AI video analysis highly accurate in detecting infantile epilepsy spasms
Key takeaways:
- The model correctly identified 83% of movements that were infantile spasms.
- Researchers developed a digital tool for parents to upload videos of their infants’ movements to a secure platform for AI analysis.
LOS ANGELES — An AI model accurately detected epileptic spasms in infants through smartphone videos, representing a “highly accessible approach” to accelerating diagnosis, according to researchers.
The findings were presented at the American Epilepsy Society annual meeting.
“Early identification of infantile spasms is crucial because it enables timely treatment, ideally within 1 week of the onset of the spasms, which may significantly improve outcomes for these infants,” Gadi Miron, MD, lead author and researcher at Charité-
Medical University Berlin, said in a press release related to the study. “We believe this technique could empower parents by capturing critical information about their child’s symptoms.”
Smartphone recordings have previously shown to aid in the diagnosis of epileptic spasms, but analysis by an epilepsy specialist was still required. So, Miron and colleagues sought to examine whether an AI-based video analysis is effective in detecting epileptic seizures from home recordings.
Their phase 2, retrospective study featured videos of 168 infants aged younger than 2 years with epileptic seizures and 216 infants without seizures. The researchers trained an AI model that was designed for video analysis to detect epileptic seizures, according to the release.
Miron and colleagues reported that the AI model correctly identified 83% of movements consistent with epileptic spasms and was 95% accurate at identifying movements that were not epileptic spasms, with an overall accurate detection rate of 85%.
Due to the positive data, Miron and colleagues developed a digital tool that parents can use to upload videos to a secure platform and log their children’s movements, the release said. If the AI model identifies concerns, parents are notified and receive recommendations for further evaluation. The researchers are currently conducting a study to determine whether it can be used in a health care setting.
“This technique could also aid doctors by offering additional data for patient evaluations and potentially expedite referrals for diagnostic tests, leading to quicker treatment decisions,” Christian Meisel, MD, PhD, principal study investigator and neurologist at Charité-Medical University, Berlin, said in the release.
Reference:
- AI analysis of home videos could help identify infantile epileptic spasms to speed up diagnosis and treatment. https://aesnet.org/about/aes-press-room/press-releases/ai-analysis-of-home-videos-could-help-identify-infantile-epileptic-spasms--to-speed-up-diagnosis-and-treatment. Published Dec. 7, 2024. Accessed Dec. 7, 2024.
Collapse
Miron G, et al. Automatic detection of epileptic spasms from home videos. Presented at: American Epilepsy Society annual meeting; Dec. 6-10, 2024; Los Angeles.
Read more from American Epilepsy Society
View all
Read more about