Epilepsy Essentials: Seizure Detection & Prediction

Why Should We Detect Seizures?

The first goal is the safety of the person with epilepsy or other seizure disorders. Seizure detection devices can alert caregivers that a seizure is occurring. This can prevent injury and life-threatening events that may occur with prolonged seizures. Sudden unexpected death in epilepsy (SUDEP) correlates most strongly with having seizures during sleep, sleeping in the prone position, and having more than 3 tonic-clonic seizures per year.¹ Detecting seizures during sleep, especially if there are changes in heart rate or respiration, may allow a caregiver to rouse or stimulate the person having a seizure. Such monitoring and intervention, even simply sharing a bedroom, makes SUDEP less likely.²

The second goal of seizure detection is to understand how seizures occur.³ It is very difficult for people who have seizures to accurately count and report seizure frequency, not because they are poor reporters, but because of the nature of seizures that may prevent a person from being aware of every seizure they have. A person with epilepsy may not know they had a seizure unless it was observed by someone else. Objectively measuring seizure frequency can identify when a person’s seizure control is not optimal and guide treatment decisions (eg medication changes or adherence, or referral for epilepsy surgery or implantable neuromodulation devices.)

What Seizure Detection Is Available?

EEG is still the most reliable method of seizure detection because it measures the electric activity that defines a seizure.⁴ Ambulatory EEG, however, is not easily done for extended periods, because the quality of recording diminishes with time and it is cumbersome for patients to wear an electrode cap. Ambulatory EEG records for a maximum of 72 hours before reapplication of electrodes is needed. The advantages of EEG are the amount of data gathered and the direct and sensitive measurement of neurophysiologic activity, although this also brings the disadvantage of a higher artifact rate making interpretation more challenging.

Other external devices detect tonic-clonic seizures by measuring other physiologic changes that occur with those seizures. Changes in heart rate, sweating, and movements typical of tonic-clonic seizures can be measured with wearable devices. Movements can also be observed with detectors in the environment, including mattress devices and video cameras. Devices that measure multiple parameters tend to be more sensitive and accurate.³ The ability to wear devices or have them integrated into the environment are advantages of this technology. The detection of seizure, however, is less precise. Currently these devices err or the side of false positives when it comes to convulsions because it is better to rouse someone unnecessarily than to miss a seizure that could be life-threatening.³ As accuracy of the devices improves, it may become possible to alert EMS that someone is having a seizure, but the high rate of false-positive alarms make that unfeasible at present. In contrast, seizures that are not associated with abnormal movements tend to escape detection.

Smaller EEG devices including a lead with 4 electrodes that can be implanted under the skin are in development.⁴ These provide the reliability of EEG, without the inconvenience and obviousness of EEG caps and use of less battery power. Use however, requires an invasive procedure with infection risk. Also, these EEG leads could easily miss seizures if they were not placed on the appropriate part of the head taking into account where an individual’s seizures are occurring. Another device in development adheres 2 or more patches of 2 electrodes to the scalp,⁵ which takes away the risks of an invasive procedure but still has the issue of appropriate placement to resolve.

Although more invasive and reserved only for those with drug-resistant epilepsy, the responsive neurostimulation (RNS) device treats epilepsy with implanted electrodes that deliver neuromodulatory current to the brain in response to seizure detection. Advantages are that RNS is therapeutic in addition to detecting seizure activity.³ Disadvantages are that it is highly invasive, only appropriate for a subset of people with epilepsy, and has limited data storage capacity.

Who Benefits Most From Seizure Detection?

SUDEP risk reduction is the largest benefit of seizure detection. Individuals at risk because of nocturnal seizures, prone sleep position, or a history of 3 or more tonic-clonic seizures per year can benefit from seizure detection.^1,2 Some people with epilepsy are aware of unobserved seizures because of an injury (eg, a tongue laceration or a bump on the head from falling), and these should be included in the count of tonic-clonic seizures. Clinicians should consider referral to a tertiary epilepsy care center when they consider seizure detection devices, because those at higher risk of SUDEP benefit from this level of care.⁶ Other than EEG, seizure detection devices do not detect focal seizures, so people without tonic-clonic seizures are not yet benefitting from the this technology.

What’s Next in Seizure Detection?

To detect a seizure, devices must be in place before it occurs. Consequently, seizure detection devices are recording data of what happens in the hours leading up to a seizure. This data, often with the use of artificial intelligence, is being used to research seizure prediction. Knowing when a seizure is going to occur would be truly life changing for people with epilepsy and seizure disorders because the unpredictability of seizures limits people in so many ways. These limits include not be allowed to drive because a seizure during driving could be fatal or avoiding going to a grocery store or a party because seizures tend to be embarrassing, alarming, and dramatic. People also may lose their jobs because they are felt to be a liability in the workplace.

With seizure prediction, people could receive an alert that they are likely to have a seizure within a certain number of hours, which would allow planning and risk-reduction around that likelihood. Perhaps someone could even reduce risk by taking medication to prevent or stop a seizure or ensure they will be in a safe spot in the company of people who know what to do in the event of a seizure.

What Is Most Important for Clinicians to Know?

Seizure detection may be useful for anyone with epilepsy who has more than 3 tonic-clonic seizures in a 12-month period, nocturnal seizures, or a prone sleep position. Seizure detection can save lives by reducing risk of SUDEP and guide treatment decisions by accurately counting seizures. Seizure prediction would improve quality of life and reduce disability for people with epilepsy. This may be another reason for people to consider seizure detection devices—the more data we have on changes leading up to seizures, the closer we may get to be able to predict them.