Presentation Title: Construction of a Machine Learning-based Neural Interface
Section: Poster Session Group 2 2
Location: McCardell Bicentennial Hall, Great Hall
Date & Time: Friday, April 19, 2013 - 2:45pm - 3:30pm
Once considered expensive research hardware, electroencephalograph (EEG) neural scanners are now mass-produced and commercially available. Using one such commercial EEG headset, we have constructed a brain-computer neural input interface. This neural interface detects electromagnetic brain waves emanating from a userâ€™s scalp and decodes them into a set of mental commands to be sent to a computer. Previous neural interface system have expected users to learn how to produce the specific brain states recognized by the neural interface; these systems rely strictly on the user's learning ability and, thus, require lengthy training times. In contrast, a neural interface based on machine learning (ML) principles can greatly shorten this training period by learning to recognize mental states specific to each user. Using ML and commercial hardware, we have produced an adaptive neural interface that reduces user training time while allowing hands-free mental input.
Type of Presentation: Individual poster
Number of presenters:
Presenter(s): Horn, Colby Ansel
Major(s): Computer Science
Class Year(s): 2013
Sponsor(s): Kauchak, David R.
Dept(s): Computer Science