Cognitive-based Tuning of 6-DOF Devices for SCI Rehabilitation
Samuel Wilder, MS in Biomedical Engineering | BME High Academic Achievement Award Department of Biomedical Engineering, Summer 2020 Title: Sense of Agency and Performance in Using 6-DOF Devices
Abstract: In this preliminary study (n = 4), we characterized sense of agency and performance in the use of six degrees of freedom (6-DOF) devices. We observed how agency, or perception of control, and performance changes with devices whose manual interface is either fixed (SpaceMouse) or compliant (Stewart-Gough Platform). We investigated the covariations in agency versus performance of a simple circle tracing task across multiple operation-sensitivity levels for each device. The presented approach in this study may be relevant to the design of better motor rehabilitation platforms utilized in physical therapy after neurotraumas such as spinal cord injury. Computerized interfaces are increasingly used in rehabilitation since they are cognitively engaging and encourage skillful hand function. Rehabilitation methods that adapt the device interface to promote greater user agency and performance may facilitate better motor outcomes. In this study, subjects performed a circle tracking task while agency was implicitly assessed using time-interval estimation based on intentional binding. Our findings are currently inconclusive with completion of only 4 subject data collections. Our preliminary results indicate that performance dependence on agency is greater with the fixed interface. This finding suggests that motion at the manual interface introduces uncertainties to the user that degrade agency-based performance. Future work should include additional data collections, evaluation of neurophysiological (electromyography, electroencephalography) signals as signatures for agency-based performance, and identification of device design features that are generalizable to any human-machine interface (HMI).
2020 Johnson and Johnson Engineering Showcase
Sean Sanford and Mingxiao Liu competed in the 2020 Johnson and Johnson Engineering Showcase held in New Brunswick, NJ. Sean Sanford finished 1st place in the University Poster competition.
2019 Biomedical Engineering Society Annual Meeting
Sean Sanford and Mingxiao Liu attended the 2019 Biomedical Engineering Society Annual Meeting held in Philadelphia, PA. Mingxiao Liu presented his work on Cognitive Agency while Sean Sanford presented his own research and the research of fellow MOCORE member, Kevin Walsh.
2019 Innovation and Entrepreneurship Summer Scholars
Brian Collins and Daniel Kang represented the MOCORE laboratory in the 2019 Office of Innovation and Entrepreneurship Summer Scholarship program. Brian Collins developed work for a vibrotactile stimulation device geared towards virtual reality rehabilitation research. Daniel Kang worked to further develop a prototype glove for the Cognitive Agency research project.
Hoboken Summer School Community Outreach
Hoboken Summer school students of all ages toured various laboratories at Stevens. The tour group stopped by the MOCORE laboratory for a short presentation given by Sean Sanford and Mingxiao Liu on neuromuscular pathologies and assistive devices. One student had the opportunity to demonstrate a new virtual reality task developed by Sam Wilder.
Hoboken Middle School Community Outreach
Local Hoboken Middle School students had the opporunity to visit the MOCORE laboratory and take part in interactive workshops geared towards our research in user-device integration.
Mount Sinai Spinal Cord Injury Community Fair
The MOCORE laboratory presented a booth at the 2nd Annual Spinal Cord Injury Research Community Fair located at Mount Sinai in Manhattan, NY.
Adaptive Machine Learning for Myoelectric Control
Kevin Walsh, MS in Bioengineering | Outstanding Dissertation Award Department of Biomedical Engineering, Spring 2019 Title: Simulated Nervous System Lesions in Machine Learning for Myoelectric Control
Abstract: Neural disorders or traumas can significantly impact an individual’s quality of life, due to impaired motor function. Electromyography (EMG) signals arising from muscle activation have become popular for providing inputs to advanced prosthetics that help these patients. Machine learning (ML) methods such as artificial neural networks (ANN) and adaptive boosting can recognize patterns in EMG signals and classify them into useful command inputs. Myoelectric systems are applied to advanced prostheses and other rehabilitation devices as a means for natural user control. This thesis demonstrates 91% accuracy in classifying movements by orthogonal movement direction and force exerted during the movement using ANN and 96% accuracy using adaptive boosting ML classifiers. However, the muscles that are available for EMG recordings may be limited or compromised depending on the specific nature of each movement disability. The objective of this study was to investigate what classification accuracy may be achieved in the prediction of directional and force magnitude controls from the muscle recordings presumed available following various neural lesions. Simulated nervous system lesions lowered classification accuracy to between 35% for a data set with a simulated C5 spinal cord lesion that receives input only from the trapezius and 88% for a data set with a simulated ulnar nerve lesion. All classification accuracies were greater than chance, showing that these simulated deficiencies still may allow for some measure of device control. These ML methods will be further investigated for ‘functional’ control of a prosthetic device or virtual reality (VR) arm in real time. Our research group is currently developing capabilities for real-time streaming of EMG data into MATLAB and VR for this purpose.
ISPE Regional Conference
Sean Sanford was awarded 1st place in the University poster competition at the International Society of Pharmaceutical Engineering conference and was the regional representative for the 2019 ISPE Annual Meeting and Expo
Role of Agency in Hand Reach and Grasp Rehabilitation
David Hollinger, MEng Department of Biomedical Engineering, Fall 2018 Title: Accelerating Neuromotor Learning with Reward Feedback
Abstract: Approximately 85% of stroke survivors suffer hemiparesis, resulting in loss of motor function on the contralateral side of the damaged brain (Nakayama et al., 1994). 46-95% of hemiparesis survivors still experience problems 6 months post stroke (Kong et al., 2011). Interventions focusing on high-intensity and repetitive task-specific practice showed the most promise for improving motor recovery (Ávila et al., 2012). The rate of motor learning depends on task type and feedback display. Feedback following ‘good’ trials has shown to enhance motor learning and retention during a sequential timing task (Chiviacowsky & Wulf, 2002). Previous studies using virtual reality (VR) gaming have leveraged desired rehabilitation movement outcomes. The element of gamification heightens user interaction during repetitive rehabilitation movement tasks (Merians et al., 2002; Jack et al., 2001; Deutsch et al., 2001). Patients in control of their movements gain greater motor learning benefits than those without control. Therefore, it is imperative to incorporate methods which maximize user control, or agency, to enhance motor learning. Our research study will test how reward and punishment impacts motor performance (e.g. path length, end effector accuracy) and agency during an upper arm reach-to-touch VR task. Results indicate that participants who receive reward outperformed participants receiving punishment. Reward enhanced cognitive agency and contact accuracy indicating that knowledge of positive results led to improved self-efficacy and motor performance. The results agree with OPTIMAL theory of motor learning where positive motivation and external focus of attention facilitate automaticity of movement control and overall motor learning (Lewthwaite & Wulf, 2016).
Brain-Computer Interface Seminar
Co-sponsored by renowned Christoph Guger, CEO, g.tec, Austria
2018 Innovation and Entrepreneurship Summer Scholar Award
Office of Innovation and Entrepreneurship Scholar Award presented to Corrine Rybarski for 1st Place in the Elevator Pitch Competition
2018 American Society of Biomechanics Annual Meeting
42nd Annual Meeting of the American Society of Biomechanics, August 8th-11th 2018
Stevens Team Uses Virtual Reality to Help Control Prosthetic Limbs and Restore Post-Amputation Sense of Motion Click to view article
44th Annual Northeast Bioengineering Conference, March 28th-30th 2018