Classification Techniques in Machine Learning for Age-related Electroencephalography Data Analysis

Abstract

Electroencephalography (EEG) is used to measure event-related potentials in neuroscience. Age-related changes can alter the EEG signals as well as neurological diseases. Understanding EEG signals is beneficial to the diagnosis, prediction and prevention of neurological disorders, including neurological rehabilitation and the brain-com­puter interface. EEG data analytic application is a new frontier in neuroscience and neuroengineering. In this review article, EEG analysis during the resting state, working memory tasks and brain aging is briefly discussed. Several clas­sification techniques in machine learning are discussed and compared in terms of aging, including the support vector machine, K-nearest neighbor, decision tree, random forest, multilayer perceptron, logistic model tree and Naïve Bayes. Dealing with big data analysis using machine learning will be a mega trend in the future, including EEG data.