Students built a 'virtual neurologist' tool

A stroke cuts off blood flow to the brain, and doctors race to tell whether a clot or a bleed is causing it. Student interns at xSpring in Napa, California, helped build an artificial-intelligence prototype meant to guide that first assessment faster. (cdc.gov, govtech.com) The nonprofit’s project is called the “Virtual Neurologist,” and NeuroSpring, the group’s earlier name, describes it as an artificial-intelligence medical device for neurological emergencies. GovTech reported on April 10, 2026 that interns helped develop the tool as part of xSpring’s research program. (idealist.org, govtech.com) The basic problem is speed: the Centers for Disease Control and Prevention says stroke treatment begins when emergency medical services arrive, and the agency tells people to call 9-1-1 at the first sign. The American Heart Association said in January 2026 that updated ischemic-stroke guidance puts added weight on rapid diagnosis, imaging and coordinated emergency care. (cdc.gov, heart.org) Artificial intelligence in this setting works like a pattern-finding assistant: the American Academy of Neurology says machine learning can spot signals in clinical data, while deep learning can analyze complex inputs such as images, speech and electronic records. A 2025 review in *Current Neurology and Neuroscience Reports* said stroke tools already help with imaging alerts, workflow orchestration and outcome prediction. (aan.com, pubmed.ncbi.nlm.nih.gov) That same review said commercial stroke systems have reported double-digit reductions in door-to-needle times, the hospital interval between arrival and clot-busting treatment. It also said wider adoption still runs into dataset bias, opaque reasoning and limited external validation. (pubmed.ncbi.nlm.nih.gov) xSpring’s pitch is access as much as speed. NeuroSpring’s public description says the device is being developed to diagnose neurological emergencies without requiring a neurologist to be physically involved at the point of care. (idealist.org) That matters in places that do not have round-the-clock stroke specialists on site. National Institutes of Health StrokeNet says the United States still relies on a network of 27 regional centers and about 500 hospitals to run stroke trials and build treatment infrastructure, a sign of how dependent care remains on specialized systems. (nihstrokenet.org) The internship angle is central to the story. GovTech reported that xSpring grew out of NeuroSpring, a Napa-based 501(c)(3), and expanded its internship pipeline through ties with the Napa Valley Unified School District and several local high schools, including New Technology High School, Justin-Siena, Vintage and American Canyon. (govtech.com) The same report said Grace Del Carmen, an exchange student from Nicaragua who later enrolled at Northeastern University, became the program’s first intern after connecting through Queen of the Valley Medical Center and neurologist Mark Borsody. Borsody also served as NeuroSpring’s senior scientist, according to the report. (govtech.com) The tool is still presented publicly as a prototype rather than a deployed standard of care. The larger field is moving fast, but the medical literature and neurology groups are still asking for prospective trials, explainable systems and tighter validation before artificial intelligence can be trusted to make high-stakes stroke decisions at scale. (pubmed.ncbi.nlm.nih.gov, aan.com)