Time-Frequency Analysis of Frontal EEG Channels for Alzheimer's Disease Detection Using Deep Learning

Alzheimer's disease (AD) is a progressive neurodegenerative disorder that primarily affects the elderly population. Electroencephalography (EEG) offers a non-invasive and cost-effective approach for real-time monitoring of brain activity in AD patients. This study proposes a novel EEG-based framework for AD detection, in which frontal EEG signals are transformed into time-frequency representations using continuous wavelet transform (CWT) and encoded as red-green-blue (RGB) spectrograms. A ResNet18 neural network, enhanced with a parallel convolutional block attention module (CBAM), is employed for feature extraction and classification. The proposed method achieves an accuracy of 89.23% using leave-one-subject-out cross-validation (LOSOCV) on a publicly available dataset, outperforming previously reported methods under the same validation protocol. To evaluate the generalization ability of the model, we further test it on a related neurodegenerative condition-frontotemporal dementia (FTD)-and observe promising classification performance. These results suggest that the proposed approach not only provides accurate AD detection but also holds potential for broader applications in neurodegenerative disorder analysis.Clinical Relevance: This study highlights the practicality and effectiveness of utilizing non-invasive EEG signals from frontal brain regions for AD detection. The proposed framework has the potential to assist clinicians in developing timely intervention strategies, thereby improving patient outcomes and quality of life.