Keywords: Scaffolds in tissue engineering, Translational issues in tissue engineering and biomaterials - Bioreactors, Electric fields - Tissue regeneration
Abstract: One of the major constraints of a successful implementation of tissue engineering therapies is the rate of conversion of cells to tissues within a scaffold in a bioreactor. Tissue engineering (TE) is a growing facet of regenerative medicine, in which patient’s own cells can be extracted and seeded in a biodegradable scaffold together with growth factors. The scaffold is placed inside a bioreactor, which mimics body internal conditions to provide adequate support to the transformation of cells into viable tissue. The scaffold plus tissue is then implanted in in vivo models at lesion site to repair damaged tissue or organ. Eventually, the scaffold degrades away, providing adhesion of tissue at lesion site to restore normal function of the damaged region. The use of mechanical/electrical/magnetic field stimulation, either isolated or combined, offers the exciting prospect of greatly enhanced rates of cells proliferation and differentiation. Optimization of the design of systems to achieve this is a multiscale problem and is best approached using computer simulation. The simulation involves the optimization of the scaffold properties and the materials used, the design of the bioreactor and the flow of nutrients, the application of stimulating fields and the integrative model of the different biological systems considered. Furthermore, the design of a system (scaffolds and bioreactor) can take advantage of the possibilities afforded by direct digital manufacturing. This MiniSymposium brings together researchers connected to the components of this problem to advance the different multiscale aspects involved.

Keywords: Cardiovascular and respiratory signal processing - Lung Sounds, Pulmonary and critical care - Pulmonary function testing & instrumentation, Cardiovascular and respiratory signal processing - Complexity in cardiovascular or respiratory signals
Abstract: Based on EUROSTAT, the diseases for the respiratory system, including conditions such as chronic obstructive pulmonary disease (COPD), pneumonia or asthma, are accountable for the 8.2% of all deaths in the European Union . In addition, the financial impact, translated into hospitalization days, medication costs, days spent in Intensive Care Units (ICUs), disability adjusted life years (DALYs), is also very high. Adding to these problems the last nine months brought the need to cope with the COVID-19 epidemic in the whole world. This virus predominantly affects lung function, and is extremely dangerous for elderly and people who have lung diseases substrate. In the context of the above mentioned problems the EU funded WELMO project is developing new wearable sensors for bioimpedance estimation based on the EIT, as well as the recording of lung sounds, and thus the possibility of feature extraction in relation with lung diseases such as COPD or asthma. Additionally, a new system (CoCross) has been developed which is capable of recording lung and heart sounds, X-Rays and ultrasounds from COVID-19 patients in the ICU, thus adding much needed data and information that can be used for advanced digital diagnostics in lung diseases. Thus, the main objectives of this special session are: • Presentation of new PoC and wearable sensor array based technologies • Presentation of the health care models applicable in the precision medicine era • Translation of user requirements in technical tools from sensors, systems, data analytics, user interfaces for achieving accurate lung digital diagnostics • New large scale databases for setting benchmarks for HR-MD development • New AI/DL/ML based analytics and feature extraction in lung biosignals and bioimages • New benchmark annotated database from COVID-19 ICU patients

Keywords: Nonlinear dynamic analysis - Biomedical signals, Physiological systems modeling - Signal processing in physiological systems, Principal and independent component analysis - Blind source separation
Abstract: A better understanding of the EHG signal before and along pregnancy, as well as at preterm and term labor, is of great interest for developing advanced methods to be use in the clinical field.

The session will include contributions of researches from different approaches, methods, institutions and countries.

Keywords: Biologically inspired robotics and micro-biorobotics - Biologically inspired locomotion, Biologically inspired robotics and micro-biorobotics - Modeling
Abstract: This paper illustrates the use of an electronic lateral superior olive (LSO) to create map of obstacles in the environment.

Keywords: Instruction and learning, Biomedical engineering curricula, Novel approaches to BME education
Abstract: Targeted reaching experimentation is the bread and butter of motor control research, such experiments are typically implemented using expensive and complicated tools such as robots. This makes them inaccessible to the average student and impractical for classroom projects. This paper presents a step-by-step tutorial as an educational tool for implementing a visuomotor adaptation experiment using Unity® game-engine, computer, and a mouse device. By the end of the tutorial, the student will know how to set up a complete visuomotor adaptation experiment that demonstrates neuroplasticity in a hand-eye coordination task. We found that this simple tool resulted in real evidence of motor learning, with significant after-effects of adaptation. This fundamental building block can be expanded to other experiments and other interface devices and lends itself well to online education, rehabilitation, and computational neuroscience. This tutorial can function as a building block for a future device or software developments that require motor training surgery, piloting, and neurorehabilitation.

Keywords: General and theoretical informatics - Unsupervised learning method, General and theoretical informatics - Machine learning
Abstract: Characterization and risk prediction of adverse events from anesthesia is an important challenge. This study uses unsupervised machine learning with the wake up safe database to create risk categories to characterize patients at risk of experiencing critical events.

Keywords: General and theoretical informatics - Supervised learning method, Health Informatics - Decision support methods and systems, General and theoretical informatics - Machine learning
Abstract: Perioperative adverse events are severe life threatening outcomes from anesthesia. We use the APRICOT dataset to build a machine learning model to identify patients at low risk for these events. Feature ranking also shows factors important to the model.

Keywords: Neurorehabilitation, Brain-computer/machine interface, Motor neuroprostheses - Robotics
Abstract: We describe the design and development of a platform for training neural activation patterns throughout a movement, from initiation to termination, to promote hand rehabilitation in stroke survivors. Brain and muscle activation are used to control a soft-hard hybrid hand exoskeleton. The platform will be used both for therapy and to carry out motor control studies to better understand hand impairment following stroke.

Keywords: Wearable sensor systems - User centered design and applications, Physiological monitoring - Instrumentation, Wearable wireless sensors, motes and systems
Abstract: The ALAS system is based on neuroengineering and machine learning tools, applied in educational settings. It records different physiological signals, such as electroencephalography, heart rate, temperature, and others, estimating mental fatigue and cognitive performance of students to identify the best learning strategies. All these features are collected using a Raspberry Pi, and then transferred to a web platform for visualization at real time.

Keywords: Mechanics of locomotion and balance, New technologies and methodologies in biomechanics, New technologies and methodologies in human movement analysis
Abstract: Abstract— This abstract summarizes user-centered design approaches in progress to develop sonified biofeedback that facilitates older adults (re)learning balance strategies during walking and turning maneuvers.

Clinical Relevance— This abstract shares trends from focus group interviews with physical therapists who provided insights and impressions about sound-based rehabilitation approaches.

Introduction— Our goal is to develop rehabilitation technology that honors an older adult’s ability to (re)learn dynamic balance strategies in partnership with a clinician. This goal will be achieved if the technology is aligned with end-user priorities and capabilities. Sonified biofeedback through sonification – displaying data through non-verbal sound – has emerged as a promising modality for balance training. It may leverage auditory-motor coupling and it allows the visual system to navigate environments during concurrent biofeedback, which is especially important during turning maneuvers. Further, sonified biofeedback can motivate patients through its potential to use musical elements including the potential to tap into the benefits of rhythmic cueing (such as singing or Rhythmic Auditory Stimulation). We started focus groups with physical therapists to learn their insights and impressions about using sound-based rehabilitation approaches with older adults to inform our design process. This abstract reports the methods used and emerging trends from this focus group research.

Keywords: Signal pattern classification, Neural networks and support vector machines in biosignal processing and classification, Physiological systems modeling - Signal processing in simulation
Abstract: The development of PVA detection algorithms is hampered by a lack of labelled data. In this study, the training of PVA detection methods was augmented with simulated data to improve the resulting performance. Although marginal, an increase in performance of about 1% was achieved by training the detection method with a ratio of 1:2 clinical to simulated data.

Keywords: Image segmentation, Image visualization, Image feature extraction
Abstract: In this article, an automatic computational method was proposed for the extraction of characteristics from thermographic images of people who practice physical activities, through digital image processing techniques, with the objective of helping the physical educator to monitor the athlete's physical performance. In the testing phase of the algorithm, satisfactory results were obtained indicating the areas most worked during training.

Keywords: Cardiovascular and respiratory system modeling - Cardiovascular control models
Abstract: Physiological closed-loop controlled (PCLC) medical devices are systems integrating medical devices with a patient's physiology through closed-loop control algorithms. Mathematical models are used throughout the development and evaluation of PCLC devices. Uncertainties about the fidelity of these models need to be addressed before achieving a reliable PCLC evaluation. To identify the best candidate model for in silico evaluation of PCLC devices, this research develops tools for assessing and comparing accuracy and predictive capability performance across multiple models.

Keywords: Physiological monitoring - Instrumentation, Health monitoring applications, Sensor systems and Instrumentation
Abstract: This paper presents the development of a biomedical device based on inertial measurement units (IMUs) that estimates (keeps tract of) the physical work demands of the upper limb by evaluating the amplitude and frequency of arm elevations. The arm elevation is obtained with a data fusion algorithm using accelerometer and gyroscope data. The physical work demands are evaluated by four different metrics which give information on the time spent with the arms in elevation and the number of elevations.

Keywords: Optical imaging, Multimodal imaging, Image analysis and classification - Machine learning / Deep learning approaches
Abstract: Accidental removal of the parathyroid glands during thyroid surgery can occur, potentially leading to postoperative hypocalcemia. We tested a computer-aided detection (CADe) using the co-registered RGB/NIR camera and ex-vivo thyroid tissue specimen. The average precision was significantly higher in the paired RGB/NIR data (0.995) than NIR (0.339) and RGB (0.954) data alone. The proposed CADe may increase the parathyroid detection rates clinically.

Clinical Relevance—Parathyroid glands control calcium levels in the body and are crucial to maintaining homeostasis. Accidental excision of the gland can bring about negative consequences, thus precise and fast detection of normal glands can improve the surgical outcome during thyroid surgery.

Keywords: Signal pattern classification
Abstract: This study conducted a pattern recognition study on human crawling movements. 10 healthy adults participated in the crawling data collection experiments. The surface electromyography (sEMG) signals were recorded from 16 limbs and trunk muscles related to crawling movement and the pressure signal was recorded from left palm. A pattern recognition model based on sEMG signals and linear discriminant analysis (LDA) method was proposed and implemented for the 8 defined inter-limb coordination modes. High accuracies were obtained for participant-dependent, multi-participant, and participant-independent schemes.

Keywords: Image analysis and classification - Machine learning / Deep learning approaches, Novel imaging modalities
Abstract: This paper introduces an ophthalmic disease detection system2 that allows users to take a fundus image and detect common eye diseases using a smartphone. The detection is based on a convolutional neural network to classify the various retinal diseases by fundus images. The overall accuracy was 74%, and AUC was 0.93. Grad-CAM was generated to provide heatmaps with visual explanations of the prediction.

Keywords: Modeling and analysis, IoT sensors for health monitoring, Sensor systems and Instrumentation
Abstract: Breast cancer risk assessment based on the patient health information results in a breast cancer risk probability. The smartphone compression-induced sensing (SCIS) system generates tactile images, and these images estimate the deformation index ratio of a lesion. We estimate the breast cancer risk using the patient health record and the deformation index ratio. For a small number (eight) of patients, the malignancy of the tumor was classified based on relative risk scoring and patient health information with the sensitivity and specificity of 100%.

Keywords: Brain-computer/machine interface, Motor neuroprostheses, Neurological disorders - Stroke
Abstract: We propose a novel Event-related Desynchronization (ERD) detection method that provides a motor imagery cue when the mu band power at rest is kept larger than a predetermined threshold for a certain period. In this study, we investigated the optimal resting threshold and its duration, and its effect on improving the measured ERD. In both healthy adults and patients, the ERD power of the proposed method was significantly increased compared to the conventional method without a threshold, and the fluctuation of the ERD power across trials was also significantly reduced.

Keywords: Image feature extraction, Image analysis and classification - Machine learning / Deep learning approaches
Abstract: Remote health monitoring of elderly adults is important. Since Body Mass Index (BMI) is highly correlated to many diseases, it is a good indicator for health monitoring. This paper presents new approaches to estimate BMI from 2D body images and videos. Besides proposing new anthropometric features, we create two visual-body-to-BMI datasets. The simulation results confirm the excellent performance of the proposed method comparing to other methods.

Keywords: Time-frequency and time-scale analysis - Empirical mode decomposition in biosignal analysis, Adaptive filtering
Abstract: Electrocardiogram (ECG) signal is an important physiological signal. In this paper, we first present a novel method based on EMD and adaptive filter for the removal of baseline wander (BW) and power line interference (PLI) in ECG signal. We then extend the method to the complex scenario where ECG is contaminated by four most common noises, PLI, BW, electrode motion artifact (EM) and muscle artifact (MA). Several efficient methods are proposed and proposed Parallel EMD adaptive filter structure yields the best SNR improvement on the MIT-BIH arrhythmia database.

Keywords: Cardiovascular and respiratory signal processing - Lung Sounds, Cardiovascular, respiratory, and sleep devices - Diagnostics, Respiratory transport, mechanics and control - Pulmonary mechanics in disease
Abstract: Respiratory illness is experienced by many patients infected with SARS-CoV-2 (COVID-19), with a dry cough one of the most common symptoms. Analysis of cough sounds provides a way to diagnose, assess and monitor respiratory diseases. This study analyzes differences in cough signals recorded using a smartphone in 24 patients hospitalized with COVID-19 at different disease stages. Frequency variability and the percentage power in three bands (0-250 Hz, 1000-1250 Hz and 3750-4000 Hz) in the first cough sound were higher in patients at their most severe stage of the disease compared to patients prior to this point in the disease trajectory. Several frequency-based features of the second cough sound were higher in female patients compared to males. Results presented support the prognostic potential of cough sounds.

Keywords: Cardiovascular and respiratory signal processing - Blood pressure measurement, Vascular mechanics and hemodynamics - Arterial pressure in cardiovascular disease, Cardiovascular and respiratory signal processing - Cardiovascular signal processing
Abstract: Parametric functions of the brachial artery compliance curve of an Oscillogram model were compared using extensive patient data. The exponential-linear function was found to be the best amongst eight functions in terms of accuracy and convenience.

Keywords: Physiological systems modeling - Signal processing in physiological systems, Physiological systems modeling - Signal processing in simulation, Kalman filtering
Abstract: This paper develops an extended Kalman filter (EKF)-based location tracking method for a wireless capsule endoscope (WCE) in order to achieve precise localization accuracy and computation cost acceptable for real-time tracking. This study compares two kinds of EKF based on constant velocity(CV) and interactive multiple models (IMM) transition functions and evaluates the performances based on computer simulations. Our evaluation results demonstrate that the proposed EKF-based tracking achieves the estimation accuracy of around 8 mm with both models, which means that the proposed method is applicable to WCE movement tracking.

Keywords: Health Informatics - Human factors (ergonomics) in health information systems, Health Informatics - Technology and services for home care and assistedl living, Sensor Informatics - Wireless sensors and systems
Abstract: This study aimed to examine an effective reading aloud method for Braille learning materials with a reading aloud function. We have developed a prototype system that can identify the position of Braille and finger using a tablet computer, and obtain audio information when the user touches Braille with a finger. One simple check of the speechreading accuracy of the prototype system was conducted. As a result, it was confirmed that Braille could be read aloud using the prototype system.

Keywords: Haptic interfaces, Tactile displays and perception
Abstract: This study aimed to examine stiffness discrimination characteristics of elastic materials in qualified acupuncturists for establishing an objective evaluation method for palpation skills. As a result, it was confirmed that the discrimination rate for elastic materials in a certain range of hardness tended to be low.

Keywords: Modeling and analysis
Abstract: Remote, sensor-based tests show promise in augmenting standard clinical assessments of functional ability in neurological diseases. Ensuring accordance with test instructions remains, however, a challenge for such tests. We present here a novel methodology to assess accordance with test instructions for improving the collection of reliable data with unsupervised, sensor-based tests in digital health studies.

Keywords: Health Informatics - Electronic health records, Health Informatics - Personal health records, Health Informatics - Patient tracking
Abstract: Urination prediction is critical for elderly care, and it has been shown that urinary volume influences significantly vital signs. However, no studies have reported the relationship between different urination desire levels and vital signs, nor the predictability of urination desire level by vital signs and HRV (Heart Rate Variability). This paper presents the preliminary results that fills up the gap above.

Keywords: Cancer detection and diagnosis technologies, Diagnostic devices - Physiological monitoring, Clinical engineering - I think this is a bit too gneric, but if it was ours to start with, then no worries
Abstract: Tissue biopotential offers a method of identifying ovarian and breast cancer, which affects 1 in 78 women and 1 in 9 women respectively. Since cancerous cells differ from normal cells in their electrical properties, the construction of a biopotential sensing device shaped like a catheter can be used with endoscopes for various minimally-invasive procedures.

Keywords: Computer-assisted surgery, Robot-aided surgery - Remote surgery systems / telesurgery, Surgical robotics
Abstract: A new voice interactive control system to drive a surgical assistant manipulator in clock position performed by a surgeon in a sterilized area has developed. The system was constructed using speech recognition engine Julius and forceps manipulator, and the tracking performance was evaluated.

Keywords: Biomedical engineering curricula, Teaching design
Abstract: Advances in new technologies have changed traditional education methods in the field of biomedical engineering dramatically in the past few decades. The expansion of technologies into different disciplines of engineering sciences and biology attracts many students with different educational backgrounds to pursue higher education in biomedical engineering. Directed Research Study is a methodology of teaching that allows students to choose and focus on a specific knowledge area they interest to learn while understanding other core concepts of biomedical engineering principles integrated into one teamwork project.

Keywords: Data mining and big data methods - Machine learning and deep learning methods, Physiological systems modeling - Signal processing in physiological systems
Abstract: This work presents an end-to-end DNN-based system for autism severity estimation of Hebrew-speaking children. The system includes a speaker diarization algorithm for identification of child speech segments and an algorithm for estimating the severity of core autism symptoms in individual children. Audio recordings of 132 children who completed an Autism Diagnostic Observation Schedule (ADOS) assessment were analyzed. The ADOS severity scores predicted by the system were significantly correlated with the true severity scores, yielding an average correlation of 0.581 and a Normalized Root Mean Square Error (NRMSE) of 0.254.

Keywords: New technologies and methodologies in medical robotics
Abstract: Endoscopes are essential equipment for the screening of gastrointestinal tract. Due to their lower cost and complexity, single-use endoscopes are recently more and more adopted in the clinical practice. However, the use of disposable endoscopes arises environmental impact and a few components, e.g. camera and optics, could be reused also due to their cost if of high quality. To mitigate this issue, we developed a customized waterproof magnetic connector (WaMaCo), acting as possible element to be utilized in between a disposable and a reusable portion of new generation endoscope.

Keywords: Models of medical devices
Abstract: Pressure recording equipment with different dynamic properties may influence parameters derived from blood pressure waveforms and pressure-derived cardiac output measurements. In this work, we propose a computational approach for modeling different components of a pressure recording system to evaluate their effect on derived hemodynamic measurements.

Keywords: Neural stimulation - Deep brain
Abstract: High-frequency stimulation (HFS) of electrical pulses within brain has been developed for treating various brain diseases. However, the underlying mechanisms of the therapy are not completely clear. To investigate the effect of HFS on local inhibitory circuits, we applied antidromic-HFS (A-HFS) at the efferent fibers (alveus) of pyramidal neurons and applied test stimulations of paired-pulse at the afferent fibers (Schaffer collaterals) to induce orthodromic population spikes (OPS). Results showed that the paired-pulse depression (PPD) of OPS disappeared during A-HFS, indicating a suppression of local inhibitions by axonal HFS of efferent fibers.

Keywords: Neural stimulation - Deep brain
Abstract: High-frequency stimulation (HFS) of electrical pulses has been used in deep brain stimulation (DBS) to treat certain brain disorders successfully. However, the effects of HFS on neurons are not completely clear yet. In this study, we utilized recordings of electrode array and the method of current source density (CSD) analysis to investigate the changes of neuronal reactions around somata when their axons were under sustained antidromic HFS (A-HFS). The result showed that, in addition to the axonal blockage reported previously, the propagation speed of current sink around somata significantly decreased during the A-HFS. The finding indicated a change at the neuronal soma caused by sustained axonal HFS.

Keywords: Neural stimulation, Brain physiology and modeling
Abstract: High-frequency stimulation (HFS) of pulses has been used in deep brain stimulation (DBS) to treat neurological diseases. The inter-pulse-interval (IPI) of HFS is one of the critical parameters to dominate the stimulation efficacy, not only because IPIs determine pulse rates but also because the IPIs are commonly too short to allow neuronal state to recover completely during sustained HFS. Therefore, varying IPI can generate various neuronal reactions. However, the recovery of neuronal excitability along with the time of IPI is not clear. In this study, we investigated the recovery progress by utilizing a computational model of a myelinated neuron. The results show a non-linear recovery of the neuron to obtain re-firing ability during sustained HFS, which is dominated by the non-linear dynamics of sodium channels in the axonal membrane.

Keywords: Neural stimulation - Deep brain, Neural signal processing
Abstract: Electrical pulse stimulation with varying inter-pulse-interval (IPI) provides a new stimulation mode to improve the therapy of deep brain stimulation. However, the fidelity of neuronal reactions to varying IPI remains unclear. In this study, we investigated the differences in the reproduction of neuronal responses to an identical pulse sequence with randomly varying IPI applied at efferent and afferent fibers of rat hippocampal CA1 region in-vivo. Results showed that the stimulation of efferent fiber was able to induce a sequency of antidromically-evoked population spikes (APS) with a high fidelity among different rats, whereas the stimulation of afferent fiber induced various sequences of orthodromically-evoked population spikes (OPS) due to the participation of synaptic transmissions. The results suggest that the immediate neuronal reactions to stimulations of varying IPI are determinative, but the reactions of post-synaptic neurons are variable.

Keywords: Neural stimulation - Deep brain, Brain physiology and modeling - Neuron modeling and simulation
Abstract: High-frequency stimulation (HFS) of pulses has been used to treat some neurological disorders, but its action mechanisms are not completely clear. Here we built a computational model to investigate the role of GABAergic inhibitions on neuronal reactions in the initial period of orthodromic HFS in rat hippocampal CA1 region. Results showed that the computational simulation reproduced the neuronal responses in real rat experiments. The change of GABA-A current from inhibition to excitation and the partial inactivation of sodium channels induced by HFS inputs resulted in a triphasic change of the firing of pyramidal cells: inhibition-excitation-inhibition. The study provides new clues to reveal the mechanisms of HFS.

Keywords: Biomaterial-cell interactions - Engineered vascular tissue, BioMEMS/NEMS - Angiogenesis and vasculogenesis, Microfluidic techniques, methods and systems
Abstract: Abstract— A novel, double vessel chip exposed to three dimensional cyclic stretch, varying extracellular matrix stiffness and biochemical signals is presented in this work, in an effort to better mimic the in-vivo tissue environment and to provide a platform combining multiple mechanical and biochemical stimuli.

Clinical Relevance— The integration of the blood microvasculature is one of the key challenges for the next generation Organs-on-Chip (OOC) platforms to enable the recapitulation of intricate biological processes involving the circulation of immune or cancer cells.

Keywords: Magnetic resonance imaging - Pulse sequence, Image reconstruction and enhancement - Parametric image reconstruction, Iterative image reconstruction
Abstract: Abstract— This paper investigates the effect of different sequence designs on T1, T2 and proton density (PD) estimation within a simulation-based MR reconstruction framework. Clinical Relevance— This approach offers an accurate estimation of parameter maps using fast scans.

Keywords: Instruction and learning, Teaching design, BME and global health
Abstract: Biomedical Engineering(BME) is a branch of applied sciences that combines a wide range of scientific knowledge areas and engineering principles to apply and obtain engineering solutions in the field of medicine and biology which attract several learners in bachelor and master programs to learn and explore distinct topics of sciences and engineering technologies that are associated with the field of biomedical engineering. Focusing on main subjects between this wide range of different fields can be overwhelming sometimes for many learners to choose what may help them to focus on a specific knowledge area with the ability to have a full understanding of the integrated processes in the field of biomedical engineering.[1]

Keywords: Models of medical devices, Models of organs and medical devices - Inverse problems in biology, Organ modeling
Abstract: A biophysical model is envisioned to assist the radiologist in the percutaneous thermal ablation procedure of liver tumors, by predicting in real time the thermal deposition from a microwave (MW) applicator. This model is integrated in the imaging platform, in addition to the segmentation and automatic registration tools. A retrospective analysis on a cohort of 21 patients shows an improvement of model prediction compared to manufacturer data and highlights the cooling effect from vasculature on the thermal dose. It also demonstrates the importance of using the appropriate contrast phase selection and dosage for an accurate vasculature segmentation. At the end, this integrated model provides a solution to tackle multiple contrast enhanced scans for vasculature segmentation, real time adjustment of the applicator in 3D and visualization of the predicted ablation overlaid on the CT images. Hence it can help in the standardization of the thermal ablation procedure towards less experienced radiologists.

Keywords: Medical technology - Design and development, Medical technology - Product development process, Medical technology - Safety
Abstract: Full-face protective glasses was redesigned and tested their aerosol protective capability and usability in medical operations. With better face coverage and protective angle adjustment, the new design can significantly improve aerosol protective capability.

Keywords: Image classification, Cardiac imaging and image analysis
Abstract: Hypoxia Inducible Factor-1α (HIF-1α) is thought to be a critical factor associated with heart failure, but its function remains unclear. In this study, we developed a machine learning approach to evaluate the effect of HIF-1α in cardiomy- ocytes of HIF-1α knockout mice under heart failure. Our model suggests that HIF-1α affects the contraction of cardiomyocytes when heart failure occurs.

Keywords: Health monitoring applications, Wearable wireless sensors, motes and systems
Abstract: This study aimed to investigate the changes in electromyography (EMG) and mechanomyography (MMG) due to the changes in pedaling cadence using an MMG/EMG hybrid transducer. The EMG signal increased and MMG signal decreased with increasing cadence, and their signals showed an inverse relationship.

Keywords: Medical technology - Design and development, Medical technology - Innovation
Abstract: Leadless pacemakers are self-contained devices implanted entirely within the target heart chamber for single-chamber pacing. This feasibility study demonstrates leadless dual-chamber pacing using wireless, bidirectional communication between paired leadless pacemakers implanted in the right atrium and right ventricle.

Clinical Relevance— Leadless pacemakers can mitigate complications associated with transvenous, single-chamber devices. Using a novel wireless communication modality, true dual-chamber leadless pacing is now possible.

Keywords: Physiological systems modeling - Signal processing in physiological systems, Physiological systems modeling - Signal processing in simulation, Physiological systems modeling - Systems identification
Abstract: Estimation of changes in arterial resistance and compliance during a drug intervention can give insights on its differential action on-resistance versus compliance vessels. This way every drug can be reassessed for its differential action. The study demonstrates an objective method of quantifying arterial resistance and compliance from simultaneous intra-arterial pressure (from radial artery) and Doppler-based volume flow recordings (from the distal end of the radial artery) in an ICU setting. The study was performed in a stable ICU patient who already had an arterial cannulae placed as a standard of care. Intra-arterial pressures from the radial artery, its Doppler images, and the Doppler audio signal were captured synchronously on a multi-channel validated data acquisition system (CMCdaq). The pressures and the audio were sampled at 4kHz (> twice the Doppler-shift frequency). A video of Doppler imaging of the radial artery was recorded on CMCdaq using an externally mounted camera at a fixed frame rate of 30 fps. The ultrasound probe insonation angle was 45°. From the Doppler-shift frequency (from DFT of Doppler audio signal), the linear velocity of blood flow was calculated. Product of linear-velocity and area of radial artery cross-section gave volume flow. Arterial vascular impedance (Z) was then computed as the ratio of harmonic terms of intra-arterial pressure to the corresponding harmonic terms of volume flow in the frequency domain. Blood flow in arteries was modeled as a linearized lumped electrical circuit (three-element Windkessel) with (a) time-varying pressure pump (heart), (b) compliant vessels as a parallel combination of resistance (R1) and capacitance (C), (c) small artery/arteriolar load by a pure resistance (R2), and inertance of blood as (L). From the Laplace domain analysis of the model, impedance Z = f (R1, R2, C, L). Four impedance spectral peaks were then used to solve for R1, R2, L, and C.

Keywords: Physiological systems modeling - Signal processing in physiological systems, Adaptive filtering, Parametric filtering and estimation
Abstract: Acquisition of high signal-to-noise (SNR) electrically-evoked Compound Muscle Action Potentials (CMAP) during acute, invasive experiments can be a challenge due to motion and stimulation artifacts in addition to biological noise contamination, and electromagnetic interferences. This study compares the gold-standard bandwidth limit filter (4th order Bandpass Butterworth) against a combined strategy using an adaptive filter with a wavelet denoising approach to process electromyographic (EMG) signals, extracting their CMAPs from a dataset of Pelvic floor muscle (PFM) muscle contraction evoked via electrical stimulation in a rabbit.

Keywords: Neural interfaces - Implantable systems, Neural interfaces - Microelectrode technology, Brain-computer/machine interface
Abstract: Abstract— A method is presented to fabricate complex 3-dimensional (3D) silicon objects using a simple three step process composed of commonly available fabrication procedures. The combination of greyscale photolithography, silicon etch, and Silicon-on-Insulator wafer lift-off allows for near-arbitrary 3D structure fabrication without overhangs with a vertical resolution as fine as 60 nm and horizontal resolution of 600 nm using the equipment described. Clinical Relevance— The fabrication of stiffening devices for neural implants is time consuming due to the complexity of 3D shanks designed to reduce insertion force. This method allows a faster fabrication process for such stiffeners as well as similar objects.

Keywords: Neural stimulation - Deep brain, Neurological disorders, Neural stimulation
Abstract: The effects of different deep brain stimulation (DBS) frequencies on lower limb movement in patients with Parkinson’s disease are not clear. In this study, we analyzed the clinical evaluation scores and biomechanical paraments in lower limb motor examinations under different DBS frequency settings. Both high-frequency and variable-frequency stimulations showed obvious improvements in motor symptoms. Considering the effectiveness in controlling axial symptoms, variable-frequency stimulation might be a potential stimulation setting in practice.

Keywords: Wearable robotic systems - Orthotics and Exoskeletons, New technologies and methodologies in medical robotics, Biologically inspired robotics and micro-biorobotics - Modeling
Abstract: Abstract—In the past years, artificial muscles (and or soft-actuators) have gained attention to be used in rehabilitation robots due to their inherent compliance. Twisted and coiled polymers (TCP) artificial muscles are one of the prominent alternatives. A recently introduced method to activate the TCP is to use an embedded metallic wire. This work analyses the strain and the response time of TCP actuators with different Nichrome (NiCr) wire diameters.

Keywords: General and theoretical informatics - Natural language processing, General and theoretical informatics - Data mining, General and theoretical informatics - Machine learning
Abstract: Bias in letters of recommendation can influence applications to surgical fellowships. Machine learning can be used to quantify differences in letters of recommendation by race and gender. We find differences in polarity and intensity of emotions, with differences in anger being statistically significant.

Keywords: Ultrasound imaging - Cardiac, Image reconstruction and enhancement - Machine learning / Deep learning approaches
Abstract: Abstract— Fast ultrasound imaging is required for cardiac ultrasound imaging. Divergent wave imaging (DWI) can be used to increase imaging frame rate but produces lower quality images. In this work, a Convolutional Neural Network (CNN) is used to improve DWI images to have a similar quality to the conventional single-line acquisition (SLA) images. The CNN was demonstrated to improve DWI images by a mean MS-SSIM of 0.26 for 500 testing images.

Clinical Relevance— Faster and higher quality cardiac ultrasound images can provide more information enabling more accurate assessments of cardiac structure and function.

Keywords: Parametric filtering and estimation, Adaptive filtering, Physiological systems modeling - Signal processing in simulation
Abstract: We present a source localization method for electroencephalography (EEG) data based on the structure of the generalized sidelobe canceler (GSC) in which we use an adaptive blocking matrix (ABM) to determine the optimum reduction of the rank. Our realistic simulations show that we can achieve a more consistent source localization than the full-rank multi-source activity index (MAI) when brain sources are embedded in high background activity.

Keywords: Neural stimulation
Abstract: A sensitivity analysis of the main parameters of cardiac-synchronized vagus nerve stimulation was performed to determine their importance on the acute provoked chronotropic effects in bilaterally vagotomized rabbits. The preliminary results suggest a preeminent importance of current amplitude and pulse width, while other parameters may play a minor role.

Keywords: Health Informatics - Healthcare modeling and simulation
Abstract: The Japanese government provides health guidance for metabolic syndrome to prevent chronic diseases, mainly weight loss instruction to reduce visceral fat. When health guidance recipients undergo a visceral fat measurement test by X-ray CT, they are shown the state of their visceral fat accumulation and are motivated to lose weight. However, it is difficult to easily measure visceral fat because an X-ray CT scanner is large and expensive. Therefore, We developed a visceral fat simulator that generates and displays the shape of visceral fat before and after weight change by analyzing abdominal CT images. We confirmed that this simulator can support effective health guidance that focuses on visceral fat by simulating realistic visceral fat shapes before and after weight changes.

Keywords: Microfluidic techniques, methods and systems, Microfluidic applications
Abstract: Minimally invasive cell sorting techniques have a lot of application. Microfluidics is one of techniques to be applied to cell sorting. Cell sorting technology has a lot of application. Detection of target cells leads to diagnostic technology. Collection of target cells can be applied to regenerative medicine by tissue engineering. Asymmetric surface electrodes were manufactured in the flow channel by micromachining techniques for analyzing dielectrophoretic kinetics of biological cells for micro-invasive sorting. A micro flow-channel with a pair of titanium-coated (200 nm thick) surface electrodes (a triangular electrode and a rectangular reference electrode) was made by photolithography technique. The suspension of myoblasts (C2C12: mouse myoblast cell line) was injected into the channel, and the movement of each flowing cell was analyzed with the microscopic movie image. Experimental results show that the movement of each cell depends on several parameters: the alternating electric field (waveform, frequency, and amplitude), the diameter of cell, the deformation ratio, and the direction of the major axis.

Keywords: Integrated sensor systems, Textile-electronic integration, IoT sensors for health monitoring
Abstract: Monitoring body core temperature is one of the effective ways to identify heat-related risks of conditions such as heat stroke. However it is difficult to measure it directly during activity. We have developed a wearable device that can monitor a user's heart rate and the temperature and humidity inside their clothing during activity. Here, we investigated a method for estimating body core temperature from sensor data, and compared the estimated and actual values through clinical experiments.

Keywords: Neural stimulation, Sensory neuroprostheses - Visual, Sensory neuroprostheses
Abstract: Cortical voltage responses to repetitive micro-stimulation pulses in the primary visual cortex layer II/III were examined in the mouse cerebral slices. The stimulus-induced excitations were composed of the accumulated and the responsive components, which were attributable to those with the non-synaptic and synaptic mechanisms.

Keywords: Health Informatics - Mobile health, Health Informatics - Patient tracking, Health Informatics - Information technologies for healthcare delivery and management
Abstract: Prior studies have shown that personalized, tablet-based home practice is effective in the recovery of word finding in aphasia patients. A mobile application was developed which is used as a plugin-platform for the test of different algorithms used in patient feedback and collection of objective parameters (i.e. naming latency (NL)). The application provides direct feedback to the patient and collects training data which the therapist can review and use to adapt the therapy-plan.

Keywords: New technologies and methodologies in human movement analysis, Modeling and identification of neural control using robotics, Mechanics of locomotion and balance
Abstract: A continuous movement does not mean that the inherent control is continuous; the central nervous system may control consequent movements in an intermittent way. This study presents a continuous motion of the ankle in the steady-state running of high-class athletes and highlights the predictive event-driven regulation of ankle equilibrium and ankle impedance. The ankle is the critical joint between the leg and the foot that helps to control physical interactions for propulsion, shock absorption, and balance of the body. Information on the intermittent control at the ankle of high class athletes would be beneficial for mechanical interventions (for example, footwear) that enable faster/economical running by beginner and middle-class runners. The findings would be useful not only for intervention coaching to develop talents but also from clinical and injury perspectives. Our results reaffirm the significance of the foot-strike and toe-off events at the level of motor commands.

Keywords: Sensor systems and Instrumentation, Integrated sensor systems, Modeling and analysis
Abstract: At present, few studies have been conducted on the relationship between playing electronic sports (esports), cognitive skills, and heart rate variability (HRV) indices. In this study, 20 healthy adult males were monitored to examine the changes in cognitive skills and HRV indices using the Stroop test and RR intervals whilst playing esports. The results suggest that playing esports could improve cognitive skills, and the changing LF/HF ratio could provide insight into the trend of improvement. Further investigation, including other HRV indices, is therefore warranted.

Keywords: Neuromuscular systems - EMG processing and applications, Neuromuscular systems - EMG models
Abstract: With the aim of constructing a myoelectric control system that can handle the movements of daily tasks to improve the controllability of prosthetic devices, electromyography (EMG) signals of the forearms of four subjects with normal limbs during typing movement were acquired, and the best combinations of feature and classifier configurations were investigated. When Smoothing EMG signals were selected as features, and support vector machines (SVM) were used as discriminators, the best key identification accuracy (66% for one subject and 48% for the mean) was obtained.

Keywords: Magnetic resonance imaging - Diffusion tensor, diffusion weighted and diffusion spectrum imaging, Brain imaging and image analysis, Fetal and Pediatric Imaging
Abstract: In our preliminary study, we use diffusion tensor MRI to evaluate the corticospinal tract (CST) and a quantitative force measurement assessment to characterize hand impairment severity in individuals with hemiplegic cerebral palsy (HCP). There were significant correlations between diffusivity measures of the CST and severity of paretic hand weakness and mirror movements. These preliminary findings reveal a link between atypical CST development and hand impairment severity in HCP.

Keywords: Cardiovascular and respiratory signal processing - Heart Rate and Blood Pressure Variability, Cardiovascular regulation - Heart rate variability, Cardiovascular, respiratory, and sleep devices - Wearables
Abstract: Heart rate variability (HRV) analysis in wristband PPG device is gathering interest for emotion related and well-being application; however, motion artifact is critical issue. We verified robust framework against motion artifacts for detecting HRV by motion-tolerant instant heart rate (IHR) detection algorithm [1]. The results suggested high correlation between HRV calculated from PPG by our method and HRV calculated from ECG, and the improvement in stress/resting classification in PPG even when motion artifact occurred.

Keywords: Signal pattern classification, Physiological systems modeling - Signal processing in physiological systems, Time-frequency and time-scale analysis - Time-frequency analysis
Abstract: In this study, a support vector machine (SVM) system for the detection of COVID-19, based on patients' coughing signals extracted from remote audio recordings is presented. The system includes cough detection and segmentation, followed by separation between the different cough phases. Acoustic features from each individual phase were extracted and used for the classification of Corona patients. Audio recordings were collected from crowdsourced VOCA dataset and include recordings from 2,954 subjects, of whom 26 are diagnosed as positive for COVID-19. The predicted results from the SVM classifier yield sensitivity of 0.796 and F1 score of 0.656.

Keywords: Image analysis and classification - Machine learning / Deep learning approaches
Abstract: Microelectronic retinal prostheses can evoke artificial visual percepts to blind individuals. Among several factors, expression of original images using limited resources (i.e., small number of pixels and low gray scales) is critical for improved recognition of artificial vision. To assess effectiveness of new image processing techniques, prosthetic researchers have performed psychophysical tests with normally-sighted subjects. Here, we evaluated facial recognition task performances of two machine learning (ML) models for phosphene images in various conditions, representing several electrical stimulation cases. Both ML models showed that similar trends in correct response ratios, comparable to those of human subjects reported in an early psychophysical study. Our results suggest ML approaches would be useful for evaluating newly-developed artificial vision image processing techniques. In particular, the use of ML approaches is expected to substantially reduce time and cost needed for psychophysical testing of phosphene images.

Keywords: Physiological monitoring - Novel methods, Wearable body sensor networks and telemetric systems, New sensing techniques
Abstract: Photoplethysmography (PPG) -based heart rate measurement at anterior region of forearm (palm side) was explored for wrist-worn PPG device. Our motion artifact reduction framework showed high robustness at near the wrist where the motion artifact had large impact on PPG signal. Furthermore, our implemented framework on edge device was verified running in real-time processing. These results suggested realizing various applications such as healthcare by PPG-based heart rate measurement at anterior region of forearm.

Keywords: Cardiovascular regulation - Heart rate variability, Cardiovascular regulation - Autonomic nervous system
Abstract: Truck accidents caused by health-related conditions such as fatigue are increasing. Therefore, we are developing technology that uses autonomic nervous functions to monitor the health condition of active drivers and detect dangerous conditions that lead to accidents. To handle autonomic nerve functions (ANF) in the dynamic environment of driving, we propose a time-series accident risk prediction method that shortens the time range for calculation, normalizes to eliminate individual differences, and indexes environmental factors. We achieved an 88.8% accuracy of Recall and a 0.84 area under the curve (AUC).

Keywords: Image visualization
Abstract: To visualization the flow of neo-sinus after aortic valve implantation, we created a new particle image velocimetry technique and pulsatile flow and pressure circulation system.

Keywords: Cardiovascular and respiratory system modeling - Cardiovascular Disease, Vascular mechanics and hemodynamics - Vascular Disease
Abstract: We developed an accelerated durability test system which can apply radial diameter-change load to stents by producing physiologic aortic pulsatile pressure differences.

Keywords: BioMEMS/NEMS - Tissue engineering and biomaterials
Abstract: To produce a decellularized small-caliber aortic vascular graft utilizing microwave and pulsatile circulation decellularization treatment, and to evaluate residual DNA and mechanical properties of the decellularized bovine arteries.

Keywords: Neural networks and support vector machines in biosignal processing and classification, Data mining and big data methods - Machine learning and deep learning methods, Data mining and big data methods - Biosignal classification
Abstract: This work focuses on predicting near-term onset of hypotension prior to onset using convolutional neural networks. Based solely on the arterial blood pressure curve, our initial attempt can predict an onset with 60% sensitivity and 80% specificity 5-15 minutes before onset. Clinical relevance: Hypotension is common during large surgery. By identifying and treating hypotensive episodes early, preferably even before onset, hypotension and its associate post-surgery complications are reduced. Even a prediction with 80% sensitivity/specificity is valuable for the anesthesiologist.

Keywords: Image analysis and classification - Machine learning / Deep learning approaches, Image feature extraction, Machine learning / Deep learning approaches
Abstract: Tracking the movement of the muscle-tendon junction is important for quantifying how muscles and tendons drive human movement. However, this remains an arduous task. Deep learning methods have recently shown promise for tracking, but there are limits to how well these networks generalize to novels subjects and tasks. Muscle-tendon junction movement is dependent upon muscle activation and joint-level motion. Therefore, we sought to determine if supplementing current networks with ankle angle and electromyographical (EMG) data would improve tracking. We observed a 20% increase in tracking accuracy for novel subjects when we incorporated EMG and ankle angle data into a current deep-learning network. Our results indicate that extending current methods to include data beyond the ultrasound image can significantly improve their accuracy.

Keywords: Neural interfaces - Implantable systems, Brain-computer/machine interface, Neural stimulation
Abstract: Neural interfaces have been widely used to understand the function and connectivity of the brain. Many different types of neural electrodes have been developed for the purpose from single neuron recording to cortical activity recording. In this study, we fabricated a hybrid electrode array, in which two different types of microelectrodes are combined for multi-site recordings. We also conducted preliminary animal experiments for validation of the developed hybrid electrodes.

Keywords: Brain imaging and image analysis, Image feature extraction, Multivariate image analysis
Abstract: Earlier studies indicated that fMRI preprocessing methods influence properties of intrinsic connectivity networks (ICNs). In this pilot study, we examined the effects of spatial smoothing on the network dimensionality and spatial maps. Resting state BOLD fMRI data were acquired from healthy participants with a 3.0T MRI scanner. During preprocessing, various levels of spatial smoothing were applied to the data using an isotropic Gaussian kernel with full width at half maximum (FWHM) sizes 0 to 12 mm with a step of 2 mm. Independent component analysis (ICA) was applied to derive ICNs. Results revealed that the level of spatial smoothing clearly affects the network dimensionality, intensities of spatial maps, and peak voxel location. Using minimum description length (MDL) criteria, dimensionality generally decreased as smoothing kernel size increased. In contrast, entropy-rate based order selection indicated a general increase in model order as smoothing kernel size increased. Intensities of spatial maps, which are associated with the cohesiveness and connectivity inside the network, decreased in most ICNs, including the default-mode and salience networks, as the smoothing kernel size decreased. These findings provide a preliminary insight into the effects of spatial smoothing on model order and spatial maps.

Keywords: Brain imaging and image analysis, Functional image analysis, Multivariate image analysis
Abstract: An important practical question, given various sources of noise in resting state fMRI images, is how much data to acquire to get identifiable and stable independent component analysis (ICA) networks in MRI scans. Here, we investigate how ICA dimensionality, stability of the estimated independent components, and the features of the extracted networks are affected by scan length. Our results show that the length of scan clearly influences the network dimensionality, spectral power, and functional network connectivity (FNC). Using minimum description length criterion, we observed that dimensionality decreased as the scan length decreased from 12 to 2 minutes. In addition, due to decreased dimensionality in shorter scans, some networks such as the ventral attention and subcortical networks could not be extracted. But the networks that could be identified in shorter scans including the anterior and posterior default-mode, salience, and self-referential networks were stable (cluster quality index Iq > 0.8). Spectral parameters (the dynamic range and fractional of amplitude of low frequency fluctuation; fALFF), and FNC also changed with the scan length with shorter scans having lower values. Future studies using larger sample sizes are warranted to further assess the implications of the scan length on ICA outcome measures.

Keywords: Wearable low power, wireless sensing methods, Bio-electric sensors - Sensing methods, Chemo/bio-sensing - Biological sensors and systems
Abstract: This work demonstrated using vibrational motors to provide the double-click stimulus signal via the peripheral nerve system to measure the sensory gating (SG) effect. The system is portable and benefits from doing SG monitoring without adding workloads to the visual or auditory system.

Keywords: Time-frequency and time-scale analysis - Time-frequency analysis
Abstract: This paper compares whether PPG signals can replace ECG signals. With that method, the Welch's method, Lomb-Scargle, and Autoregressive algorithms are compared to calculate the PRV from the PPG signal and compare the similarity with the HRV-PSD.

Keywords: Nonlinear dynamic analysis - Biomedical signals, Independent component analysis
Abstract: In recent years, measuring PPG using a smartphone camera has been actively researched. In clinical practice, PPG is used to monitor Heart Rate Variability(HRV) and Blood Pressure (BP) [1, 2]. Therefore, there are many studies that improve the accuracy of HRV from PPG detected from smart phone [3]. However, the shape of PPG extracted from the video is different from PPG measured by a medical device. It is difficult to evaluate the elasticity of blood, which can be seen in the shape of PPG. Therefore, we proposed an algorithm that reconstructs the morphology of PPG.

Keywords: Wearable low power, wireless sensing methods, Wearable sensor systems - User centered design and applications, Wearable wireless sensors, motes and systems
Abstract: The paper introduces an embedded flow rate measurement system with a user interface to visualize data. The sensor transfers flow data to the microcontroller through I2C protocol. The microcontroller then wirelessly transmits data to a laptop by Bluetooth low energy. Qt creator is used to develop a user interface to visualize the flow rate waveform and Bluetooth profile of the device.

Keywords: Sensor systems and Instrumentation, Physiological monitoring - Instrumentation, Bio-electric sensors - Sensor systems
Abstract: We present the further development of the Data Acquisition (DAQ) system that uses Multi-Electrode Arrays (MEAs) for neural activity monitoring and electrical stimulation. The system is entirely operational at its full capacity (512 input/output channels) thanks to the refreshed design of its electronics.

Keywords: Image segmentation, Machine learning / Deep learning approaches
Abstract: This study developed a segmentation method for pressure ulcers. To manage the wound in telemedicine approaches, we suggest a segmentation for automated size measurement of different tissues in the wound. We discovered U-Net structures are appropriate for the multi-way classification of pressure ulcers and show high enough accuracy.

Keywords: Ambulatory diagnostic and therapeutic devices - Ambulatory and ADL technologies, Ambulatory Therapeutic Devices - Biofeedback and related technologies, Clinical engineering -Verification and validation of diagnostic & therapeutic systems / technologies
Abstract: Training that simultaneously targets motor and cognitive domains may lead to improved ambulation recovery. The objective of the investigation was to evaluate the efficacy of simultaneous motor & cognitive training (SMCT) to improve ambulation, using biomechanical and functional outcomes. Preliminary data for a single subject with traumatic brain injury (TBI) who utilized a virtual reality integrated treadmill to provide SMCT during the chronic phase is shown. Clinical Relevance —Preliminary data provides initial evidence for SMCT as a therapeutic intervention for gait and balance rehabilitation in young adults with TBI.

Keywords: Computational modeling - Analysis of high-throughput systems biology data, Synthetic biology
Abstract: We performed computational fluid dynamics (CFD) simulations for the vessel of experimental animal models before an aneurysm initiation and observed tissue samples of the model which had the aneurysm by immunostaining. The stretching force on the vessel wall due to the blood flow degenerates both endothelial cells and smooth muscle cells, and the stagnation causes low wall shear stress, which illustrates inflammation, resulting in the initiation of a cerebral aneurysm.

Keywords: Neural networks and support vector machines in biosignal processing and classification, Data mining and big data methods - Biosignal classification, Data mining and big data methods - Machine learning and deep learning methods
Abstract: This work presents an SVM-CNN-based system for detecting COVID-19 using speech signals in a non-invasive and remote way. Speech recordings of 42 diagnosed and control subjects from the Voca.ai database were analyzed. The system achieved an average accuracy of 74.7%.

Keywords: Physiological systems modeling - Signal processing in physiological systems, Adaptive filtering, Signal pattern classification
Abstract: We proposed a method to utilize EDA from wrist to counter its lack of sensitivity in estimating significant SCR. Comparing the estimated time of SCR at the transition from baseline to task, a delay between EDA from wrist and finger were confirmed with correlation to slope of SCL. Adaptive thresholding that controls threshold of discriminating SCR by the value of past slope of SCL showed ability to estimate transition with high sensitivity.

Keywords: Clinical engineering -Verification and validation of diagnostic & therapeutic systems / technologies
Abstract: Radiofrequency catheter ablation (RFCA) is an effective treatment for tachyarrhythmia. Catheter-tissue coupling during RF energy application is important to create sufficient lesions and block tachycardia circuits. Contact force (CF) value was reported to correlate with lesion size, and CF monitoring is widely used in RFCA. However, suitable CF would be different among tissue and it might be excessive at vulnerable sites, so the risk of serious complications such as cardiac tamponade would not be solved using only CF monitoring. The evaluation of myocardial mechanical properties, such as stiffness and thickness, might be useful to solve these problems. In recent years, catheters that can measure local impedance (LI) between the tip electrode and proximal ring have developed. The dependence between LI and the distance from the tissue to the catheter tip was reported. The information about myocardial mechanical properties would be obtained by the combination of LI measurement for strain estimation and CF measurement. In this study, we measured LI and CF simultaneously using left ventricular (LV) and right atrium (RA) of porcine hearts in vitro to develop novel transcatheter monitoring. A 4.5 mm open-irrigated ablation catheter was used. The ablation catheter has 3 microelectrodes incorporated within the tip electrode. LI was measured between the microelectrodes and the proximal ring of the ablation catheter. The catheter tip was pushed into LV and RA epicardium at a constant velocity while measuring LI, CF, and catheter displacement in vitro. We distinguish myocardial tissue with different mechanical properties by combined transcatheter measurement of CF and LI. Simultaneous measurement of LI and CF may enable transcatheter investigation of myocardial properties and provide useful information for determining the optimal ablation conditions for the tissue.

Keywords: Human machine interfaces and robotics applications
Abstract: Despite the increasing popularity of endoscopic surgery due to its minimal invasiveness for the patient, the associated increase in the physical burden on surgeons is a critical problem. In our previous study, we developed an upper limb support suit and evaluated the effect of this suit on reducing the physical burden on surgeons. Thus far, there has been no study of the physiological burden of wearing the device. In this study, we examined the physiological effects of this device on subjects during endoscope retention. In this experiment, four adult males of non-medical association used an endoscope to monitor a fixed point with their dominant hand until they could no longer raise their upper arm due to arm fatigue. This task simulating the motion of a camera driver was carried out two times for each subject. The influence on the autonomic nervous system was evaluated by measuring the RR intervals. The results obtained with and without the proposed device were compared. The results indicate that there was no significant difference whether the device was worn or not worn. These results suggest that the physiological burden is not increased by wearing the device. In the future, we will measure the burden for medical staff and verify the results in a real environment.

Keywords: Optical imaging and microscopy - Diffuse optical tomography, Optical imaging and microscopy - Near infra-red spectroscopy, Regularized image Reconstruction
Abstract: We developed a novel brain-wide diffuse optical tomography framework (Fig. 1) that integrates a cap-based whole-head optode placement system with multiple computational and data-driven approaches to reconstruct resting-state networks in dual contrasts of oxygenated (HbO) and deoxygenated (HbR) hemoglobin.

Keywords: Cardiac imaging and image analysis
Abstract: Non-contact methods of determining the human body’s heart rate are of interest for clinical use. This research used a video magnification technique on the individual frames from a 15-second video taken using a digital single-lens reflex (DSLR) camera. It was possible to determine the beats per minutes of the heart rate by extracting the green spectrum from a region of interest information from the video frames. In this paper, three methods are presented using this colour change between the frames transform as a signal to find the heart rate.

Keywords: Data mining and big data methods - Machine learning and deep learning methods, Data mining and big data methods - Biosignal classification, Signal pattern classification
Abstract: Depression is one of the fatal mental diseases that not only affect daily activities but also can lead to suicide. Diagnostic methods of depression have not been developed by quantitative methods so far. For the past decades, many researchers have tried to develop quantified diagnosis methods by using bio-signal. However, because of the complexity of depression, biosignal-based diagnosis methods have yet to be used. In this study, we have researched if we can effectively diagnose depression using multi-modal bio-signals. We found that features extracted from multi-modal bio-signals could estimate depressive patients by 70 percent. These results indicate multi-modal bio-signals based features can be used for classifying depressive participants

Keywords: Data mining and big data methods - Machine learning and deep learning methods, Data mining and big data methods - Biosignal classification, Signal pattern classification
Abstract: Anxiety is known as a natural response of the body to stress. It is normal to experience anxiety occasionally, however, frequent episode of intense anxiety which interferes with daily activities might be diagnosed as anxiety disorder. There are several self-reporting questionnaires to examine the level of anxiety, such as Beck’s Anxiety Inventory. However, these self-reports might be biased by repetitive examination or individual personalities. For the past decade, various attempts were made to establish a quantitative diagnostic method for anxiety disorder using bio-signals such as electroencephalogram (EEG), electrocardiogram (ECG), and photoplethysmography (PPG). In this study, we have investigated the performance of a multi-modal biosignal-based classifier to estimate the anxiety level of 137 participants.

Keywords: Time-frequency and time-scale analysis - Time-frequency analysis
Abstract: We investigated whether light exposure during sleep affects the sleep quality of shift workers. Our finding suggests that when shift workers sleep during the daytime, exposure to light can degrade the quality of sleep in the late sleep cycle.

Keywords: Health Informatics - Behavioral health informatics, Health Informatics - Precision medicine, General and theoretical informatics - Machine learning
Abstract: Discriminating psychological stress (PS) in daily life based on passively sensed heart rate variability (HRV) is useful to monitor/manage PS in a constant manner with low burden. We classify PS levels from HRV continuously measured in daily life. Using a smartphone application and chest-mounted heart rate monitor, PS (6 times/day) and continuous HRV were obtained from ten participants over two weeks. A “gradient boosting classifier” was used to predict whether PS is higher or lower than the individual mean using HRV features in time intervals (~10, 30, 60 minutes) preceding PS measurement. Overall classification showed a high accuracy of 79%.

Keywords: CT imaging, CT imaging applications
Abstract: Recently, the Intraoperative Computed Tomography (iCT) system is very useful in operating room because it can improve the surgical accuracy, reduce complications, and greatly improve patient treatment results. In addition, it is providing real-time 3D information and navigation, roadmap, and tracking functions in conjunction. The purpose of this study is to develop an integrated embedded controller of mobile iCT that provides easy movement and fast CT scan in hospital.

Keywords: Wearable sensor systems - User centered design and applications, Health monitoring applications, Wearable body-compliant, flexible and printed electronics
Abstract: The world is currently underprepared to deal with viral pandemics such as COVID-19 in many aspects, but we are improving. The use of face masks and other wearable devices has become the new normal to reduce spread of the coronavirus, with the help of technology we open a new world of improving prevention and health monitoring. Our platform focuses on temperature detection and monitoring with low power, compact-ness, at a small cost to mass produce. This device is designed with its own custom PCB, with capabilities to be easily manipulated, changed, and improved. Lastly the addition of a motion data acquisition device was installed for person to person contact tracing via Bluetooth Low Energy Communication system. In this paper, we developed an affordable platform that allows for smart contact tracing that can be used for multipurpose data collection. This data will be used for contact tracing, temperature logs, and many more applications that break into this new field and allows for future improvements.

Keywords: Wearable body sensor networks and telemetric systems, Bio-electric sensors - Sensing methods
Abstract: Abstract— The purpose of this study was to investigate the changes in lower limb muscle activities and cardiopulmonary metabolic energy cost during treadmill walking with different inclination grades and a correlation between these two measures in older adults. Twenty-four healthy older adults participated, simultaneous measurements of lower-limb muscle activity and cardiopulmonary metabolic energy cost during inclined treadmill walking were collected. Clinical Relevance— The key muscles showing a significant relationship with increased net cardiopulmonary metabolic energy cost during inclined treadmill gait were the SOL, GCM, and VM muscles. These results can be used as basic data for various gait training programs and also as an indicator for the development of assistive algorithms of wearable walking robots for older adults.

Keywords: Wearable robotic systems - Orthotics and Exoskeletons, Rehabilitation robotics and biomechanics - Exoskeleton robotics
Abstract: The purpose of this study was to investigate effect of functional and gait training with Gait Enhancing and Motivating System (GEMS) on gait and physical function in elderly persons. Thirty-three elderly adults received 40-minute functional and gait training session with the GEMS in various environment, total 24 sessions for consecutive 8 weeks. Clinical Relevance— After 24-sessions of robot-assisted functional and gait training with GEMS, significant improvement was demonstrated in the 10MWT, FRT, FSST, TUG, BBS, 6MWT, and muscle strength (hip/knee/ankle). A newly developed wearable hip assist robot, the GEMS, is a potentially useful training device for improving gait and physical function in elderly persons.

Keywords: Dynamics in musculoskeletal biomechanics, Joint biomechanics, Modeling and simulation in musculoskeletal biomechanics
Abstract: Gait analysis using a musculoskeletal model was conducted with 3,000 cases of open-gait database. We propose a product of knee joint reaction force and angular velocity of knee joint as a knee load index (KLI) during gait movement. Greater KLI values were observed with subjects of bended knee joint.

Keywords: Modeling and simulation in musculoskeletal biomechanics, Joint biomechanics, Dynamics in musculoskeletal biomechanics
Abstract: Turning patient for frequent reposition on a bed caused lower back pain (LBP) among caregivers. Previous studies mentioned that adjusting bed height and foot position of caregiver contributed to reduce lumbar loads of turning patient. However, suitable combination of bed height and foot position were not clarified. The objective of this study is to find suitable combination of bed height and foot position for preventing LBP while turning patient. The relationship between compression force of L4-L5 as lumbar load and combination of two bed heights and nine foot-positions was investigated via musculoskeletal simulation. The results showed that compression force of L4-L5 became smallest in combination of high-height bed (51% of body height) and short anteroposterior length foot-position (15 % of body height). Therefore, a caregiver had better use the above combination when turning a patient on a bed.

Keywords: Sensor Informatics - Wearable systems and sensors, Sensor Informatics - Intelligent medical devices and sensors, Health Informatics - Mobile health
Abstract: Elderly people experience fall accidents due to tripping by lack of foot clearance (FC). Thus, FC should be continually monitored in daily life for preventing fall accidents. Smartwatch is considered as suitable wearable device for FC measurement since users preferred to wear sensors on the wrist in daily life. However, smartwatch cannot directly measure FC because smartwatch is mounted on wrist. Therefore, FC prediction from wrist position is necessary. The objective of this study is to propose the prediction method for FC using wrist position. The proposed method predicts FC by machine learning-based regression model and 3D positions of wrist. The proposed method was tested by public gait dataset. The results showed that the proposed method could accurately predict FC. These results indicates that the proposed method can be used for fall prevention system using smartwatch.

Keywords: Public Health Informatics - Health risk evaluation and modeling, Public Health Informatics - Outcomes research, General and theoretical informatics - Statistical data analysis
Abstract: In order to provide efficient health services, it is necessary to predict individual health conditions. In this study, we constructed a method that predicts the probability of hospitalization and its timing by applying the Joint model, which models both past data transitions and future hospitalization occurrence timing. Experimental result shows that the discrimination accuracy of the proposed method is 0.81. From the result, it is suggested that an effective health service planning is possible using the proposed method.

Keywords: Drug Delivery Systems and Carriers, Smart implanted drug delivery systems
Abstract: Microbubbles have potential for applications as drug and gene carriers, and the drug release can be triggered by ultrasonication in blood vessels. The acoustic response of microbubbles coated by surfactants have been investigated by several groups, indicating the viscoelasticity of the surrounding molecular shell and the internal gas affect the behavior of the bubble under ultrasound irradiation. Considering the practical drug delivery system, the molecular concentration on the bubble surface is one of the important factors for the quantitative drug administration. In this paper, the molecular desorption from a single microbubble under ultrasound irradiation was observed, and the amount of molecules desorbed from the bubble surface was estimated quantitatively. Tens-micrometer-sized microbubbles coated with DMPC were fabricated using DMPC solution and fluorocarbon gas. The optical observation system consists of a high-speed camera and an ultrasound cell. The microbubbles attached on a glass plate placed in the cell were irradiated once with̷ ; ; 9;50-cycle pulsed ultrasound with 20 kPa at 39 kHz. The desorption of the DMPC molecules from the surface of the bubble was evaluated by measuring the contact angle between the bubble and the glass plate; a decrease of the contact angle indicates desorption of the DMPC molecules from the bubble surface. Time change of the contact angle of the bubble was measured before and after ultrasound irradiation. The molecular concentration on the bubble surface was decreased significantly by ultrasound irradiation. The average values of the molecular concentration

Keywords: Brain-computer/machine interface, Human performance - Cognition, Neurological disorders - Epilepsy
Abstract: Resective surgery for mesial temporal lobe epilepsy (MTLE) carries the potential risk of memory decline. Neurofeedback (NF) which regulates brain activity and function, may be a promising solution for improving memory function. We constructed a memory-NF system using theta band power of electrocorticograms from the parahippocampal gyrus (PHG) and evaluated it with a preliminary study. As a result, we observed changes in theta-range oscillatory activity in PHG during NF training, and gained important insights for further developing the memory-NF system for memory enhancement.

Keywords: Micro- and nano-technology
Abstract: Microbubbles used as contrast agents in ultrasound diagnosis have many attractive potensials, and they are expected to apply for therapeutic applications such as drug derivery system and gene delivery. The microbubbles were required to hold on blood vessel for long periods of time, since they flow in systemic circulation. Ultrasound is submitted to the microbubbles for controlling the release of drug, targeting and monitoring existence. Although the stability of microbubbles was reported in many articles experimentally and theoretically, the few reports have been written in relationship between the molecular film coated and lifetime under irradiating ultrasound. This paper discussed the effect of a self assembled-shell to the persistence of microbubble during ultrasound exposure. Pluronic F-68 used in this experiment, a poloxamer surfactant, is routinely being used as a dispersing and foaming agent to facilitate phospholipid-based microbubble preparation. Adsorption kinetics at gas/water interface was investigated to analyze shell characteristics at several concentration (0 to 10-2 molL-1) using a tensiometer, and at least two successive first-order phase transitions are shown. The volume change of microbubble decorated with Pluronic F-68 were evaluated by optical observation system. Simultaneously, a laser Doppler vibrometer employed in an optical microscope measured the oscillation characteristics by ultrasound. An ultrasound transducer was excited with a continuous sinusoidal wave at a resonance frequency of 38.8 kHz to generate an ultrasound standing-wave field with the sound pressure of 7.2 kPa in a transparent cell. The resonance bubble radius was approximately 65 μm in each concentration, hardly have any relation with concentration. The bubbles around the resonance radius shrinked remarkably, and the persistence was improved with incleasing the surfactant concentration, meaning the moleculer film on the bubble surface prevented diffusion of the inner ga

Keywords: Magnetic resonance imaging - MR neuroimaging, Brain imaging and image analysis
Abstract: Traumatic brain injury (TBI) can arise from events such as concussion, brain contusion or axonal injury, and result in cerebral microhemorrhages or ischemic lesions. To detect these injuries, 1.5-Tesla (1.5T) and 3T MRI visualize hyper- or hypo-intense lesions of pathology. Because 7T MRI was approved for clinical use by the FDA only recently in 2017, we conducted a systematic review of direct comparisons between 7T MRI and lower field analysis in the diagnosis and management of TBI. Our search yielded a total of three reports with 29 research subjects. The three reports describe 7T as providing higher resolution and the ability to detect a greater number of lesions in pathologic cases. However, it is unclear whether these advantages provide a significant clinical benefit that would justify its widespread adoption in assessing TBI.

Keywords: Brain-computer/machine interface, Motor learning, neural control, and neuromuscular systems, Neurorehabilitation
Abstract: Concentration while rehabilitation training is critical for positive outcome. Therefore, concentration measurement and its feedback during exercise rehabilitation could help to maximize the effect of exercise rehabilitation. In this study, we investigated whether the concentration level could be detected by using the modulation effect of SSVEP amplitude by the external focus on exercise and be superiorly classified the concentration level. For this, we used dual-task paradigm that made users distracted in focusing on performing an exercise task. The experiment was conducted on 22 healthy adults. As results, it was shown that the distract modulated SSVEP amplitude, which means it could be influenced by the user’s concentration on body movements. Also, the accuracy was superior when using SSVEP feature. Therefore, this SSVEP paradigm could provide an effective rehabilitation method of providing feedback and it could induce continuous focus on motor rehabilitation.

Keywords: Public Health Informatics - Infectious disease outbreak modeling, General and theoretical informatics - Machine learning, General and theoretical informatics - Predictive analytics
Abstract: Abstract— Modelling of COVID-19 as time series data for machine learning problems for greater prediction accuracy is a challenging problem. This problem is further compounded when we attempt to extrapolate the data. In the present work popular ensemble strategies are used to accurately model Florida State COVID-19 positive cases as a time series data. Experimentally it was found that the optimized XGBoost model was superior to Random Forest in terms of Root Mean Square Error (RMSE) which indicated data around best fit, by 60.97% and to CatBoost by 50.69%. In terms of Mean Absolute Error (MAE) which predicted the average of deviation from true values, XGBoost outperformed Random Forest by 61.5% and CatBoost by 51.25%. XGBoost fared relatively poorer for a single step extrapolation and it deviated from the observed value as a regression error by 22.01%. However it was found that the Multi-Layer Perceptron (MLP) as a Deep Learning algorithm outperformed XGBoost in terms of extrapolation regression error by 6.629% and indicated more promising results. This clearly indicates us to use the Deep Learning Model for generalization and extrapolation of time series values.

Clinical Relevance— This modeling of healthcare time series data is useful to extrapolate COVID 19 infections as time series data and prepare emergency medications and health services.

Keywords: Signal pattern classification, Neural networks and support vector machines in biosignal processing and classification, Signal pattern classification - Genetic algorithms
Abstract: Auditory-perceptual evaluation is a common assessment used for evaluating the voice quality of patients; however, it suffers from inter- and intra-rater reproducibility problems. To this end, we propose an objective evaluation metric utilizing multi-input self-attention technology to evaluate voice quality using the grade (G), roughness (R), and breathiness (B) scale. The results showed that our proposed system has better performance accuracies of 69.25% for G, 77.5% for R, and 82.25% for B, as compared to that reported in previous studies. The findings also suggests that our system serves as a potential approach for voice quality evaluation applications in the future.

Keywords: Data mining and big data methods - Biosignal classification, Data mining and big data methods - Machine learning and deep learning methods, Physiological systems modeling - Signal processing in physiological systems
Abstract: Recent studies have explored seizure detection with bracelets or EMG sensors but little attention has been given to connected shirts. In this work, we analyzed respiratory signals recorded with the Hexoskin smart wear. Breathing rate (BR) and minute ventilation (MV) were extracted. A statistical analysis was performed to evaluate the existence of a significant difference between the ictal (i.e. seizure) and interictal (i.e. non-seizure) epochs. Four different classifier models were trained. A total of 67 seizures from 13 patients were analyzed. A significant difference was observed in BR and in MV between the ictal and interictal epochs (p < 0.01). The RandomForest classifer performed the best with an average accuracy of 91.1%.

Keywords: Signal pattern classification, Time-frequency and time-scale analysis - Time-frequency analysis, Physiological systems modeling - Signal processing in physiological systems
Abstract: Waveform analysis of oscillatory signals is crucial in the study of electrical processes in biomedical engineering. FMM is a simple parametric model that provides accurate parameter estimators, which are physically interpretable. The FMM parameters potential is shown in this work by solving two classification problems regarding neurons' action potential curves in a simulated experiment.

Keywords: Image enhancement, Image reconstruction and enhancement - Filtering
Abstract: High-intensity focused ultrasound (HIFU) is an emerging tool for effective thermal ablation of tissue. However, real-time monitoring of HIFU treatment is challenging, as HIFU application causes strong interference for imaging. Thus, this paper introduces singular value decomposition-based filtering capable of removing the HIFU interference from harmonic images without undesirable spectral distortion. The results experimentally validated with a custom-made phantom indicate that our approach eliminates HIFU-induced artifacts effectively, which is essential for real-time monitoring of therapeutic process.

Keywords: Image enhancement, Multimodal image fusion
Abstract: Ultrasound beamforming relies on a constant sound speed to reconstruct images. However, the sound speed varies in heterogeneous tissues, resulting in beamforming errors and thus image quality degradation. Therefore, this study investigates estimation of sound speed from computed tomography (CT) scan to improve the beamforming accuracy. We have demonstrated that our approach can be successfully applied to ultrasound/CT fusion imaging with improved ultrasound image quality.

Keywords: Neural interfaces - Implantable systems, Neural stimulation, Neuromuscular systems - Peripheral mechanisms
Abstract: Implantable magnetic stimulation is a technique that is free from the degradation of an implanted stimulator by immune responses in the body, and the optimal size, design, and stimulation intensity of the stimulator have been studied continuously. In this study, we applied magnetic stimulation for peripheral nerves in-vivo. We observed that magnetic stimulation using a small coil with a diameter of 4 mm could elicit neuronal responses in the sciatic nerve. These responses were elicited in all subjects when the stimulus intensity was above 5.5 A.

Keywords: Therapeutic robotics in rehabilitation
Abstract: In this paper, it is presented a trajectory tracking control by using PD Computed Torque Control (PD CTC) strategy for a nonlinear two-link robotic hand exoskeleton. The control strategy performance analysis was carried out through simulations in MATLAB/SIMULINK. Good performance and a very small error was obtained.

Keywords: Data mining and big data methods - Machine learning and deep learning methods, Connectivity, Signal pattern classification
Abstract: There is broad interest in discovering objective biomarkers for psychiatric disorders, which currently are primarily diagnosed based on subjective self-reports. The discovery of new biomarkers have the potential to support diagnosis and expand our knowledge about the pathophysiological processes underlying the disorders, hence opening new avenues for targeted treatment. The present study aims to investigate potential resting-state electroencephalographic (EEG) biomarkers for post-traumatic stress disorder (PTSD). Data from 221 combat-exposed veterans was analyzed (53% with PTSD and 47% controls). We developed a comprehensive machine learning framework to robustly identify features capable of classifying PTSD among commonly studied EEG features. Random forest significantly classified patients from controls with an accuracy of 62.7%. Evaluation of the selected features revealed Granger causality from the frontoparietal control network to dorsal and ventral attention networks in the gamma and delta frequency bands were particularly important for the classification, suggesting that these networks are abnormal in PTSD. Given the modest accuracy, we speculate that only a portion of the PTSD patients is characterized by the selected features, hence there is not enough evidence that they represent a clinically relevant biomarker.

Keywords: Brain functional imaging - fMRI, Human performance - Cognition, Neurological disorders
Abstract: In this research, we examined the effect of low-intensity stretching exercise intervention in Alzheimer’s disease over a period of 12 months---consisting of a 6-month intervention period and a 6-month follow-up period. Our analysis indicated that stretching exercises may help reduce the progressive decline in, or may even slightly increase, the neuronal activity, functional connectivity, and the global cognition in older adults with mild-to-moderate AD dementia, especially during the 6-month follow-up period.

Keywords: Modeling and analysis, Sensor systems and Instrumentation, Novel methods
Abstract: This paper presents an automatic markerless system for quantitative gait analysis. The system enables tracking of 9 body parts of freely walking rat subjects to achieve kinematic gait parameters. The system features real time analysis with a comparable performance to marker-based evaluation solution.

Keywords: Ultrasound imaging - Vascular imaging
Abstract: In our accompanying paper [1], for achieving high accuracy deep learning (DL) segmentation on an ultrasound echo image, we propose to perform (i) speckle reduction and (ii) superresolution as preprocessing. For the speckle reduction, the Auto-Encorder (AE) model is effectively used, whereas for the superresolution, an effective model must be searched for. In this report, the performances of new models such as the Deep Denoising Super Resolution CNN (DDSRCNN) [2] and the Temporal Coherence Generative Adversarial Network (TecoGAN) [3] are evaluated for human in vivo carotid echo images.

Keywords: General and theoretical informatics - Deep learning and big data to knowledge, General and theoretical informatics - Machine learning, General and theoretical informatics - Causality analysis and case-based reasoning
Abstract: We propose a prediction model which combines time series learning and feature relationship learning on a cohort of aging trends study. We also propose an autoencoder-based method for missing data imputation and salient feature extraction. Evaluation experiments by using 1,645 samples of NHATS (National Health and Aging Trends Study) dataset show that the proposed method achieves high accuracy of physical capacity and cognition activities of elderly people. The concordance correlation coefficient (CCC) is higher than 0.89.

Keywords: Neurological disorders - Diagnostic and evaluation techniques, Brain functional imaging - EEG, Neurological disorders - Mechanisms
Abstract: Electroencephalographic (EEG) activity during a period known as the “burst suppression phase” following resuscitation from cardiac arrest (CA), has been shown to correlate favorably with neurological outcomes. However, since individual EEG sub-band activities may correspond more specifically to neuronal mechanisms, further analysis of specific EEG sub-bands during this period could facilitate better prognosis. Therefore, using a rodent model of cardiac arrest and resuscitation, we evaluate EEG theta band activity for its capacity to predict post-CA neurological recovery. Here, theta band activity was quantified using its spectral power while neurological outcomes 4 hours post-resuscitation were measured via a scale known as the neurological deficit score (NDS). Our analysis reveals that theta band spectral power is negatively correlated with NDS (-0.75, p<0.05), showcasing the potential for using it as a predictor of neurological outcome.

Keywords: Public Health Informatics - Public health management solutions
Abstract: Ambulance environments are associated with a high risk of infection. We have developed surface disinfector using low concentration hydrogen peroxide and plasma. To verify the performance of the development device, disinfection was performed in 2 different ambulances. Surfaces were swabbed before and after disinfection for counting the CFU of bacteria. Plasma hydrogen peroxide mist disinfection resulted in significant decreases in the CFU values of bacteria on ambulance surfaces.

Keywords: Public Health Informatics - Health risk evaluation and modeling, General and theoretical informatics - Machine learning, General and theoretical informatics - Data mining
Abstract: This paper presents our endeavor to fuse positive unlabeled and ensemble learning to shape a better ADR signal detector from spontaneous reporting data, like FAERS. The results show that the proposed PU-ensemble methods surpass the non-PU ensemble with 18%, 144%, 19%, and 80% improvement on accuracy, precision, recall, and F-measure, respectively, of the detected ADR signals.

Keywords: Image reconstruction and enhancement - Machine learning / Deep learning approaches, Ultrasound imaging - Breast, Image reconstruction and enhancement - Tomographic reconstruction
Abstract: This paper introduces U-Net for solving the speed of sound reconstruction on sparse view ultrasound computed tomography. It is trained and validated on the ellipse simulation dataset and in vitro experiment dataset. Both qualitative and quantitative results show a significant improvement in the speed of sound reconstruction for sparse views.

Keywords: Physiological monitoring - Novel methods, Thermal sensors and systems
Abstract: We have been developing a wearable deep body thermometer based on the zero-heat-flux method. In this method, to obtain the zero-heat-flux conditions, an electrical heating element has been usually used. However, considering the use in hot environments such as outdoor jobs in the summertime, air temperature usually rises higher than deep body temperature. To solve this problem, we tried to utilize a Peltier module not only for a “heating” element, but also for a “cooling” element. After simulation and building up our prototype thermometer, simultaneous measurements using a commercial device revealed that the prototype thermometer has the same accuracy as the commercial device and could maintain accurate measurement even in the high-temperature environment.

Keywords: Wearable body sensor networks and telemetric systems, Wearable sensor systems - User centered design and applications, Smart textiles and clothings
Abstract: When the World Health Organization declared the COVID-19 outbreak a pandemic, physical distancing measures (most recognizable through visual cues) were implemented to lower transmission rates. Lacking access to these visual cues, physical distancing requirements have become a significant source of anxiety and an obstacle to independence for visually impaired people. In response, a COVID-19 physical distancing design concept for visually impaired adults called the iSensor was developed. Consisting of a hat within an embedded LiDAR sensor system, paired with a belt containing vibration motors that communicate proximity and direction of people, it would allow users to self-regulate their distance from others in public. Here we present the configuration, methodology, and expectations of early experimental evaluation of the performance characteristics of the iSensor system.

Keywords: Rehabilitation robotics and biomechanics - Exoskeleton robotics, Biomechanics and robotics in physical exercise, Home robots
Abstract: Considering movement characteristics of human finger, a novel mechanism for rehabilitation of a hand finger was developed. The kinematics of exoskeleton was analyzed in SolidWorks and validated using SimMechanics. The results showed that the exoskeleton functions properly and provides good kinematic compatibility with the movement of the finger.

Keywords: Signal pattern classification
Abstract: Abstract— In this work, we investigate the use of machine learning as an alternative to manual review and annotation of electroencephalography (EEG) data in preclinical mouse models of brain injury. Preliminary results from classifiers trained and tested on independent datasets from two preclinical mouse models of brain injury show high performance in predicting labels of signal patterns of interest. These results demonstrate the potential utility of an automated tool to assess the quality of EEG data in preclinical models of brain injury.

Clinical Relevance— High quality EEG data ensures the development and assessment of reliable EEG-based biomarkers based on EEG features that reflect brain function and not noise.

Keywords: General and theoretical informatics - Data mining
Abstract: Abstract— In this work, we investigate the feasibility of a biomarker of brain injury severity extracted from electroencephalography (EEG) data. Preliminary analysis of features extracted from EEG data from a preclinical mouse model of traumatic brain injury shows strong tonotopic clustering of EEG features by injury severity level and future outcome. These promising findings demonstrate the potential to develop an EEG-based biomarker of brain injury severity.

Clinical Relevance— An accurate biomarker of brain injury severity will have a significant impact on neurocritical care by providing clinically meaningful information to optimize care for an individual patient and to facilitate identification of cohorts of patients with similar injury severity levels for clinical trials.

Keywords: Biomimetic materials
Abstract: In long bones, c-axes of Hydroxyapatite (HAp) crystallites are mainly oriented along the bone axis. The HAp coating film with same orientation may accelerate the bonding of the implant to the bone. In this study, fabrication of a c-axis tilted HAp film was performed using a RF magnetron sputtering system. The fabricated film sample was evaluated by the X-ray diffraction (XRD) and Transmission Electron Microscope (TEM) analyses.

Keywords: Optical imaging
Abstract: Fixational eye moments, such as microsaccades, are a useful phenomenon to understand human internal states. Conventional detection methods of such eye vibration, however, have a limitation of a fixed head due to a trade-off between camera resolution and angle of view. In this paper, we propose a new method to detect small eye vibration using high-speed optical control. The high-speed optical control enables to observe dynamic eyes at high resolution, and to measure the small eye vibration from the unrestrained human using pupil center corneal reflection. Experimental results have showed the detected small eye vibration with the unrestrained head.

Keywords: Thermal sensors and systems, Wearable sensor systems - User centered design and applications, Sensor systems and Instrumentation
Abstract: A miniaturized probe of a non-invasive core body thermometer was fabricated by optimization using parametric studies to minimize convection errors. It is robust against convection changes, based on in-vivo experiments for seven subjects with showing the measurement accuracy of ±0.2 ºC.

Keywords: Brain functional imaging - EEG, Brain functional imaging - Source localization, Human performance
Abstract: The effects of trait anxiety on the brain activity in response to pain is unknown, although differences in trait anxiety are known to cause different pain sensitivities. The present study examined the difference in brain activity in response to pain between individuals with particularly high and low trait anxiety. We found that participants with high trait anxiety expressed significantly higher activity in the Orbitofrontal Cortex during the moderate intensity stimulus compared to that in the low trait anxiety people.

Keywords: Cardiovascular and respiratory system modeling - Cardiovascular control models, Cardiovascular regulation - Baroreflex, Cardiac mechanics, structure & function - Ventricular assist devices
Abstract: A baroreflex model with an unstressed venous volume chamber was implemented on a biventricular mock circulation loop to mimic a facet of native physiological hemodynamic control and maintain a set level of arterial pressure. The model was tested with volume shifts on different arterial pressure targets; the model effectively maintained a set level of pressure with lower error than normal operation.

Keywords: Translational biomedical informatics - Data processing, Systems biology and systems medicine - prediction of disease related regulator, Translational biomedical informatics - Comparative effectiveness research
Abstract: Based on the empirical data from the formulae of Traditional Chinese Medicine, this study screened out the natural products with antibacterial potential. The efficacy of natural products was evaluated by regression and classification models as candidates for new antibiotics.

Keywords: Model building - Algorithms and techniques for systems modeling, Computational modeling - Analysis of high-throughput systems biology data, High throughput data - Machine learning and deep learning
Abstract: The correct classification of cancer subtypes is of great significance for the in-depth study of cancer pathogenesis and the realization of accurate treatment for cancer patients. In this paper, the cascade flexible neural forest (CFNForest) model was proposed to accomplish cancer subtype classification.Experiments on RNA-seq gene expression data showed that CFNForest effectively improved the accuracy of cancer subtype classification.

Keywords: Health Informatics - Health data acquisition, transmission, management and visualization, Health Informatics - Assessment of health information systems, Health Informatics - Electronic health records
Abstract: Implementation gaps exist for secondary use of EHR-generated data. We Leverage HL7-FHIR resources served from EPIC to populate an internally developed acute-care multi-patient viewer (AMP). The aim is to identify census errors during sprints of application development for iterative improvements. Between 4/15/20-5/26/20, census errors between EPIC and AMP are communicated to IT team. We observe a reduced average percentage from 7.93% to 2.53%. Validation of the census provides an indispensable feedback-loop for the developers for fix deployment. This highlights the need for an automatic validation of FHIR messages on the receiving /user side for a complete and more accurate census.

Keywords: Nonlinear dynamic analysis - Biomedical signals, Physiological systems modeling - Signal processing in physiological systems, Time-frequency and time-scale analysis - Empirical mode decomposition in biosignal analysis
Abstract: Resting-state electroencephalography (EEG) has been demonstrated to be a promising tool in pinpointing connectivity disruption in ALS [1]. A dynamic approach of brain networks could further improve the understanding and detection of networks with disrupted functional balance, as evidenced by findings in other neurodegenerative diseases [2]. Transient, quasi-stable and recurrent brain state episodes, also called microstates, are hypothesized to represent ‘building blocks of spontaneous thinking’ [3]. This study reports preliminary findings on abnormalities of EEG microstates in ALS patients, as potential prognosis biomarkers of cognitive and motor network function.

1] S. Dukic et al., “Patterned functional network disruption in amyotrophic lateral sclerosis,” Hum. Brain Mapp., vol. 0, no. 0, Jul. 2019, doi: 10.1002/hbm.24740. [2] P. L. Faber et al., “Fundamentally altered global- and microstate EEG characteristics in Huntington’s disease,” Clin. Neurophysiol., vol. 132, no. 1, pp. 13–22, Jan. 2021, doi: 10.1016/j.clinph.2020.10.006. [3] D. Lehmann, W. K. Strik, B. Henggeler, T. Koenig, and M. Koukkou, “Brain electric microstates and momentary conscious mind states as building blocks of spontaneous thinking: I. Visual imagery and abstract thoughts,” Int. J. Psychophysiol., vol. 29, no. 1, pp. 1–11, Jun. 1998, doi: 10.1016/S0167-8760(97)00098-6.

Keywords: Wearable sensor systems - User centered design and applications, Health monitoring applications, Sensor systems and Instrumentation
Abstract: 25 - 35% of older adults experience a near-fall every month, and these near-falls have been shown to be an independent predictor of future fall risk. The objective of this study is to compare a multiclass thresholding algorithm with a novel algorithm for near-fall detection.

We collected data from a participant wearing an inertial measurement unit on the lower back collecting tri-axial linear accelerations at 1000Hz. The participant simulated various falls, near-falls, and activities of daily living (ADLs). 90 falls and 78 near-falls were identified from video and sectioned into 1s segments. 2-minute ADL signals were partitioned into 1s windows resulting in 840 ADL segments.

Algorithm #1: Upper peak values were extracted for each fall based on the magnitude of acceleration. An upper fall threshold (UFT) was established based on the smallest peak. The same procedure was used to identify an upper near-fall threshold (UNFT). Signals that crossed the UFT were labelled as falls. Signals that crossed the UNFT but not the UFT were labelled as near-falls. Signals that did not cross either were labelled as ADLs.

Algorithm #2: Signals that did not pass an established UNFT were labelled as an ADL. Signals that passed the UNFT were further analyzed. We extracted frequency and time-domain features in the anterior-posterior (AP), superior-inferior (SI), and left-right (LR) axes. We used forward feature selection to select the top features and trained a multiclass Support Vector Machine.

Algorithm #1: The UFT was set at 24.87 m/s^2 and the UNFT was set at 12.75 m/s^2. Algorithm #2: The forward selected features were power in the 1-2Hz, 2-3Hz, 10-15Hz, 15-20Hz, and 12-13Hz bands (AP); power in the 0-5Hz band (LR); and power in the 4-5Hz band (SI). Algorithm #2 performed better with an accuracy of 97.01% compared to Algorithm #1 with an accuracy of 65.67%.

The development of a multiclass algorithm allows for a quantitative assessment of fall risk in daily environments.

Keywords: Human performance - Engineering, Brain physiology and modeling - Neural dynamics and computation, Brain functional imaging - EEG
Abstract: Brain activity during creative thinking of engineering design can be different from that during general creativity tasks. This study revealed that the default network (DNCORE, DNMTL, DNSUB3), and the frontoparietal control network (FPCN) are active during engineering design and these areas are more active in creative thinking than in divergent thinking because of specific functional requirements and technical problems.

Keywords: Cardiovascular regulation - Heart rate variability, Cardiovascular regulation - Autonomic nervous system, Cardiovascular and respiratory signal processing - Heart Rate and Blood Pressure Variability
Abstract: To prevent occupational driving accidents, we previously proposed a method of detecting worsening physiological states, such as fatigue, by measuring autonomic nerve function (ANF) before drivers start their shifts. Although increasing measurement times (e.g., while a driver is working) will improve the applicability to various worsening physiological states, little is known about how often we should measure ANFs for occupational driving due to their temporal variations within a day. We investigated how ANF indicators show diurnal variations compared to the baseline ANFs before drivers started their shifts.

Keywords: Human performance - Cognition, Human performance - Attention and vigilance, Human performance
Abstract: To explore the unconscious preference formation mechanism, we have investigated activity of the autonomic nervous system during subliminal mere exposure by measuring pupil diameter, HRV and respiratory. Smaller size and constriction of pupil was found for participants with mere exposure effect and for stimuli with high preference, respectively.

Keywords: Brain functional imaging - Evoked potentials, Brain functional imaging - EEG, Neural signal processing
Abstract: The spectral discrimination capacity is essential to recognize speech. In this study, the feasibility of the Ear-EEG approach as an electrophysiological method to evaluate spectral discrimination capacity was determined. Acoustic change complex evoked by spectral ripple reversal was acquired in most subjects. The amplitude was reduced in the hearing loss group.

Keywords: Health Informatics - Behavioral health informatics, Health Informatics - Clinical information systems, Health Informatics - Health data acquisition, transmission, management and visualization
Abstract: The impact of visual cues on speech recognition in noise is explored in individuals with normal hearing and hearing loss (HL). Use of visual cues statistically significantly influenced the HL group’s speech understanding in noise.

Keywords: Connectivity, Coupling and synchronization - Coherence in biomedical signal processing
Abstract: We investigated the strength of resting-state EEG functional connectivity of the default mode network (DMN) between 21 healthy participants and 2 cancer patients with chronic pain to explore the effect of chronic cancer pain on the intrinsic cortical network. Functional connectivity was assessed with the lagged coherence between pairs of 6 regions of interests for DMN within 7 frequency bands of the delta, theta, alpha 1, alpha 2, beta 1, beta 2, and beta 3 ranges. Cancer patients showed augmented functional connectivity between the anterior-to-posterior DMN regions relative to healthy participants in alpha 1 and beta 2 bands. These results support the disruptions of the DMN that were observed in patients with other types of chronic pain such as back pain and fibromyalgia.

Keywords: Brain-computer/machine interface, Neural signals - Machine learning & Classification, Neurological disorders
Abstract: To enable to move or turn a wheelchair to any place at once without the need for physical manipulation, we constructed an audiovisual BCI system. Virtual markers and sound sources were placed in front of and around the user in a mixed reality space. The subjects counted the number of the stimuli from the target marker or the sound. The accuracy of choosing the target marker or sound were obtained from EEG analysis, and they were 60 % and 20 % at maximum.

Keywords: Computer-assisted surgery, Robot-aided surgery - Remote surgery systems / telesurgery, Surgical robotics
Abstract: A new 5-DOFs intra-abdominal type master device to control a surgical assistant robot for teleoperated laparoscopic surgery performed by a surgeon in a sterilized area has developed. The master-follower system using the proposed master device was constructed, and the performance of the device was evaluated.

Keywords: Cardiovascular and respiratory signal processing - Blood pressure measurement, Vascular mechanics and hemodynamics - Arterial pressure in cardiovascular disease, Cardiovascular, respiratory, and sleep devices - Monitors
Abstract: The challenges in evaluating cuffless blood pressure (BP) measurement devices are described, and proposals are made for overcoming them.

Clinical Relevance—This commentary can facilitate understanding of emerging cuffless BP measurement devices.

Keywords: Physiological monitoring - Novel methods, Wearable sensor systems - User centered design and applications, Modeling and analysis
Abstract: There is a need for tools that objectively measure professional interest. Evidence supports personality prediction through physiological signals. Given that professional interests have a close relation with personality, it may be predicted using the same means. The objective of this study is to develop a Machine Learning algorithm that estimates engineering interest in children. It can be used to guide education policy, study interest in other professional fields, improve vocational counseling services in educational institutions.

Keywords: Brain functional imaging - EEG, Brain-computer/machine interface, Brain physiology and modeling - Sensory-motor
Abstract: Event-related de-synchronization (ERD) is a feature of the EEG that reflects sensory-motor processing in the brain. ERD is generated not only by motor execution but also by motor imagery, but the mechanism is not well understood. In this study, we hypothesize that the generation of ERD is related to the motor planning process in the brain based on our previous work which reported that ERD was maintained during isometric grasping with visual force feedback. To clarify the effect of visual force feedback on the resultant ERD, we compared the generation of the ERD in four conditions; visual force Feedback (FB), No Visual (NV), visual stimuli Playback (PB), and Visual Only (VO). As a result, we confirmed significant difference in ERD generation between the conditions PB and FB in the somatosensory cortex, but not in the motor cortex.

Keywords: Signal pattern classification, Time-frequency and time-scale analysis - Nonstationary analysis and modeling
Abstract: The main purpose of this study is to investigate the performance of the specified dynamical system model called Collective Almost Synchronization (CAS) features in the classification of Motor Imagery (MI) states. A methodology for pattern recognition of EEG signals was established based on the coefficient parameters of the CAS-based linear regression model. Compared with state-of-the-art methods, our proposed method achieved better performance on two-class MI classification.

Keywords: Diagnostic devices - Physiological monitoring
Abstract: In this study, we developed a toy to evaluate the tongue movements of infants during sucking by directly measuring the tongue force using a glove-shaped device equipped with 2 rows of miniaturized pressure sensors. By simple insertion into the oral cavity of an infant, the device can assess the tongue movements and display the evaluation results in real time.

Keywords: Human machine interfaces and robotics applications, Design and development of robots for human-robot interaction, Assistive and cognitive robotics in aided living
Abstract: In this study, we aimed to evaluate the mental workload caused by handwriting characters with a stylus on tablets. We conducted an experiment to investigate the effects of the screen angle on the mental workload when handwriting numeric characters using the NASA task load index. The relationship between the screen angle and the mental workload was revealed in this study.

Keywords: Wearable sensor systems - User centered design and applications, Wearable body sensor networks and telemetric systems, Wearable wireless sensors, motes and systems
Abstract: A highly accurate, compact, and inexpensive patch type wireless R-R interval (RRI) telemeter is being developed to monitor heart rate with a smartphone application. In order to perform stable RRI measurements for a long time using such a device, it is necessary to determine the optimal electrode configuration. However, people with little or no expertise in analyzing heart rate variability (HRV) may have difficulty in attaching a patch type sensor, where the electrode placement is already fixed, to an appropriate position according to individual differences. In this study, we developed a system as a smartphone application for selecting the optimal position of a patch type measurement device to improve usability for non-experts. The RRIs of 10 male subjects in four posture were measured by attaching the device at the position determined by an optimal attachment position selection system. By comparing the RRIs of the patch type device and the reference electrocardiogram measurement system, the measurement accuracy of the proposed devices was confirmed to be sufficient for HRV analysis.

Keywords: Diagnostic devices - Physiological monitoring, Cancer detection and diagnosis technologies, Clinical engineering -Verification and validation of diagnostic & therapeutic systems / technologies
Abstract: The gold standard for evaluating the digestive disease is histology. We propose an original thermographic endoscope (TE) equipped real time image processing for evaluating gastrointestinal (GI) lesions. We have found that the transient thermal impulse response (TTR) by spraying cooling water reflected histological alteration in the GI lesion. Even if TE could not reveal any static temperature differences between normal mucosa and tumorous or inflammatory area, our prototype of TE could pick up the lesion clearly by mean of image processing applied dynamic subtraction technique to the alteration of TTR.

Keywords: CT imaging
Abstract: The etiology of primary ankle osteoarthritis is not enough elucidated. X-ray images and Computed Tomography (CT)-scan images are usually used for the classification of ankle osteoarthritis (OA), however, the classification by the 2D measurements of the X-ray images and CT images can vary between physicians. The aim of this study is the 3D geometrical deformity characteristics quantitative evaluation of the tibial plafond according to the ankle OA classification. The 3D geometric statistical analysis was conducted for the 18 ankle OA feet. Although the 3D analysis results in tibial plafond have not been shown consistent deformity along with the progress of ankle OA, the deformity characteristics of high severity ankle OA revealed.

Keywords: Optical and photonic sensors and systems, Implantable sensors, Chemo/bio-sensing - Biological sensors and systems
Abstract: Epilepsy is one of the most prevalent neurological disorders in the world, with temporal lobe epilepsy (TLE) being the most common type. Despite this, the mechanism of how epilepsy occurs is unclear because of the difficulty in observing large neuronal populations across multiple layers in the brain during seizures. Therefore, new methods and technology are needed to solve this problem. Here, we present a brain-implantable CMOS image sensor that is tiny, energy-efficient, and bio-compatible for studying epileptiform signals in a freely behaving animal.

Our CMOS image sensor can visualize and measure neuronal calcium fluorescence changes from transgenic GCaMP6s mice. The micro-imaging sensor is 450 x 1500 µm in size, containing 40 x 120 pixels and an imaging area of 7.5 µm2 per pixel. Then, 305 × 280 µm blue µLEDs were added as a source of excitation light. To keep the device small and lensless, a thin-film absorption filter was placed on the imaging array to block the excitation light and transmit fluorescence from GCaMP6s. All components were attached and wire-bonded onto a flexible printed circuit board that connects to the peripheral electronics. The device was only 0.05 g in weight and was small enough to allow multiple sites of implantation. Furthermore, the vertical implantation of the device enabled simultaneous imaging of multiple layers in the hippocampus. Compared to head-mountable microscopes, our implantable device is smaller, lighter, requires fewer optical components, and can image varying depths simultaneously.

Using our device, we were able to observe and characterize different calcium waveforms during kainic acid-induced epilepsy. Specifically, we found three patterns of neuronal activity that closely resembled previously published results. These can then be correlated with behavior since the mouse was freely moving. Therefore, our newly developed device can provide insight into the neural mechanisms of epilepsy by linking the brain and behavior.

Keywords: Therapeutic devices and systems - Pulmonary disease, Formulation design, engineering and evaluation, Models and simulations of therapeutic devices and systems
Abstract: In this research we have designed a low cost, oxygen concentrator based on pressure swing absorption that can produce up to 85% pure oxygen streams. It is largely used in patients suffering from chronic respiratory diseases during COVID-19 pandemic. A pulse oximeter is also installed on the system to evaluate the oxygen treatment. Lungs of COVID-19 patient loses its ability to exchange oxygen. The proposed oxygen concentrator will be used to increase oxygen level in the blood.

Keywords: Integrated sensor systems, Sensor systems and Instrumentation, Modeling and analysis
Abstract: The effectiveness of exercise can be optimized using a normobaric hypoxic environment. However, the safety and effectiveness of exercise in such an environment have rarely been investigated in nonathlete subjects. In this study, we clarify the safety of normobaric hypoxic exercise and evaluate useful measurement indices for a real-time monitoring system. The results of this study suggest that the proposed system is capable of real-time monitoring of hypoxic load during normobaric hypoxic exercise and may contribute to achieving safe exercise.

Keywords: Neuromuscular systems - EMG models
Abstract: This paper proposes a visible feature to capture the hand motion from the information of myoelectric signals obtained from multiple electrodes. This feature is the center of the myoelectric signals, the average position of multiple muscle electrodes weighted by myoelectric signals. We preliminary conducted a basic experiment to confirm the center of myoelectric signals corresponding to hand motions by measuring the EMG signals from the electrodes attached around a cross-section of the forearm and confirmed that this feature could visually separate wrist motions and grasps.

Keywords: Cardiac signal remote monitoring devices and technologies, Cardiovascular assessment and diagnostic technologies, Ambulatory diagnostic and therapeutic devices - Wireless telemetric systems
Abstract: Abstract— We present a technology that improves accuracy of diagnosing acute myocardial infarction based on ECG signals in the Emergency Department setting. It achieves such improvements without altering the standard of care workflow or signal collection methodology.

Clinical Relevance— Presented tool may contribute to a reduction in the delay to treatment for patients with acute myocardial infarction as well as in reducing the rate of missed acute myocardial infarction.

Keywords: Time-frequency and time-scale analysis - Empirical mode decomposition in biosignal analysis, Signal pattern classification
Abstract: Face masks prevent viral spread. Their effectiveness depends on intrinsic material composition and user fit. Our research shows that Word Error Rate (WER) calculations for a custom speech test on two common Automatic Speech Recognition (ASR) platforms correlates with intrinsic mask filtration.

Keywords: Neural interfaces - Implantable systems, Sensory neuroprostheses, Neural interfaces - Microelectrode technology
Abstract: In this work, we fabricated an ultra-flexible mesh-structured electrodes array using parylene-C, and conducted in vivo experiments to record neuronal signals from peripheral nerves using the developed electrodes. The developed mesh electrodes successfully recorded evoked afferent action potentials.

Keywords: Motor learning, neural control, and neuromuscular systems, Neural stimulation, Neuromuscular systems - Learning and adaption
Abstract: Planning and execution of a finger reaching task to control finger aperture depends on proprioceptive feedback, especially when vision is occupied by other tasks or occluded. To address this visual-proprioceptive mismatch and improve finger control, we introduce electro-prosthetic proprioception (EPP) that provides a more accurate sensory reference. EPP was rendered by electrical stimulation applied to the fingertip with a frequency inversely proportional to finger aperture. We measured how EPP could enhance accuracy without vision, and found it to significantly improve, even after EPP was removed. This effect was retained 24-hours post training, while the visuomotor training control group showed no retention. This new type of proprioception offers a new pathway for feedback in any tasks that require accuracy enhancement.

Keywords: Health Informatics - Behavioral health informatics, Health Informatics - Mobile health, Bioinformatics - Bioinformatics for health monitoring
Abstract: Individual characterization of the finger movement can be applied in a variety of areas, such as security in E-commerce or early detection of neurodegenerative diseases. However, the way of characterizing finger movement has been understudied. The finger movement is a sophisticated motor task, highly variable by multiple factors like posture and joint position. Therefore, it is important to identify the best way to obtain reliable data for characterization. In this paper, we tested the use of a smartphone pattern lock application to identify the best posture to characterize and distinguish the finger movement patterns between individuals.

Keywords: Magnetic resonance imaging - MR neuroimaging, Image analysis and classification - Machine learning / Deep learning approaches, Functional image analysis
Abstract: Eating behaviors are related to both obesity phenotypes and brain function, but brain function does not per se mediate eating behaviors in individuals with obesity. Here, we identified subgroups of eating behaviors using functional connectivity and autoencoder-derived latent features. We found three subgroups in eating behaviors, where disinhibition and hunger factors showed significant differences among the groups.

Keywords: Ultrasound imaging - High-frequency technology, Ultrasound imaging - Other organs, Ultrasound imaging - Doppler
Abstract: To suppress receive grating lobe artifacts we propose virtual receive elements with a reduced pitch size. A virtual RF data is created using physical RF data and plane-wave beamforming is performed with reduced pitch size.

Keywords: Time-frequency and time-scale analysis - Time-frequency analysis
Abstract: In this work, the classic indices of pulse/heart rate variability extracted by the photoplethysmographic signal recorded with the Polar OH1 sensor have been compared with those extracted from an electrocardiographic signal with a high sampling rate. Currently a large number of smart wearable that record photopletysmographic signal (PPG) are hitting the market. The use of this technology is currently spreading remarkably in many different everyday applications, such as sports performance, stress indicator, heart rate control, etc. Polar OH1 records the subject’s PPG signal at approximately 135 Hz. Recent studies have shown that using the interpolated midpoint of the PPG as fiducial point allows the PPG sample rate to be reduced to 50 Hz without causing significant changes in the PRV indices. A total of 22 subjects (12 males and 10 females) were recorded in this study. During the recordings subjects stayed relaxed and sitting comfortably, remained in silence and without performing movements for four minutes. In PPG signal, the interpolated medium point defined as the one in which the amplitude has reached the 50% of the pulse amplitude, was used as fiducial point. A previous linear interpolation at 1000 Hz of the PPG up-slope was performed. Classical time and frequency domain indices from the HRV and PRV signals were computed The relative error obtained in the estimation of the PRV indices with respect to their corresponding HRV indices was calculated. The absolute error was obtained for P_{LFn} and R_{LF/HF} indices. Although the results show a greater error than expected and could make this sensor not suitable for the analysis of the response of the autonomous nervous system, further study should be needed.

Keywords: Neural stimulation, Neuromuscular systems - Central mechanisms, Neural signal processing
Abstract: The amplitude of the rate-dependent depression (RDD) of the H-reflex was evaluated at 0.1, 1, 5 and 10 Hz in a 10 consecutive pulse protocol recorded in the soleus muscle, before, during and after cathodal and anodal Galvanic Vestibular Stimulation (GVS) in a single subject in a prone position. Although, RDD slightly decreased during and after GVS compared to before (no GVS), this spinal mechanism was preserved in all stimulation frequencies and GVS polarity.

Keywords: Systems modeling - Decision making, Computational modeling - Biological networks, High throughput data - Machine learning and deep learning
Abstract: Artificial neural network (ANNs) models have emerged as an important tool to understand the brain function. By fitting real neural data to an ANN, we created a powerful test-bed model for evaluating brain function. We further causally tested the effects of lesions in the network and their effects on the predicted responses and performance.

Keywords: Clinical engineering - Health technology / system management and assessment
Abstract: Athletes that participate in contact sport are at risk of suffering Sports-Related Concussion (SRC). A traditional approach of SRC testing relies on the 5th version of the pen-and-paper based sports concussion assessment tool (SCAT5). An open, digital equivalent may facilitate more efficient and transparent assessment. We describe a co-created developed iOS SCAT5 app to enable an Internet of Things (IoT) assessment.

Keywords: Bio-electric sensors - Sensing methods, Physiological monitoring - Novel methods, Health monitoring applications
Abstract: Cochlear implantation has successfully restored the perception of hearing for nearly 200 thousand profoundly deaf adults and children. More recently, implant candidature has expanded to include those with considerable natural hearing which, when preserved, provides an improved hearing experience in noisy environments. But more than half of these patients lose this natural hearing soon after implantation. To reduce this burden, biosensing technologies are emerging that provide feedback on the quality of surgery. Here we report clinical findings on a new intra-operative measurement of electrical impedance (4-point impedance) which, when elevated, is associated with high rates of post-operative hearing loss and vestibular dysfunction. In vivo and in vitro data presented suggest that elevated 4-point impedance is likely due to the presence of blood within the cochlea rather than its geometry. Four-point impedance is a new marker for the detection of cochlear injury causing bleeding, that may be incorporated into intraoperative monitoring protocols during CI surgery.

Keywords: Micro- and nano-technology, Translational issues in tissue engineering and biomaterials, BioMEMS/NEMS - Tissue engineering and biomaterials
Abstract: This study aims to obtain the nanoscale variations of topographical and mechanical properties of articular cartilage (AC) and subchondral bone (SB) in the knee joint of guinea pigs from the early to moderate osteoarthritis (OA) and to investigate how the mechanical properties and topographical parameters in AC and SB relates to OA level. Male Dunkin Hartley strain guinea pigs were grouped according to age (1, 3, 6, and 9 months, with 10 rats in each group). The topographical characterizations and mechanical properties of the AC and SB in the tibial plateau of the guinea pigs were determined through atomic force microscopy and nanoindentation. In addition, the structure of collagen in AC was observed by histological and Immunofluorescence analyses. Our findings indicated that the indentation moduli and hardnesses of the subchondral bone plate (SP) and subchondral trabecular bone (ST) in the tibial plateau of 3-, 6-, and 9-month-old guinea pigs increased compared with those of 1-month-old guinea pigs. There was no significant difference between the 1- and 3-month groups in the mean roughness of the ST. The roughness of AC exhibited a higher correlation with OA score than those of SP and ST. The moduli of SP and ST had significant correlations with OA score (r = 0.6610, r = 0.7390, respectively; P＜0.001). This study indicated that the quantitative surface characterization results may be utilized to develop an objective OA assessment method for clinical application and assist in the cartilage tissue engineering.

Keywords: Signal pattern classification - Genetic algorithms, Time-frequency and time-scale analysis - Time-frequency analysis, Data mining and big data methods - Machine learning and deep learning methods
Abstract: This work seeks to determine a group of neural biomarkers that a classification algorithm could use in an adaptive deep brain stimulator using rodent animal models. To overcome the variability introduced from the small sample size, this work proposes a novel method for combining and running Genetic Feature Selection and Stepwise Feature Selection. Three separate classification algorithms, Logistic Regression (LR), k-Nearest Neighbor (KNN), and Random Forest (RF) are used to verify the proposed method. For LR, the method finds the set of Alpha Power, High Beta Power (20-30 Hz), and 55-95 Hz Power to have the best performance in classification. For KNN, it finds Low Beta Power (12-20 Hz), High Beta Power, All Beta Power (12-30 Hz), 55-95 Hz Power, and 95-105 Hz Power. For RF, the results are High Beta Power, All Beta Power, 55-95 Hz Power, 95-105 Hz Power, and 300-350 Hz Power.

Keywords: Models and simulations of therapeutic devices and systems, Neuromodulation devices, Drug Delivery Systems and Carriers
Abstract: Drug delivery to the cerebral cortex by infusion was modeled using finite element analysis. The results support that this method of drug delivery could be effective in treating a relevant volume of brain tissue over short time scales.

Keywords: X-ray - Interventional radiology, Functional image analysis
Abstract: Contrast reflux, which is the retrograde movement of contrast proximal to the catheter-tip, is a commonly observed phenomenon during standard 2D angiography. The hemodynamics underlying contrast reflux has not been properly investigated in the literature. We conducted an in vivo study to evaluate the effect of injection conditions and baseline hemodynamics on reflux. Several parameters (catheter size, injection rate, Reynolds number of flow) were significantly correlated to contrast reflux length, while injection time had no effect. These correlations have never been noted before and may have clinical utility. For example, once this methodology is validated, a simple measurement of the length of contrast reflux during standard angiography can provide an estimate of baseline (pre-injection) mean blood flow rate in the patient’s vessel of interest.

Keywords: Neuromuscular systems - Stroke therapy devices / technologies, Models and simulations of therapeutic devices and systems
Abstract: Animal models used to test mechanical thrombectomy devices mostly use autologous blood to prepare thrombi ex vivo that are then injected back into the animal. In this study, we compared the thrombi produced from multiple thrombus preparation methods in terms of their histological characteristics, injectability, and radiopacity. Of 8 different clot-production protocols tested, two protocols produced clots that were most similar in histological structure to thrombi extracted from ischemic stroke patients. Brief (order of seconds) rotation of blood in a syringe, incubation at room temperature, the addition of fibrinogen, and use of catgut sutures (collagen) may produce experimental thrombi closest to that of stroke. Further testing is needed to verify our results.

Keywords: Cardiac electrophysiology - Simulation for cardiac arrhythmia, Cardiac electrophysiology - Ventricular arrhythmia mechanisms, Cardiac electrophysiology - Atrial fibrillation
Abstract: A cardiac pacemaker cell was modeled through the use of MATLAB. Modifications to the individual parameters of the model were studied and it was subsequently determined that the funny current directly affected the frequency of the action potential wave, ultimately impacting the calculated heart rate.

Keywords: Health Informatics - Emerging IT for efficient/low-cost healthcare delivery, Health Informatics - Health information systems and convergence of healthcare, Health Informatics - eHealth
Abstract: Insomnia is a highly prevalent sleep disorder. A simple way to improve sleep quality has great importance. In this study, we analyzed EEG data recoded from subjects during sleep in three types of sound environments: silence, environmental noise, and relaxing music. We used an automatic spindle detection algorithm to evaluate the sleep quality from the viewpoint of spindle occurrence.

The results showed that the number of sleep spindles increased in seven of eleven subjects. Among five subjects with poor sleep efficacy, four subjects had increased number of spindles under the relaxing music environment.

This study indicated that the relaxing music environment could improve the sleep quality, particularly in subjects with insomnia.

Keywords: Ultrasound imaging - Other organs, Image feature extraction
Abstract: In vivo determination of skeletal muscle architecture is often implemented with Diffusion Tensor Imaging (DTI). In comparison to ultrasound imaging, DTI requires longer acquisition times, is costlier, and is much less portable than ultrasound. This work aims to determine pennation angles from 3D freehand ultrasound images. To this end, we reconstructed 3D volumes by simultaneously obtaining ultrasound images and optical marker data. We computed fiber orientations using a multiscale vessel enhancement filtering method.