Related Experiment Video
Updated: Sep 17, 2025
14:27
Identification of Disease-related Spatial Covariance Patterns using Neuroimaging Data
Published on: June 26, 2013
15.8K
INVESTIGATING CORRELATIONS BETWEEN MENTAL DISORDERS AND FUNDUS IMAGING DATA USING DEEP LEARNING: A Study From the UK
Jiadi Dong1,2,3,4, Lisheng Wang1, Zhi Zheng2,3,4
1Department of Automation, Shanghai Jiao Tong University, Shanghai, China.
Retina (Philadelphia, Pa.)
|July 2, 2025
Summary
Deep learning models can identify mental disorders using eye scans (fundus imaging and OCT). This research reveals potential links between retinal biomarkers and mental health conditions, paving the way for early detection.
Area of Science:
- Ophthalmology
- Neuroscience
- Psychiatry
Background:
- Mental and behavioral disorders pose a significant global health challenge.
- Early detection and intervention are crucial for managing mental health conditions.
- Fundus imaging offers a non-invasive method to assess ocular health.
Purpose of the Study:
- To develop an automated method for identifying mental and behavioral disorders using fundus imaging data.
- To explore potential associations between mental diseases and specific fundus biomarkers.
Main Methods:
- A deep learning-based multimodality training approach was employed.
- The model was trained and evaluated on fundus images and optical coherence tomography (OCT) data from 1,494 UK Biobank participants.
- A five-fold cross-validation strategy was used for hyperparameter selection and model optimization.
Main Results:
- The multimodality model achieved an Area Under the ROC Curve (AUC) of 0.8490, with 0.7702 sensitivity and 0.8552 specificity.
- Separate classifiers (Random Forest, Linear Classifier) using OCT measures yielded AUCs of 0.8121 and 0.8094, respectively.
- A negative correlation was found between retinal nerve fiber layer/ganglion cell-inner plexiform layer thickness and mental disorders.
Conclusions:
- The study demonstrates the potential of using fundus imaging for identifying mental disorders.
- This non-invasive approach shows promise for early detection and intervention strategies in mental healthcare.
