Eye movement and EEG features in generalized anxiety disorder.

Background

Generalized anxiety disorder (GAD) is a common psychiatric condition that severely impacts individuals' mental health and quality of life. However, its diagnosis and treatment remain challenging in clinical practice, largely relying on subjective assessments and lacking objective, quantitative biological markers. Therefore, this study aimed to compare eye movement and electroencephalography (EEG) features between GAD patients and healthy controls, and to evaluate the classification performance of eye movement, EEG, and their combined features using machine learning.

Methods

Twenty-nine GAD patients and 27 healthy controls were recruited. Clinical symptoms, cognitive function, eye movement, and EEG data were collected. Descriptive statistics and partial correlation analyses were conducted to compare group differences and explore feature associations. A linear support vector machine (SVM) model was constructed using eye movement, EEG, and combined features for classification.

Results

GAD patients exhibited eye movement abnormalities, including reduced fixation duration, increased saccade velocity, enlarged vertical saccade amplitude, and prolonged antisaccade latency, as well as abnormal beta power during Go/Nogo tasks. The model combining eye movement and EEG features achieved the best overall performance, outperforming models based on eye movement or EEG features alone.

Limitations

The relatively small sample size warrants further validation in larger cohorts.

Conclusion

Multimodal neurophysiological features integrating eye movement and EEG improve classification performance for GAD and may serve as potential objective biomarkers.