Lilly Richards
Pronouns: she/her
Research Mentor(s): Chandramouli Krishnan
Research Mentor School/College/Department: Physical Medicine & Rehabilitation / Medicine
Program:
Authors: Lilly Richards, Thomas Augenstein, Chandramouli Krishnan
Session: Session 4: 1:40 pm – 2:30 pm
Poster: 83
Abstract
Survivors of severe neurological injuries often have impaired motor control and require assistive devices to perform day-to-day activities. Body-machine interfaces, or BoMIs, allow these individuals to control assistive devices (e.g., robotic arms) with their residual motor control using wearable sensors. However, it is unclear how different mappings between sensors and body movements affect an individual’s ability to learn and control assistive devices. For example, BoMIs are often controlled by the user’s trunk/shoulder motion, which can create a nonintuitive mapping between body and robot movements. Further, individuals will frequently don and doff the sensors in their daily lives, and because these sensors are position-sensitive, individuals will need to frequently adapt to new mappings. Here we studied how changing the mapping between body movements and assistive device control affects motor learning in healthy participants in order to advance rehabilitation methods for individuals with neurological disorders. In this experiment, IMU sensors were attached to a human subject’s shoulders, and they were asked to use their trunk and shoulder movements to control a virtual robotic arm on a screen in front of them. Participants learned the task either with an intuitive mapping (moving the trunk left or right will move the robot left or right) or with a nonintuitive mapping (moving the trunk left or right will move the robot forward or back). Transfer tests were performed by assessing their ability to perform the task in the opposite, mirrored mapping. This mirrored mapping was meant to represent the largest possible change in mapping. Data collection is ongoing, however, pilot results indicate that participants improve their ability to control the robot using both mappings, but make smaller improvements with the nonintuitive mapping.
