Eye Movement Patterns Can Distinguish Schizophrenia From the Major Affective Disorders and Healthy Control Subjects.

Area of Science:

• Neuroscience
• Psychiatry
• Computational Biology

Background:

• Objective diagnostic tests for major psychiatric disorders are lacking.
• Eye movement behavior is explored as a potential objective biomarker.

Purpose of the Study:

• To evaluate the potential of eye movement patterns to differentiate schizophrenia from major affective disorders and healthy controls.
• To develop a predictive model using machine learning for psychiatric disorder diagnosis.

Main Methods:

• Recorded eye movements from training and validation cohorts including schizophrenia, bipolar disorder, major depressive disorder, and control groups.
• Extracted 98 eye movement features and employed a gradient boosted decision tree classifier.
• Calculated the area under the curve (AUC) to assess model performance.

Main Results:

• The predictive model achieved AUC estimates of 0.85 for schizophrenia and 0.85 for controls in one-versus-all comparisons.
• External validation showed an AUC of 0.89 for schizophrenia and 0.65 for controls.
• Free viewing, fixation duration, and smooth pursuit tasks were identified as key discriminators.

Conclusions:

• Eye movement patterns demonstrate significant potential for discriminating schizophrenia from major mood disorders and controls.
• The findings suggest eye movement analysis could offer a novel, objective approach to psychiatric diagnosis.
• The model's robustness was confirmed against potential confounding factors.