James Weimer, assistant professor of computer science and noted medical-device entrepreneur, has received a National Science Foundation CAREER Award to improve the safety and efficacy of systems that use advanced machine learning and artificial intelligence techniques.

Learning-enabled medical cyber-physical systems (LE-MCPS) are smart medical systems that provide actionable feedback to caregivers. They use machine learning and AI to enhance their functionality and decision-making capabilities.

One example is a stroke detection device in hospitals that processes accelerometer data to not only alert nurses when a stroke is suspected, but also personalize it to the patient such that a maximum number of false alarms is guaranteed. Another example is a postpartum hemorrhage predictor that provides timely insight into the risk of hemorrhage for a woman prior to childbirth, such that clinical care teams can plan ahead and be prepared to take prompt action after delivery.

However, LE-MCPS rely heavily on experimentation to generate data for design and assurance, which can lead to inefficiencies and increased patient risk. Weimer and his team plan to use the more than $570,000 grant to develop technology that will improve efficiency and reduce the possibility of errors.

The CAREER Award project will leverage Weimer’s expertise—specifically his co-founding of companies Neuralert and Vasowatch—to study the LE-MCPS problem in the context of stroke detection and postpartum hemorrhage prediction.

“We’re addressing foundational technical challenges I’ve identified through my experience in building real-world medical devices,” says Weimer, who is also faculty affiliate of the Vanderbilt Institute for Surgery and Engineering (VISE) and the Institute for Software Integrated Systems (ISIS).

In 2022, Weimer was one of two Vanderbilt faculty who received the distinction of Best Invention by Time magazine editors. He created Neuralert, a lightweight, noninvasive wristband device that automates stroke detection and alerting. Neuralert’s stroke detection device can identify the onset of asymmetric movement in as little as 15 minutes, even if the wearer is asleep. In 2021, the FDA named Neuralert a Breakthrough Device.

In an episode of WebMD’s YOUR HEALTH ON TECH with John Whyte, MD, Steven Messe, MD, discusses an innovation in stroke detection: a wristband that could alert hospital professionals for earlier intervention.

TRANSCRIPT OF VIDEO

JOHN WHYTE
A person has a stroke nearly every 40 seconds in the United States. Despite recent advances in treatment, there are still significant challenges after suffering a stroke. That’s why diagnosis is so important. We tend to think about strokes occurring in the community. But strokes can also occur in the hospital setting. And when they do, they often go undetected for over 4 hours.

My guest today has the device to change that, using a special wristband. It received FDA breakthrough designation last year and was also named one of Time magazine’s 200 Best Inventions of 2022. Joining me is Dr. Steven Messe. He’s a Professor of Neurology at the Hospital the University of Pennsylvania, as well as Associate Director of the Vascular Neurology Fellowship. Steve, it’s nice to see you.

STEVEN MESSE
Thanks, Dr. Whyte. Appreciate the opportunity to speak with you today.

JOHN WHYTE
Well, I have to start off because a lot of viewers are probably thinking strokes occur in the hospital? What’s going on here? Because you’re not talking about people who came in with stroke. Is that correct?

STEVEN MESSE
Absolutely correct. So about 10% overall of all the strokes that are occurring are happening under our noses in the hospital. And the issue is that people come in with diseases or are undergoing procedures that we know are high risk for stroke. And so they definitely can occur while we are taking care of these patients. And in particular, procedures of the heart, cardiovascular procedures, and the aorta, which is the main artery leaving the heart, you can have debris or a clot that can go up into the head and cause a stroke. So it definitely happens all the time.

JOHN WHYTE
But Steve, people are in a hospital setting, where there’s a lot of monitoring and a lot of oversight. So it still seems a bit surprising, doesn’t it, that your data shows it can take over 4 hours in the current environment if someone has a stroke in the hospital to actually be detected? Is that right?

STEVEN MESSE
That is. Yeah, that’s absolutely correct. And there are a number of challenges. Patients are in bed or bedbound, oftentimes sedated, coming out of anesthesia. And when you have a stroke, it’s not like a heart attack, where you have a lot of pain and maybe will wake up and say, I’m in pain. When you have a stroke, pain is not usually involved. But you may not be able to speak. You may not be able to move.

So the patients can’t necessarily tell the staff they’re having trouble. And as we all know, the nurses and doctors are working very hard in the hospital. But they can’t be in the room all the time. And so oftentimes things will go undetected for hours. And with stroke, as you mentioned, we have treatments that are amazing, but time is the most critical factor to getting a good outcome, even to be eligible for treatment. If you’re too late, you may not be eligible to be treated.

JOHN WHYTE
And that’s where diagnosis comes into play. We always talk about recognizing the signs and symptoms, which sometimes can be confusing, and then quick diagnosis. But when we think about the diagnosis, we often think about fancy imaging, right? We’re going to do a CT scan. We’re going to do an MRI, PET scan, and all these other type of scans. But your device involves a wristband. So tell us how you thought about that and how it works.

STEVEN MESSE
Yeah. So we recognized that this was a problem, that we weren’t catching the strokes fast enough. And we wanted to really work on that. And one of the hallmarks of stroke, one of the cardinal signs of stroke is weakness on one side of the body. And so we knew that’s what we wanted to tackle as sort of a low-hanging fruit for stroke detection. And accelerometers are ubiquitous and really sort of basic technology now, things that can measure movements very easily. We wear them on our smartwatches and our Fitbits.

And we knew, again, that asymmetric movement is a hallmark of stroke. So we put accelerometers on our patients’ wrists, patients who had a stroke and patients who didn’t have a stroke so that we could develop an algorithm that can differentiate between the two and detect when somebody has that weakness that’s indicative of stroke, to alert the staff to come check the patient and identify that stroke faster. So this has been going on for– something we’ve been working on for years now. And it took a long time to get the algorithm to work as we need it to because we can’t have it alerting the nurses all the time.

JOHN WHYTE
Sure.

STEVEN MESSE
You don’t want to have a lot of false alarms. But we also want to catch these strokes quickly.

JOHN WHYTE
Is it looking for subtle changes? Because let’s be fair. In the hospital setting, you’re not moving around too much. You’re kind of confined, for the most part, to that hospital bed, which is tight, to be honest, in moving your arms. So is it just looking for these small changes in terms of movement?

STEVEN MESSE
Yeah, that’s a great question, and it was something that we also were concerned about when we were developing this. Are we going to be able to have this work when people are sleeping or not? And in our validation study, we did 200 patients prospectively, patients with stroke or without stroke, it worked equal well whether it was daytime or nighttime. And we were thrilled to see that.

It was engineered to be that way. And when we saw that it actually worked as we hoped it would, it was very exciting for us. So it does pick up subtle findings and subtle differences to make that detection of stroke.

JOHN WHYTE
What about an outpatient setting? I’m sure that’s what everyone’s thinking about right now. How can I have this at home for a loved one? Or what’s its potential usage in a skilled nursing facility or a nursing home? Where do you see that developing? Is it going to be ready for that at some point in time?

STEVEN MESSE
Absolutely. We wanted to start in the hospital because we knew that it was an issue there, and we knew the patients who are at greatest risk for stroke. We knew when they were going to have their stroke. And we knew the nurses would be there to check the patient quickly. And we would use that, then, to help patients, but also to show that it works.

From there, obviously we’ll be able to develop it to go outside of the hospital and, exactly as you said, going to the rehab facilities and the nursing facilities. We know those patients are high risk. And then outside of those facilities and in the home as well, there’s no question that we could adapt this technology for those environments.

JOHN WHYTE
And is it currently being used in the hospital setting or primarily in a research setting right now?

STEVEN MESSE
It’s still in the research setting. We’re not selling it for clinical use yet. We need to do the final pivotal trials to prove that it is useful. But we hope to be initiating those very soon and getting those done as quickly as possible so that we can get them into clinical use.

JOHN WHYTE
And this doesn’t replace the fancy imaging that we’re referencing, correct, in terms of CT or MRI. It’s really about those early signs of diagnosis that then can help determine what level of care. Is that right?

STEVEN MESSE
Yeah, that is a really important point. So this, by no means, is not a diagnostic test. It doesn’t tell you if you had a stroke or not. It just tells you that you are moving in a way that could be indicative of stroke, and you need to have somebody come and check you out quickly. So the nurse would come by the bedside and do a neurologic exam, as they do intermittently throughout the day.

But the key is to do it at the right time. And from there, if they do see signs of a stroke that look concerning, you would then initiate a call to the stroke team in the hospital. And they would come by, and you would get all the studies that you mentioned, the CT scan, perfusion scans, and blood vessel studies, and if a stroke is confirmed, treat them as quickly as possible. But we would be the tip of the spear, then, to try to identify the movement that could be indicative of stroke.

JOHN WHYTE
And on this series Your Health On Tech, we’re talking about how more and more tools are going to be able to be used in the home, right? So we always focused on everything has to be done in the ER setting or the hospital setting. Is this a pattern that you think is going to accelerate in the next few years, partly because we’re more use to be able to do things at home? The same for the wristband, people are more accepting of having something on their wrist that can help impact their health?

STEVEN MESSE
Yeah. I think I think that is a reasonable expectation that we will be getting better and better at managing people’s diseases at home and treating them at home and diagnosing them outside of the hospital. And the goal, again, is to do it faster and better. So I think both people’s understanding and willingness to wear these things that can be so helpful to them is real. And then our ability to use these tools to help people, obviously, is growing rapidly. So it’s a very exciting time.

JOHN WHYTE
What’s stroke detection going to look like five years from now?

STEVEN MESSE
Wow. Five years is a relatively short time frame. But I really do hope that for the people that are at risk of stroke that we will have something that is easily worn and well tolerated that can help them identify the stroke quickly so that we can treat it. Obviously, we are continuing to develop those treatments and ways to prevent stroke as well. So we’re going to work at it from both ends.

JOHN WHYTE
And let’s just remind patients, what are those signs of strokes that they should call 911 if they or a loved one experience?

STEVEN MESSE
Great question. So if you have an unexplained severe headache like you’ve never had before, that can be signs of a stroke. If you develop sudden double vision or loss of vision, difficulty speaking or understanding, room spinning, incoordination, imbalance, numbness, tingling, or weakness, which typically occurs on half of your body, any of those signs could indicate a stroke. And again, stroke is a very time-critical disease. Every minute counts. So if that happens, call 911 is the right answer.

JOHN WHYTE
Good reminder. And if you have any question about it, you call 911. You don’t try to diagnose it yourself. Dr. Messe, thank you very much for taking the time today to talk to us about your technology and where we are with stroke detection.

STEVEN MESSE
Thank you. I appreciate it. Nice to talk you.

View the Published Article

One of the hallmark indicators of stroke is asymmetrical arm movement (e.g., weakness, neglect or less movement on one side). Neuralert’s stroke detection device can identify the onset of asymmetric movement in as little as fifteen minutes, even if the wearer is asleep. The device requires no time-wasting baseline development, significantly speeding stroke detection, assessment, and treatment; which can saves lives, improve outcomes, and help patients to continue to live full and independent lives.

“We are honored that the correspondents and editors at TIME have chosen our stroke detection device for its List of the Best Inventions of 2022,” said Eric Corkhill, CEO, Neuralert Technologies. “Our mission is to end the devastation caused by a stroke that is not detected in time for effective treatment. Neuralert is the first, non-invasive, continuous wearable monitor that automates stroke detection, and will lead to more patients getting proven treatments, faster.”

Of the new list, TIME’s editors write, “The result is a list of 200 groundbreaking inventions (and 50 special mention inventions)—including life-mapping artificial intelligence, diamonds made from excess carbon in the air, and the most powerful telescope ever—that are changing how we live, work, play, and think about what’s possible.”

“We are thrilled that the Neuralert device made it to TIME’s List of the Best Inventions of 2022.” said Steven Messé, MD, Neuralert Co-Founder. “Our device represents a breakthrough in technology that offers significant advantages over current alternatives. Stroke is a terrible disease which causes death and major long-term disability, so being able to detect it at a stage when it is most treatable is incredibly exciting, as there will be a tremendous benefit for both patients and their families.”

One out of four people living today will have a stroke, making it the number one cause of long- term, severe disability and the number five cause of death. Neuralert’s mission is to accelerate the treatment for stroke via faster detection and automated alerting. Recent advances in treatment have led to improved outcomes from stroke–but only if discovered in time. Delays in stroke identification result in lower likelihood of receiving time critical treatment resulting in poorer outcomes and even death. The Neuralert device constantly monitors for stroke and automates an alert resulting in a faster stroke detection process than manual methods used today. It has undergone rigorous studies which have shown that the device is able to detect stroke in as quickly as 15 minutes. Reduction in stroke detection time will save patient lives, improve outcomes, and reduce hospital costs and liability.

See the full list of TIME’s Best Inventions of 2022 at time.com/best-inventions-2022.

About Neuralert Technologies

Founded in 2019 as a spin-out from the University of Pennsylvania, Neuralert Technologies’ mission is to transform the method of monitoring for stroke symptoms in hospitalized patients. By combining lightweight, non-invasive, wearable devices with a patented algorithm, the Neuralert system constantly monitors patients for stroke symptoms and automatically alerts hospital staff when detected. Neuralert’s device speeds treatment, saves lives, and improves outcomes. Neuralert Technologies is headquartered in Philadelphia, Pennsylvania.

Philadelphia, PA – Neuralert Technologies, makers of a lightweight, non-invasive wristband device that automates stroke detection and alerting, has announced that it has achieved Minimum Viable Product (MVP) status. To achieve this milestone, recently developed purpose-built wearable devices are now combined and working in conjunction with its patented algorithm and alerting software, completing the functional arc of stroke detection, data analysis, and medical staff alerting.

One of the hallmark indicators of stroke is asymmetrical arm movement (e.g., weakness, neglect or less movement on one side). Neuralert’s device can detect the onset of asymmetric movement in as little as fifteen minutes even if the wearer is asleep. The stroke detection devices require no time wasting baseline development, greatly speeding stroke detection, assessment, and treatment, which can saves lives, improve outcomes, and helps enable patients to continue to live full and independent lives.

“Our mission is to end the devastation caused by a stroke that is not detected in time for effective treatment,” said Eric Corkhill, Neuralert CEO. “With the delivery of working wristband devices and development of the software that utilizes our stroke detection algorithm, we are extremely proud to have achieved this major milestone. Neuralert will be the first, non-invasive, continuous wearable monitor that automates stroke detection, and will lead to more patients getting proven treatments, faster.”

Neuralert’s mission is to accelerate the treatment for stroke via faster detection and automated alerting. One out of four people living today will have a stroke, making it the No.1 cause of long- term severe disability and the No. 5 cause of death. Recent advances in treatment have led to remarkably improved outcomes from stroke …but only if discovered in time. Delays in stroke identification result in lower likelihood of receiving time critical treatment resulting in worse outcomes and even death. In the hospital, identifying stroke symptoms is challenging, as patients are bed bound and/or sedated, making strokes harder to detect, and nurses are not able to assess patients frequently enough to insure that it is identified in time for treatment. Neuralert constantly monitors for stroke and automates an alert resulting in a faster stroke detection process than manual methods used today. Neuralert has undergone rigorous studies which have shown that the device is able to detect stroke in as quickly as 15 minutes. Reduction in stroke detection time will save patient lives, improve outcomes, and reduce hospital costs and liability.

“We are thrilled that Neuralert has achieved Minimum Viable Product status.” said Steven Messé, MD, Neuralert Co-Founder. “Our device represents a breakthrough in technology that offers significant advantages over current alternatives. Stroke is a terrible disease which causes death and major long-term disability, so being able to detect it at a stage when it is most treatable is incredibly exciting, as there will be a tremendous benefit for both patients and their families.”

About Neuralert Technologies

Founded in 2019 as a spin-out from the University of Pennsylvania, Neuralert Technologies’ mission is to transform the method of monitoring for stroke symptoms in hospitalized patients. By combining lightweight, non-invasive, wearable devices with a patented algorithm, the Neuralert system constantly monitors patients for stroke symptoms and automatically alerts hospital staff when detected. Neuralert’s device speeds treatment, saves lives, and improves outcomes. Neuralert Technologies is headquartered in Philadelphia, Pennsylvania.

Neuralert is a spin-out from the University of Pennsylvania, founded by Steven Messé, MD (Department of Neurology) and James Weimer, PhD (Department of Computer and Information Science). The company’s novel invention addresses the problem of late or undetected stroke in hospitalized patients. The system uses non-invasive, wearable devices which continuously monitor at-risk hospitalized patients for stroke symptoms and automatically alerts clinical staff resulting in more rapid assessment and treatment of the patient. Neuralert’s automated detection process may be up to ten times faster when compared to the current standard of care which relies on manual detection processes. Reduction in stroke detection time will save patient lives, improve outcomes, and reduce hospital cost and liability.

“Currently, hospital stroke detection requires consistent, timely, and effective manual assessment. Unfortunately, the majority of in-hospital strokes are not detected in a timely manner resulting in treatment delays which adversely impact patient outcomes and hospital cost” said Eric Corkhill, CEO of Neuralert. “Neuralert is an automated stroke symptom detection system which constantly monitors the patient for stroke-based asymmetry, a revolutionary change in managing stroke. FDA granting Neuralert Breakthrough Device designation is a major milestone in our efforts to bring this valuable and revolutionary technology to market.”

Learn More

Neuralert is a spin-out from the University of Pennsylvania, founded by Steven Messe, MD (Department of Neurology) and James Weimer, PhD (Department of Computer and Information Science). The company’s novel invention addresses the problem of late or undetected stroke in hospitalized patients. The system uses non-invasive, wearable devices which accurately detect stroke symptoms and then alerts clinical staff to more rapidly initiate stroke mitigating treatments. Neuralert decreases stroke detection time by over 400% compared to the current standard of care saving patient lives, improving outcomes, and reducing hospital costs and liability.

“Currently, hospital stroke detection requires consistent, timely, and effective manual assessment. Unfortunately, the majority of in-hospital strokes are not detected in a timely manner resulting in treatment delays which adversely impacts patient outcomes and hospital cost” said Eric Corkhill. CEO of Neuralert. “Neuralert is an automated stroke detection system which constantly monitors the patient for stroke-based asymmetry, a revolutionary change in managing stroke. Funding from the Science Center’s Launch Lane Accelerator will significantly help us to operationalize the product. In addition to their investment, I look forward to participating programming tailored to our growth needs and access to the Science Center’s vast network of connections and resources.”

Learn More

Neuralert is a spin-out from the University of Pennsylvania, founded by Steven Messe, MD (Department of Neurology) and James Weimer, PhD (Department of Computer and Information Science). The company addresses the problem of late or undetected stroke in hospitalized patients using non-invasive, wearable devices. These devices accurately detect stroke symptoms and alert clinical staff to more rapidly initiate stroke mitigating treatments. Neuralert dramatically reduces the time to detect stroke symptoms, saving patient lives, improving outcomes, and reducing hospital costs and liability.

“Neuralert can help end the devastation caused by in-hospital strokes. Automating detection decreases the current delay between a stroke event and treatment allowing the hospital to utilize proven lifesaving and stroke mitigating treatments” said Eric Corkhill, CEO of Neuralert Technologies. “The $25,000 investment and ongoing support from Buck Built will help us to perform additional testing which will refine our product and position us for near term commercial growth”.

Learn More

Neuralert is a spin-out from the University of Pennsylvania, founded by Steven Messe, MD (Department of Neurology) and James Weimer, PhD (Department of Computer and Information Science). The company focuses on early detection of stroke symptoms in hospitalized patients via smart, non-invasive, wearable devices. Its innovative platform can accurately detect these symptoms and immediately alert clinical staff to assess the patient and initiate stroke mitigating treatments. Neuralert dramatically reduces the time to detect stroke symptoms, saving patient lives, improving outcomes and reducing hospital costs and liability.

“We believe we can end the devastation caused by in-hospital strokes. Using an automated method like Neuralert to monitor post-surgical patients as well as patients with elevated stroke risk, decreases the delay between a stroke event and treatment” said Eric Corkhill, CEO of Neuralert Technologies. “Ben Franklin PRIMA Grant funding will help us to port our technology to a stable, secure infrastructure utilizing Amazon Web Services.

Learn More

Neuralert is a spin-out from the University of Pennsylvania, founded by Steven Messe, MD (Department of Neurology) and James Weimer, PhD (Department of Computer and Information Science). The company focuses on early detection of stroke symptoms in hospitalized patients via smart, non-invasive, wearable devices. Its innovative platform can accurately detect these symptoms and immediately alert clinical staff to assess the patient and initiate stroke mitigating treatments. Neuralert dramatically reduces the time to detect stroke symptoms, saving patient lives, improving outcomes and reducing hospital costs and liability.

As part of the RAPID Accelerator’s Fall 2020 cohort, Neuralert will receive hands-on technical product development support, seed capital, business coaching, and access to manufacturing and prototyping equipment at NextFab.

“We believe we can end the devastation caused by in-hospital strokes. Using an automated method like Neuralert to monitor post-surgical patients as well as patients with elevated stroke risk, decreases the delay between a stroke event and treatment” said Eric Corkhill, CEO of Neuralert Technologies. “NextFab will help us to create purpose-built, wearable devices designed to operate within the complexities of the clinical environment while ensuring high adoption by both patients and staff”.

Learn More