A Novel CNN-Based Framework for Alzheimer's Disease Detection Using EEG Spectrogram Representations.

Area of Science:

• Neuroscience
• Medical Imaging
• Artificial Intelligence

Background:

• Alzheimer's disease (AD) and frontotemporal dementia (FTD) are progressive neurodegenerative disorders with increasing global prevalence.
• Current diagnostic methods for AD and FTD are slow and resource-intensive, highlighting the need for automated solutions.
• Early and accurate diagnosis is crucial for managing dementia effectively.

Purpose of the Study:

• To develop and evaluate a novel deep learning methodology for classifying Alzheimer's disease (AD), frontotemporal dementia (FTD), and control (CN) signals using electroencephalography (EEG).
• To assess the cross-subject generalizability and performance of the proposed deep learning model against existing state-of-the-art methods.

Main Methods:

• A deep learning approach utilizing Convolutional Neural Networks (CNNs) for EEG signal classification.
• Incorporation of advanced preprocessing techniques and Fast Fourier Transform (FFT)-based spectrograms for feature extraction.
• Evaluation using leave-N-subjects-out cross-validation to ensure robust generalizability across subjects.

Main Results:

• The proposed deep learning methodology achieved high accuracy in classifying dementia types from EEG signals.
• Achieved 79.45% accuracy for Alzheimer's disease (AD) versus control (CN) classification.
• Achieved 80.69% accuracy for the combined AD and FTD versus control (CN) classification.

Conclusions:

• Deep learning models applied to EEG data show significant potential for early dementia screening.
• The developed methodology offers a more efficient, scalable, and accessible approach to dementia diagnosis.
• This research paves the way for improved diagnostic tools for neurodegenerative disorders like AD and FTD.