Sleep Stage and Heat Stress Classification of Rodents Undergoing High Environmental Temperature

Abstract

Stress is one of the major concerns originated from neuronal activities which may lead to mental health problems, such as anxiety, depression, and personality disorders. Physiological studies have also been carried out to explore the application of computing techniques to predict “Heat Stress”—stress which develops due to high environmental temperature. Prerecorded data has been synthesized and analyzed to detect the changes in sleep electroencephalogram (sleep EEG) under heat stress. This work presents a technique to detect the heat stress by employing linear discriminant analysis (LDA) followed by continuous wavelet transform (CWT). Through wavelet decomposition, different frequencies embedded in the EEG signal were analyzed and features were extracted to detect the changes in stressed data with respect to control. The comparison of LDA with Adaptive neuro-fuzzy system (ANFIS) has also been addressed, where LDA shows good accuracy in stressed REM pattern as compared to other two stages of sleep EEG. An increase of 7.5% has been observed in LDA while detecting REM patterns.