Electroencephalogram-based time-frequency analysis for Alzheimer's disease detection using machine learning.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Machine Learning

Background:

• Alzheimer's disease (AD) is the most prevalent form of dementia, posing a significant global health concern due to its progressive nature and lack of effective treatments.
• The increasing prevalence and debilitating effects of AD necessitate the development of accurate and early diagnostic tools.

Purpose of the Study:

• To develop and evaluate an algorithm for differentiating between early-stage Alzheimer's disease (mild cognitive impairment [MCI]), moderate AD, and healthy controls (C) using electroencephalogram (EEG) signals.
• To assess the efficacy of machine learning classifiers in analyzing time-frequency features extracted from EEG data for AD diagnosis.

Main Methods:

• Utilized a publicly available EEG database, selecting seven recordings per group (MCI, AD, C) for a balanced dataset.
• Extracted 43 time-frequency features from 1-s EEG segments, compressed them using 10 statistical measures, and applied 15 classifiers with 7-fold cross-validation.
• Employed a novel strategy for feature analysis and classification to enhance diagnostic accuracy.

Main Results:

• Achieved 100% accuracy in binary classifications distinguishing healthy controls (C) from mild cognitive impairment (MCI) and moderate Alzheimer's disease (AD).
• Demonstrated a 2% accuracy increase for C versus MCI and a 4% increase for C versus AD compared to state-of-the-art methods.
• Outperformed prior work on the same database by 4.8% for the AD versus MCI comparison, despite a minor decrease in AD versus MCI performance.

Conclusions:

• EEG signals show significant potential as a tool for the early diagnosis of Alzheimer's disease.
• The proposed algorithm demonstrates high accuracy in differentiating AD stages, offering a promising avenue for clinical application.
• Future research should focus on utilizing larger datasets to improve the generalizability of the findings.