CMU and Pitt develop AI tools to target brain cells more precisely

Aiming for more precise brain research

The work is being carried out by researchers from CMU's Kenneth C. Griffin School of Computer Science and the University of Pittsburgh, with support from a multiyear grant through the National Institutes of Health's Armamentarium for Precision Brain Cell Access initiative. The initiative funds research into molecular genetic reagents that can target specific neural circuits.

Researchers are seeking tools that can identify and manipulate particular types of brain cells, rather than affecting broad systems. Andreas Pfenning, an associate professor in CMU's Ray and Stephanie Lane Computational Biology Department, said that the brain contains many cell subtypes and that treatments can affect several systems at once.

“The main problem in studying the brain is its complexity,” Pfenning said, according to CMU. He said that modulating dopamine, for example, can influence systems involved in more than mood and lead to side effects.

From brain maps to usable molecular tools

Scientists can now identify many distinct types of brain cells, but the researchers say reliable ways to reach and manipulate most of them are still lacking. Existing targeting sequences can be slow and costly to test one at a time. They may also work differently in different brain regions or experimental conditions, activate the wrong cells, or reach too few of the intended cells.

Data, protocols and materials linked to successful tools are also often spread across scientific publications and laboratories. According to CMU, that makes findings harder to reproduce or apply, widening the gap between mapping the brain's cellular diversity and understanding how individual cells and circuits contribute to health and disease.

The funded research aims to create a scalable pipeline rather than a single invention. It will combine AI guided design, multiplexed screening, spatial measurement, functional validation and open distribution. The intended result is a practical toolkit for investigating brain circuits and their roles in disease.

AI design will be tested through experiments

Pfenning's lab will combine AI methods with newly developed wet lab genomic technologies to create the molecular tools. Research partners including William Stauffer, an assistant professor in neurobiology at Pitt, will validate the tools and develop them for laboratory use.

Instead of relying on slow trial and error, the team plans to use AI to design customized molecular tools, including specialized nucleic acids, for specific brain cells. Technology developed at CMU allows dozens of designs to be tested simultaneously, which researchers say will help them refine the tools and share them with the scientific community.

The work is still in its early stages. Researchers have begun training AI models while growing the cells needed for their first experiments. The process is intended to create an iterative loop in which experimental data improves computational models, which then optimize tools for particular cell types.

If the approach works, it could improve basic understanding of how different brain cells function and potentially support more targeted gene therapies for rare diseases. CMU said the project could also contribute to treatments for complex disorders such as depression, although the researchers are still developing and validating the underlying tools.