Related Experiment Video
Updated: Sep 8, 2025
07:48
Eye Tracking During A Complex Aviation Task For Insights Into Information Processing
Published on: April 4, 2025
563
Artificial Intelligence for the Analysis of Workload-Related Changes in Radiologists' Gaze Patterns
IEEE Journal of Biomedical and Health Informatics
|June 15, 2022
Summary
Artificial intelligence (AI) can detect fatigue in radiologists by analyzing gaze patterns. This study found AI metrics correlate with decreased performance, offering a novel way to predict reading pattern changes due to radiologist fatigue.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Human Factors in Medicine
Background:
- Radiological errors, particularly overlooked abnormalities, increase towards the end of work shifts, impacting patient safety.
- Understanding radiologist fatigue is crucial for mitigating diagnostic errors and improving healthcare outcomes.
- Existing methods for assessing radiologist fatigue are limited in real-time applicability.
Purpose of the Study:
- To investigate the potential of artificial intelligence (AI) in identifying radiologist fatigue through gaze pattern analysis.
- To correlate eye-tracking metrics with cumulative work done (CWD) to quantify performance decline.
- To develop a novel AI-based metric for predicting fatigue-related changes in image reading patterns.
Main Methods:
- Retrospective analysis of lung X-ray images from 400 subjects read by four radiologists.
- Eye movements of radiologists were recorded during image interpretation, alongside concentration tests.
- A U-Net neural network was utilized to annotate lung anatomy and extract gaze-based features (coverage, information gain) correlated with CWD.
Main Results:
- Gaze-traveled distance, X-ray coverage, and lung coverage significantly deteriorated with increasing CWD for most radiologists (p < 0.01).
- Reading time and information gain over lung fields showed statistically significant deterioration for all four radiologists.
- A novel AI metric integrating reading time, speed, and organ coverage was developed, demonstrating predictive capability for fatigue-related pattern changes.
Conclusions:
- AI-driven analysis of radiologist gaze patterns can effectively detect performance decrements associated with fatigue.
- The developed AI metric offers a promising tool for real-time monitoring of radiologist fatigue and potential error risk.
- This approach has the potential to enhance patient safety by proactively managing radiologist workload and fatigue.
