EEG Signals as a Biomarker for Depression

Proposal to extract novel and robust features from low cost and portable electroencephalogram (EEG) signals for depression detection, using different signal processing techniques.

Project Motivation

The project relies on the use of EEG signals primarily, as when compared to other physiological signals, for their good time resolution, low maintenance cost, portability, and simple operating method, making them more suitable to be used in a simple clinical setup. The EEG signals have been used to study the symptoms of depression including insomnia and sleep disorders and for diagnosis of epilepsy which historically shown their effectiveness in directly monitoring regions of the brain for changes in amplitude of power at various frequencies.

The aim for this project is to extract novel features from the EEG signals taken from Database for Emotional Analysis using Physiological Signals (DEAP) dataset, having a direct impact on valence and arousal metric as given in to aid in the building of a pre-evaluation screening methodology for clinical use. As noticed across a variety of studies, we also find a strong correlation between a depression addled brain and its associated EEG spectral power for specific frequency sub-bands.