Some methods for dynamic analysis of the scalp recorded EEG

This paper describes methods for quantifying the spatiotemporal dynamics of EEG. Development of these methods was motivated by watching computer-generated animations of EEG voltage records. These animations contain a wealth of information about the pattern of change across time in the voltages observed across the surface of the scalp. In an effort to quantify this pattern of changing voltages, we elected to extract a single quantifiable feature from each measurement epoch, the highest squared voltage among the various electrode sites. Nineteen channels of EEG were collected from subjects using an electrode cap with standard 10-20 system placements. Two minute records were obtained. Each record was sampled at a rate of 200 per second. Thirty seconds of artifact-free data were extracted from each 2 minute record. An algorithm then determined the location of the channel with the greatest amplitude for each 5 msec sampling epoch. We quantified these spatio-temporal dynamics as scalars, vectors and cluster analytic plots of EEG activity for finger tapping, cognitive effort (counting backwards) and relaxation to illustrate the utility of the techniques.