Systematic review of machine learning and deep learning models for EEG-based detection of depression
Machine learning shows promise for depression detection from brain scans, but larger, rigorous studies are needed before clinical use becomes realistic.
This PRISMA-compliant systematic review of 42 ML/DL studies for EEG-based depression detection finds that both approaches show diagnostic potential but near-perfect reported accuracy values are unreliable, typically associated with small samples, subject-dependent validation, and methodological weaknesses revealed by QUADAS-2 assessment. The field requires larger, diverse samples with external validation and standardized reporting before clinical translation is feasible.
What the study was
- Study design
- Systematic review (PRISMA 2020; 42 eligible studies)
- Population
- Patients with depression; studies using EEG for ML/DL-based detection
- Category
- Diagnostics
- Maturity
- Exploratory
- Journal
- Journal of Psychiatric Research
Why it surfaced
Systematic review with QUADAS-2 quality assessment; identifies critical methodological limitations in a high-volume field; relevant to AI/ML in clinical diagnostics watchlist topic.
A plain-language summary of published research — not medical advice. Talk to a clinician about your own care.