A microcomputer-based system for mapping EEG signals after source derivation

Abstract

The EEG (electroencephalogram) data are obtained as a set of up to 16 unipolar derivations, sampled and displayed in real time, before storing on disk. The source derivation technique counteracts the smearing of the electrical fields as measured on the scalp and thus improves the spatial selectivity of mapping. It is calculated using the finite difference method, assuming a uniform grid of electrodes or alternatively a non-uniform grid. Source derivation is applied to the data before estimating the power spectral density function (using FFT) from selected epochs of the signals. Spectral parameters can then be extracted and mapped on a surface corresponding to the lateral projection of a brain hemisphere. Spatial interpolation is employed to estimate the parameter values at points where no electrodes are located. The brain topograms are then exhibited on the computer monitor. The software was written in PASCAL with time-critical routines in assembly language.<>