Wearable Data in Epilepsy Care

Abstract

Seizure monitoring by combined diary and wearable data: A multicenter, longitudinal, observational study.

Vieluf S, Tomioka S, Zhang B, Krishnan V, Bosl WJ, Grinnell T, Loddenkemper T. Epilepsia. 2025 Nov;66(11):4259–4271. doi:10.1111/epi.18550. Epub 2025 Jul 12. PMID: 40650883.

Objective: In patients with intractable epilepsy, accurate diaries of seizure occurrence and timing can substantially inform management. Wearable devices that provide confirmation of seizure occurrence complement such diaries, which are frequently incomplete and/or inaccurate. Here, we combined seizure diaries and longitudinally deployed wrist-worn device recordings to evaluate whether wearable recordings contain information that can discriminate between days containing seizure-related activity and those without.

Methods: Patients with focal seizures were prospectively enrolled in a clinical trial to test the effectiveness of eslicarbazepine acetate as an adjunct to levetiracetam or lamotrigine (phase IV clinical trial NCT03116828). One hundred two patients maintained a seizure diary and wore a biosensor for >30 weeks. Based on diaries, we labeled days as either “seizure” versus “no-seizure” or “preseizure” versus “no-preseizure.” We compared patterns obtained by harmonic 24-h modeling between conditions. Best-ranking wearable markers and seizure diary variables were fed into a fully connected neural network, with several hidden layers and depth as hyperparameters that classified between seizure day conditions.

Results: The final sample contained 70 patients (median age = 42.5 years, 43 female) with 5437 recorded patient-days, including 557 seizure days and 537 preseizure days. Twenty-four- hour patterns in electrodermal activity and accelerometry differentiated no-seizure versus seizure days, as well as no-preseizure versus preseizure days (p < .01). Classification between no-seizure and seizure days (weighted F1 = .81, sensitivity = .82, specificity = .67) as well as between no-preseizure and preseizure days (weighted F1 = .82, sensitivity = .80, specificity = .66) revealed good performance.

Significance: Wearable data capture seizure-related differences with daily resolution, differentiating between days with lower and higher seizure susceptibility. Combining diary-based clinical and wearable data bears the potential for developing a dynamic seizure detection and prediction system with daily resolution.

Commentary

This study by Vieluf and colleagues adds to a growing body of work examining how physiologic data from wearable devices complement patient-reported seizure diaries. ¹ Seizure frequency, the cornerstone outcome in clinical care and trials, has relied heavily on self or caregiver report. Yet inaccuracy, especially underreporting of nocturnal and focal impaired awareness seizures, is well recognized as a major challenge. ² The authors combine diary entries with wrist-worn biosignals and apply pattern analysis to distinguish seizure, pre-seizure, and no-seizure days, signaling detection and prediction capability.

In this retrospective observational cohort study, the authors evaluated wrist-based Embrace (Empatica) sensors, which recorded electrodermal activity, peripheral wrist temperature, and three-axis accelerometry; heart rate was not measured. Multimodal biosignals were used to derive seizure-risk metrics and assess temporal associations with patient-reported events. The key analyses revealed overall good performance. However, a fundamental flaw is in the use of seizure diaries as the reference comparator in several analyses. Anchoring validation on a flawed standard can distort the resulting statistics. The authors acknowledge their methodological constraint.

What strikes me most is that this very limitation illustrates the challenges of relying heavily on seizure diaries in both clinical practice and research. It highlights the need for objective tools that can serve as reliable proxies for seizure outcomes, guiding both treatment decisions and trial endpoints.

What Devices are Currently Available?

Interest in wearable seizure-detection devices emerged in the late 2000s alongside widespread of smartphones, with early work at Massachusetts Institute of Technology and Boston Children's Hospital highlighted by an influential study demonstrating wrist-worn seizure detection. ³ The expansion of technological ecosystems creating the infrastructure for biosensors, cloud storage, machine learning, and real-time alerts has enabled the convergence of consumer electronics and regulated medical devices.

Nowadays, wearable seizure detection is no longer theoretical. Several devices are in clinical use, including Embrace2, EpiMonitor, NightWatch, and EpiWatch—all FDA-cleared systems that monitor physiologic signals such as motion, electrodermal activity, temperature, or heart rate to detect convulsive or nocturnal seizures and promptly alert caregivers through smartphone notifications or bedside alarms. Implantable systems such as the NeuroPace RNS System provide chronic intracranial EEG recordings in selected patients. Beyond commercially available tools, research platforms continue to evolve including surface EMG patches and behind-the-ear EEG systems.

Validation studies show reasonable sensitivity for generalized tonic–clonic seizures but more modest performance for nonmotor events. ⁴ Forecasting studies suggest seizure risk follows temporal patterns detectable with long-term monitoring.^5,6 The technical groundwork is now established, but the greater challenge lies in translating these capabilities into meaningful integration within routine clinical care and trials.

Parallels with Other Fields

Many industries have adopted long-term physiologic and performance surveillance. Elite sports use biometric tracking to guide training decisions, aviation employs fatigue monitoring to reduce risk, and industrial sectors rely on continuous sensors to anticipate system failure. In healthcare, cardiology moved from episodic electrocardiograms to continuous rhythm surveillance. The Apple Heart Study demonstrated large-scale atrial fibrillation detection using a smartwatch. ⁷ Continuous glucose monitoring reshaped practice by replacing intermittent finger-stick data with round-the-clock metrics, resulting in real-time alerts that improves safety and glycemic control. ⁸

The broader shift has been toward continuous data capture, which enables predictive analytics and ultimately informs operational decision-making, even in real time. Epilepsy care, by contrast, still relies largely on episodic EEG and patient or caregiver recall. This work by Vieluf and team reflects that epilepsy is entering its own phase of continuous physiologic monitoring. As in other fields, adoption of these technologies must be grounded in validation, defined clinical use cases, and structured integration.

A Forward Path

Recognizing and addressing persistent gaps at clinical, research, patient-centered, and system-level, is essential for successful implementation. The work by Vieluf and team reinforces the value of integrating subjective and physiologic data. Their findings support a hybrid model in which diary reports are contextualized by continuous biosignals.

Epilepsy management has been reactive, guided largely by breakthrough events rather than continuous risk assessment. Wearable monitoring does not replace EEG or patient narrative but opens the possibility of preventive adjustment based on risk signals. By generating continuous, high-volume data, wearables provide an opportunity to improve seizure quantification, identify temporal risk patterns, and guide individualized therapy. Additionally, it can help improve accuracy in research endpoint assessments.

Adoption should proceed with rigor: transparent validation standards, defined clinical use cases, seamless connection to clinical and research operations, and ongoing assessment of patient experience. System-level challenges including data interpretation, electronic health record integration, standardized reporting, reimbursement, and governance must be addressed in parallel.

The promise lies not in the devices themselves, but in disciplined integration grounded in evidence and clinical judgment.

COI Statement: The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.