A Data-Driven Algorithm for Identification of Tonic and Phasic Sleep Phases based on EEG Data

Abstract

Rapid Eye Movement (REM) sleep is a crucial phase of sleep, which is responsible for memory consolidation, emotional regulation, and cognitive function [1]. REM sleep consists of phasic and tonic microstates that exhibit distinct neural activity patterns. Our research endeavors to investigate the currently limited knowledge regarding neural activity within subcortical structures during both tonic and phasic REM sleep phases [2]. The primary objective of this study is to devise a data-driven algorithm that can effectively discern between tonic and phasic sleep phases with high accuracy, leveraging EEG data as its foundation. A dataset of 20 participants, professionally coded with tonic and phasic sleep phase labels, have been collected. The analysis will involve several steps, including data preprocessing, feature extraction, feature selection, model training using machine learning algorithms, model evaluation, algorithm validation on an independent dataset, refinement and optimization, and interpretation of results. The proposed algorithm will be designed to capture specific characteristics of tonic and phasic REM sleep phases by extracting relevant features and utilizing machine learning techniques. The algorithm will be validated and optimized to ensure its robustness and generalizability. The ultimate goal is to gain insights into the neurophysiological mechanisms underlying tonic and phasic REM sleep phases. This research aims to contribute to the understanding of sleep physiology and potentially facilitate the development of novel sleep stage classification methods.

References

[1] D. Tempesta, V. Socci, L. De Gennaro, and M. Ferrara, "Sleep and emotional processing," Sleep Medicine Reviews, vol. 40, pp. 183-195, 2018.

[2] P. Simor et al., “REM sleep microstates in the human anterior thalamus,” The Journal of Neuroscience, vol. 41, no. 26, pp. 5677–5686, 2021. doi:10.1523/jneurosci.1899-20.2021