NeuroVista, Emerging from Stealth Mode, Unveils Technology to Predict Epileptic Seizures

(Page 2 of 2)

to a pacemaker-style device, about the size of a book of matches, implanted under the collarbone. That component, called a telemetry device, receives the brain wave data and beams it to another device about the size of a cell phone, that clips onto a belt or fits in a purse.

That receiver, programmed with proprietary mathematical algorithms, transforms the information into a real-time warning system. It has three lights, which can tell the patient if they have a low, moderate, or high likelihood of suffering a seizure soon. Sort of like a traffic light, the settings are color-coded. Blue light for low susceptibility, white light for moderate probability, and red light for high.

This is all easier said than done, of course. To make it happen, NeuroVista gathered hundreds of high-resolution EEG readouts from patients who had severe enough epilepsy that they were hospitalized for more than a week and continuously monitored. NeuroVista’s computers mined the massive reams of data to detect statistical abnormalities that can’t be spotted by a neurologist’s eye, and built them into its algorithms. Then comes the engineering challenge. Getting the electrode to be sensitive and durable enough for long-term use, making a telemetry device small enough to be implantable yet with enough battery power to transmit a lot of data, and then having an easy-to-wear receiver that can warn patients of the coming storm.

Diabetics already have automatic insulin pumps that can monitor and control their disease, and heart patients have implantable defibrillators that can shock their heart back into a normal rhythm in a cardiac arrest. But there is nothing out there like this real-time warning system for neurological disorders that can have unpredictable flare-ups, like migraine headaches, bouts of severe depression, schizophrenia, or psychotic episodes. “This device may actually be a platform for a number of conditions,” Litt says.

OK, before we go too far down that road, where’s the proof? So far, the company has yet to run a clinical trial to see how patients use its system in real life and how accurate it is. The algorithms that determine the seizure risk are proprietary and aren’t being published in the scientific literature for prying eyes to see.

NeuroVista plans to run its first feasibility clinical trial in 2009, which will be followed later by a pivotal trial that could lead to FDA approval, Harris says. “The proof will be in the pudding,” Litt says.

Harris already has a reputation as an innovator in Seattle medical devices, as a co-founder of Heartstream, which developed automatic defibrillators like the ones posted at airports in case of a sudden cardiac arrest. He also co-founded Northstar Neuroscience (NASDAQ: NSTR), which didn’t have such good luck, as its electrical stimulation device failed in January to restore arm movement in a trial of stroke patients.

Harris left Northstar before it went public, and did a stint as an entrepreneur-in-residence at Versant Ventures in Menlo Park, CA. That’s when he heard about the seizure prediction concepts that led him to start NeuroVista in January 2005. It now has 47 employees, mainly with backgrounds in engineering, biostatistics, mathematics, and clinical research.

Without overpromising to patients, Harris can get bullish in a hurry on the subject of the company. “This will be a very important company,” he says. “We’re not just creating a widget, we’re creating a platform that will affect all sorts of neurological conditions, and will lead to a revolution in therapies.”

Single PageCurrently on Page: 1 2 previous page

Trending on Xconomy

By posting a comment, you agree to our terms and conditions.

4 responses to “NeuroVista, Emerging from Stealth Mode, Unveils Technology to Predict Epileptic Seizures”

  1. Nick says:

    Most likely they’re looking at using global dimensional complexity calculations of the EEGs that will give a 15-20 minute warning before onset of a seizure. Or that’s my best guess anyway.

  2. I saw a presentation on this last year and saw one big problem with this approach. I don’t recall all of the details, but as I remember it, in order to collect the EEG data from their subjects with epilepsy, they need to heavily sedate or drug them. It was this data, collected when they are drugged, that was used to derive the algorithms that are supposed to predict the coming attacks. The big question I had was whether or not the data collected when the subjects are drugged accurately mimics what happens in the brain of a real (i.e. non-drugged) epileptic when they are about to have a seizure. If this is not the case, then their algorithms will not be useful in predicting seizures, and the technology will fail. Without this “proof of principle” piece of data I would have concerns about investing in this technology. Please post a correction if I misunderstood this, or if this basic premise has somehow changed since I heard the presentation in 2007. Stewart Lyman, Lyman BioPharma Consulting