Files in this item



application/pdfECE499-Fa2009-mccormick.pdf (1MB)Restricted to U of Illinois
(no description provided)PDF


Title:Signal Processing for Brain-Computer Interfacing Applications
Author(s):McCormick, Martin
Contributor(s):Coleman, Todd
Subject(s):human-computer interfaces
brain-computer interfaces
signal processing
Abstract:Direct brain-computer interfacing allows for new types of human interaction-augmenting the ability of both healthy individuals for gaming, military and productivity purposes and disabled persons with locked-in syndromes. We propose a new classification method for EEG-based brain-computer interfaces, the Common Spatial Analytical Pattern (CSAP) in combination with a hidden Markov model, which significantly improves the current prevailing technique employed for binary motor imagery classification. The CSAP method solves a blind source separation problem and estimates discriminative source signal variance from the instantaneous envelope of a narrow-band signal centered at the task-relevant mu-rhythm frequency (10 to 14 Hz) using the magnitude of an analytic signal. The hidden Markov model is formulated to perform classification by belief propagation algorithm, and to account for the interdependence between consecutive classification estimates. Experiments showed information transfer rates between the subject and the computer as high as 81 bits/minute, exceeding the best rates published in the literature. This channel is then used to spell sentences, specify wheelchair paths and even fly a remote control plane.
Issue Date:2009-12
Publication Status:unpublished
Peer Reviewed:not peer reviewed
Date Available in IDEALS:2014-01-22

This item appears in the following Collection(s)

Item Statistics