A novel framework for cognitive state identification using resting-state EEG.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Artificial Intelligence

Background:

• Cognitive impairment research is advancing, with electroencephalography (EEG) showing promise for early detection in the elderly.
• Changes in neural activity and functional connectivity patterns correlate with cognitive decline.
• Identifying cognitive states is crucial for timely intervention and management of cognitive impairment.

Purpose of the Study:

• To introduce PowerSyncNet, a novel deep learning framework for cognitive state identification using EEG functional connectivity.
• To develop a framework that effectively extracts and analyzes functional connectivity features across different frequency bands.
• To enhance the accuracy of cognitive state identification compared to existing deep learning methods.

Main Methods:

• Developed PowerSyncNet, a framework comprising three modules: Channel-Pair Feature Sequences Builder, Encoder4Band, and Classifier.
• Extracted features characterizing functional connectivity across various frequency bands.
• Utilized temporal-frequency representations and cross-band information for improved feature clarity.

Main Results:

• PowerSyncNet demonstrated superior performance in cognitive state identification on both the CAUEEG and ECED datasets.
• The framework effectively captured temporal-frequency representations indicative of cognitive states.
• Results indicate enhanced feature clarity through combined cross-band information.

Conclusions:

• PowerSyncNet offers a powerful tool for accurate cognitive state identification based on EEG functional connectivity.
• The framework facilitates early assessment and timely intervention for individuals with cognitive impairment.
• This approach holds significant potential for improving patient outcomes in cognitive decline research.