Advances in EEG-based detection of Major Depressive Disorder using shallow and deep learning techniques: A systematic...

The contemporary diagnosis of Major Depressive Disorder (MDD) primarily relies on subjective assessments and self-reported measures, often resulting in inconsistent and imprecise evaluations. To address this issue and facilitate early intervention, there is a growing interest in utilizing objective criteria such as Electroencephalography (EEG) features analyzed through Artificial Intelligence (AI) techniques. This systematic review explores the advances in EEG-based detection of MDD using both shallow and deep learning methods, with the aim of enhancing understanding of the neural mechanisms underlying the disorder and identifying potential biomarkers for its diagnosis. Following PRISMA guidelines, a comprehensive search of the Scopus, IEEE Xplore, and ScienceDirect databases was conducted. The initial search yielded 5,603 articles; after rigorous screening and application of inclusion and exclusion criteria, 22 studies were deemed most relevant for this review. Key EEG markers, including frequency band power, EEG asymmetry, event-related potential (ERP) components, and functional and non-linear connectivity metrics, were examined. The findings affirm the utility of these measures in differentiating individuals with MDD from healthy controls. Despite these promising results, the review warrants the need for further research to enhance the interpretability of EEG metrics in the context of MDD. Future research directions are outlined to support the continued development of this emerging field.