New Delhi: People living with persistent symptoms after a concussion may soon have a way to anticipate when their headaches are likely to worsen. A new study suggests that an AI model could predict moderate-to-severe post-traumatic headaches up to 24 hours before they occur, potentially helping patients plan their activities and manage their symptoms.
Published in the Journal of Head Trauma Rehabilitation, the study, titled ‘Foundation Model Forecasting of Headache Days in People With Persisting Postconcussion Symptoms’, explored whether AI could identify patterns in patients’ daily symptoms and physiological measurements to forecast headache episodes.
Concussions, also known as mild traumatic brain injuries, can cause headaches, fatigue, dizziness, sensitivity to light, and difficulty concentrating. While most people recover within days or weeks, some continue experiencing symptoms for months or longer.
How did the AI model work?
Researchers recruited 20 participants with persistent post-concussion symptoms and asked them to track their symptoms daily for 28 days using health applications. They also collected physiological measurements, including heart rate, skin temperature, and tension in the trapezius muscle, which connects the neck and shoulders.
The researchers used two machine learning models to analyse the data and predict whether participants would experience moderate-to-severe headaches the following day. The foundation model achieved an area under the curve (AUC) score of 0.69, slightly outperforming a conventional model, which scored 0.68. It also showed promising results when predicting fatigue, achieving an AUC of 0.81.
Also Read: Can blood sugar be controlled without drugs or insulin? A new wireless implant offers hope
What could this mean for patients?
The findings suggest that symptoms following a concussion may follow identifiable patterns rather than occurring randomly.
“There are strong indications that the symptoms experienced the following day do not occur randomly in people struggling with persistent post-concussion issues,” said Antonios Danelakis, a researcher at the Norwegian University of Science and Technology (NTNU) and the Norwegian Centre for Headache Research, speaking to Newswise.
“The symptoms can follow specific patterns. AI can learn these patterns by training on physiological data collected over time,” he added.
A future application could send patients notifications through a smartwatch or mobile phone, warning them when a headache or period of fatigue is likely. This could help them plan their work, social activities and daily routines.
Danelakis said that more than 17,000 people in Norway visit emergency clinics or hospital emergency departments annually with minor head injuries. The actual figure is likely at least twice as high when consultations with general practitioners and people who do not immediately seek medical attention are included.
However, the researchers describe their work as a methodological proof of concept, not a clinically validated forecasting system. With only 20 participants, the findings require testing in larger populations.
The research team plans to launch another study in spring 2027, aiming to recruit at least 200 people with persistent post-concussion symptoms.
(Edited by Maryam Hassan)