Wavelet-Neural Classification of Sleep EEG under Stressful Condition

Alterations in transients during awake, slow wave sleep (SWS), and rapid eye movement (REM) sleep stages due to an exposure to high environmental heat have been studied using continuous wavelet transform (CWT) method and artificial neural network (ANN). After two hours long EEG (Electroencephalogram) recordings from healthy rats, EEG data representing three sleep states was visually selected and further subdivided into 2 seconds long epoch. After extracting features in terms of wavelet coefficients for all the epochs multilayer perceptron neural network (MLPNN) has been trained to detect changes in the vigilance states of the subjects exposed to environmental heat stress. It reveals that, the classifications of wavelet coefficients of EEG signals in acute as well as chronic heat conditions along with the control data show the overall accuracy of 94.5% in SWS, 91.75% in REM sleep and 91.75% in AWAKE state.