New approach for Alzheimer’s disease classification using topographic maps and deep learning model

With an aging global population, age-related illnesses like Alzheimer’s disease (AD) are a growing research priority. Electroencephalogram (EEG) offers a low-cost and safe way to collect brain activity for AD detection. However, analyzing EEG is labor-intensive, and experts’ visual inspection is subjective and prone to errors, especially with lengthy recordings containing signal artifacts. This study proposes a novel method for diagnosing AD, Frontotemporal Dementia (FTD), and healthy control (HC) subjects using brainwave patterns from EEG. Our approach involves analyzing the different frequencies within the EEG signal. We then create topographic maps, which are special images that capture the strength of these brainwaves across different brain areas. Finally, we apply a deep learning method to these images to distinguish between AD, FTD, and HC subjects. The results demonstrate an improvement in classification performance compared to traditional hand-crafted feature methods, achieving accuracy of 83.0%, 85.5%, and 87.3% for AD/FTD, AD/HC, and FTD/HC, respectively. Our approach also achieves results comparable to state-of-the-art methods on the same dataset, highlighting its potential as an effective tool for the diagnosis of AD and FTD.