SeizeIT - Discrete personalized epileptic seizure detection

From 01-10-2016 to 30-09-2018
Website: https://www.iminds.be/en/projects/SeizeIT

Description

Epilepsy, affecting about 1% of the world population, is a neurological disease state marked by its heterogeneity resulting in unique patient journeys connected only through the presence of recurrent seizures. Patients are significantly affected in their daily activities and emotional life (fear and anxiety), as seizures typically occur unexpectedly. Because of loss of control and consciousness, patients often experience serious injuries. Moreover, each year 1 out of 1000 epilepsy patients die suddenly, of unexplained reasons (SUDEP).

A wearable seizure detection system would be of high benefit to epilepsy patients, to give them control in daily life.  First, such a system could warn the patient and the caretaker who can, in turn, provide assistance to the patient. Second, as patients do not report accurately their seizures to the physicians, no objective information is available about what happens in between visits. The objective assessment of seizure occurrences could help to determine the exact seizure frequency and the severity of the disease, facilitate better follow-up and evaluate treatment options in a personalized manner. Finally, continuously recorded physiological signals may provide information to help explain the occurrence of SUDEP.

Wearable sensor technology has shown a rapid development in the past years. Indeed, several devices that claim seizure detection functionality have reached the market or are in development phase. Nevertheless, none of these devices have so far fulfilled the needs of the patient nor are of true clinical benefit. Indeed, a great challenge is posed by the inherent trade-off between wearability and detection performance, due to considerations such as device size, placement or battery life. The optimal choice is very disease-specific and requires an application-driven design methodology, where the patient-specific needs are identified in an early stage and translated in both technical and design requirements for a tailor-made wearable solution.

The majority of current wearable solutions, however, pursue wearability as a primary objective and assess seizures by a single wrist-worn device. Such a device-driven approach lacks sensitivity and specificity. In fact, epilepsy comprises a broad spectrum of seizure types and diverse clinical manifestations. In order to capture relevant information, it is crucial to record various complementary physiological signals and develop detection algorithms which can be personalized.

There is clearly an interdependence between the hardware, software and design specifications of such a dedicated seizure detection device driven by patient needs. Therefore, its development and valorization must involve an interdisciplinary team from the earliest stage on, including clinical neurologists working in close contact with epilepsy patients (Faculty of Medicine), hardware designers (Byteflies), signal processing experts (Stadius), industrial designers (pilipili) and an industrial partner in the healthcare/pharmaceutical sector with a proven track-record in bringing medical quality products to the market (UCB).

Team

• Sabine Van Huffel, Promoter
• Borbála Hunyadi, Research Leader
• Steven Vandeput, Team member
• Kaat Vandecasteele, Team member

Financing

Funding: imec - imec

Program/Grant Type: imec ICON - imec-ICON