Related Experiment Video
Updated: Jan 18, 2026
07:08
Using Optical Coherence Tomography and Optokinetic Response As Structural and Functional Visual System Readouts in Mice and Rats
Published on: January 10, 2019
10.5K
Self-AttentionNeXt: Exploring schizophrenic optical coherence tomography image detection investigations.
Mehmet Kaan Kaya1, Sermal Arslan1, Suheda Kaya2
1Universal Eye Clinic, Elazig 23100, Türkiye.
World Journal of Psychiatry
|September 11, 2025
Summary
A novel AI model, Self-AttentionNeXt, accurately detects schizophrenia (SZ) using retinal images from optical coherence tomography (OCT). This technology aids in early diagnosis by identifying subtle retinal biomarkers linked to central nervous system changes.
Area of Science:
- Ophthalmology and Neuroscience
- Artificial Intelligence in Healthcare
- Biomedical Imaging Analysis
Background:
- Optical coherence tomography (OCT) provides high-resolution, non-invasive retinal imaging.
- Retinal layer alterations may indicate central nervous system changes in psychiatric disorders like schizophrenia (SZ).
Purpose of the Study:
- To develop an advanced deep learning model for classifying OCT images.
- To distinguish schizophrenia patients from healthy controls using retinal biomarkers.
Main Methods:
- A novel convolutional neural network, Self-AttentionNeXt, was designed with grouped self-attention and residual blocks.
- The model was trained and validated on custom SZ OCT data and the public OCT2017 dataset.
Main Results:
- Self-AttentionNeXt achieved 97.0% accuracy on the custom SZ dataset and over 95% on the OCT2017 dataset.
- Gradient-weighted class activation mapping confirmed attention to clinically relevant retinal regions.
Conclusions:
- Self-AttentionNeXt effectively integrates attention mechanisms and CNNs for early schizophrenia detection via OCT.
- This AI-driven approach shows promise for psychiatric diagnostics and clinical decision support.
