Mixed-Input Deep Learning Approach to Sleep/Wake State Classification by Using EEG Signals.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Artificial Intelligence

Background:

• Accurate sleep stage classification is crucial for diagnosing sleep disorders and related health conditions.
• Manual sleep staging is labor-intensive and subject to inter-scorer variability.
• Advancements in automatic sleep staging leverage polysomnography data, particularly electroencephalogram (EEG) signals.

Purpose of the Study:

• To develop and evaluate a hybrid deep learning model for classifying sleep and wake states using single-channel EEG.
• To improve the accuracy and efficiency of automatic sleep stage classification.

Main Methods:

• A hybrid deep learning model combining an Artificial Neural Network (ANN) and a Convolutional Neural Network (CNN) was developed.
• The ANN processed statistical features from EEG epochs.
• The CNN analyzed Hilbert spectrum images derived from EEG epochs.
• The model was trained and tested on single-channel Pz-Oz EEG data from the Sleep-EDF database Expanded.

Main Results:

• The proposed hybrid model achieved approximately 96% accuracy in classifying sleep and wake states.
• Performance was evaluated on EEG recordings from four individuals.

Conclusions:

• The hybrid ANN-CNN model demonstrates high efficacy for automatic sleep-wake classification from single-channel EEG.
• This approach offers a promising, accurate, and potentially more efficient alternative to manual sleep staging.