Predicting Seizures


You probably know someone living with epilepsy. It’s a fairly common neurological condition, affecting about 1% of the population, which means there are more than 3 million people in America and 60 million people worldwide with epilepsy. My wife is one of them, and the few times I’ve seen her have a seizure were pretty scary - she passes out cold, then comes to after 15 seconds or so, with a horrible headache and no recollection of what just happened. Her seizures are fairly tame compared to what you might find combing YouTube, but the fact is that all seizures are dangerous: most of the injuries and deaths attributed to epilepsy result from events that are a consequence of being unconscious or out of control during a seizure, like falls or car accidents.

My wife is relatively lucky, as her seizures are well-controlled with medication, and I’ve only seen her seize a handful of times over the last decade. Some people with epilepsy have much more severe symptoms, though, and can suffer dozens of seizures in a single day. What’s worse, the vast majority of people with epilepsy don’t know what triggers their seizures, or even when one is about to happen. Some people report experiencing an “aura” or other sensation immediately prior to a seizure, but this isn’t the norm.

Figuring out a way to predict when a seizure is likely to happen would be really useful to people with epilepsy, and researchers working with a company in Seattle have developed a system that they claim can do just that. The system is comprised of several components: a grid of electrodes is surgically implanted on a small part of the patient’s brain, specifically the part of the brain that doctors believe trigger the seizures (this place is slightly different for everyone, as our brains are all slightly different from one another); the electrode grid is wired to a little computer that is implanted in the patient’s shoulder, which reads the electrical signals from the brain; finally, the little computer wirelessly transmits the data to a handheld device that the patient carries with them, which is programmed to warn the patient (with vibrations, flashing lights, and sounds) when it gets a signal that electrical activity in the brain that might signal an impending seizure.

For the study, the researchers implanted the system in 15 patients with drug-resistant epilepsy, and tracked how well it predicted seizures over 4 months. In 11 of the 15 patients, the system worked, and predicted seizures with 65% to 100% accuracy. This may not seem impressive, but it is a really big deal - predicting seizures at better than chance (50% accuracy) has never been done before with an implantable device.  Of course, the study had plenty of shortcomings. There were several adverse events for participants in the study, and a couple of them had to have their implants removed for safety reasons. And, though 65% is better than chance, it leaves a lot of room for improvement.

Systems like this are not for everyone - my wife will not be lining up for an implantable seizure detector anytime soon. However, for people with seizures that aren’t treatable with medication or other therapies, this system could be a lifesaver. And if this system follows the same kind of developmental trajectory as other innovative products, we can expect the next generation of seizure predictors to be safer and more accurate than the system described here. I imagine that someday, we’ll even be able to do this without all the surgical implantation, maybe using sensors built into a hat or pair of eyeglasses, but that’s really just speculation. The point is, this represents an important development for people with epilepsy, and is just another example of science being awesome.

Image: Neural signaling in the brain (Wikimedia Commons).


Epilepsy surgery can be

Epilepsy surgery can be especially beneficial to patients who have seizures associated with structural brain abnormalities, such as benign brain tumors, malformations of blood vessels (including disorders known as arteriovenous malformations, venous angiomas, and cavernous angiomas), and strokes.

Add new comment

Filtered HTML

  • Web page addresses and e-mail addresses turn into links automatically.
  • Allowed HTML tags: <a> <em> <strong> <cite> <blockquote> <code> <ul> <ol> <li> <dl> <dt> <dd> <p> <div> <br> <sup> <sub>
  • Lines and paragraphs break automatically.

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.
This question is for testing whether or not you are a human visitor and to prevent automated spam submissions.