Wei Probes into Spatial Working Memory and Aging Process
When Fangzhou “Vicky” Wei ’20 first came to F&M, she couldn’t have envisioned launching a journey into the mind: examining how aging affects working memory with mathematical models. Spatial working memory allows one to temporarily store recent information for future use. It plays an important role in concentration and in following instructions.
In her study, “Reducing Firing Rates of Inhibitory Neurons in a Model of Age-Related Working Memory Impairment,” Vicky simulated a network of neurons performing the Delayed Response Task, a spatial working memory task administered to rhesus monkeys, to assess the effects of aging on the pyramidal neurons in the dorsolateral prefrontal cortex (dlPFC). To complement scenarios modeled in Professor Christina Weaver’s recent paper (Ibañez et al., Frontiers in Computational Neuroscience 2020), the student researcher used a less excitable activation function for the inhibitory neurons.
“Specifically,” explains Vicky, “we compared how many points maintained tuned to persistent activity to the original stimulus location at 0 (TPA-S), and compared the synaptic weight distributions across a six-dimensional Latin hypercube sample exploration of parameter space, under different conditions.”
“By doing research, students can get a sense of how math can contribute to real-world applications, not just homework questions.”
“I previously had no neuroscience background. This research really gave me a chance to work in a project with a combination of math, neuroscience, computer science and a little bit of statistics,” the student notes. The project equipped Vicky -- and also fellow researcher Alexa Gordon '20 -- with the background to consider options to pursue computational neuroscience in future, and how to apply math in interdisciplinary areas.
“The brain is incredibly complex, with many kinds of neurons interacting with one another. Mathematical modeling is one of many essential tools that neuroscientists use to understand the brain better,” notes Prof. Weaver, Vicky’s research advisor. “Like humans, rhesus monkeys generally experience cognitive decline as they age. We have empirical data about how aging affects some aspects of the prefrontal cortex. Dr. Sara Ibañez and I developed our network model to predict how several of these changes might work together during a working memory task (the DRT).”
The faculty researcher continues, “Mathematical models can always be refined further, and Vicky focused on how inhibitory neurons (called interneurons) functioned in our network model. She learned how to run and make changes to our network model (written in MATLAB), and tested how the network performed when we made the model interneurons behave more realistically. Her results will help my research team build stronger models in the future.”
“I had some moments when I felt like, ‘Oh, I'm stuck in my research. How is it possible to fix this?’ It was very helpful to understand my professor’s way of thinking to address a problem, then attack it again.”
“It wasn't until graduate school that it really "clicked" that researchers and policy makers actually use math modeling all the time, to make decisions or learn new things,” Prof. Weaver reflects. “Engaging students in research lets them see what it's like to get your hands "dirty" with modeling, so they can picture themselves actually doing it in a future career.”
Also, I have an active grant from the National Institutes of Health: a partnership with neuroscientists from Boston University School of Medicine and the Icahn School of Medicine at Mount Sinai (NYC). The grant has a strict timeline and well-defined research aims. Sometimes students are able to run some modeling experiments complementing the main aims of the project, to see whether it's a viable avenue that my research team should spend more time exploring. That was the case for both Vicky and Alexa this year.