapies.

Keywords: Motor neuroprostheses, Motor neuroprostheses - Prostheses, Motor neuroprostheses - Robotics
Abstract: The success of pattern recognition based upper-limb prostheses control is linked to their ability to extract appropriate features from the electromyogram (EMG) signals. Traditional EMG feature extraction (FE) algorithms fail to extract spatial and inter-temporal information from the raw data, as they consider the EMG channels individually across a set of sliding windows with some degree of overlapping. To tackle these limitations, this paper presents a method that considers the spatial information of multi-channel EMG signals by utilising dynamic time warping (DTW). To satisfy temporal considerations, inspired by Long Short-Term Memory (LSTM) neural networks, our algorithm evolves the DTW feature representation across long and short-term components to capture the temporal dynamics of the EMG signal. As such the contribution of this paper is the development of a recursive spatio-temporal FE method, denoted as Recursive Temporal Warping (RTW). To investigate the performance of the proposed method, an offline EMG pattern recognition study with 53 movement classes performed by 10 subjects wearing 8 to 16 EMG channels was considered with the results compared against several conventional as well as deep learning-based models. We show that the use of the RTW can reduce classification errors significantly, paving the way for future real-time implementation.

Keywords: Human performance - Attention and vigilance, Neural stimulation
Abstract: Rhythmic visual stimulation (RVS) has been demonstrated to modulate ongoing neuronal oscillations which might be greatly involved in attention processes and thus bring some behavioral consequences. However, there was little knowledge about the effective frequency parameter of RVS which could impact task performance in visuo-spatial selective attention. Thus, here, we addressed this question by investigating the modulating effects of RVSs in different attention-related frequency bands, i.e., alpha (10 Hz) and gamma band (40 Hz). Sixteen participants were recruited to perform a modified visuo-spatial selective attention task. They were required to identify the orientation of target-triangle in visual search arrays while undergoing different RVS backgrounds. By analyzing the acquired behavioral and EEG data, we observed that, compared with control group (no RVS), 40 Hz RVS led to significantly shorter reaction time (RT) while 10 Hz RVS did not bring obvious behavioral consequences. In addition, although both 10 and 40 Hz RVS led to a global enhancement of SSVEP spectrum in the gamma band, 40 Hz RVS led to even larger 40 Hz SSVEP spectrum in prefrontal cortex. Our findings indicate that 40 Hz RVS has an effectively enhancing effect on selective attention and support the crucial role of prefrontal area in selective attention.

Keywords: Motor neuroprostheses - Neuromuscular stimulation, Human performance - Fatigue, Neuromuscular systems - Peripheral mechanisms
Abstract: The rapid onset of muscle fatigue during functional electrical stimulation (FES) is a major challenge when attempting to perform long-term periodic tasks such as walking. Surface electromyography (sEMG) is frequently used to detect muscle fatigue for both volitional and FES-evoked muscle contraction. However, sEMG contamination from both FES stimulation artifacts and residual M-wave signals requires sophisticated processing to get clean signals and evaluate the muscle fatigue level. The objective of this paper is to investigate the feasibility of computationally efficient ultrasound (US) echogenicity as a candidate indicator of FES-induced muscle fatigue. We conducted isometric and dynamic ankle dorsiflexion experiments with electrically stimulated tibialis anterior (TA) muscle on three human participants. During a fatigue protocol, we synchronously recorded isometric dorsiflexion force, dynamic dorsiflexion angle, US images, and stimulation intensity. The temporal US echogenicity from US images was calculated based on a gray-scaled analysis to assess the decrease in dorsiflexion force or motion range due to FES-induced TA muscle fatigue. The results showed that a monotonic reduction in US echogenicity change along with the fatigue progression for both isometric (R2 = 0.817 +- 0.073) and dynamic (R2 = 0.804 +- 0.035) ankle dorsiflexion. These results implied a strong linear relationship between US echogenicity and TA muscle fatigue level. The findings indicate that US echogenicity may be a promising computationally efficient indicator for assessing FES-induced muscle fatigue and help design muscle-in-the-loop FES controllers that consider the onset of muscle fatigue.

Keywords: Brain-computer/machine interface, Neural signals - Machine learning & Classification, Brain functional imaging - EEG
Abstract: Emotion recognition based on electroencephalography (EEG) plays a pivotal role in the field of affective computing, and graph convolutional neural network (GCN) has been proved to be an effective method and made considerable progress. Since the adjacency matrix that can describe the electrode relationships is critical in GCN, it becomes necessary to explore effective electrode relationships for GCN. However, the setting of the adjacency matrix and the corresponding value is empirical and subjective in emotion recognition, and whether it matches the target task remains to be discussed. To solve the problem, we proposed a graph convolutional network with learnable electrode relations (LR-GCN), which learns the adjacency matrix automatically in a goal-driven manner, including using self-attention to forward update the Laplacian matrix and using gradient propagation to backward update the adjacency matrix. Compared with previous works that use simple electrode relationships or only the feature information, LR-GCN achieved higher emotion recognition ability by extracting more reasonable electrode relationships during the training progress. We conducted a subject-dependent experiment on the SEED database and achieved recognition accuracy of 94.72% on the DE feature and 85.24% on the PSD feature. After visualizing the optimized Laplacian matrix, we found that the brain connections related to vision, hearing, and emotion have been enhanced.

Keywords: Human performance - Gait, Human performance - Modelling and prediction, Neurological disorders
Abstract: Digital gait measures derived from wearable inertial sensors have been shown to support the treatment of patients with motor impairments. From a technical perspective, the detection of left and right initial foot contacts (ICs) is essential for the computation of stride-by-stride outcome measures including gait asymmetry. However, in a majority of studies only one sensor close to the center of mass is used, complicating the assignment of detected ICs to the respective foot. Therefore, we developed an algorithm including supervised machine learning (ML) models for the robust classification of left and right ICs using multiple features from the gyroscope located at the lower back. The approach was tested on a data set including 40 participants (ten healthy controls, ten hemiparetic, ten Parkinson's disease, and ten Huntington's disease patients) and reached an accuracy of 96.3% for the overall data set and up to 100.0% for the Parkinson's sub data set. These results were compared to a state-of-the-art algorithm. The ML approaches outperformed this traditional algorithm in all subgroups. Our study contributes to an improved classification of left and right ICs in inertial sensor signals recorded at the lower back and thus enables a reliable computation of clinically relevant mobility measures.

Keywords: Brain-computer/machine interface, Brain functional imaging - EEG, Neural signals - Machine learning & Classification
Abstract: Electroencephalogram (EEG)-based emotion recognition has made great progress in recent years. The current pipelines collect EEG training data in a long-time calibration session for each new subject, which is time consuming and user unfriendly. To reduce the time required for the calibration session, there have been many studies using domain adaptation (DA) approaches to transfer knowledge from existing subjects (source domain) to the new subject (target domain) for reducing the dependence on the calibration session. Existing DA methods usually require substantial unlabeled EEG data of the new subject. However, the real scenario is that there are a small number of labeled samples in the calibration session of the target. Motivated by this, we introduce a novel domain adaptation architecture based on adversarial training to learn domain-invariant feature representations across subjects. To improve the performance when there are few labeled EEG data in the calibration session, we add a soft label loss to the architecture, which can ensure that the inter-class relationships learned from the source domain are transferred to target domain. We evaluate the method on the SEED dataset, and the experimental results show that our method uses only 15 examples per trial in the calibration session to achieve an average accuracy of 87.28%, indicating the effectiveness of our framework.

Keywords: Neural interfaces - Bioelectric sensors, Neural stimulation, Motor neuroprostheses
Abstract: Implantable neuromodulation devices that interface with the peripheral nervous system are a promising approach to restore functions lost to nerve damage. Existing nerve stimulation electrodes require direct contact with the target nerve and are associated with mechanical nerve damage and fibrous tissue encapsulation. Endovascularly delivered electrode arrays may provide a less invasive solution. Using a hybrid tissue conductor-neuron model and computational simulations, this study demonstrates the feasibility of delivering electrical stimulation of a peripheral nerve from a blood vessel in the vicinity of the target and predicts that the stimulation intensity required strongly depends on nerve-vessel distance and relative orientation, which are important factors to consider when screening candidate blood vessels for electrode implantation.

Keywords: Neural signal processing, Neurological disorders - Epilepsy, Brain physiology and modeling
Abstract: Cross-frequency coupling in general and phase-amplitude coupling (PAC) as a particular form of it, provides an opportunity to investigate the complex interactions between neural oscillations in the human brain and neurological disorders such as epilepsy. Using PAC detection methods on temporal sliding windows, we developed a map of dynamic PAC evolution to investigate the spatiotemporal changes occurring during ictal transitions in a patient with intractable mesial temporal lobe epilepsy. The map is built by computing the modulation index between the amplitude of high-frequency oscillations and the phase of lower frequency rhythms from the intracranial stereoelectroencephalography recordings during the seizure. Our preliminary results show early abnormal PAC changes occurring in the preictal state prior to the occurrence of clinical or visible electrographic seizure onset, and suggest that dynamic PAC measures may serve as a potential clinical technique for analyzing seizure dynamics.

Keywords: Brain-computer/machine interface, Human performance - Attention and vigilance, Neural signals - Machine learning & Classification
Abstract: Brain-computer interface (BCI) is a communication system that allows a direct connection between the human brain and external devices. With the application of BCI, it is important to estimate vigilance for BCI users. In order to investigate the vigilance changes of the subjects during BCI tasks and develop a multimodal method to estimate the vigilance level, a high-speed 4-target BCI system for cursor control was built based on steady-state visual evoked potential (SSVEP). 18 participants were recruited and underwent a 90-min continuous cursor-control BCI task, when electroencephalogram (EEG), electrooculogram (EOG), electrocardiography (ECG), and electrodermal activity (EDA) were recorded simultaneously. Then, we extracted features from the multimodal signals and applied regression models to estimate vigilance. Experimental results showed that the differential entropy (DE) feature could effectively reflect the change of vigilance. The vigilance estimation method, which integrates DE and EOG features into the support vector regression (SVR) model, achieved a better performance than the compared methods. These results demonstrate the feasibility of our methods for estimating vigilance levels in BCI.

Keywords: Neuromuscular systems - EMG processing and applications, Neuromuscular systems - Computational modeling, Neuromuscular systems - Locomotion
Abstract: Current clinical decision-making is based on rapid and subjective functional tests such as 10 m walking. Moreover, greater accuracy can be achieved at the expense of rapidity and costs. In biomechanical laboratories, advanced technologies and musculoskeletal modeling can quantitatively describe the biomechanical reasons underlying gait disorders. Our work aims to blend clinical rapidity and biomechanical accuracy through multi-channel (MC) electromyography (EMG) clustering and real-time neuro-musculoskeletal (NMS) modeling techniques integrated into a sensorized wearable garment that is quick to set up. Here we present a unique pipeline that goes from MC EMG signals to ankle torque estimation following two steps: (1) non-negative matrix factorization (NNMF)-based EMG clustering for the extraction of muscle-specific activations and (2) subject-specific EMG-driven NMS modeling. The results show the potential of NNMF as an electrodes clustering tool, as well as the ability to predict joint torque during movements that were not used for the EMG clustering.

Keywords: Brain-computer/machine interface, Neural signals - Machine learning & Classification, Brain functional imaging - EEG
Abstract: Brain-Computer Interfaces (BCI) provide effective tools aimed at recognizing different brain activities, translate them into actions, and enable humans to directly communicate through them. In this context, the need for strong recognition performances results in increasingly sophisticated machine learning (ML) techniques, which may result in poor performance in a real application (e.g., limiting a real-time implementation). Here, we propose an ensemble approach to effectively balance between ML performance and computational costs in a BCI framework. The proposed model builds a classifier by combining different ML models (base-models) that are specialized to different classification sub-problems. More specifically, we employ this strategy with an ensemble-based architecture consisting of multi-layer perceptrons, and test its performance on a publicly available electroencephalographybased BCI dataset with four-class motor imagery tasks. Compared to previously proposed models tested on the same dataset, the proposed approach provides greater average classification performances and lower inter-subject variability.

Keywords: Brain functional imaging - EEG, Neuromuscular systems - Computational modeling, Neural signal processing
Abstract: The analysis of electroencephalographic (EEG) series associated with movement performance is important for understanding the cortical neural control on motor tasks. While the existence of long-range correlations in physiological dynamics has been reported in previous studies, such a characterization in EEG series gathered during upper-limb movements has not been performed yet. To this end, here we report on a fractional integrated autoregressive analysis of EEG series during different functional classes of motor actions and resting phase, and data were gathered from 33 healthy volunteers. Results show significant differences in EEG long-range correlations on EEG series from characteristic topography.

Keywords: Brain functional imaging - EEG, Brain physiology and modeling - Sleep, Neural signals - Information theory
Abstract: The cortical activation and the interaction between cortical regions were considered to exist a strong correlation in recent neuroscience researches. However, such association during sleep was still unclear. The aim of the present work was to further investigate this association according to an activation-interaction association network. This study included 24 healthy individuals and all of them underwent overnight polysomnography. The absolute spectral powers of three frequency bands and the phase transfer entropy were extracted from six electroencephalogram channels. For each frequency band and sleep stage, activation-interaction association networks were built and correlation analysis was conducted by using Pearson correlation test. Results revealed the evident association between features derived from the two approaches during sleep, and as the sleep deepened, these correlation values attenuated in the alpha band, whereas the inversion happened in the delta band. This study exposed more detailed information of cortical activity during sleep, which will facilitate us to conduct research from a more comprehensive perspective, helping us make a more appropriate evaluation and explanation.

Keywords: Neural signal processing, Brain functional imaging - Connectivity and information flow, Brain functional imaging - EEG
Abstract: Olfactory hedonic perception involves complex interplay among an ensemble of neurocognitive systems implicated in sensory, affective and reward processing. However, the mechanisms of these inter-system interactions have yet to be well-characterized. Here, we employ directed functional connectivity networks estimated from source-localized EEG to uncover how brain regions across the olfactory, emotion and reward systems integrate organically into cross-system communities. Using the integration coefficient, a graph theoretic measure, we quantified the effect of exposure to fragrance stimuli of different hedonic values (high vs low pleasantness levels) on inter-systems interactions. Our analysis focused on beta band activity (13-30 Hz), which is known to facilitate the integration of cortical areas involved in sensory perception. Higher-pleasantness stimuli induced elevated integration for the reward system, but not for the emotion and olfactory systems. Furthermore, the nodes of reward system showed more outward connections to the emotion and olfactory systems than inward connections from the respective systems. These results suggest the centrality of the reward system—supported by beta oscillations—in actively coordinating multi-system interactivity to give rise to hedonic experiences during olfactory perception.

Keywords: Neural signals - Machine learning & Classification, Brain functional imaging - Classification, Brain functional imaging - EEG
Abstract: Consumer neuroscience is a rapidly emerging field, with the ability to detect consumer attitudes and states via real-time passive technologies being highly valuable. While many studies have attempted to classify consumer emotions and perceived pleasantness of olfactory products, no known machine learning approach has yet been developed to directly predict consumer reward-based decision-making, which has greater behavioral relevance. In this proof-of-concept study, participants indicated their decision to have fragrance products repeated after fixed exposures to them. Single-trial power spectral density (PSD) and approximate entropy (ApEn) features were extracted from EEG signals recorded using a wearable device during fragrance exposures, and served as subject-independent inputs for 4 supervised learning algorithms (kNN, Linear-SVM, RBF-SVM, XGBoost). Using a cross-validation procedure, kNN yielded the best classification accuracy (77.6%) using both PSD and ApEn features. Acknowledging the challenging prospects of single-trial classification of high-order cognitive states especially with wearable EEG devices, this study is the first to demonstrate the viability of using sensor-level features towards practical objective prediction of consumer reward experience.

Keywords: Brain functional imaging - fMRI, Brain functional imaging - Spatial-temporal dynamics, Brain physiology and modeling - Cognition, memory, perception
Abstract: For the last several decades, emotion research has attempted to identify a ``biomarker" or consistent pattern of brain activity to characterize a single category of emotion (e.g., fear) that will remain consistent across all instances of that category, regardless of individual and context. In this study, we investigated variation rather than consistency during emotional experiences while people watched video clips chosen to evoke instances of specific emotion categories. Specifically, we developed a sequential probabilistic approach to model the temporal dynamics in a participant's brain activity during video viewing. We characterized brain states during these clips as distinct state occupancy periods between state transitions in blood oxygen level-dependent (BOLD) signal patterns. We found substantial variation in the state occupancy probability distributions across individuals watching the same video, supporting the hypothesis that when it comes to the brain correlates of emotional experience, variation may indeed be the norm.

Keywords: Neurological disorders - Communication, Brain-computer/machine interface, Neural signals - Machine learning & Classification
Abstract: In this paper, we propose a deep learning-based algorithm to improve the performance of automatic speech recognition (ASR) systems for aphasia, apraxia, and dysarthria speech by utilizing electroencephalography (EEG) features recorded synchronously with aphasia, apraxia, and dysarthria speech. We demonstrate a significant decoding performance improvement by more than 50% during test time for isolated speech recognition task and we also provide preliminary results indicating performance improvement for the more challenging continuous speech recognition task by utilizing EEG features. The results presented in this paper show the first step towards demonstrating the possibility of utilizing non-invasive neural signals to design a real-time robust speech prosthetic for stroke survivors recovering from aphasia, apraxia, and dysarthria. Our aphasia, apraxia, and dysarthria speech-EEG data set will be released to the public to help further advance this interesting and crucial research.

Keywords: Neurological disorders - Stroke, Neurorehabilitation, Human performance - Gait
Abstract: Post-stroke hemiparesis often impairs gait and increases the risks of falls. Low and variable Minimum Toe Clearance (MTC) from the ground during the swing phase of the gait cycle has been identified as a major cause of such falls. In this paper, we study MTC characteristics in 30 chronic stroke patients, extracted from gait patterns during treadmill walking, using infrared sensors and motion analysis camera units. We propose objective measures to quantify MTC asymmetry between the paretic and non-paretic limbs using Poincare analysis. We show that these subject independent Gait Asymmetry Indices (GAIs) represent temporal variations of relative MTC differences between the two limbs and can distinguish between healthy and stroke participants. Compared to traditional measures of cross-correlation between the MTC of the two limbs, these measures are better suited to automate gait monitoring during stroke rehabilitation. Further, we explore possible clusters within the stroke data by analysing temporal dispersion of MTC features, which reveals that the proposed GAIs can also be potentially used to quantify the severity of lower limb hemiparesis in chronic stroke.

Keywords: Neurorehabilitation, Human performance - Gait, Neuromuscular systems - Central mechanisms
Abstract: Exoskeleton-assisted gait rehabilitation is a promising complement to traditional motion rehabilitation programs for afflictions such as stroke or spinal cord injury. However, some challenges persist that hinder the translation of this approach to the clinical practice. One of these aspects is the objective assessment of patients' progress from information collected during exoskeleton-assisted therapy sessions with minimal hardware setup. In order to carry out an objective assessment with the data collected during the sessions, in this work: (1) we implement and compute a set of metrics (Harmonic Ratio, Joint Trajectory Correlation, and Intralimb Coordination) from data provided by the exoskeleton and two inertial motion units (IMUs) while subjects walked during their rehabilitation sessions, (2) we evaluate the capacity of the metrics to discriminate between the different patients' physical conditions, and (3) assess the correspondence of the patient evaluations using the mentioned metrics and traditional clinical scores. Our results show that Intralimb Coordination has the greatest capacity to discriminate between different physical states of the patients and presents the best correlation with their clinical assessment.

Keywords: Neural stimulation, Brain physiology and modeling - Neuron modeling and simulation
Abstract: Transcranial electrical stimulation (tES), which modulates cortical excitability via electric currents, has attracted increasing attention because of its application in treating neurologic and psychiatric disorders. To obtain a better understanding of the brain areas affected and stimulation’s cellular effects, a multi-scale model was proposed that combines multi-compartmental neuronal models and a head model. While one multi-scale model of tES that used straight axons reported that the direction of electric field (EF) is a determining factor in a neuronal response, another model of transcranial magnetic stimulation (TMS) that used arborized axons reported that EF magnitude is more crucial than EF direction because of arborized axons' reduced sensitivity to the latter. Our goal was to investigate whether EF magnitude remains a crucial factor in the neuronal response in a multi-scale model of tES into which an arborized axon is integrated. To achieve this goal, we constructed a multi-scale model that integrated three types of neurons and a realistic head model, and then simulated the neuronal response to realistic EF. We found that EF magnitude was correlated with excitation threshold, and thus, it may be one of the determining factors in cortical neurons' response to tES. Clinical Relevance—This multi-scale model based on biophysical and morphological properties and realistic brain geometry may help elucidate tES's neural mechanisms. Moreover, given its clinical applications, this model may help predict a patient’s neuronal response to brain stimulation effectively.

Keywords: Brain-computer/machine interface, Neuromuscular systems - EMG processing and applications, Neurorehabilitation
Abstract: EEG-EMG based hybrid Brain Computer Interface (hBCI) utilizes the brain-muscle physiological system to interpret and identify motor behaviors, and transmit human intelligence to automated machines in AI applications such as neurorehabilitations and brain-like intelligence. The study introduces a hBCI method for motor behaviors, where multiple time series of the brain neuromuscular network are introduced to indicate brain-muscle causal interactions, and features are extracted based on Relative Causal Strengths (RCSs) derived by Noise-assisted Multivariate Empirical Mode Decomposition (NA-MEMD) based Causal Decomposition. The complex process in brain neuromuscular interactions is specifically investigated towards a monitoring task of upper limb movement,whose 63-channel EEGs and 2-channel EMGs are composed of data inputs. The energy and frequency factors counted from RCSs were extracted as Core Features (CFs). Results showed accuracies of 91.4% and 81.4% with CFs for identifying cascaded (No Movement and Movement Execution) and 3-class (No Movement, Right Movement, and Left Movement) using Naive Bayes classifier, respectively. Moreover, those reached 100% and 94.3% when employing CFs combined with eigenvalues processed by Common Spatial Pattern (CSP). This initial work implies a novel causality inference based hBCI solution for the detection of human upper limb movement.

Keywords: Neuromuscular systems - Computational modeling, Brain physiology and modeling - Neural dynamics and computation
Abstract: We have investigated selective electrical stimulation of myelinated nerve fibers using a computational model of temporal interfering (TI) fields. The model consists of two groups of electrodes placed on the outer bundle surface, each group stimulated at a different frequency. We manipulated the stimulus waveform, magnitude and frequency of short-duration stimuli (70 ms), and investigated fiber-specific stimulus-elicited compound action potentials. Results show that under 100Hz & 200Hz TI stimulation with 0.6 mA total current shared by the electrodes, continuous action potentials were generated in deeper nerve fibers, and that the firing region was steerable by changing individual electrode currents. This study provides a promising platform for non-invasive nerve bundle stimulation by TI fields.

Keywords: Neuromuscular systems - Locomotion, Neuromuscular systems - EMG models, Neurorehabilitation
Abstract: Abstract: Rehabilitation promoting 'assistance-as-needed' is considered a promising scheme of active rehabilitation, since it can promote neuroplasticity faster and thus reduce the time needed until restoration. To implement such schemes using robotic devices, it is crucial to be able to predict accurately and in real-time the intention of motion of the patient. In this study, we present an intention-of-motion model trained on healthy volunteers. The model is trained using kinematics and muscle activation time series data, and returns future predicted values for the kinematics. We also present the results of an analysis of the sensitivity of the accuracy of the model for different amount of training datasets and varying lengths of the prediction horizon. We demonstrate that the model is able to predict reliably the kinematics of volunteers that were not involved in its training. The model is tested with three types of motion inspired by rehabilibation tasks. In all cases, the model is predicting the arm kinematics with a Root Mean Square Error (RMSE) below 0.12m. Being a non person-specific model, it could be used to predict kinematics even for patients that are not able to perform any motion without assistance. The resulting kinematics, even if not fully representative of the specific patient, might be a preferable input for a robotic rehabilitator than predefined trajectories currently in use.

Clinical relevance: This model predicts intention of motion for use as a setpoint for robotic rehabilitators. This can be useful for patients that are not able to perform motions without assistance.

Keywords: Neural stimulation - Deep brain, Neural signal processing, Smart neural implants - Neurostimulation
Abstract: Adaptive deep brain stimulation (aDBS) promises a significant improvement in patient outcomes, compared to existing deep brain stimulation devices. Fully implanted systems represent the next step to the clinical adoption of aDBS. We take advantage of a unique longitudinal data set formed as part of an effort to investigate aDBS for essential tremor to verify the long term reliability of electrocorticography strips over the motor cortex as a source of bio-markers for control of adaptive stimulation. We show that beta band event related de-synchronization, a promising bio-marker for movement, is robust even when used to trigger aDBS. Over the course of several months we show a minor increase in beta band event related de-synchronization in patients with active deep brain stimulation confirming that it could be used in chronically implanted systems.

Keywords: Brain-computer/machine interface
Abstract: Neurological disorders can lead to significant impairments in speech communication and, in severe cases, cause the complete loss of the ability to speak. Brain-Computer Interfaces have shown promise as an alternative communication modality by directly transforming neural activity of speech processes into a textual or audible representations. Previous studies investigating such speech neuroprostheses relied on electrocorticography (ECoG) or microelectrode arrays that acquire neural signals from superficial areas on the cortex. While both measurement methods have demonstrated successful speech decoding, they do not capture activity from deeper brain structures and this activity has therefore not been harnessed for speech-related BCIs. In this study, we bridge this gap by adapting a previously presented decoding pipeline for speech synthesis based on ECoG signals to implanted depth electrodes (sEEG). For this purpose, we propose a multi-input convolutional neural network that extracts speech-related activity separately for each electrode shaft and estimates spectral coefficients to reconstruct an audible waveform. We evaluate our approach on open-loop data from 5 patients who conducted a recitation task of Dutch utterances. We achieve correlations of up to 0.80 between original and reconstructed speech spectrograms, which are significantly above chance level for all patients (p < 0.001). Our results indicate that sEEG can yield similar speech decoding performance to prior ECoG studies and is a promising modality for speech BCIs.

Keywords: Brain-computer/machine interface, Neural stimulation, Neurorehabilitation
Abstract: Developing closed-loop brain stimulation systems can benefit the treatment of neurological and neuropsychiatric disorders and facilitate brain functions. Current designs of closed-loop controllers have used time-invariant linear models of brain activity to devise non-adaptive controllers. However, unmodeled nonlinear dynamics can happen during real-time closed-loop control, leading to nonlinear uncertainty in the brain activity model. Current non-adaptive controllers cannot track the nonlinear model uncertainty and are not robust to noise, both of which can compromise their control performance. Here, within an L1 adaptive control framework, we develop a new discrete-time robust and adaptive closed-loop control algorithm that addresses a general form of nonlinear model uncertainty. We conduct Monte Carlo simulations to validate the robust and adaptive control algorithm and show that it significantly outperforms existing closed-loop control algorithms. Our results can facilitate future designs of precise and safe closed-loop brain stimulation systems to treat neurological and neuropsychiatric disorders and modulate brain functions.

Keywords: Brain functional imaging - EEG, Neurological disorders - Stroke, Neural signals - Nonlinear analysis
Abstract: A valid evaluation of neurological functions after stroke may improve clinical decision-making. The aim of this study was to compare the performance of EEG-related indexes in differentiating stroke patients from control participants, and to investigate pathological EEG changes after chronic stroke. 20 stroke and 13 healthy participants were recruited, and spontaneous EEG signals were recorded during the resting state. EEG rhythms and complexity were calculated based on Fast Fourier Transform and the fuzzy approximate entropy (fApEn) algorithm. The results showed a significant difference of alpha rhythm (p = 0.019) and fApEn (p = 0.003) of EEG signals from brain area among ipsilesional, contralesion hemisphere of stroke patients and corresponding brain hemisphere of healthy participants. EEG fApEn had the best classification accuracy (84.85%), sensitivity (85.00%), and specificity (84.62%) among these EEG-related indexes. Our study provides a potential method to evaluate alterations in the properties of the injured brain, which help us to understand neurological change in chronic strokes.

Keywords: Neural stimulation
Abstract: In this study, the neural response to pulse-width modulated (PWM) transcranial magnetic stimulation (TMS) is estimated using a computational neural model which simulates the response of cortical neurons to TMS. The recently introduced programmable TMS uses PWM to approximate conventional resonance-based TMS pulses by fast switching between voltage levels. The effect of such stimulation on the six cortical layers is modelled by estimating the activation threshold of the neurons. Modelling results are compared between the novel device and that of conventional TMS stimuli generated by Magstim stimulators. The neural responses to the PWM pulses and the conventional stimuli show a high correlation, which validates the use of pulse-width modulated pulses in TMS.

Keywords: Neural signals - Machine learning & Classification, Human performance - Cognition, Brain functional imaging - Classification
Abstract: This paper analyzes local field potentials (LFP) from 10 human subjects to discover frequency-dependent biomarkers of cognitive conflict. We utilize cortical and sub-cortical LFP recordings from the subjects during a cognitive task known as the Multi-Source Interference Task (MSIT). We decode the task engagement and discover biomarkers that may facilitate closed-loop neuromodulation to enhance cognitive control. First, we show that spectral power features in predefined frequency bands can be used to classify task and non-task segments with a median accuracy of 88.1%. Here the features are first ranked using the Bayes Factor and then used as inputs to subject-specific linear support vector machine classifiers. Second, we show that theta (4–8 Hz) band, and high gamma (65–200 Hz) band oscillations are modulated during the task performance. Third, by isolating time-series from specific brain regions of interest, we observe that a subset of the dorsolateral prefrontal cortex features is sufficient to decode the task states. The paper shows that cognitive control evokes robust neurological signatures, especially in the prefrontal cortex (PFC).

Keywords: Motor learning, neural control, and neuromuscular systems, Neuromuscular systems - EMG processing and applications, Neurological disorders
Abstract: Effective physical human-robot interaction (pHRI) depends on how humans can communicate their intentions for movement with others. While it is speculated that small interaction forces contain significant information to convey the specific movement intention of physical human-human interaction (pHHI), the underlying mechanism for humans to infer intention from such small forces is largely unknown. The hypothesis in this work is that the sensitivity to a small interaction force applied at the hand is affected by the movement of the arm that is affected by the arm stiffness. For this, a haptic robot was used to provide the endpoint interaction forces to the arm of seated human participants. They were asked to determine one of the four directions of the applied robot interaction force without visual feedback. Variations of levels of interaction force as well as arm muscle contraction were applied. The results imply that human’s ability to identify and respond to the correct direction of small interaction forces was lower when the alignment of human arm movement with respect to the force direction was higher. In addition, the sensitivity to the direction of the small interaction force was high when the arm stiffness was low. It is also speculated that humans lower their arm stiffness to be more sensitive to smaller interaction forces. These results will help develop human-like pHRI systems for various applications

Keywords: Human performance - Engineering, Brain functional imaging - fMRI, Neural signal processing
Abstract: This paper investigates the effect of filtering (or modulating) the functional magnetic resonance imaging (fMRI)time-series on intelligence metrics predicted using dynamic functional connectivity (dFC). Thirteen brain regions that have highest correlation with intelligence are selected and their corresponding time-series are filtered. Using filtered time-series, the modified intelligence metrics are predicted. This experiment investigates whether modulating the time-series of one or two regions of the brain will increase or decrease the fluid ability and fluid intelligence among healthy humans. Two sets of experiments are performed. In the first case, each of the thirteen regions is separately filtered using four different digital filters with pass-bands: i) 0-0.25π, ii) 0.25π-0.5π, iii)0.5π-0.75π and iv) 0.75π–π, respectively. In the second case, two of the thirteen regions are filtered simultaneously using a low-pass filter of pass-band 0 - 0.25π. In both cases, the predicted intelligence declined for 45-65% of the subjects after filtering in comparison with the ground truths. In the first case, the low pass filtering process achieves the highest predicted intelligence among the four filters. In the second case, it was noticed that the filtering of two regions simultaneously resulted in a higher prediction of intelligence for over 80% of the subjects compared to the individual low pass filtering performed in the first case.

Keywords: Smart neural implants - Neurostimulation, Neural interfaces - Tissue-electrode interface, Neural stimulation
Abstract: A design method of electrode-tissue interface equivalent circuit for the voltage injection test of active implantable neurostimulator (INS) is presented and analyzed. In this proposed method, characteristic frequencies of the equivalent circuit are determined, based on the published data of human tissue permittivity and conductivity. The equivalent circuit structure is defined, according to "electrode-tissue" interface model. Appropriate values of electronic components are matched by simulation. In addition, a method of replacing the electrode-tissue interface equivalent circuit with purely resistance is also proposed. According to ISO14708-3, voltage injection tests are carried out with these different equivalent circuits and INS. Results showed that these design methods can meet test requirements with no significant difference. This study explored convenient and universal methods for the voltage injection test of INS, which is useful to improve the guarantee of the electromagnetic compatibility (EMC) safety of the INS.

Keywords: Human performance - Vestibular functions, Human performance - Sensory-motor
Abstract: To orient and move efficiently in the environment, we need to rely on multiple external and internal cues. Previous studies reported the combined use of spatialized auditory cues and self-motion information in spatial navigation and orientation. In this study, we investigated the feasibility of a setup composed of a motion platform and an acoustic virtual reality tool with sighted and visually impaired participants. We compared the performance in a self-motion discrimination task with and without auditory cues. The results revealed good usability of the setup and increased precision with auditory cues for visually impaired people.

Keywords: Human performance, Human performance - Activities of daily living, Neural signal processing
Abstract: Lumbar paraspinal muscles are heavily involved in daily and work-related activities including trunk bending, trunk twisting, and lifting. Repetitive or inappropriate activation of the lumbar muscles while performing these activities can lead to low back pain. The aim of this preliminary study was to quantify the activation patterns of multiple lumbar muscles when participants performed three different trunk movement tasks, including sustained lumbar flexion posture, dynamic lumbar flexion and extension, and left-right twisting movements. Two 8×8 high-density electromyogram (HD-EMG) electrode arrays were used to record the lumbar muscle activity during these movements. We observed a symmetric and rapid increase in the amplitude of EMG in the erector spinae muscles during the sustained flexion or oscillation tasks. Asymmetric activation patterns were observed in bilateral lumbar muscles during the trunk twisting task. In addition, we observed substantial bilateral co-activation of the lumbar muscles for both twisting directions. These preliminary results demonstrated the potential feasibility of using HD-EMG as a tool to monitor spatial activation patterns of the lumbar muscles during different trunk movements. This approach can also be further developed to assess lumbar muscle function in individuals with low back pain.

Keywords: Brain-computer/machine interface, Human performance, Neural signal processing
Abstract: In the study of an electroencephalography (EEG)- based brain computer interface (BCI) using the P300, there have been many reports on computer algorithms that identify the target intended by a user from multiple candidates. However, because the P300 amplitude depends on the subject's condition and is attenuated by physical and mental factors, such as fatigue and motivation, the performance of the BCI is low. Therefore, we aim to improve performance by introducing a feedback mechanism that provides the user with an evaluation calculated by the computer during EEG measurement. In this study, we conducted an experiment in which the user input one character from four characters on the display. By changing the character size according to the evaluation score calculated by the computer, the computer's current evaluation was fed back to the user. This is expected to change the consciousness of the user to enable them to execute a task by knowing the evaluation; that is, if the evaluation is high, the user needs to maintain the current state, and if the evaluation is low, a behavioral change, such as increasing attention, is required to improve the evaluation. As a result of comparing 10 subjects with and without feedback, accuracy improved for seven subjects that were given feedback.

Keywords: Neural signals - Machine learning & Classification, Neural signal processing, Brain functional imaging - EEG
Abstract: In clinical examination, event-related potentials (ERPs) are estimated by averaging across multiple responses, which suppresses background EEG. However, acquiring the number of responses needed for this process is time consuming. We therefore propose a method for shortening the measurement time using weighted-average processing based on the output of deep learning. Using P300 as a representative component, here we focused on the shape of the ERP and evaluated whether our method emphasizes the P300 peak amplitude more than conventional averaging, while still maintaining the waveform shape and the P300 peak latency. Thus, using either CNN or EEGNet, the correlation coefficient reflecting the waveform shape, the peak P300 amplitude, and the peak latency were evaluated and compared with the same factors obtained from conventional waveform averaging. Additionally, the degree of background EEG suppression provided by our method was evaluated using the root mean square of the pre-stimulation waveform, and the number of fewer responses required for averaging (i.e., the reduction in measurement time) was calculated. The results showed that compared with P300 values obtained through conventional averaging, our method allowed for the same shape and response latency, but with a higher amplitude, while requiring a smaller number of responses. Our method showed that by using EEGNet, measurement time could be reduced by 13.7%. This corresponds to approximately a 40-second reduction for every 5 minutes of measurement time.

Keywords: Brain-computer/machine interface, Neural signals - Machine learning & Classification, Brain functional imaging - EEG
Abstract: As an important element in the human-machine interaction, the electroencephalogram (EEG)-based emotion recognition has achieved significant progress. However, one obstacle to practicality lies in the variability between subjects and sessions. Although several studies have adopted domain adaptation (DA) approaches to tackle this problem, most of them treat multiple data from different subjects and different sessions together as a single source for transfer. Since different EEG data have different marginal distributions, these approaches fail to satisfy the assumption of DA that the source has a certain marginal distribution. We therefore propose the multi-source EEG-based emotion recognition network (MEERNet), which takes both domain-invariant and domain-specific features into consideration. Firstly we assume that different EEG data share the same low-level features, and then we construct multiple branches corresponding to multiple sources to extract domainspecific features, and then DA is conducted between the target and each source. Finally, the inference is made by multiple branches. We evaluate our method on SEED and SEED-IV for recognizing three and four emotions, respectively. Experimental results show that the MEERNet outperforms the single-source methods in cross-session and cross-subject transfer scenarios with an accuracy of 86:7% and 67:1% on average, respectively.

Keywords: Brain-computer/machine interface, Neural signals - Machine learning & Classification
Abstract: Brain-Computer Interfaces (BCIs) that decode a patient's movement intention to control a prosthetic device could restore some independence to paralyzed patients. An important step on the road towards naturalistic prosthetic control is to decode movement continuously with low-latency. BCIs based on intracortical micro-arrays provide continuous control of robotic arms, but require a minor craniotomy. Surface recordings of neural activity using EEG have made great advances over the last years, but suffer from high noise levels and large intra-session variance. Here, we investigate the use of minimally invasive recordings using stereotactically implanted EEG (sEEG). These electrodes provide a sparse sampling across many brain regions. So far, promising decoding results have been presented using data measured from the subthalamic nucleus or trial-to-trial based methods using depth electrodes. In this work, we demonstrate that grasping movements can continuously be decoded using sEEG electrodes, as well. Beta and high-gamma activity was extracted from eight participants performing a grasping task. We demonstrate above chance level decoding of movement vs rest and left vs right, from both frequency bands with accuracies up to 0.94 AUC. The vastly different electrode locations between participants lead to large variability. In the future, we hope that sEEG recordings will provide additional information for the decoding process in neuroprostheses.

Keywords: Neural interfaces - Neuromorphic engineering, Neural signals - Machine learning & Classification, Brain-computer/machine interface
Abstract: Accurate and low-power decoding of brain signals such as electroencephalography (EEG) is key to constructing brain-computer interface (BCI) based wearable devices. While deep learning approaches have progressed substantially in terms of decoding accuracy, their power consumption is relatively high for mobile applications. Neuromorphic hardware arises as a promising solution to tackle this problem since it can run massive spiking neural networks with energy consumption orders of magnitude lower than traditional hardware. Herein, we show the viability of directly mapping a continuous-valued convolutional neural network for motor imagery EEG classification to a spiking neural network. The converted network, able to run on the SpiNNaker neuromorphic chip, only shows a 1.91% decrease in accuracy after conversion. Thus, we take full advantage of the benefits of both deep learning accuracies and low-power neuro-inspired hardware, properties that are key for the development of wearable BCI devices.

Keywords: Neural signals - Machine learning & Classification, Human performance - Activities of daily living, Neurological disorders - Stroke
Abstract: Objective and accurate activity identification of physical activities in everyday life is an important aspect in assessing the impact of various post-stroke rehabilitation therapies and interventions. Since post-stroke hemiparesis affects gait and balance in individuals with stroke, activity identification algorithms that consider stroke-specific movement irregularities are needed. While wearable physical activity monitors provide the means to detect activities in the free-living, algorithms using their data are specific to the wear location of the device. This pilot study builds, validates, and compares three machine learning algorithms (linear support vector machine, Random Forest, and RUSBoosted trees) at three popular wear locations (wrist, waist, and ankle) to identify and accurately distinguish mobility-related activities (sitting, standing and walking) in individuals with chronic stroke. A total of 102 minutes of data from two lab visits of three-stroke participants was used to build the classifiers. A 5-fold cross-validation technique was used to validate and compare the accuracy of classifiers. RUSBoosted trees using data from waist and ankle activity monitors, with an accuracy of 99.1%, outperformed other classifiers in detecting three activities of interest.

Clinical Relevance— One of the major aims of post-stroke rehabilitation is improving mobility, which may be facilitated by understanding the structure and pattern of everyday mobility through real-world, objective outcomes. Accurate activity identification, as shown in this pilot investigation, is an essential first step before developing objective outcomes for monitoring mobility and balance in everyday life of these individuals.

Keywords: Neurorehabilitation
Abstract: Research using nonhuman primate models for human disease frequently requires behavioral observational techniques to quantify functional outcomes. The ability to assess reaching and grasping patterns is of particular interest in clinical conditions that affect the motor system (e.g., spinal cord injury, SCI). Here we explored the use of DeepLabCut, an open source deep learning toolset, in combination with a standard behavioral task (Brinkman Board) to quantify nonhuman primate performance in precision grasping. We examined one male rhesus macaque (Macaca mulatta) in the task which involved retrieving rewards from variously-oriented shallow wells. Simultaneous recordings were made using GoPro Hero7 Black cameras (resolution 1920 x 1080 at 120 fps) from two different angles (from the side and top of the hand motion). The task/device design necessitates use of the right hand to complete the task. Two neural networks (corresponding to the top and side view cameras) were trained using 400 manually annotated images, tracking 19 unique landmarks each. Based on previous reports, this produced sufficient tracking (Side: trained pixel error of 2.15, test pixel error of 11.25; Top: trained pixel error of 2.06, test pixel error of 30.31) so that landmarks could be tracked on the remaining frames. Landmarks included in the tracking were the spatial location of the knuckles and the fingernails of each digit, and three different behavioral measures were quantified for assessment of hand movement (finger separation, middle digit extension and preshaping distance). Together, our preliminary results suggest that this markerless approach is a possible method to examine specific kinematic features of dexterous function.

Keywords: Brain-computer/machine interface, Neural stimulation, Brain functional imaging - EEG
Abstract: Brain-computer interface (BCI) based rehabilitation has been proven a promising method facilitating motor recovery. Recognizing motor intention is crucial for realizing BCI rehabilitation training. Event-related desynchronization (ERD) is a kind of electroencephalogram (EEG) inherent characteristics associated with motor intention. However, due to brain deficits poststroke, some patients are not able to generate ERD, which discourages them to be involved in BCI rehabilitation training. To boost ERD during motor imagery (MI), this paper investigates the effects of high-frequency repetitive transcranial magnetic stimulation (rTMS) on BCI classification performance. Eleven subjects participated in this study. The experiment consisted of two conditions: rTMS + MI versus sham rTMS + MI, which were arranged on different days. MI tests with 64-channel EEG recording were arranged immediately before and after rTMS and sham rTMS. Time-frequency analysis were utilized to measure ERD changes. Common spatial pattern was used to extract features and linear discriminant analysis was used to calculate offline classification accuracies. Paired-sample t-test and Wilcoxon signed rank tests with post-hoc analysis were used to compare performance before and after stimulation. Statistically stronger ERD (-13.93±12.99%) was found after real rTMS compared with ERD (-5.71±21.25%) before real rTMS (p<0.05). Classification accuracy after real rTMS (70.71±10.32%) tended to be higher than that before real rTMS (66.50±8.48%) (p<0.1). However, no statistical differences were found after sham stimulation. This research provides an effective method in improving BCI performance by utilizing neural modulation.

Keywords: Neurological disorders - Epilepsy, Brain functional imaging - Connectivity and information flow, Neural signals - Information theory
Abstract: Transfer entropy (TE) is used to examine the connectivity between nodes and the roles of nodes in epileptic neural networks during rest, moments before seizure, during seizure, and moments after seizure. There is a set of nodes that dominate information flow to epileptogenic zone (EZ) nodes, regions that trigger seizure, and non-EZ nodes during rest. The TE from the dominant to the EZ nodes decreases shortly before a seizure event and reaches a minimum during seizure. During the seizure, the dominant nodes cease or only weakly interact with the EZ nodes. This supports the hypothesis that seizure occurs when some nodes stop inhibiting the EZ nodes. The TE from the dominant to the EZ nodes peaks immediately after seizure, suggesting that seizure may stop when the brain exerts the highest level of information flow/activation/communication to the EZ nodes. The information flow from the dominant to EZ nodes is different from that to non-EZ nodes. This TE dynamics entering and exiting seizures may identify more accurately the EZ nodes, which may improve surgical planning.

Keywords: Neuromuscular systems - Computational modeling, Brain physiology and modeling - Neuron modeling and simulation, Neural signals - Blind source separation (PCA, ICA, etc.)
Abstract: The in vivo estimation of α-motoneuron (MN) properties in humans is crucial to characterize the effect that neurorehabilitation technologies may elicit over the composite neuro-musculoskeletal system. Here, we combine biophysical neuronal modelling, high-density electromyography and convolutive blind-source separation along with numerical optimization to estimate geometrical and electrophysiological properties of in vivo decoded human MNs. The proposed methodology implements multi-objective optimization to automatically tune ionic channels conductance and soma size of MN models for minimizing the error between several features of simulated and in vivo decoded MN spike trains. This approach will open new avenues for the closed-loop control of motor restorative technologies such as wearable robots and neuromodulation devices.

Keywords: Sensory neuroprostheses - Visual, Neurorehabilitation, Human performance - Modelling and prediction
Abstract: Retinal prosthesis (RP) is used to partially restore vision in patients with degenerative retinal diseases. Assessing the quality of RP-acquired (i.e., prosthetic) vision is needed to evaluate RP impact and prospects. Spatial distortions caused by electrical stimulation of the retina in RP, and the low number of electrodes, have limited the prosthetic vision: patients mostly localize shapes and shadows rather than recognizing objects. We simulate prosthetic vision and evaluate vision on image classification tasks, varying critical hardware parameters: total number and size of electrodes. We also simulate rehabilitation by re-training our models on prosthetic vision images. We find that electrode size has little impact on vision while at least 400 electrodes are needed to sufficiently restore vision (more than 65% classification accuracy on a complex visual task after rehabilitation). Argus II, a currently available implant, produces a low-resolution vision leading to low accuracy (21.3% score after rehabilitation) in complex vision tasks. Rehabilitation produces significant improvements (accuracy improvement of up to 30% on complex tasks, depending on the number of electrodes) in the attained vision, boosting our expectations for RP interventions and motivating the establishment of rehabilitation procedures for RP implantees.

Keywords: Neurological disorders - Traumatic brain injury, Neural signals - Machine learning & Classification, Human performance - Sleep
Abstract: Traumatic Brain Injury (TBI) is a highly prevalent and serious public health concern. Most cases of TBI are mild in nature, yet some individuals may develop following-up persistent disability. The pathophysiologic causes for those with persistent postconcussive symptoms are most likely multifactorial and the underlying mechanism is not well understood, although it is clear that sleep disturbances feature prominently in those with persistent disability. The sleep electroencephalogram (EEG) provides a direct window into neuronal activity during an otherwise highly stereotyped behavioral state, and represents a promising quantitative measure for TBI diagnosis and prognosis. With the ever-evolving domain of machine learning, deep convolutional neural networks, and the development of better architectures, these approaches hold promise to solve some of the long entrenched challenges of personalized medicine for uses in recommendation systems and/or in health monitoring systems. In particular, advanced EEG analysis to identify putative EEG biomarkers of neurological disease could be highly relevant in the prognostication of mild TBI, an otherwise heterogeneous disorder with a wide range of affected phenotypes and disability levels. In this work, we investigate the use of various machine learning techniques and deep neural network architectures on a cohort of human subjects with sleep EEG recordings from overnight, in-lab, diagnostic polysomnography (PSG). An optimal scheme is explored for the classification of TBI versus non-TBI control subjects. The results were promising with an accuracy of 95% in random sampling arrangement and 70% in independent validation arrangement when appropriate parameters were used using a small number of subjects (10 mTBI subjects and 9 age- and sex-matched controls). We are thus confident that, with additional data and further studies, we would be able to build a generalized model to detect TBI accurately, not only via attended, in-lab PSG rec

Keywords: Neurological disorders - Stroke, Neurorehabilitation, Neural signal processing
Abstract: Electroencephalogram (EEG) is a basic physiological signal of human body, which can effectively record the nervous system activities of the brain and contains rich information. The synchronization of EEG signals is not only the key to the exchange of information between different brain regions, but also reflects the neural activity of the brain, which in turn can infer people's cognitive activities. Therefore, studying the phase synchronization of EEG signals after stroke is of great significance for understanding the communication and neuroplasticity of neurons after brain injury. In this paper, the changes of EEG phase synchronization in bilateral, cyclical ankle movements alternately after stroke were studied by Hilbert transform. Ten stroke patients and six healthy adults participated in the test. The results showed that the inter-hemisphere phase synchronization index (inter-PSI) and the global PSI of patients were significantly lower than that of the healthy subjects during the task. The PSI between Cz and the affected sensory cortex associated with lower limb movements was also significantly lower than that in the control group. There was a significant negative correlation between National Institutes of Health Stroke Scale (NIHSS) and cortical synchronization. The above results indicated that PSI under ankle alternating movements may be used as a new biomarker to evaluate the recovery of patients’ brain neurons.