Danielle Bassett: Using graph theory to better understand how we think.

When I walked into an office in Hayden Hall last week, I expected to see Danielle Bassett, a systems neuroscientist researcher and professor at the University of Pennsylvania. However, I didn’t expect to see her sitting at a table whose right side was covered in hundreds of small, colorful, lego-like pieces and ball and stick models. Bassett was the youngest recipient of the MacArthur Fellowship in 2014 for her research on network neuroscience. Specifically, she utilizes a fusion of graph theory and statistical mechanics to better understand how different parts of the brain are connected with one another.

 

Bassett studied physics and mathematics at Pennsylvania State University. Towards the end of her undergraduate education, she had the choice of going to graduate school or work in industry. Realizing she was just scratching the surface of her fields of interest, Danielle was eager to learn more.

“I wanted to keep learning. The only way I knew how to do that was to go to graduate school. I guess, at every step of the way, when I had the choice to keep learning or get a regular job, I chose to keep learning. And that’s still what I’m doing.”

With her strong physics and math background, there were a number of different research directions Bassett could take in her pursuit of learning. Among these different research directions, the analysis of networks and complex systems stood out.

“I think I’m fascinated by problems that look complex from the outside, but you can get a very simple intuition about how they work. It’s almost the disconnect between the apparent complexity and the actual simplicity that I find really interesting.”

While this definitely narrowed down the research possibilities to a certain extent, networks can be used to represent virtually any system that contained interactions between individual sub-components. Bassett chose to focus on analyzing networks and systems that model brain functionality.

“I’ve been interested in mental health for a very long time, but I decided not to go directly into medicine. I liked physics and math more. When I was looking at graduate schools, I was wondering if there was a place I could combine the tools and intuitions from physics and mathematics with something that would actually make an impact on someone who has an mental health disorder.”

When trying to represent neuronal networks, Bassett uses a combination of statistical mechanics and physics for a good reason.

“In statistical mechanics, it’s assumed that all bodies interact with each other in a homogenous way. Graph theory enables you to say, ‘hey, these things are not interacting in a homogenous way, they are actually interacting in a really, very specific way, and I’m going to track every single connection.’”

While statistical mechanics may generalize the interactions between individual bodies to the average interaction of the system, graph theory helps specify the exact connections in the system which allows for a better understanding of the underlying patterns in a system.

 

She addressed the other side of her desk, picking up a structure made from interlocking balls and sticks. It contained two complex clusters of balls connected to one another by a single rod.

“Here’s an example of a system that’s not homogeneous. You can really tell they’re different. One cluster has a starlike structure, one has a closed, cluster-like structure. In the cluster-like structure, information will be more densely passed among the individual units. In the starlike structure, information will be passed from these ‘collector-like’ nodes to the center of the star. So the way in which these two parts of the system work and interact with each other depends on the makeup of their individual connections.”

However, while graph theory may help specify these individual theories, it is still possible that the intricacies of complex networks in the brain may be over-simplified. Theoretical models always make assumptions, which can run the risk of obscuring reality. But for now, Bassett’s research has remained undisturbed by these assumptions.

“One of the interesting questions that needs to be asked, is: how far will a simple model take you? It turns out that in my field of research, you can actually get very far in gaining useful intuitions from understanding these simpler patterns of conductivity while ignoring where things sit in the brain, what types of neurons are here, or what types of neurotransmitters are packed there. It’s surprising, but I think that’s why it’s so exciting.”

Bassett’s research has led to a better understanding of a plethora of phenomenon. One such phenomenon relates to how behavior and cognitive abilities in humans change throughout the aging process. Results from Bassett’s research suggest that that the more modular the structure of an individual's neuronal networks, the easier it is for that individual to learn. In other words, it’s beneficial for your brain to consist of several smaller, densely packed ‘cities,’ each connected to each other by a single path. Contrast this structure with one large city, containing lots of entangled paths to between its small subcomponents.

“When two parts of a system are separate, it allows each part of the system to...change independently of other parts. This is a hallmark of an adaptive system. We’ve been finding that people who have this more segregated organization are those that can flexibly arrange their brain conductivity to help them learn better.”

Taking these results a step further, Bassett and her colleagues recently published a paper which demonstrated a new procedure that, using a specific brain scan, allows researchers to predict how much an individual will learn during a 6-week training session focused on picking up a new skill.

 

When Bassett isn’t spending her time trying to understand how the brain works, she teaches students how to better understand the brain from a network science perspective, in her Network Neuroscience class (BE 566). Bassett’s attitude towards teaching this class summarizes her love for research and exploration.

“I really like teaching this Network Neuroscience class because it gives many students their first exposure to research. I don’t think people know what it’s really like until they’ve tried it. It’s sort of like… you can get really obsessed with it. It’s a place where you feel like you’re at the edge of discovery and you’re constantly discovering something that nobody else knew before. It just very exciting. When I can see that in a student… when I see that they’ve caught the bug, it’s very satisfying.”