Performing Dynamic Weight-Shifting Balance Exercises With a Smartphone-Based Wearable Telerehabilitation System for Home Use by Individuals With Parkinson's Disease: A Proof-of-Concept Study.

Telerehabilitation technology often helps individuals with Parkinson's disease (PD) to control their balance and improve postural stability. This proof-of-concept study describes the redesign of a smartphone-based wearable balance rehabilitation system, or Smarter Balance System (SBS) intended for in-home use, and determines the number of exercise sessions required to achieve steady-state balance exercise performance by people with PD who performed 6 weeks of in-home dynamic weight-shifting balance exercises. The redesigned SBS supplied real-time multimodal (visual and vibrotactile) biofeedback during dynamic weight-shifting balance exercises (WSBEs). A Technology Acceptance Model (TAM) questionnaire completed by participants validated its acceptability and use. The results of regression analyses of participants' balance exercise performance, based on the average cross-correlations and absolute position errors between the target motion and the exerciser's motion, showed exponential trends, a performance plateau after 3 weeks, and a quasi-steady state performance by the end of 6 consecutive weeks.

Effective Data Augmentation, Filters, and Automation Techniques for Automatic 12-Lead ECG Classification Using Deep Residual Neural Networks.

Automatic electrocardiogram (ECG) analysis plays a critical role in early detection and diagnosis of cardiac abnormalities and diseases. Data augmentation and automation strategies have been proposed to enhance the robustness of the machine and deep learning model for the classification of cardiac abnormalities. Here we propose 15 data augmentation and 6 filters, and an automation method using an end-to-end deep residual neural network (ResNet) model for automatic cardiac abnormalities detection from 12-lead ECG recordings. We evaluate the effectiveness of data augmentation/filtering and automation techniques using the proposed ResNet-based model on the China Physiological Signal Challenge (CPSC) dataset consisting of 9 diagnostic classes. The average F1 scores across 9 classes on the CPSC dataset trained with three data augmentation (baseline wander addition, dropout, and scaling) and a filter (sigmoid compression) were significantly higher than that without using augmentation/filters (baseline). The highest average F1 score with sigmoid compression method was significantly higher (relative improvement of 2.04 %) than the baseline while horizontal and vertical flipping augmentations were detrimental to the classification performance. Additionally, the results show that the random combination of four selected data augmentation and filter using the modified RandAugment technique provided a significantly higher average F1 score (relative improvement of 2.54 %) compared to the baseline. The proposed data augmentation, filters, and automation techniques provide an effective solution to improve the classification performance of the end-to-end deep learning model from ECG recordings without changing the model hyperparameters and structure.

Aging Affects Lower Limb Joint Moments and Muscle Responses to a Split-Belt Treadmill Perturbation.

Age-related changes cause more fall-related injuries and impede the recoveries by older adults compared to younger adults. This study assessed the lower limb joint moments and muscle responses to split-belt treadmill perturbations in two groups (14 healthy young group [23.36 ± 2.90 years] and 14 healthy older group [70.93 ± 4.36 years]) who performed two trials of unexpected split-belt treadmill perturbations while walking on a programmable split-belt treadmill. A motion capture system quantified the lower limb joint moments, and a wireless electromyography system recorded the lower limb muscle responses. The compensatory limb's (i.e., the tripped limb's contralateral side) joint moments and muscle responses were computed during the pre-perturbation period (the five gait cycles before the onset of a split-belt treadmill perturbation) and the recovery period (from the split-belt treadmill perturbation to the baseline gait relying on the ground reaction forces' profile). Joint moments were assessed by maximum joint moments, and muscle responses were quantified by the normalization (%) and co-contraction index (CCI). Joint moments and muscle responses of the compensatory limb during the recovery period were significantly higher for the YG than the OG, and joint moments (e.g., knee flexion and extension and hip flexion moments) and muscle responses during the recovery period were higher compared to the pre-perturbation period for both groups. For CCI, the older group showed significantly higher co-contraction for biceps femoris/rectus femoris muscles than the young group during the recovery period. For both groups, co-contraction for biceps femoris/rectus femoris muscles was higher during the pre-perturbation period than the recovery period. The study confirmed that older adults compensated for muscle weakness by using lower joint moments and muscle activations and increasing muscle co-contractions to recover balance after split-belt treadmill perturbations. A better understanding of the recovery mechanisms of older adults who train on fall-inducing systems could improve therapeutic regimens.

Effects of a Smartphone-Based Wearable Telerehabilitation System for In-Home Dynamic Weight-Shifting Balance Exercises by Individuals with Parkinson's Disease.

This paper describes the effects of a smartphone-based wearable telerehabilitation system (called Smarter Balance System, SBS) intended for in-home dynamic weight-shifting balance exercises (WSBEs) by individuals with Parkinson's disease (PD). Two individuals with idiopathic PD performed in-home dynamic WSBEs in anterior-posterior (A/P) and medial-lateral (M/L) directions, using the SBS 3 days per week for 6 weeks. Exercise performance was quantified by cross-correlation (XCORR) and position error (PE) analyses. Balance and gait performance and level of fear of falling were assessed by limit of stability (LOS), Sensory Organization Test (SOT), Falls Efficacy Scale (FES), Activities-specific Balance Confidence (ABC), and Dynamic Gait Index (DGI) at the pre-(beginning of week 1), post-(end of week 6), and retention-(1 month after week 6) assessments. Regression analyses found that exponential trends of the XCORR and PE described exercise performance more effectively than linear trends. Range of LOS in both A/P and M/L directions improved at the post-assessment compared to the pre-assessment, and was retained at the retention assessment. The preliminary findings emphasize the advantages of wearable balance telerehabilitation technologies when performing in-home balance rehabilitation exercises.

Electrocortical activity changes in response to unpredictable trip perturbations induced by a split-belt treadmill.

This study explored the contributions of cortical activity in the primary sensorimotor cortex (SMC) and the posterior parietal cortex (PPC) to recovery responses following unpredictable trip perturbations. A technology platform equipped with a programmable split-belt treadmill induced unpredictable trip perturbations while walking. 128-channel non-invasive electroencephalography (EEG) signals were collected. Power spectral analysis was performed to quantify the electrocortical activity of two clusters in the SMC and PPC during quiet standing, steady state walking, and recovery periods. Alpha (8-13 Hz) power of the SMC and PPC was significantly suppressed during the recovery period compared to the standing and walking periods. The main finding of this study could inform the future development gait perturbation paradigms that facilitate the recovery responses in different populations, based on motor learning by repetition.

Reward Expectation Modulates Local Field Potentials, Spiking Activity and Spike-Field Coherence in the Primary Motor Cortex.

Reward modulation (M1) could be exploited in developing an autonomously updating brain-computer interface (BCI) based on a reinforcement learning (RL) architecture. For an autonomously updating RL-based BCI system, we would need a reward prediction error, or a state-value representation from the user's neural activity, which the RL-BCI agent could use to update its BCI decoder. In order to understand the multifaceted effects of reward on M1 activity, we investigated how neural spiking, oscillatory activities and their functional interactions are modulated by conditioned stimuli related reward expectation. To do so, local field potentials (LFPs) and single/multi-unit activities were recorded simultaneously and bilaterally from M1 cortices while four non-human primates (NHPs) performed cued center-out reaching or grip force tasks either manually using their right arm/hand or observed passively. We found that reward expectation influenced the strength of α (8-14 Hz) power, α-γ comodulation, α spike-field coherence (SFC), and firing rates (FRs) in general in M1. Furthermore, we found that an increase in α-band power was correlated with a decrease in neural spiking activity, that FRs were highest at the trough of the α-band cycle and lowest at the peak of its cycle. These findings imply that α oscillations modulated by reward expectation have an influence on spike FR and spike timing during both reaching and grasping tasks in M1. These LFP, spike, and spike-field interactions could be used to follow the M1 neural state in order to enhance BCI decoding (An et al., 2018; Zhao et al., 2018).

Manipulator-driven selection of semi-active MR-visible markers.

BACKGROUND: A method for the identification of semi-active fiducial magnetic resonance (MR) markers is presented based on selectively optically tuning and detuning them. METHODS: Four inductively coupled solenoid coils with photoresistors were connected to light sources. A microcontroller timed the optical tuning/detuning of coils and image collection. The markers were tested on an MR manipulator linking the microcontroller to the manipulator control to visibly select the marker subset according to the actuated joint. RESULTS: In closed-loop control, the average and maximum were 0.76° ± 0.41° and 1.18° errors for a rotational joint, and 0.87 mm ± 0.26 mm and 1.13 mm for the prismatic joint. CONCLUSIONS: This technique is suitable for MR-compatible actuated devices that use semi-active MR-compatible markers.

Near Perfect Neural Critic from Motor Cortical Activity Toward an Autonomously Updating Brain Machine Interface.

We are developing an autonomously updating brain machine interface (BMI) utilizing reinforcement learning principles. One aspect of this system is a neural critic that determines reward expectations from neural activity. This critic is then used to update a BMI decoder toward an improved performance from the user's perspective. Here we demonstrate the ability of a neural critic to classify trial reward value given activity from the primary motor cortex (M1), using neural features from single/multi units (SU/MU), and local field potentials (LFPs) with prediction accuracies up to 97% correct. A nonhuman primate subject conducted a cued center out reaching task, either manually, or observationally. The cue indicated the reward value of a trial. Features such as power spectral density (PSD) of the LFPs and spike-field coherence (SFC) between SU/MU and corresponding LFPs were calculated and used as inputs to several classifiers. We conclude that hybrid features of PSD and SFC show higher classification performance than PSD or SFC alone (accuracy was 92% for manual tasks, and 97% for observational). In the future, we will employ these hybrid features toward our autonomously updating BMI.