Abstract
Major depressive disorder (MDD) is a common and highly debilitating condition that threatens the health of millions of people. However, current diagnosis of depression relies on questionnaires that are highly correlated with physician experience and hence not completely objective. Electroencephalography (EEG) signals combined with deep learning techniques may be an objective approach to effective diagnosis of MDD. This study proposes an end-to-end deep learning framework for MDD diagnosis based on EEG signals. We used EEG signals from 29 healthy subjects and 24 patients with severe depression to calculate Accuracy, Precision, Recall, F1-Score, and Kappa coefficient, which were 90.98%, 91.27%, 90.59%, and 81.68%, respectively. In addition, we found that these values were highest when happy-neutral face pairs were used as stimuli for detecting depression. Compared with exiting methods for EEG-based MDD classification, ours can maintain stable model performance without re-calibration. The present results suggest that the method is highly accurate for diagnosis of MDD and can be used to develop an automatic plug-and-play EEG-based system for diagnosing depression.
Highlights
To achieve the above requirements, we present a novel end-toend architecture, supervised EEG-based event-related potential (ERP) classification
Experiment 1 (All): Subjects were stimulated by all three types of face pairs (480 trials for each subject); Experiment 2 (Fcue): Subjects were only stimulated by fearneutral face pairs (160 trials for each subject); Experiment 3 (Scue): Subjects were only stimulated by sadneutral face pairs (160 trials for each subject); Experiment 4 (Hcue): Subjects were only stimulated by happyneutral face pairs (160 trials for each subject)
With the preprocessed signal used as input, the highest average classification accuracy (90.98%) obtained by leave-one-subject-out cross-validation (LOSOCV) was for Experiment 4
Summary
Depression is one of the most prevalent mental disorders. Patients with depression experience a severely impaired quality of life and are at increased risk of suicide [1–3]. Variations in voltage resulting from changes in ionic current within the brain’s neurons contribute to EEG signals and might help to diagnose mental disorders like depression. EEG signals are frequently affected by different types of noise due to eye blinking and body motion [20]. It is needed a deep learning technique that can effectively learn brain activity patterns from EEG signals. The EEG database used here is small and does not require complex EEG pre-processing This method successfully extracts information across different subjects for ERP decoding, and accomplishes three tasks simultaneously. The remainder of this article is structured as follows: We firstly provide background and introduce the database, we describe the structure of the proposed method, experimental results are presented and discussed
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
