EpiWatch App Shows High Sensitivity for Tonic-Clonic Seizures

The EpiWatch app achieved high sensitivity for detecting tonic-clonic seizures, according to coverage from Neurology Advisor, positioning smartwatch-based seizure detection as a credible adjunct to conventional scalp EEG monitoring in the epilepsy evaluation pathway.

EpiWatch, developed through research at Johns Hopkins, uses accelerometer and heart-rate data from a wrist-worn device — specifically the Apple Watch — to detect the characteristic motor signatures of tonic-clonic seizures. Unlike scalp video-EEG, which remains the diagnostic gold standard in the epilepsy monitoring unit, the app offers continuous, ambulatory detection outside hospital walls.

The reported finding matters for one practical reason: tonic-clonic seizures are the seizure type where delayed detection carries the greatest clinical risk, including suffocation, injury, and sudden unexpected death in epilepsy (SUDEP). A wrist sensor that reliably flags these events could extend monitoring into the home, where most seizures actually occur and where electrodes are rarely in place.

What did the study measure?

The evaluation centered on sensitivity — the proportion of tonic-clonic seizures the app correctly detected. High sensitivity for this seizure type is the benchmark that matters most, because motion-sensing wearables are inherently better suited to detecting convulsive events than non-motor seizure types such as absence or focal aware seizures, which lack the rhythmic motor output an accelerometer can capture.

The detection approach relies on two streams of data:

• Accelerometry, capturing the rhythmic clonic movements of a tonic-clonic event
• Heart-rate measurement, exploiting the tachycardia that typically accompanies these seizures

Combining motor and autonomic signals is what separates serious detection tools from generic fitness trackers; ictal tachycardia provides a second, physiologically independent channel that helps filter out non-seizure movement such as exercise or falls.

How does this fit into clinical practice?

For epileptologists and neurophysiologists, the implications are procedural rather than theoretical. Outpatient seizure diaries are notoriously unreliable — patients frequently under-report or fail to recognize events. An app that logs detections automatically could improve:

• Seizure-frequency estimates used to titrate antiseizure medication
• Candidacy assessments for epilepsy surgery, which hinge on accurate event counts
• Nocturnal monitoring, where caregivers may sleep through events
• SUDEP risk counseling, since nocturnal tonic-clonic seizures are a recognized risk factor

Clinicians should note the distinction: EpiWatch is an adjunctive detection and logging tool, not a diagnostic instrument. Confirmation of seizure type and classification still requires scalp video-EEG in the monitoring unit, and in selected cases intracranial EEG. The app detects events; EEG characterizes them.

What are the limitations?

Any wearable detection system confronts the same constraints. Detection performance depends on the patient wearing the device consistently, including during sleep. Specificity remains the counterpart metric to sensitivity — false alarms erode caregiver trust — and the headline finding of high sensitivity should be read alongside whatever false-alarm rate the study reported.

The technology also cannot address non-convulsive seizure types or status epilepticus without prominent motor features, which is why home EEG and wearable detection are increasingly seen as complementary rather than competing layers of ambulatory monitoring.

What comes next?

The EpiWatch findings add to a growing evidence base supporting wearables as a bridge between episodic inpatient monitoring and the unsupervised reality of daily life with epilepsy. If sensitivity findings hold up in larger, prospective cohorts, smartwatch seizure detection could become a routine part of the discharge conversation in epilepsy monitoring units — a low-cost sensor array already sitting on the patient's wrist.