Can Embrace detect any type of seizure?



Embrace at present is optimized to detect seizures involving convulsive movements and a spike in autonomic stress, with a duration of at least 20 seconds. These characteristics are generally associated with tonic clonic seizures.

In Europe Embrace is a CE Medical class 2a device according to ISO 13485:2012. This means that it meets the essential requirements of the European General Medical Devices Directive, and furthermore that it is safe to use. In the United States, The Embrace and Alert app system for seizure detection is for investigational use only. The system is not clinically proven to detect seizures. Evidence of seizure detection using the technology in the Embrace has been gathered in a number of clinical trials since the initial discovery of skin conductance changes during a seizure was published in 2012 (Poh et al., 2012a). Empatica is in the process of conducting a Seizure Characterization clinical trial to evaluate seizure detection of generalized tonic clonic seizures (GTCS) also known as grand mal, or convulsive seizures, using the Embrace watch and Alert app.

What is Embrace designed to detect?

The software that is being tested is optimized to detect events that involve both unusual  motor activity and autonomic activation. Unusual motor activity and autonomic activation accompany GTCS or convulsive seizures.

Can Embrace detect complex partial seizures?

Complex partial seizures usually do not generate convulsions, so we would not expect Embrace to detect these types of seizures.  

Can Embrace detect absence seizures?

Absence seizures do not generate significant autonomic or motion responses, therefore we would not expect Embrace to detect these seizures.

Can Embrace detect nonconvulsive seizures?

Currently, Embrace cannot detect nonconvulsive seizures as the specific movement signature of convulsions takes precedence in our algorithm. EDA is also a contributing thread of information, but given that EDA is more prone to fluctuation, relying too heavily on it would yield a lower accuracy rate. 

Have more questions?