Our window into the brain has grown exponentially over the last few decades. Researchers can now record electrical signals from thousands of neurons at once, gather fMRI data twice as fast as they could just five or ten years ago, and incorporate behavioral data into their analyses more easily than ever before – all of which creates a major data processing challenge for anyone trying to make sense of it all.

At the same time, there’s been growing interest in the brain from all corners of science. Once the domain of biology, medicine, and psychology, neuroscience today incorporates engineering, physics, statistics, genetics, and even ethics and education.

The new Stanford Center for Neural Data Science aims to take advantage of the moment. By bringing Stanford’s community of scholars closer together, its leaders hope to tackle hard problems in neuroscience and advance the burgeoning field of data science at the same time.

The center – playfully dubbed “The Center for NeurDS” – is led by Wu Tsai Neurosciences Institute faculty scholar and Stanford Data Science faculty member Laura Gwilliams, an assistant professor of psychology in the School of Humanities and Sciences, and Wu Tsai Neurosciences Institute faculty scholar Scott Linderman, an assistant professor of statistics also in H&S. As the pair prepare to host the center’s inaugural symposium on Oct. 22, Gwilliams and Linderman spoke about how the center came together and what they hope to accomplish.

Stanford affiliates can register for the symposium here.

How did this center get started?

Scott Linderman: Over the last few years, Stanford Data Science has been doing a phenomenal job of creating these faculty-led centers that are explicitly about building communities of scholars that go beyond departmental boundaries. And a couple of years back, the leadership of SDS and of Wu Tsai came to us and said, “We’d like to develop a center for neural data science. This seems like a natural application domain for data science.”

Building such a community here at Stanford would first of all advance neuroscience, but also surely point to new directions for data science research as well. Neuroscience is full of challenging data scientific questions, and in many cases, the tools to answer those questions don’t yet exist. Hopefully, developing a community like this will lead to new research to address both of those issues.

One of the things that makes this a good moment to start a center is this explosion in neuroscience data. Can you tell me what that looks like in practical terms?

Linderman: One measure that I like to point to is the number of single neurons that we can record from simultaneously, which has been growing exponentially. Labs around Stanford now are putting tiny electrical probes into the brains of model organisms like mice and recording thousands of cells at a time. That gives us just a remarkable ability to look inside the brain and measure activity across time and across brain regions. Hopefully, that will give us new insights into how patterns of activity across the whole brain are contributing to the sophisticated computations that go on in the brain.

Laura Gwilliams: You can also look at non-invasive neuroscience with humans. FMRI technology, for example, is advancing in a way that gives you much higher resolution images, which translates into orders of magnitude more data per scan. That allows you to more precisely say where the brain is active, for example.

So I think that across the board, at different scales of neuroscience from the whole brain level all the way down to single cells, we have this trend that as the hardware advances, it is possible to record larger and higher-fidelity data as a result.

And it’s not just more of the same data, right?

Linderman: It’s not. It’s more and different. There’s a lot of interest, for example, in how we relate those measurements of neural activity to measures of behavior, because one of the fundamental questions in neuroscience is how does the collective activity of large populations of cells give rise to complex behavior? And so that adds a whole other dimension to this question in terms of the types of data that we’re working on analyzing. We think the center could really help us to develop better tools for that.

How will the center develop those kinds of tools?

Gwilliams: I think that our role here is not to put forward, “We want to develop X method,” but more so to bring together all the people in this amazing community we already have – many of whom are already developing really powerful methods – so that they can learn from each other and work together. It’s possible that new methods naturally emerge from that, but our goal really is to facilitate the connections and collaborations that make the most sense given the expertise and shared interests that we already have here at Stanford. We want to make it so that it’s a very low bar to entry for people crossing paths and discovering what their shared interests are and what those possible entry points for collaborations could be.

What are some of the specific things you’re planning?

Linderman: One of the things we’re working on is an internal seminar series where we’ll hear from scientists on campus. We’re envisioning giving senior graduate students and postdocs an opportunity to give talks to the community here about their work and their ideas. That may very naturally lead to new collaborations from people across campus who may not have heard about it otherwise. So that’s the internal focus.

To complement that, we’re also planning a reading group or a journal club where faculty can learn about ongoing methodological and scientific work from outside of Stanford in a community of like-minded scholars and really have deep discussions about the state of the art and research in this field.

I’ve heard the word “community” a lot. Who do you want to be part of your community?

Gwilliams: There are scientists in the School of Medicine, statistics, psychology, engineering, and numerous others – spread across the whole of campus – whose work sits solidly within the neural data science community.

By virtue of the scale and heterogeneity of the data, the scientific questions of interest often lead to challenges that go beyond what you can easily do with your existing toolkit. And one of the big goals for building this community is to surface more of those challenges and to bring together people from across disciplines who are looking for exactly that type of challenge and are interested in building methods to tackle them.

You’re about to host the Center’s inaugural symposium on Oct. 22. What are you looking forward to the most?

Gwilliams: We have four fantastic keynote speakers confirmed, spanning different aspects of and approaches to neural data science. We have Kalanit Grill-Spector, who studies the human visual system using a range of computational techniques, Andreas Tolias, who is collecting large-scale neural recordings in animal models in order to build foundation models of the brain.

Linderman: We’ve also got E.J. Chichilnisky, who’s working on neural data science with a very translational focus [as part of the Wu Tsai Neuro-supported Artificial Retina Project].

Gwilliams: And Anne Brunet, who studies the molecular mechanisms of longevity and aging. So we have a lot of breadth in the different topics that we’ll be hearing about, and that was very much by design. We want this center to span as many different topics in neuroscience where data science can be applicable.

Linderman: Transformative even. I think all of these applications are ones where advances on the data science and methodological side have really enabled these types of work. The keynote speakers are representative of that.

And then, of course, the center is about building this community here at Stanford. And so intermixed with these talks, there’ll be plenty of opportunities to meet other people who are interested in this subject. This is really the kickoff.

What else are you excited about?

Gwilliams: One thing that we’re both excited about is connecting with the other Stanford Data Science centers. On the surface, the astrophysicists, for example, are doing something totally different from what the neuroscientists are doing, but actually, we’re both working with high-dimensional spatial and time-series data. So perhaps there are methods or approaches that we have in common that we haven’t discovered yet. And through collaborations, this could be a real opportunity to discover something that we wouldnt otherwise.

Linderman: I think that one of the big goals for building this center is to surface more of those challenges – and to bring together people who are looking for exactly that type of challenge and are interested in building methods to tackle them.