Mild cognitive impairment detection with optimally selected EEG channels based on variational mode decomposition and supervised machine learning

Majid Aljalal, Marta Molinas, Saeed A. Aldosari, Khalil AlSharabi, Akram M. Abdurraqeeb, Fahd A. Alturki

Abstract: Highlights•Using VMD and entropy-based approaches is effective for detecting MCI from resting-state EEG signals.•Multi-objective optimization methods (NSGA-II and R-NSGA-II) and greedy algorithms are employed and compared for EEG channel selection.•NSGA-II and R-NSGA-II succeed in finding a few suitable channels that are able to reach and exceed full-channel accuracy.•Results demonstrate the clear superiority of optimization methods over greedy methods in finding a few suitable channels.•The selected channels are influenced by several factors, including feature extraction methods, classifiers and their parameters, channel selection approach, and so on.