Automatic recognition of Alzheimer's disease using genetic algorithms and neural network

Abstract

We propose an Alzheimer's disease (AD) recognition method combined the genetic algorithms (GA) and the artificial neural network (ANN). Spontaneous EEG and auditory ERP data recorded from a single site in 16 early AD patients and 16 age-matched normal subjects were used. We made a feature pool including 88 spectral, 28 statistical and 2 nonlinear characteristics of EEG and 10 features of ERP. The combined GA/ANN was applied to find the dominant features automatically from the feature pool, and the selected features were used as a network input. The recognition rate of the ANN fed by this input was 81.9% for the untrained data set. These results lead to the conclusion that the combined GA/ANN approach may be useful for an early detection of the AD. This approach could be extended to a reliable classification system using EEG recording that can discriminate between groups. © Springer-Verlag Berlin Heidelberg 2003.