Wearable seizure detection devices demonstrated high sensitivity with some achieving up to 100% sensitivity and significantly reduced false alarm rates, highlighting their potential for real-world application in epilepsy management.
Do wearable devices and associated AI algorithms accurately detect seizures compared to standard EEG in people with epilepsy?
23 studies evaluating people with epilepsy
Wearable seizure detection devices (e.g., wrist- and ear-based systems) utilizing artificial intelligence (AI) algorithms (e.g., machine learning, neural networks, support vector machines)
Standard EEG modalities (e.g., video EEG, scalp EEG, intracranial EEG) evaluated by human experts
Seizure detection performance (sensitivity and false alarm rate/false positives per 24 hours)
Wearable devices combined with AI algorithms show high sensitivity for automated seizure detection, though reducing false alarm rates remains a key challenge for everyday clinical integration.
Epilepsy affects millions of people worldwide, driving the need for advanced methods to monitor patients’ health and seizure activity. Recent advances in wearable technologies have enabled continuous collection of physiological data to support real-time seizure detection in the real-world. This review presents a targeted synthesis of 23 studies evaluating wearable devices and their associated artificial intelligence (AI) algorithms for automated seizure detection. Both wrist- and ear-based systems demonstrate high sensitivity, with performance influenced by device design, signal reliability, and analytic approach. The main challenges include reducing false alarms and maintaining data integrity during everyday use. More recent studies highlight the ability to anticipate seizures before they occur, marking a promising step toward improving safety and well-being for people living with epilepsy. Ongoing efforts to identify reliable physiological markers and to evaluate device performance across diverse populations are key to integrating wearable technologies for seizure detection into routine medical care.
Building similarity graph...
Analyzing shared references across papers
Loading...
T. L. Ho
Bridget E.L. Ostrem
James Hillis
Frontiers in Neurology
SHILAP Revista de lepidopterología
Harvard University
University of California, Berkeley
University of California, San Francisco
Building similarity graph...
Analyzing shared references across papers
Loading...
Ho et al. (Mon,) conducted a review in Epilepsy (n=23). Wearable seizure detection devices vs. No wearable device was evaluated on Sensitivity of seizure detection. Wearable seizure detection devices demonstrated high sensitivity with some achieving up to 100% sensitivity and significantly reduced false alarm rates, highlighting their potential for real-world application in epilepsy management.
www.synapsesocial.com/papers/69b3aaa802a1e69014ccb656 — DOI: https://doi.org/10.3389/fneur.2026.1756895