Related Experiment Video
Updated: Jul 7, 2026
Measurement of Fronto-limbic Activity Using an Emotional Oddball Task in Children with Familial High Risk for Schizophrenia
Published on: December 2, 2015
Facial expression analysis using convolutional neural network for drug-naive and chronic schizophrenia
Tongxin Li1, Xiaofei Zhang2, Conghui Wang3
1Institute of Mental Health, Tianjin Anding Hospital, Mental Health Center of Tianjin Medical University, Tianjin, China; Brain Assessment & Intervention Laboratory, Tianjin Anding Hospital, Mental Health Center of Tianjin Medical University, Tianjin, China.
Facial image analysis using convolutional neural networks (CNNs) can accurately detect schizophrenia patients. Specific facial expressions in drug-naive schizophrenia patients correlate with psychiatric symptoms, suggesting potential diagnostic indicators.
Area of Science:
- Computational psychiatry
- Machine learning in mental health
- Facial expression analysis
Background:
- Facial images can reflect mental health conditions.
- Schizophrenia diagnosis often relies on clinical interviews and symptom assessment.
- Objective markers for schizophrenia detection are needed.
Purpose of the Study:
- To differentiate drug-naive schizophrenia (DN-SCZ), chronic schizophrenia (C-SCZ), and healthy controls (HCs) using facial images.
- To investigate differences in facial expressions among these three groups.
- To explore the relationship between facial expressions and psychiatric symptoms in schizophrenia patients.
Main Methods:
- Recruited 45 DN-SCZ, 106 C-SCZ, and 101 HCs.
- Used convolutional neural networks (CNNs) for group classification (90.99% accuracy) and facial expression prediction (up to 95.95% accuracy).
- Assessed psychiatric symptoms using the Positive and Negative Syndrome Scale.
Main Results:
- CNNs accurately classified participants and predicted basic facial expressions.
- Facial images of HCs were rated higher in valence, arousal, and attractiveness compared to DN/C-SCZ patients.
- Specific facial expressions (e.g., sad, surprised, happy, angry) differed significantly between groups and correlated with symptom scores in DN-SCZ patients.
Conclusions:
- Facial expressions show objective differences among schizophrenia patients and healthy controls.
- Facial expression analysis holds potential as an indicative sign for detecting DN-SCZ and C-SCZ.
- Certain facial expressions in DN-SCZ patients are associated with their psychiatric symptoms.
