ABSTRACT
Introduction
The problem of determining brain electrical sources from potential patterns recorded on the scalp surface is mathematically underdetermined. Most efforts to identify EEG sources have focused on performing simultaneous spatial segregation and localization of source activity. Recently, we have applied the ICA algorithm of Bell and Sejnowski [1] to the problem of EEG source identification (What?) considered apart from source localization (Where?) [2]. By maximizing the joint entropy of a set of output channels derived from input signals by linear filtering without time delays, the ICA algorithm attempts to derive independent source waveforms from highly correlated scalp EEG signals without regard to the physical locations or configurations (focal or diffuse) of the source generators.
