Balancing Spectral, Temporal and Spatial Information for EEG-based Alzheimer's Disease Classification

Area of Science:

• Neuroscience
• Biomedical Engineering
• Medical Imaging

Background:

• Cost-effective screening for Alzheimer's disease (AD) is essential for future treatments.
• Electroencephalography (EEG) is an economical neuroimaging modality with potential for AD diagnosis.
• Recent EEG analysis focuses on spatial information using graph-based frameworks.

Purpose of the Study:

• To investigate the relative importance of spatial, spectral, and temporal information in EEG for AD classification.
• To evaluate how varying the proportion of each dimension impacts classification accuracy.
• To propose a resolution-based feature extraction method for improved EEG-based AD diagnosis.

Main Methods:

• Systematic testing of different dimension resolution configurations on two EEG datasets.
• Analysis of EEG data by varying the proportion of spatial, spectral, and temporal features.
• Utilizing graph signal processing or graph neural networks for spatial information extraction.

Main Results:

• Spatial information is more critical than temporal information for AD classification using EEG.
• Spatial information holds equal importance to spectral information.
• Substituting spectral with spatial information increased classification accuracy by 1.1% on a larger dataset.

Conclusions:

• Spatial information plays a vital role in EEG-based Alzheimer's disease classification.
• The proposed resolution-based feature extraction can enhance AD classification accuracy.
• This approach has potential for improving the diagnosis of neurodegenerative diseases using multivariate signal analysis.