The Healthcare Engineering Blog is the
place for conversation and discussion about Infrastructure, Healthcare Devices and Products, Information Management, and Patient Safety and Security. Here, you'll find
everything from application ideas, to news and industry trends, to hot topics
and cutting edge innovations.
Engineering students at Rice University recently developed a system designed to detect and possibly stop epileptic seizures before they occur. The system first feeds neurological signals into a customized piece of hardware via intracranial electrodes. The hardware uses a custom machine-learning algorithm to monitor for signs of seizures. If a seizure’s detected, the hardware relays neurostimulation signals back to the electrode to stop the seizure from occurring. The team’s project was ultimately successful, detecting every seizure at least two minutes before its occurrence, with 3.9 false positives per hour.
The research took place as part of a competition Rice calls the Engineering Design Showcase, in which the students took the top prize and won $5,000. A team member called the project a “vertically integrated” one, meaning that the team mixes upperclass and lowerclass undergraduates and uses grad students as mentors.
On a personal note, my daughter was diagnosed with epilepsy last year, and prior to that I knew almost nothing about it. I’ve since learned that it’s one of the most common neurological diseases (it affects 1 in 26 Americans) and is often difficult to treat. Many cases respond to medication, but the side effects of anticonvulsant drugs can be extreme and include personality changes, mental deterioration, and hormonal problems. Up to 40% of epilepsy cases are refractory or “intractable,” meaning the seizures do not respond to medications. If alternative therapies like ketogenic diet and cannabinoids are ineffective, these patients face the removal of offending parts of the brain via surgery, which is completely successful in only 60-70% of cases.
In light of these statistics the system developed by the Rice team is a remarkable breakthrough. Interestingly, it’s not the first time Rice students have made developed an innovative epilepsy tool. In 2013 students developed a seizure detection belt that monitors electrical conductance on the skin and respiratory rate to alert a caregiver that a seizure is occurring. Since that time a few companies like Empatica have commercialized similar seizure detection watches that many parents of pediatric epilepsy patients find useful.
Still, these devices only alert a caregiver that a seizure is happening in real-time. Although the new system might be five or ten years from commercialization, having a system to detect seizures and stopping them before they occur would be a game-changer for the millions living with an untreatable neurological condition.
Re: Engineering Undergrads Develop A Revolutionary Seizure Detection System
05/04/2017 8:20 AM
I have read where they can smell cancers... before they are diagnosed.
I don't know of any other sense that they could be using.
I'll need to reread the article and find more background how they are detecting these 'neurological signals'... It appears when that happens, is already on its way...
if one can detect what sets up these 'neurological signals' like a dog can, you can gain an extra couple of minute to prepare.
__________________
“ When people get what they want, they are often surprised when they get what they deserve " - James Wood
Re: Engineering Undergrads Develop A Revolutionary Seizure Detection System
05/04/2017 8:43 AM
Seizure dogs are becoming popular as a kind of last line of defense for kids with untreatable epilepsy.
I've witnessed my own pets exhibiting behavior that makes me think they can somehow detect illness or distress. My grandma absolutely hated cats, but starting about two years before she died my cat would not stop laying on and pawing her stomach whenever she sat down in my house. She ended up dying of stomach cancer, and the docs estimated that she'd had it for several years starting about the time my cat exhibited that behavior. We think he somehow sensed it.
__________________
Nobody realizes that some people expend tremendous energy merely to be normal. -Camus
Re: Engineering Undergrads Develop A Revolutionary Seizure Detection System
05/04/2017 4:32 PM
The apparatus is a belt that measures skin conductivity and breathing rate. I can see how higher skin conductivity (perspiration) and scent picked up by a dog could be connected.
Re: Engineering Undergrads Develop A Revolutionary Seizure Detection System
05/04/2017 4:42 PM
Are we going to start back up on that argument about measuring body fat content through EC measurements, AND stop epilepsy in its tracks in one swell foop?
__________________
If it ain't broke, don't fix it. Just build a better one.
Re: Engineering Undergrads Develop A Revolutionary Seizure Detection System
05/08/2017 10:06 AM
No, I was referring to this giant argument over whether one can or cannot measure body fat content by EC (apparently some fitness palace gadget claims this).
What a crock! I think my dog could detect a seizure coming on, and this weekend, I met a dog that is suffering from epilepsy. Did you know that dogs can get that condition? They are using a panel of the usual drugs to treat his condition.
__________________
If it ain't broke, don't fix it. Just build a better one.
Re: Engineering Undergrads Develop A Revolutionary Seizure Detection System
05/04/2017 1:25 PM
"If a seizure’s detected, the hardware relays neurostimulation signals back to the electrode to stop the seizure from occurring. The team’s project was ultimately successful, detecting every seizure at least two minutes before its occurrence, with 3.9 false positives per hour."
This might be good for a detection system, but with ~4 false positives/hour, that is a lot of unnecessary shocks to the brain.
__________________
TANSTAAFL (If you don't know what that means, Google it - yourself)
Re: Engineering Undergrads Develop A Revolutionary Seizure Detection System
05/06/2017 3:06 AM
3.9 false positives per hour does not mean 3.9 false positives per hour per person. Few epileptics have seizures more than once per day and many much less frequently, sometimes going years between episodes. This statistic is very poorly worded and gives no indication how it was derived.
Re: Engineering Undergrads Develop A Revolutionary Seizure Detection System
05/06/2017 9:11 AM
I agree that 3.9 false positives per hour is very circumspect. I suspect the time unit of an hour is the error. This numeric error per week, month, semester or in total would be worthy of bragging rights. I greatly doubt the number of test subjects mitigate this error rate for six subjects would still mean someone had a false positive every day. It also seems very unlikely an undergraduate engineering team in a competition would be allowed even six known epileptic subjects they could gather data and shock with daily false positives triggering a shock.
__________________
"Don't disturb my circles." translation of Archimedes last words
Re: Engineering Undergrads Develop A Revolutionary Seizure Detection System
05/04/2017 3:05 PM
Remarkably, the ketogenic diet is quite similar to the new diabetic type II diet I am currently using. Coconut oil (MCT's) are a key component, along with all other healthy fats (discluding transfats), limitation of carbohydrates to a target of 50 g per day (regardless of the type), protein in normal amounts, plenty of vegetables, just very little potatoes.
Interesting indeed!
I find it somewhat difficult to leave off the types of carbs that used to find me. Donuts, hamburger buns, wheat bread, so now, it is just oats with berries and milk in the morning (loaded with MCT extra, and plenty of butter, even walnuts)
At least, I can have all the bacon and ham I want.
__________________
If it ain't broke, don't fix it. Just build a better one.
Re: Engineering Undergrads Develop A Revolutionary Seizure Detection System
05/04/2017 5:01 PM
With 3.9 false positives per hour (per subject?) this leads me to a few questions.
How many true positives per hour happened? If hundreds of false positives were detected for each true positive I then wonder if just random luck of triggering an alarm every 15 minutes lands within 2 minutes of a real seizure.
What effect did the electrodes have on the subject during a false positive event?
How was a false positive result determined? Possibly some of the false positive results were actually successful interruptions of a seizure.
I suspect these questions are answered in the scholastic report.
__________________
"Don't disturb my circles." translation of Archimedes last words
Re: Engineering Undergrads Develop A Revolutionary Seizure Detection System
Comments rated to be "almost" Good Answers: