CLASSIFICATION OF DRUG-INDUCED BEHAVIORS USING A MULTILAYER FEEDFORWARD NEURAL-NETWORK

Measurement of laboratory animal motor behavior is an important part o f many studies of experimental manipulations of the central nervous sy stem. Current automated data collection and analysis systems are limit ed in the number of behaviors that can be monitored and quantified sim ultaneously. This paper describes a signal analysis technique that whe n used to analyze the data from a modified Stoelting electronic activi ty monitor is capable of classifying multiple behavior categories auto matically. In this technique, the output signal from the motility moni tor is fixed-length segmented and feature extraction is performed, cal culating the Fourier transform and power spectrum of each data segment . An error back-propagation neural network, implemented on a microcomp uter, is used to perform behavior classification of the segment power spectra. The technique provides a high degree of accuracy in automatic behavior classification and should prove useful in the quantitative a ssessment of behavior.