EEG-based clinical decision support system for Alzheimer's disorders diagnosis using EMD and deep learning techniques

Area of Science:

• Neuroscience
• Biomedical Engineering
• Artificial Intelligence

Background:

• Early detection of Alzheimer's disease (AD) is crucial for effective management.
• Electroencephalogram (EEG) signals offer a promising, non-invasive method for neurological disorder assessment.
• Existing clinical techniques may not always identify AD in its earliest stages.

Purpose of the Study:

• To develop a reliable and accurate clinical decision support system for early AD detection using EEG.
• To leverage EEG signal processing and artificial intelligence for distinguishing between neurotypical individuals and those with mild to moderate AD.
• To compare the efficacy of various AI approaches in analyzing EEG features for AD diagnosis.

Main Methods:

• Utilized a dataset comprising EEG recordings from neurotypical individuals and patients with mild and moderate AD.
• Applied band-pass filtering and Empirical Mode Decomposition (EMD) for EEG signal feature extraction.
• Employed artificial intelligence classifiers and evaluated performance using k-fold and leave-one-subject-out (LOSO) cross-validation.

Main Results:

• Achieved a maximum classification accuracy of 99.9% with k-fold cross-validation.
• Attained a classification accuracy of 94.8% using the leave-one-subject-out (LOSO) cross-validation method.
• Demonstrated the effectiveness of combined signal features and AI for distinguishing AD severity.

Conclusions:

• The proposed EEG-based diagnostic support system shows significant potential for early AD detection.
• Findings suggest the developed methodologies can aid in identifying novel diagnostic biomarkers for AD.
• This approach offers a promising supplementary tool for early clinical diagnosis of Alzheimer's disease.