Enhancing Alzheimer’s detection from EEG using MVMD and feature optimization algorithm

Alzheimer’s disease (AD) is a neurodegenerative disorder causing progressive cognitive decline, often associated with slower electroencephalogram activity (EEG) and reduced neural synchronization. Although EEG is important for diagnosing AD, distinguishing signals from individuals with mild cognitive impairment (MCI) or AD is challenging. Therefore, a novel framework to leverage multivariate variational mode decomposition (MVMD) subbands is developed. A key innovation of this paper lies in incorporating an optimization method called the Reptile Search Algorithm (RSA) to effectively choose the most important intrinsic mode functions (IMFs) produced by the MVMD. Then, the computed features are subsequently utilized to train and test various machine learning (ML) models. Among these models, the support vector machine (SVM) demonstrates superior performance across two benchmark AD-EEG datasets. Compared to current existing techniques, the introduced framework offers notable improvements in classification capability and computational efficiency, making it an effective and scalable tool for AD detection.