Identification of finger flexions from continuous EEG as a brain computer interface

Much of the research in the development of a Brain Computer Interface (BCI) has focused on differentiating between several possible commands rather than the identification of control signals from continuous EEG. This generally results in the detection of many unintended commands while the operator is trying to rest. This work implements an Asynchronous Signal Detector (ASD) capable of identifying index finger flexions from a continuous sampling of surface electrodes. Spatiotemporal features are classified using Learning Vector Quantization (LVQ). The ASD can function as a stand alone BCI capable of recognizing a single control signal or operate in conjunction with an existing BCI method to recognize multiple commands