Prospective validation of a seizure diary forecasting falls short

Area of Science:

• Neurology
• Artificial Intelligence
• Machine Learning

Background:

• Deep learning AI models have shown promise in forecasting seizure risk using retrospective data.
• Previous studies indicated higher accuracy for AI compared to random forecasts.

Conclusions:

• The AI seizure forecasting tool did not demonstrate superior performance compared to a simple moving average in this prospective cohort.
• The findings suggest that the current AI model may not be suitable for prospective seizure risk forecasting and should be reconsidered or replaced.
• Further research is needed to refine AI models for accurate and reliable seizure prediction in epilepsy.