The scholars meet repeatedly each week with members of the host lab, with Ritt and their cohort in program meetings, and individually with Ritt, who provides technical guidance. Each participant’s role is more like a consultant than a laboratory assistant, Ritt said, because they offer ideas for the team to explore solutions.

Projects from 2025 involved developing deep learning pipelines to identify and score abnormalities in lung cancer tissue and computationally merging different data snapshots of individual cells to map gene regulatory networks that guide axon development in the spinal cord.

This year, Yusef Lateef worked with two labs to use machine learning and survey results to predict elevated suicide risk, and Aida Abkenova used a specialized type of AI to analyze brainwave data and discover equations that govern the dynamics of electrical fields in the brain. Justin Park worked with three collaborating labs to develop software tools for novel fluorescence neuroimaging analysis.

Bugshan, who is concentrating in computational neuroscience, was excited about the opportunity to build her quantitative skills while working with Song’s lab. The challenge involved making sense of pupil-size recordings, which mix responses to light, mental effort, attention and memory. After discussing ideas with Ritt, Bugshan suggested using the data to train what is known as a foundation model, which, according to her research, was a new approach for measuring pupil size and reactivity. Bugshan focused the project on whether the model could learn general features of pupil dynamics that carry over to different cognitive tasks.

Bugshan found it interesting to learn how members of the lab focused on different aspects of the process, especially when things went in an unexpected direction, and how Ritt considered each roadblock a valuable learning experience.

“My experience this summer has reminded me again and again of how collaborative everybody here is at Brown,” she said.

At the end of the program, quantitative scholars share their findings in a poster presentation to their host lab, and most choose to present at Brown’s Summer Research Symposium. The scholars practiced their presentations with Ritt, and Bugshan admitted that experience was more nerve-wracking than delivering it to an audience.

“Dr. Ritt really pushes us to think through different strategies and then fully explain our thought processes,” Bugshan said. “By the time we got to the poster presentation, I felt confident because I was sure nobody was going to grill me like he did.”

Learning to love scientific process

Even when the presentation was finished, the project, by its nature, was not.

“I have to keep reminding myself that there really is no stopping point, because that’s not how research works,” Bugshan said. “I learned to really love that process. I'm very happy and grateful to everybody who's been involved with my project, and I think it has exposed me to so much of how problem-solving works. It’s also been very fun to be part of a lab.”

Bugshan has turned the assignment into an independent neuroscience project and has added applied mathematics as a second concentration. She’s now digging into two open questions: how much analysis the pretrained model can offer from pupil data alone, and whether it captures anything that traditional pupil-analysis methods can’t. The abstract from her poster presentation was accepted by the New England Psychological Association, and she’s excited to present this work at the organization’s conference in November.

Before the start of the fall semester, Song and members of her lab met with Ritt and Bugshan to talk about what could be helpful to move the project forward.

“This has really expanded the horizon for me, my graduate students and postdocs, because we never thought about this type of approach,” Song said. “Thanks to our collaboration with Haya and Jason as part of the Quantitative Scholars Program, I see more applications for that kind of thinking with our research.”