Related Experiment Video
Updated: Aug 2, 2025
09:38
Determining Pain Detection and Tolerance Thresholds Using an Integrated, Multi-Modal Pain Task Battery
Published on: April 14, 2016
12.8K
Developing an artificial intelligence-based headache diagnostic model and its utility for non-specialists' diagnostic
Masahito Katsuki1, Tomokazu Shimazu2, Shoji Kikui3
1Department of Neurosurgery, Itoigawa General Hospital, Niigata, Japan.
Cephalalgia : an International Journal of Headache
|April 19, 2023
Summary
Artificial intelligence significantly improved headache diagnosis accuracy for non-specialists. This AI model offers a promising tool to enhance diagnostic performance in headache disorders.
Area of Science:
- Neurology
- Medical Informatics
- Artificial Intelligence in Medicine
Background:
- Misdiagnosis of headache disorders presents a significant clinical challenge.
- A specialized headache hospital provided a large questionnaire database for model development.
Purpose of the Study:
- To develop and validate an artificial intelligence (AI)-based model for diagnosing headache disorders.
- To assess the impact of AI on the diagnostic accuracy of non-specialists.
Main Methods:
- An AI model was developed using a retrospective dataset of 4000 patients diagnosed by headache specialists.
- The model's diagnostic performance was evaluated on a test set.
- Non-specialists' diagnoses with and without AI assistance were compared against expert diagnoses.
Main Results:
- The AI model achieved a macro-average accuracy of 76.25% on the test dataset.
- AI significantly improved the overall accuracy of non-specialist diagnoses from 46% to 83.20% (kappa from 0.212 to 0.678).
- Key diagnostic metrics including sensitivity, specificity, and precision were enhanced with AI assistance.
Conclusions:
- AI-powered tools can substantially improve diagnostic performance for headache disorders among non-specialists.
- Further research is required to address limitations, including single-center data and accuracy for secondary headaches, necessitating broader data collection and validation.
