Support vector machine classification of patients with depression based on resting-state electroencephalography.

Area of Science:

• Neuroscience
• Computational Psychiatry
• Biomedical Engineering

Background:

• Depression diagnosis lacks objective standards, relying on symptom identification and history.
• Objective diagnostic tools are crucial for accurate and consistent depression assessment.
• Current diagnostic methods for major depressive disorder (MDD) can be subjective.

Purpose of the Study:

• To investigate resting-state electroencephalogram (EEG) differences between MDD patients and healthy controls (HCs).
• To develop an objective method for distinguishing MDD patients from HCs using EEG data and machine learning.
• To evaluate the efficacy of support vector machine (SVM) classification with different feature selection methods.

Main Methods:

• Utilized EEG data from 21 MDD patients and 21 HCs.
• Extracted various EEG features, including relative frequency power, asymmetry, coherence, and entropy measures.
• Employed t-test and receiver operating characteristic (ROC) analysis for feature selection, feeding into an SVM classifier.

Main Results:

• The SVM classifier achieved high accuracy (up to 96.66% in eye-open condition) in distinguishing MDD patients from HCs.
• Key features for classification included relative frequency power and left-right coherence.
• The classifier demonstrated over 90% mean accuracy, indicating strong discriminative power.

Conclusions:

• Machine learning analysis of EEG data offers a promising objective approach for depression diagnosis.
• The SVM classifier, particularly with t-test feature selection, shows high accuracy in differentiating MDD from HCs.
• Further research is warranted to refine this objective, efficient, and simple diagnostic method for clinical use.