Hacker News new | past | comments | ask | show | jobs | submit login
This Startup is solving epilepsy with biosensors and AI (epilert.io)
19 points by ned7 on April 5, 2021 | hide | past | favorite | 16 comments



Actual title is Epilert, the finest epilepsy detection device.


"solving epilepsy" is indeed in very bad taste.


Guh, this is truly terrible.

>CAN EPILERT PREDICT SEIZURES? >Epilert cannot predict or prevent seizures. >Our technology is designed to detect patterns that could be >associated only with motor seizures for now. However, we are >gathering EDA data to try building behavioral and emotional >profiles relative to non-motor seizures that will help >scientists better understand these types in the future.

>WHAT TYPES OF SEIZURES IS EPILERT DESIGNED TO DETECT? >Epilert Bracelet is made to detect motor seizures that can >be focal or generalized and could be tonic-clonic or from >another motor type.

So.... thanks for 'solving' the I-had-a-locomtor-seizure-and-I-guess-I've-fallen-and-can't-get-up-please help-me-problem. Thanks but not thanks; already have a toolbox of coping strategies for that. Call me back when you've got a 'solution' for: * please don't arrest me for public intoxication - I'll be better in 20minutes * here's a record of everything that happened 60minutes before and after an absence seizure (that this won't detect) 'cause I don't remember any of it * having a shitty day and I can't tell what you're feeling, what I'm feeling or how to feel about the meta. Please leave me alone but, well, don't fire me ok? * is this headache signalling a migraine or a seizure? They both suck but my best options are dramatically different

There are so many things that could be so much better with some unobtanium applied to this field... but detecting tonic clonic seizures is... underwhelming. This is no kind of solution but mightpossibly* be a tool in the toolbox; that's about it.


This product says it detects (doesn't prevent) tonic-clonic seizures. Epilepsy covers such a wide array of conditions. Tonic-clonics are the type of seizure that people envision when they think about seizures (unconscious on the ground convulsing), but there are so many other types - absence seizures, for instance, result in a person appearing to just stare off into space. There are many products that aim to detect tonic-clonic seizures.

My almost 7 year old was diagnosed with a form of epilepsy at age 3.5. It's luckily mostly controlled at the moment with a 3:1 keto diet and a cocktail of drugs, but it's been a long few years. We've dealt with many, many tonic-clonics over those years, as well as other types. By far the most common for us (dozens or more per day at times) has been myoclonic seizures, resulting in a sudden jerk of some muscle. This might result suddenly dropping whatever you're grasping in your hand, or suddenly ending up with a wet shirt after the liquid you just drank falls back out of your mouth. But the worst by far have been the atonic seizures - complete loss of muscle tone and consciousness for a brief moment resulting in a standing person falling face-first in to the pavement with no protective reflexes. It's hard to describe the feeling of helplessness as a parent when this happens to your child.

Anyway, mostly venting. It's nice to see any acknowledgment of epilepsy in more general outlets. Not sure I'd consider this solving epilepsy, though.


I have epilepsy and while this is a far cry from solving it, I love to see more products in the space.

A device that is closer to 'solving' (focal) epilepsy is Neuropace, which uses electrodes around the seizure focus to detect a developing seizure and promptly deliver electrical stimulation to counteract it and prevent its progression before it becomes noticeable. It's been quite successful so far, leading to substantial seizure reduction and boosts in cognition (which is often reduced by anti-epileptic drugs and/or regular seizures).

The research team behind Neuropace has also used data from the devices to build seizure risk forecasting models for individuals (https://www.ucsf.edu/news/2020/12/419376/seizure-risk-foreca...).

Given the device is so much less intrusive than Neuropace, it would be exciting to see Epilert successfully do the same (which seems to be one of their goals). With good forecasting, some patients could potentially prevent seizures through modulation of medication alone. Though I'm not really their 'target patient' (I have well controlled TLE), I'd happily wear the device when at home to contribute data to an effort to pull that off.


without regards to the hubbub around "AI".

> Epilert contains a monitoring system that enables epileptic patients, their caregivers and doctors to monitor patients’ vital signs, daily lifestyle, medication taking and seizures through dedicated mobile apps.

I live with someone who has epilepsy. For adults, it's largely manageable with some medications. For my partner, this enables her to work and live mostly normally.

The biggest issue for her is having an early warning system. She is unable to obtain a drivers license, though she doesn't frequently convulse at random times. It's typically cyclical for her, and of the time I have known her, it's never been "unexpected". She usually knows before hand what's going on and lies down in bed, takes a secondary anti-convulesant medication and relaxes. Most of the time this - just works.

However, I would love for her to be able to drive, as it's a huge impediment for her being able to work places, or visit places. I don't especially like being a standby taxi driver.

If she could reliably get alerts from a monitoring device, I think she would be low risk for driving. The fear of the DMV is that she has a seizure on the highway and kills herself or someone else. With a good early warning system, she could pull over, or not drive that day or whatever.

Epilepsy is manageable but also can be very rough - it's a lifelong situation and is can be triggered by a sorts of things.

Anyway, just my two cents.


if you don't mind me asking, does your partner take daily medication for it?

I discovered I have epilepsy about 2 years ago, but my case seems to be very mild. I'm now on Tegretol and haven't had any seizures since. But, of course, I know the intensity and triggers vary wildly from person to person.


yes, she does


I have no experience of epilepsy, but I'd suggest (1) health conditions are not problems to be "solved", and (2) reducing complex systems to individual "soluble" components emphasises those components at the expense of the whole. In other words, someone living with epilepsy might have a whole ton of stuff going on in relation to the epilepsy.

Monitoring, alerting, etc etc are all (potentially) great, but they're at best necessary without being sufficient.

Not to rag on Epilert. Hopefully this is a step in the right direction. The language in the submitted title just seems inappropriate. (In case it changes, it was: "This Startup is solving epilepsy with biosensors and AI").


> health conditions are not problems to be "solved",

What does this mean? Of course they are.

Leprosy was a “health condition” — something millions of people lived with and managed — until the invention of effective antibiotics. Then it was suddenly just another treatable disease.

Cataracts were another “health condition” to be managed — and then Intraocular Lens (IOL) implants were developed.

Lactose intolerance is being solved right now with gene therapy. (Pretty easily, apparently.)

I agree with you about the article’s title — that a management therapy is not the same thing as a solution, and that it’s editorial malfeasance for the article here to claim that its management-therapy tooling is part of a solution.

But that doesn’t translate to this broader claim you’re making here, of certain conditions having some inherent property of insolubility, such that even considering finding permanent treatments or prophylactics for them would be wrong-headed somehow.

“Health conditions” are just problems that don’t have any known approaches or promising avenues of research for “solving” them yet.


I didn't say they're insoluble, and I certainly didn't say we shouldn't attempt to treat and cure diseases. My issue is with seeing illness as an isolated "thing" almost as if it exists independently of the person.

Obviously, this is heavily dependent on the situation.

For example, there's the issue of fixing "problems" that are not seen as problems by the person living with them. A kinda health normalisation.

Since you mention lactose intolerance, I can certainly imagine millions of people in Asia shrugging with indifference. Plenty would be unlikely to want to change their lifestyle. What would be the point? They've been lactose intolerant for thousands of years. If they don't see a problem, does a problem exist? (I'm sure some would want to eat dairy, of course.)

Does this mean that a westerner in a dairy-consuming culture shouldn't see lactose intolerance as a problem? Nope, I know several who very much do. It's just context dependent, isn't it.

People decline treatment all the time, for a huge variety of reasons.

Even in situations where a cure is highly desirable and can be highly effective, it might not be straightforward. (I'm going to put aside religious objections, that's a whole other can of worms.) A cure might be extremely unpleasant, even disfiguring. Radical mastectomy is an obvious example.

Describing illness as a problem requires defining that problem in order to define the cure. Radical mastectomy might effectively "cure" some breast cancer, but the definitions of the problem and the cure are very limited.

So we should just move the goalposts, right? Solutions shouldn't result in severe scarring, lymphoedema, dysmorphia, depression, relationship difficulties... hmmm, moving those goalposts is starting to seem tricky.

There's plenty I'm not saying. I'm not saying we shouldn't try. I'm not saying problems don't exist (quite the opposite) and I'm not saying that health conditions should be ignored.

Health conditions are not just problems that don't yet have a solution. They can be deeply complex, involving issues of identity and ethics and more.

Like I said, I certainly don't object to curing disease. I'm just pointing out that it can be complex.


There have been many watches along these lines already, e.g. https://www.empatica.com/en-gb/embrace2/ although previous ones have mostly used motion sensing. This Epilert startup seems to be looking at motion and also other signals like heartbeat, skin temperature, electrodermal activity.

I have a close family member with severe epilepsy so I'm interested, but the usefulness of these devices comes down to reliability and minimising false alarms while also not missing any serious events, which is a very hard thing to get right.

However there's something a little odd about their slogan of "finest epilepsy detection device". I can imagine 'finest' used for chocolate or wine, but its an odd adjective for a medical device.


It looks like something written by a non-native speaker, doesn't it? They also have the following:

> With its fast-charging batteries, the device is capable of staying alive for 48h straight.

"staying alive" just sounds like a terrible choice of words for a medical device.


Indeed


In a similar vein Belgian startup FibriCheck is doing the same thing for detecting atrial fibrillation.


Theranos alert.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: