NeuroChron converts brain MRIs into predictive biomarkers by estimating brain age and identifying patterns of accelerated aging
Memory loss and cognitive decline appear after years of gradual change in the brain; at that stage, treatment can slow progression, but it cannot reverse the damage or halt ongoing decline. The challenge facing clinicians and researchers is as much about timing as treatment: how to identify risk early enough to make a significant intervention.
That’s the shared mission of Gerson Estrada, Nicholas Kim, Nicole Popenko and Ishaani Pradeep – a group of seniors at USC Viterbi’s Alfred E. Mann Department of Biomedical Engineering (BME). When the students were tasked with designing and building a biomedical prototype for their senior design project, they used the opportunity to develop a technology that could improve early detection of neurodegenerative disease – a shared interest driven both by research and personal experience with family members.
The outcome was NeuroChron, a platform that converts routine brain MRIs into predictive biomarkers by estimating brain age and identifying patterns of accelerated aging. As the prototype developed, the team soon recognized its broader clinical and commercial potential. After connecting with Reem Khan, a student in the arts, technology, and business of innovation undergraduate program at the USC Iovine and Young Academy (IYA), they decided to take the leap and build a company.
AI on the brain
L-R: NeuroChron team members: Reem Khan, Gerson Estrada, Nicole Popenko, Nicholas Kim, and Ishaani Pradeep
The concept for NeuroChron builds on a well-established relationship: when the morphology and functioning of the brain indicates an age biologically older than a person’s chronological age, the risk of neurodegenerative disease increases.
“There have been several studies linking increased divergence in brain age to much higher rates of Alzheimer’s Disease and dementia,” said Nicholas Kim, who led model integration for NeuroChron.
However, the structural changes that indicate accelerated brain aging are subtle, distributed and not easily isolated through conventional analysis – at least when relying on the all-too-human brain of a clinician. In the age of AI, that needn’t be the case. “These deep learning models can process vast amounts of data and pay attention to details people wouldn’t be able to see, which allows for much more predictive power and earlier detection,” said Kim.
NeuroChron knowhow
Rather than building a wholly new predictive model, the team focused on making existing models more intuitively usable.
Picture this: a clinician uploads an MRI scan and enters basic information. The system preprocesses the image, stripping away non-brain tissue and aligning it to a common brain atlas before passing it through a trained model. The output is a structured report: predicted brain age, the difference from chronological age, and a visual map indicating which regions contributed most to the result.
“NeuroChron assesses your predicted brain age, the difference between that and your actual age, and gives a surface-level clinical interpretation along with a saliency map shows how much each region of your MRI contributed to the prediction,” said Kim.
From senior design project to startup
From the beginning, the team operated like a super-lean startup. Estrada led deployment of the platform to production, backend and frontend development, and managed the shared codebase. Kim led preprocessing and model integration and collaborated with Estrada on deployment. Popenko worked on initial system conceptualization, data acquisition and conducted physician-guided validation to ensure the system was accurate, interpretable and aligned with clinical workflows.
Pradeep worked on project strategy, presentation development, and stakeholder-facing communication, ensuring the system was clearly positioned for clinical and translational impact. Khan introduced a focus on market definition, pitching, and long-term viability. Her entrepreneurial perspective, trained at IYA, shifted the project from a technical exercise toward something that could navigate the biomedical business landscape.
“NeuroChron is currently in the early-stage startup phase,” said Pradeep. “We have developed a working MVP that transforms standard brain MRI into predictive brain aging metrics, with a focus on clinical usability and explainability.”
The team is now exploring next steps following graduation, including formal incorporation, partnerships with clinical and research institutions, and further validation of their models to predict early progression; the starting point for taking action with preventative approaches and earlier-stage intervention.
By synthesizing routine imaging into a measurable trajectory, NeuroChron reframes decline as process that can be paced – rather than an endgame. As such, the prototype is a paradigm shift as much as a product; rethinking the timeline for intervention by changing how risk is assessed and the extent to which clinicians –and patients – have the agency to heal.
Published on May 8th, 2026
Last updated on May 11th, 2026