Epileptic seizures are believed to follow a circadian rhythm. There is an underlying relationship between Epilepsy and the sleep state. It is important to understand and evaluate the correlation between them. A robust algorithm to determine the sleep-wake states in the animals was developed and the sleep in the light and dark cycles for the epilepsy-induced animals vs the control animals was evaluated.

A quantitative study of the traditional EEG bands - delta, theta, alpha, beta and gamma were performed to understand their contribution in the sleep-wake states. The EEG band-power contribution in the epilepsy induced animal vs the control animal was evaluated in the sleep-wake states. A different set of frequency bands were evaluated to which proved more robust in differentiating the different sleep-wake states. A brute threshold algorithm and Machine learning algorithms – Logistic Regression & k-Nearest Neighbors were created to detect the sleep-wake states of the animals based on the set of EEG frequency bands.

The developed algorithms were compared for overall accuracy against the outcome of the manually marked EEG sleep segments from the Video-EEG recordings and it was observed that some of the Machine Learning algorithms outperformed the brute-threshold algorithm. Based on the detected sleep-wake states in the animals using the brute threshold algorithm, an evaluation was performed on the effect of induced epilepsy on the duration of the sleep-wake states of the animals. It was observed that the animals with induced epilepsy slept for longer durations in the light cycle as compared to the control animals. A comparative analysis was performed to determine the relationship between the sleep states in the diseaseinduced animals vs the control animals based on the light-dark cycle.