Identifying depression with mixed features: the potential value of eye-tracking features

Area of Science:

• Neuroscience
• Psychiatry
• Computational Psychiatry

Background:

• Depression with mixed features (DMF) presents diagnostic challenges due to co-occurring depressive and subsyndromal manic symptoms.
• Identifying reliable neurobiological markers for DMF is crucial for accurate diagnosis and treatment.

Purpose of the Study:

• To evaluate the efficacy of eye-tracking features as neurobiological markers for identifying DMF.
• To assess the improvement in diagnostic accuracy by integrating ocular movement data into a machine learning model.

Main Methods:

• Collected eye-tracking data from 93 participants (41 with major depressive disorder (MDD), including 20 with DMF, and 52 healthy controls).
• Utilized an infrared eye-tracker and clinical scales (MADRS, YMRS, BPRS).
• Compared an extreme gradient boosting (XGBoost) model using demographic/clinical data with one incorporating eye-tracking features.

Main Results:

• Significant differences in eye-tracking features (orienting and overlapping saccades) were found between DMF, MDD, and healthy controls.
• Integrating eye-tracking data improved the XGBoost model's predictive accuracy for DMF, increasing the area under the curve (AUC) from 0.571 to 0.679 (p < 0.05).
• Velocity of overlapping saccades and free viewing task completion time were key predictive factors.

Conclusions:

• Eye-tracking features, specifically saccade velocity and task completion time, show promise as non-invasive biomarkers for DMF identification.
• Machine learning models incorporating these ocular parameters significantly enhance DMF diagnostic accuracy.
• This approach offers a valuable tool for improving clinical decision-making in psychiatric care.