Christian Meisel, MD, PhD: Forecasting Seizures Using Multi-Modal Wristband Sensors

"Some way to tell these patients when the seizure risk is high or low would help them plan their days better and avoid certain activities. For clinicians, having a better objective measure of seizure risk would allow them to target therapies better.”

Forecasting seizures has been a goal for epilepsy specialists for many years—if perfected, may help design and coordinate more personalized treatment strategies. Christian Meisel, MD, PhD, and colleagues recently conducted a study that found multi-modal wristband sensor data from easy-to-use, non-invasive devices in combination with deep learning to provide statistically significant and clinically useful seizure forecasting.

The study, presented virtually at the American Epilepsy Society (AES) Annual Meeting, December 4–8, 2020, applied deep learning networks including long short-term memory and 1DConv to these wristband sensor data from 69 persons with epilepsy and found a mean prediction horizon of 1896 (±101) seconds. The period was deemed long enough to afford reasonable warning of seizures in advance.

Meisel, neurologist, Universitätsmedizin Berlin, and Berlin Institute of Health, sat down with NeurologyLive to discuss the findings from his study and the increased research that has gone into using devices to forecast seizures.

REFERENCE

Meisel C, El Atrache R, Jackson M, et al. Machine learning from wristband sensor data for wearable, non-invasive seizure forecasting. Presented at AES 2020 Annual Meeting; December 4–8, 2020; Abstract 241