Alzheimer's disease detection using orthogonal matching pursuit and k-nearest neighbors classifier on EEG signals

Abstract

The possible use of electroencephalography (EEG) signal as biomarker for detecting Alzheimer's disease (AD) and determining its severity is already known. EEG signals from a Florida State University study dataset were decomposed using orthogonal matching pursuit algorithm and subsequently applied k-nearest neighbors classifier. Obtained mean accuracy, sensitivity, specificity, and area under the receiver operating characteristic curve indicate that EEG signal processing can be used for screening and detecting AD in humans.