Published on

An artificial intelligence (AI) model accurately detected infantile epileptic spasm syndrome through the assessment of smartphone videos captured by families, according to a study published in a medRxiv preprint that has not been peer reviewed. Among the results, the AI model had 82% sensitivity and 90% specificity. The infantile epileptic spasm syndrome diagnosis is considered a neurologic emergency, according to MedPage Today, and experts believe the AI platform has potential as a screening tool. Infants may experience symptoms at home but not in the provider office, so video detection with AI support may be able to help providers recognize the diagnosis sooner. The AI model was trained from video segments including 141 children experiencing 991 seizures plus 597 non-seizure videos as well as 1,385 videos from 127 children without seizures. 

Better yet: Videos are easy to share among consulting specialty providers, so a clinician with a hunch and an AI suggestion may be able to screen infants remotely for referrals. According to the researchers, infantile epileptic spasm syndrome affects approximately 1 in 2,000-2,500 infants in their first year of life, and delayed diagnosis can lead to long-term disability. As with any AI tool, further validation is needed before it could be used in clinical practice.

AI Tool Correctly Detects Epileptic Syndrome in Infants By Watching Videos