Editors' ChoiceEpilepsy

Advance warning

See allHide authors and affiliations

Science Translational Medicine  27 Jan 2016:
Vol. 8, Issue 323, pp. 323ec12
DOI: 10.1126/scitranslmed.aaf0864

The invasion of wearable electronics (wearables) into consumer marketplaces carries with it an alluring appeal of continuous physiological monitoring, from how well you sleep to how many steps you take per day. Right now these data show symptoms, but the real power would be in prediction, especially for individuals suffering from epilepsy: If they could predict a seizure, they could better control it. Fujiwara et al. have developed a mathematical tool that predicts near-term onset of epileptic seizure using data routinely collected from wearable sensors.

The algorithm continually monitors variations in the RR interval—the time interval between the largest peaks of a heart wave—and determines heart rate variability (HRV) features. To develop their prediction algorithm, the authors borrowed a mathematical framework from manufacturing and process control, which is employed to understand and model process variations, such as in machining a part. This framework, called multivariate statistical process control theory, functions by examining correlations among multiple features in HRV data to create modeling data set for comparison with acquired HRV data. When HRV data do not follow expected trends for more than 10 seconds, the data are not merely considered an anomaly; rather, those data are considered to be a signal that a seizure may occur soon. The algorithm was tested using clinical data, video EEG, and ECG collected over a time period before a seizure (preictal), during a seizure, and between seizures (interictal) in 14 subjects with refractory epilepsy. By applying the prediction algorithm, seizures in ten out of eleven awakening preictal episodes could be predicted prior to the seizure onset.

This study opens doors to preventative medicine applications of wearables, demonstrating successfully that measurable signals from devices like your Apple Watch or OURA ring can provide insight into heart and brain activities and ultimately provide health and safety recommendations in real time.

K. Fujiwara et al., Epileptic seizure prediction based on multivariate statistical process control of heart rate variability features. IEEE Trans. Biomed. Eng. 10.1109/TBME.2015.2512276 (2015). [Full Text]

Stay Connected to Science Translational Medicine

Navigate This Article