Presenter(s): Sydney Bright − Human Physiology
Faculty Mentor(s): Mike Hahn
Poster 25
Research Area: Human Physiology
Motor control of the forearm flexors and extensors can be impaired due to trauma, such as a stroke, which can hinder the ability to perform daily tasks. In this study, the effects of an Artificial Intelligence (AI) controller on the ability of healthy subjects to learn a novel computer game control task were examined. By using the electromyography sensors of a Myoband in tandem with a Scratch program of Flappy bird, a program was created that adapts to player specific skills. Purpose: The purpose of this experiment was to determine the motor learning outcomes given an adaptive AI motor learning environment. Hypothesis: It was hypothesized that an adaptive AI will result in less motor learning. Methods: Subjects played a modified Flappy Bird game with a multi-channel EMG sensor that fits around the forearm (Myoband). The experiment requires two consecutive days of participation. During the first day, subjects had a 2-minute warm up period followed by 20 rounds of playing the game. The second day consisted of a similar 2-minute warm up period followed by 3 rounds of the game. Results: 46 of 48 subjects have been recorded. So far, the AI group has an average 17 point improvement, linear group has 31 points and the random group has 33 point improvement. Discussion: No statistical test have been done, but from preliminary analysis of the data, the hypothesis seems to be supported.