Empatica plots wearable-based study to develop seizure prediction algorithm

After developing a wearable device that can automatically detect seizures in people with epilepsy and immediately notify their caregivers, Empatica is planning to take the technology even further.

The Embrace wrist-worn device currently uses an artificial intelligence algorithm to spot the earliest signs of a convulsive seizure—and was proven in a study ahead of its 2018 FDA clearance to catch 100% of the 135 participants’ seizures. Its Boston-based maker now wants to expand its wearables’ AI technology to be able to detect oncoming seizures before they’ve even begun.

It’ll do so with the help of real-world patient data collected in a new study, Empatica announced Thursday.

During the study, Empatica’s wearables will gather round-the-clock data from participants—spanning both measurements of their physical health and information about lifestyle habits and patterns—and feed those into an algorithm. The machine learning AI will then parse through those data, looking for markers specific to each participant that may be linked to a heightened risk of seizure.

A seizure-predicting algorithm like Empatica’s work-in-progress is “among the most-requested features for people with epilepsy,” Rosalind Picard, Ph.D., a co-founder of the company and its chief scientist, said in the announcement.

“Patients with epilepsy understand the toll that uncertainty around seizures takes and we hope that this study will help give them better control over their life, reducing stress and perhaps also enabling early interventions that ultimately reduce or prevent seizures from happening,” Picard continued.

The study is set to begin recruitment early next year, but in the meantime interested individuals can sign up online to receive further information.

Empatica’s AI-powered wearables have been used as predictive tools before.

In early 2021, the company earned a CE mark clearance for its Aura wearable, which was programmed to send out a warning at the earliest signs of a possible case of COVID-19. It did so by continuously analyzing a wearer’s heart rate and pulse variability, as well as the electrical activity of their skin, to look for changes that could be caused by an oncoming respiratory infection.

Meanwhile, Empatica’s E4 device was used in a small 2021 study with the same goal as the company’s upcoming research: to develop seizure-predicting AI. The Epilepsy Foundation-funded study recruited only six participants, but was able to conclude that a deep learning algorithm could perform significantly better than a random predictor in forecasting seizures for five of the patients, with alerts coming in about half an hour before seizures began, on average.

At the time, the study’s authors wrote that their results “provide the first clear evidence that direct seizure forecasts are possible using wearable devices in the ambulatory setting for many patients with epilepsy.”