Related Experiment Video
Updated: Dec 3, 2025
10:31
A Multi-Modal Approach to Assessing Recovery in Youth Athletes Following Concussion
Published on: September 25, 2014
13.8K
Predicting Post-Concussion Symptom Recovery in Adolescents Using a Novel Artificial Intelligence
David E Fleck1, Nicholas Ernest2, Ruth Asch1
1Department of Psychiatry and Behavioral Neuroscience, University of Cincinnati College of Medicine, Cincinnati, Ohio, USA.
Journal of Neurotrauma
|October 29, 2020
Summary
This study used artificial intelligence (AI) with diffusion tensor imaging (DTI) to predict concussion recovery in adolescents. The AI system showed promise in forecasting symptom recovery based on objective brain imaging data.
Area of Science:
- Neuroimaging
- Artificial Intelligence
- Clinical Prediction
Background:
- Post-concussion syndrome can be challenging to predict.
- Objective biomarkers are needed for early clinical assessment.
Purpose of the Study:
- To explore the use of diffusion tensor imaging (DTI) and artificial intelligence (AI) for predicting one-week recovery from concussion symptoms.
- To evaluate a novel Genetic Fuzzy Trees (GFT) system for clinical prediction.
Main Methods:
- Forty-three adolescents (11-16 years) with mild traumatic brain injury or orthopedic injury were studied.
- Diffusion tensor imaging (DTI) scans were performed three days post-injury.
- A Genetic Fuzzy Trees (GFT) AI system was trained using DTI data and symptom scores (Post-Concussion Symptom Scale - PCSS) to predict recovery.
Main Results:
- The GFT system achieved 100% accuracy in training and 62% in validation, outperforming six other classification methods.
- Recovery prediction demonstrated 59% sensitivity and 65% specificity in the GFT validation set.
- The AI system significantly predicted recovery better than chance using objective brain measures.
Conclusions:
- Genetic Fuzzy Trees (GFT) show potential for predicting trauma symptom recovery using objective brain imaging.
- Artificial intelligence (AI) offers future promise for acute care clinical predictions.
- Further optimization is needed for widespread clinical and research applications.
