Rethinking Seizure Care Blog

Debating the Accuracy of Seizure Detectors in Ambulatory EEGs

Posted by RSC Diagnostics on Apr 14, 2021

Concept of human intelligence with human brain on blue background

Ambulatory EEG is used in an outpatient setting and is frequently associated with higher convenience and lower cost when compared to video-EEG monitoring done at a facility. Even though it is more ideal for the patient, it is quite difficult for clinicians to assess the recordings entirely and accurately. Because of this, Automatic Seizure Detectors, otherwise known as ASDs, can help assist medical practitioners when reviewing data. Many ASDs are available in different commercial formats but are not deemed entirely accurate compared to the trained eye of a licensed medical professional.

According to recent data, these Automatic Seizure Detectors return extremely varied results as far as seizure detection accuracy. A paper published in Neurology explained a study conducted by Dr. Scheuel and colleagues. The study used raw data from around 1300 peoples’ Ambulatory EEG recordings that lasted greater than 24 hours. The data was processed through one of three different Automatic Seizure Detectors to allow varied results based on the specific machine.

At the same time, the exact data being put through the ASDs was also being reviewed by an expert neurophysiologist. When the neurophysiologist detected an event in the data, such as a seizure, it was confirmed by another professional in the field. The study considered these events “expert-detected seizures.” After this, experts were able to compare the human-calculated data with the results that the Automatic Seizure Detectors were returning.

Discovery and Results of the Study

The results of the study are quite surprising. Of the 1300 people, the expert neurophysiologist detected seizure-like brain activity in almost 6% of the cases – around 80 people. The Automatic Seizure Detectors discovered events in only 3% of people, which is only half of the expert-detected seizures.

When comparing the three different types of ASD used in the study, there were no significant differences between the devices. This confirms that though they are not as accurate as many would like, at least the machines are consistent.

Another thing to note is that while comparing the types of seizures detected, the Automatic Seizure Detectors were slightly better at detecting generalized seizures rather than focal seizures. Generalized seizures begin in groups of cells located in both sides of the brain, while focal seizures typically start in a smaller area in only one group of cells. This confirms that Automatic Seizure Detectors are more accurate when looking at the full brain as a whole, while they are not so great at detecting seizure-like activity that starts in small areas.

Based on this study, experts were able to recognize that, while ASDs were mostly helpful, they cannot be compared to what a trained professional is able to detect while looking at the same data. It has been previously reported that ASDs have a high false-positive rate, and new data shows that they are also not reliable to detect all types of brain activity. While Automatic Seizure Detectors can be used at home to give more insight into brain activity, they are not recommended to be completely relied on, especially for those with an epilepsy diagnosis.

Dr. Schuele, of the study published in Neurology, believes that eventually, this technology will progress in a way that can change this area of the medical field significantly. This tool is a cost-effective and convenient way to learn more about your brain activity, but until technology advances, we must continue to constantly compare ASD data to data returned from professionals.




Topics: Epilepsy Studies