Position Statement on Noninvasive Seizure Detection/Alerting Devices

Approximately 3.4 million people in the United States live with epilepsy, a chronic neurological disorder characterized by recurrent seizures. Despite advances in antiseizure medications, roughly one-third of individuals continue to have inadequately controlled seizures, placing them at ongoing risk for traumatic injuries and sudden unexpected death in epilepsy (SUDEP).

In response, seizure detection/alerting devices (SD/AD) have emerged as a promising component of comprehensive epilepsy care. These SD/ADs include external wearables, noninvasive, home-based systems, and implanted systems that use physiologic or visual signals such as motion, electrodermal activity, audio, EEG, or other biosignals, to detect seizures and alert caregivers or healthcare providers. Beyond enabling faster intervention that may lessen the severity of seizure-related complications, these systems can help track seizure frequency and patterns over time, which may enhance clinical care, grant more independence, and support epilepsy research. Early evidence also suggests these technologies may reduce emergency healthcare utilization.

Device development in this space has rapidly expanded. Several Food and Drug Administration (FDA)–cleared seizure detection devices and platforms are now available, with growing performance and usability data, though cost and reimbursement from insurance remain a significant barrier. Emerging evidence supports promising performance, acceptance, and real-world impact. Surveys of patients and caregivers indicate increasing interest in these technologies to support independence, reduce caregiver anxiety, and promote rapid response when seizures occur, especially at night or when people with epilepsy are alone.

Although the FDA has cleared noninvasive and approved invasive (implanted) seizure detection devices, this position statement focuses on noninvasive devices.

Noninvasive SD/ADs may offer meaningful clinical benefit in a variety of scenarios. For instance, for adults with epilepsy who have tonic–clonic (convulsive) seizures and are living alone, a wrist-worn device can alert family members in real time and provide objective seizure frequency data to inform treatment decisions. Parents of children with major motor seizures can also use nocturnal monitoring and alerting devices to facilitate safer sleep when they are not in the same room. Additionally, for individuals with suspected unwitnessed tonic-clonic (convulsive) seizures (eg, waking with sore muscles or tongue biting), clinicians may recommend SD/ADs to help objectively document when or how frequently such events occur.

Across studies, noninvasive seizure detection devices demonstrate the strongest performance for detecting tonic-clonic (convulsive) seizures. Performance varies by device, sensing modality, study design, and setting, and should be interpreted using standardized metrics such as sensitivity, false alarm rate, and detection latency. Performance currently remains more limited for many seizure types, including focal impaired consciousness and absence seizures, among others.

Evidence is growing for effectiveness in home and community settings. A major benefit is real-time caregiver notification of detected seizures. These devices can also improve safety planning for individuals at elevated risk of seizure-related injury or harm and provide objective documentation of seizures to supplement self-reported diaries. The use of devices has also been shown to reduce seizure-related anxiety and social burden, particularly among caregivers.

Current noninvasive SD/ADs have important limitations. Since many devices are optimized for tonic-clonic (convulsive) events, there can be incomplete or no detection of other nonconvulsive seizure types. False positives, which can erroneously alert caretakers when a detected signal is incorrectly identified as a seizure, can create issues, and contribute to alert fatigue. The rates of false positives may vary by device and individual factors, such as levels of alertness, wakefulness, or sleep stage changes, baseline heart rate and heart rate variability, or degree of physical activity and movement. For example, clinicians should note that device performance varies between sleep and wakefulness: while false alarm rates are lower during sleep, sensitivity across sleep stages remains incompletely characterized, as electrodermal activity is suppressed during deep NREM and muscle atonia during REM alters accelerometry baselines. Additionally, there are several potential technical limitations that apply both to devices and users. Devices themselves can fail, can disconnect from Bluetooth or other communication pathways, and components can wear down over time. Users may not wear devices consistently, and children and individuals with developmental delay may find devices bothersome and remove them.

There is also insufficient data related to devices preventing SUDEP or other serious adverse outcomes. From a clinical standpoint, counseling should emphasize established risk reduction strategies, particularly for improving seizure control and reducing convulsive seizure frequency. Additionally, these limitations need to be weighed considering device cost, variable insurance coverage, and other access barriers can limit equitable use. Data generated by SA/SD devices and wearables must be protected with robust privacy and security safeguards. Patients should be informed about data ownership, secondary use, sharing practices, and risks. Optimal integration of the obtained data into clinical workflows must also be determined and implemented.

AES, EF, and the Danny Did Foundation recognize the essential role of regulatory oversight in ensuring the safety and efficacy of medical devices and Software as a Medical Device. They support a clear differentiation between regulated medical devices and consumer wellness products (eg, general baby monitors or cameras not studied in epilepsy). They support ethical marketing and communication that avoid overstating capabilities or implying unproven benefits, such as SUDEP risk reduction. They support transparency regarding algorithm updates, changes in performance over time, and postmarket surveillance.

Clinicians should help patients and caregivers understand the distinction between regulated, epilepsy-specific devices and general consumer products that do not include epilepsy-focused evidence or FDA evaluation.

Ownership of data generated by seizure detection wearable devices beyond direct care remains legally ambiguous. Clear safeguards and transparent communication are needed to build trust and protect against unintended uses, such as insurers, employers, or disability adjudicators—representing a critical and unresolved patient rights issue that we must remain cognizant of and address over time.

AES, EF, and the Danny Did Foundation acknowledge the potential of emerging technologies to advance epilepsy care, including multimodal sensing, machine learning, and closed-loop systems. Realizing this potential will require ongoing collaboration among clinicians, researchers, patients and caregivers, industry, regulators, and professional societies to ensure that innovation leads to safe, effective, and equitable improvements in the lives of people with epilepsy

Priority areas for future research should include a focus on patient-centered outcomes. Examples include tracking the impact of effective device use on quality of life, sleep, caregiver burden, injury rates, emergency response and healthcare resource utilization, along with broader health economic and outcomes research. Improving the detection of, and ability to differentiate between, a broader range of seizure types will also be important, including improved differentiation from nonepileptic events.

For device studies, it will also be important to standardize outcome measures, reporting frameworks, and benchmarking methodologies to enable meaningful comparison across devices, different populations, and real-world settings. The efficacy and accuracy of seizure detection and alerting wearables have not been adequately studied in people of color or racially and ethnically diverse populations, leaving significant gaps in our understanding of how these devices perform across all patients. For devices shown to be effective at seizure detection, additional studies are needed related to implementation, assessing adherence, human-factors considerations, alert fatigue, and optimal integration into clinician workflows. A key area of work will also need to focus on health equity and reimbursement, including coverage pathways, cost-effectiveness analyses, and strategies to reduce disparities in access and use. Actively engaging people with epilepsy and caregivers throughout study design, technology development, usability testing, and evaluation, is critical for ensuring relevance and acceptability.