Chapter 8 - Electroencephalography-Based Diagnosis of Depression

Abstract

This chapter provides a machine learning scheme with the aim of classifying a potential subject as either a depressed patient or a healthy control. The description involves efficient identification of the most relevant and discriminant features that could be potential candidates for the efficient classification of the study participants. More specifically, the chapter explains the feature selection and classification scheme, termed as the Intelligent Treatment Management System (ITMS) for unipolar major depressive disorder (MDD) patients. The ITMS intends to perform electroencephalography (EEG)-based diagnoses of depression (ITMS diagnosis) by classifying MDD patients and healthy controls recruited for this research study. This EEG-based scheme inherently involves subprocesses such as noise removal from the EEG data, EEG-based feature extraction, feature selection, classification, and validation, including a 10-fold cross-validation (10-CV). This chapter provides the technical details regarding these subprocesses.