An End-to-End Depression Recognition Method Based on EEGNet.

Area of Science:

• Neuroscience
• Artificial Intelligence
• Medical Diagnostics

Background:

• Major Depressive Disorder (MDD) is a prevalent and severe condition impacting global health.
• Current MDD diagnosis relies on subjective, questionnaire-based methods influenced by physician experience.
• Objective diagnostic methods are needed to improve the accuracy and reliability of depression assessment.

Purpose of the Study:

• To develop and evaluate an end-to-end deep learning framework for objective diagnosis of MDD using electroencephalography (EEG) signals.
• To assess the performance of the proposed framework compared to existing EEG-based depression classification methods.
• To identify optimal stimuli for enhancing depression detection accuracy using EEG.

Main Methods:

• Utilized electroencephalography (EEG) signals from 29 healthy individuals and 24 patients diagnosed with severe depression.
• Developed a novel end-to-end deep learning framework for classifying EEG data to detect MDD.
• Evaluated model performance using metrics including Accuracy, Precision, Recall, F1-Score, and Kappa coefficient.

Main Results:

• The deep learning framework achieved high diagnostic performance: 90.98% Accuracy, 91.27% Precision, 90.59% Recall, and 81.68% F1-Score.
• Optimal performance was observed when using happy-neutral face pairs as stimuli for depression detection.
• The proposed method demonstrated stable performance without requiring recalibration, outperforming existing EEG-based classification techniques.

Conclusions:

• The developed deep learning framework offers a highly accurate and objective approach for diagnosing major depressive disorder (MDD) using EEG signals.
• This method has the potential to be developed into an automated, plug-and-play system for clinical depression diagnosis.
• The findings highlight the efficacy of combining EEG with advanced AI for improved mental health diagnostics.