Autoassociative neural networks for discrimination of chronic alcoholics using Visual Evoked Potentials

The discrimination of asymptomatic chronic alcoholics from non-alcoholics using the brain activity patterns is studied in this paper. Detection of the abnormalities in the cognitive processing ability of chronic alcoholics is essential for their rehabilitation, and also in screening them for certain jobs. The brain patterns evoked in response to visual stimuli, known as visual evoked potentials (VEP), reflect the cognitive abnormalities of such alcoholics. Weighted Linear prediction cepstral coefficients, representing the spectral envelope, are extracted from the VEP and are used as discriminating feature vectors. The potential of the nonlinear autoassociative neural networks to capture the complex distribution of the feature space is explored for the discrimination task. A recognition accuracy of 52.5% is achieved. This shows that the VEP derived from all the 64 EEG channels, as used in this study, do not contain the same discriminatory attributes, causing a reduction in the accuracy levels. Studies to achieve improvements in recognition accuracy by empirically selecting channels with significant discriminatory attributes is in progress.