This innovation presents a machine learning-based diagnostic tool that uses an electrocardiogram (ECG) signal to accurately distinguish between convulsive epileptic seizures and functional/dissociative seizures.
Diagnosing epileptic seizures (ES) versus functional/dissociative seizures (FDS) is complex and traditionally depends on long-term inpatient video-EEG monitoring, which is costly, time-consuming, and often inaccessible to many patients due to geographic and financial constraints. Additionally, these events are sometimes not captured during inpatient monitoring. Consequently, current methods limit timely and accurate diagnosis, resulting in delayed or improper treatment. This challenge has spurred research into more accessible diagnostic approaches that can be performed outside hospital settings.
This technology introduces a machine learning diagnostic model leveraging electrocardiogram (ECG) recordings and ECG analytics to distinguish convulsive epileptic seizures (ES) from convulsive functional/dissociative seizures (FDS, Ryan et al. Seizure 2023, PMID: 37660533). The model accurately distinguishes convulsive ES vs. FDS. The central innovation lies in enabling reliable, non-invasive, and rapid assessment without the need for hospital admission or specialized inpatient equipment. This approach provides an initial screening tool, which has the potential to democratize access to seizure diagnostics through out-of-hospital testing. This has the potential to be especially valuable for patients in underserved or remote areas. The ECG-based algorithm delivers a scalable and affordable initial screening diagnostic method. This significantly lowers barriers to early and accurate seizure classification, facilitating timely clinical decisions and personalized patient care.
Picture for reference only, not a depiction of the invention.
• Non-invasive and out-of-hospital diagnostic capability using wearable ECG sensors.
• Improves accessibility to seizure diagnosis for patients in remote or underserved locations.
• Reduces dependency on expensive, long-term inpatient monitoring procedures.
• Provides rapid and accurate differentiation between epileptic and functional/dissociative seizures.
• Cost-effective and scalable solution leveraging advanced machine learning and ECG analytics.
• Supports improved clinical decision-making and patient treatment planning.
• Initial diagnostic screening tool for people with limited geographic or financial access to epilepsy-monitoring units.
• Initial diagnostic screening tool for people whose seizure events are not captured in the epilepsy-monitoring unit, or people with an inconclusive diagnosis.
• Remote healthcare delivery for patients lacking access to specialized epilepsy monitoring units.
• Use in wearable health technology platforms for continuous seizure evaluation.
• Support tool to classify seizure types.
• Potential integration into telemedicine services to facilitate virtual diagnosis and management.
Patent Pending 2025-0380898
This technology is at an early validation stage (TRL 3–4), with proven proof-of-concept demonstrating accurate seizure classification using real-world ECG data; ongoing work is focused on clinical validation.
This technology is available for licensing.