AI-powered project tackles epilepsy with tech and teamwork
ASU computer science researchers and business student win funds to create startup that reduces costs and improves treatment options
A toddler undergoes a brain screening test. A student entrepreneur and a team of researchers in the School of Computing and Augmented Intelligence, part of the Ira A. Fulton Schools of Engineering, are creating a startup company that will combine human expertise and artificial intelligence, or AI, to help more than 20 million people globally, including a significant number of children, who have drug-resistant epilepsy. Photo illustration by Erika Gronek/ASU
Epilepsy is a potentially debilitating neurological disorder that affects 50 million people worldwide. It’s characterized by recurrent seizures — brief episodes of involuntary movement that can occasionally involve a loss of consciousness.
While many people respond to epilepsy medication, approximately 20 million do not, facing ongoing seizures that severely affect quality of life and increase health risks. About two million new drug-resistant cases are diagnosed each year. For these patients, surgery is often the most effective treatment, yet it remains vastly underutilized. Less than 1% of eligible patients actually receive it. The affected group includes a significant number of children.
Shraddha Mittal says that entrepreneurs can create new technology with the potential to improve patient health outcomes, reduce costs and generate economic opportunity. As a student seeking her master’s degree in business administration in the W.P. Carey School of Business at Arizona State University, she is focused on meaningful applications of artificial intelligence, or AI, in the medical sector.
“I believe in doing something that makes an impact,” Mittal says. “If we can use AI to make the most viable treatment options more affordable and accessible for drug-resistant epilepsy patients who are desperate to become seizure-free, that would be game-changing for all stakeholders.”
As part of the New Venture Challenge, an application-based, eight-week course from the ASU Center for Entrepreneurship and New Business Design, Mittal received mentorship from a team of faculty members in the School of Computing and Augmented Intelligence, part of the Ira A. Fulton Schools of Engineering at ASU. The team created a proposal combining Mittal’s business acumen with Fulton Schools research. They recently received a $20,000 award from the challenge to spin their proposal into a startup company.
The EpiPrecision AI team poses with their New Venture Challenge award. From left to right: Sandeep Gupta, Fulton Schools professor of computer science and engineering; Shraddha Mittal, W.P. Carey School of Business master’s degree student; and Fulton Schools Associate Research Professor Ayan Banerjee. Photographer: Kelly deVos/ASU
AI-doctor duo can change epilepsy treatment for good
Sandeep Gupta is a professor of computer science and engineering in the Fulton Schools. He heads the IMPACT Lab where his team seeks novel ways to combine AI with expert knowledge to improve medical treatment options for typically understudied groups such as pregnant women and women with cardiovascular conditions.
Working together, Mittal and the IMPACT Lab team created EpiPrecision AI. The project targets a dire global health need: helping the more than 20 million people with epilepsy who do not respond to medications become seizure-free.
Currently, surgery for those with drug-resistant epilepsy is underutilized. Doctors face difficulty and costly challenges in identifying candidates for surgeries. Seizure onset zones must be accurately mapped using often expensive and invasive procedures such as stereotactic electric encephalography. Then, brain data must be manually analyzed and the appropriate surgical procedure must be selected.
EpiPrecision AI addresses these challenges head-on. By integrating medical data, including MRI, PET scans and scalp EEGs with a sophisticated AI model informed by expert knowledge, the platform enhances both the speed and accuracy of seizure mapping. This collaboration between machine learning and human specialists reduces mapping time from approximately 800 hours to just 160 hours, allowing more timely and precise surgeries.
The EpiPrecision AI team tapped expertise from a children’s hospital to develop their hybrid approach. Research involving 352 patients ranging from infants to seniors demonstrated that AI alone could map seizure onset zones with 65% accuracy. However, when combined with doctor expertise, that accuracy jumped to 89%. This collaborative model outperforms either method alone, showcasing the power of augmented intelligence in medical diagnostics.
Ayan Banerjee, an associate research professor in the IMPACT Lab, says this intersection of medical expertise and AI is a key component of innovation.
“It’s essential that these assistive systems are trustworthy,” Banerjee says. “A good combination of medical learning and machine learning will create reliable new technology.”
The team believes their project has the potential to deliver huge benefits to both hospitals and patients. EpiPrecision AI seeks to raise epilepsy surgery success rates to nearly 100%, dramatically improving patient outcomes.
The researchers project hospitals using EpiPrecision AI could see up to a 40% reduction in treatment costs and a threefold increase in the number of surgeries performed annually, boosting cost efficiency and revenue.
Finally, by streamlining the diagnostic process and reducing costs, the technology has the potential to make life-saving surgery accessible to more patients, including those in underserved regions where specialized expertise is scarce.
Turning tomorrow’s tech into bold business ventures
Gupta says he got involved with the New Venture Challenge because investor funds are critically needed to drive the next wave of medical innovation.
“Investor funding will be an essential part of these advancements,” Gupta says. “As part of that, it’s very important for our research team to develop relationships with entrepreneurs and ensure that our students have an entrepreneurial spark.”
Gupta notes that AI-powered medicine is a large and fast-growing field. EpiPrecision AI is positioned at the intersection of the $613 billion health AI market and the $26 billion medical imaging AI sector. Digital health technologies in neurology alone are expected to reach $303 billion by 2034. This convergence of need, technology and opportunity underscores the project’s commercial and social promise.
Now, Mittal will spearhead efforts to use the ASU funds to explore startup company options and pitch the project to investors. She says that the New Venture Challenge has been an essential element of preparing her for the job.
“It’s crucial that we learn to speak the investor’s language,” Mittal says. “We are very passionate about this research, but we have to make sure that message gets through to the people who can help us bring this technology to market.”
EpiPrecision AI exemplifies how thoughtfully applied artificial intelligence can transcend the limits of traditional medicine. By merging expert experience with machine precision, the platform not only offers a beacon of hope to millions suffering from drug-resistant epilepsy but also lays a blueprint for future AI-driven interventions in neurology and beyond.
