Tracking of the Mu Rhythm using an Empirically Derived Matched Filter

Abstract

This paper introduces a method for improved detection and tracking of cortical mu rhythm modulation for the purpose of a brain-computer interface (BCI). The cortical mu rhythm found in the EEG is of particular interest in BCIs because it can be modulated through motor imagery and can be monitored via noninvasive techniques. With proper training, a disabled person can learn to control the mu rhythm to operate a communication device. This paper discusses the extraction of the empirical mu rhythm, proposes a synthetic model for the mu rhythm, and examines the effectiveness of the synthetic model as a matched filter on two-dimensional cursor control data recorded from a BCI