Abstract Details

Title Seizure forecasting and detection with wearable devices and subcutaneous EEG - A Practical Seizure Gauge

Topic Epilepsy/Clinical Neurophysiology (EEG)

Presentation(s) N2 - Neuroscience in the Clinic: Applications of Artificial Intelligence and Machine Learning Tools in Neurology (4:50 PM-5:00 PM)

Poster/Presentation
Number 004

Background Seizure forecasting has been established using continuous intracranial EEG, however invasive devices are not appropriate for all patients. Minimally invasive wearable and subcutaneous devices may facilitate seizure forecasting and may provide accurate seizure records to support clinical decision making.

Objective To develop the ability to forecast seizures in patients with epilepsy without invasive intracranial devices.

Design/Methods Patients were recruited for ultra long term monitoring with a wearable device (Empatica E4, Fitbit Charge HR, or Fitbit Inspire) and concurrent ambulatory EEG monitoring (UNEEG SubQ, EpiMinder Subscalp, NeuroPace RNS) at three sites. Wearable and EEG data was recorded for 8 months or more. Electronic seizure diaries and periodic mood and symptom surveys were recorded by participants. Recorded data were analyzed to assess the ability to detect seizures, identify circadian and multi-day cycles, and forecast seizures.

Results

Thirty-nine patients with epilepsy recorded over 9300 days (25.5 years) of ambulatory wearable and EEG data, including over 1550 seizures. Nine patients left the study before completion due to device malfunctions, complications, poor adherence, poor data quality, or unanticipated seizure freedom. Analysis in this cohort has established the following:

· Heart rate circadian and multi-day cycles measured using Fitbit correlated with self-reported seizures in 10 of 19 patients

· Actigraphy, HR, and tonic EDA circadian and multi-day cycles measured by Empatica E4 correlated with iEEG confirmed seizures in 7, 6 and 5 of 8 patients studied, respectively.

· Seizure forecasting significantly better than chance in 5 of 6 patients using the wrist-worn Empatica E4 device using a long-short term memory (LSTM) neural network, with iEEG confirmation of events.

· Seizure forecasting using UNEEG subscalp EEG significantly better than chance in 5 of 6 patients using a Bidirectional LSTM neural network.

Conclusions This project has established the feasibility of forecasting seizures using long-term cycles, wearable devices, and subcutaneous EEG.

Authors/Disclosures
Benjamin H. Brinkmann, PhD (Mayo Clinic) PRESENTER	Dr. Brinkmann has received personal compensation in the range of $0-$499 for serving as a Consultant for Otsuka Pharmaceuticals. Dr. Brinkmann has stock in Cadence Neuroscience. The institution of Dr. Brinkmann has received research support from Epilepsy Foundation of America. The institution of Dr. Brinkmann has received research support from National Institutes of Health. The institution of Dr. Brinkmann has received research support from National Institutes of Health. The institution of Dr. Brinkmann has received research support from UNEEG A/S. The institution of Dr. Brinkmann has received research support from Seer Medical Pty. Dr. Brinkmann has received intellectual property interests from a discovery or technology relating to health care. Dr. Brinkmann has a non-compensated relationship as a Co-Investigator with Medtronic Inc. that is relevant to AAN interests or activities.
	No disclosure on file
	No disclosure on file
	No disclosure on file
	No disclosure on file
	No disclosure on file
Nicholas M. Gregg, MD (Mayo Clinic)	No disclosure on file
	No disclosure on file
Caitlin Grzeskowiak	No disclosure on file
	No disclosure on file
	No disclosure on file
Gregory A. Worrell, MD (Mayo Clinic College of Medicine)	Dr. Worrell has received stock or an ownership interest from NeuroOne Inc.. Dr. Worrell has received stock or an ownership interest from Cadence Neuroscience Inc. The institution of Dr. Worrell has received research support from NIH. The institution of Dr. Worrell has received research support from Medtronic Inc.. The institution of Dr. Worrell has received research support from Neuropace Inc,. The institution of Dr. Worrell has received research support from Epilepsy Foundation of America. Dr. Worrell has received intellectual property interests from a discovery or technology relating to health care. Dr. Worrell has received intellectual property interests from a discovery or technology relating to health care.
Andreas Schulze-Bonhage (Universitaetsklinikum Freiburg)	Andreas Schulze-Bonhage has received personal compensation in the range of $500-$4,999 for serving as a Consultant for BIAL, GW. Andreas Schulze-Bonhage has received personal compensation in the range of $5,000-$9,999 for serving on a Speakers Bureau for BIAL, GW, EISAI, UCB.
Dean Freestone	No disclosure on file
Mark Richardson, MD, MRCP (Institute of Psychiatry)	The institution of Dr. Richardson has received research support from Medical Research Council UK. The institution of Dr. Richardson has received research support from Epilepsy Research UK. The institution of Dr. Richardson has received research support from GW Pharma. The institution of Dr. Richardson has received research support from European Commission. The institution of Dr. Richardson has received research support from Autifony Therapeutics . The institution of Dr. Richardson has received research support from Epilepsy Foundation of America . Dr. Richardson has received intellectual property interests from a discovery or technology relating to health care.