Abstracts

Rationale: Noninvasive and minimally invasive devices may be able to improve patient care by detecting and forecasting seizures. The Epilepsy Foundation sponsored My Seizure Gauge project is a three-year project aimed at advancing these capabilities.

Methods: Patients have been recruited for ultra long term (more than six months) of monitoring with a wearable device (Empatica E4, Fitbit Charge HR, or Fitbit Inspire) and concurrent ambulatory EEG monitoring. Epilepsy patients at King’s College London were recruited to undergo implantation with the UNEEG SubQ subscalp EEG system. Patients at Mayo Clinic enrolled in the ongoing RC+S trial were recruited, as were patients with implanted NeuroPace RNS systems. Patients at the University of Melbourne enrolled in the EpiMinder subscalp device trial were recruited. Wearable data from enrolled patients was recorded for 8 months or more and uploaded to shared space on a cloud-based platform for analysis. Self-reported electronic seizure diaries and periodic mood and symptom surveys were kept by participants as well. Recorded data were analyzed to assess the ability to detect seizures, to identify circadian and multi-day cycles, and to forecast seizures. Selected wearable data are freely available on EpilepsyEcosystem.org to foster exploration and academic collaboration. A data science contest to predict seizures from the Empatica E4 data will be held in late 2021.

Results: To date thirty-nine patients and one volunteer have recorded a combined 6950 days (19 years) of data in the field, including over 1020 seizures. Ten patients (including one reimplanted subject) were implanted with the UNEEG subscalp device and are continuing to wear Fitbit Charge 3 devices. Ten patients have been implanted with the Minder subscalp EEG system and are using the Fitbit Inspire HR device. Three patients have been implanted with the RC+S device and have used the Fitbit Charge HR, while fourteen patients with the NeuroPace RNS device have worn the Empatica E4 device. Fifteen patients have completed over 180 days (6 months) of monitoring, and twenty-four patients continue collecting data currently. Analysis of this dataset has begun, and collectively our teams have demonstrated the ability to forecast seizures successfully with data from the Empatica E4, the UNEEG subscalp device, the Minder device, and the Fitbit inspire HR. In addition we have demonstrated a relationship between cycles of seizure risk and cyclical patterns in heart rate, actigraphy, temperature, and other signals.

Conclusions: Seizure forecasting with wearable or subcutaneous EEG devices is possible, and provide reliable ultra-long-term monitoring data.

Funding: Please list any funding that was received in support of this abstract.: Epilepsy Foundation of America, Epilepsy Innovation Institute, "My Seizure Gauge."