By the Numb3rs Spring 2020 - Faculty

Faculty

Farewell Brent

By Jonathan Rubin

Since he joined the Department of Mathematics as an Assistant Professor in 2007, Brent Doiron has been a massive figure in theoretical neuroscience at Pitt, both literally – with the 6’4’’ frame of a former Olympic-caliber fighter – and figuratively, through his research, leadership, and mentorship contributions. We are saddened to report that this summer, Brent will leave Pitt to become the director of a new center for theoretical neuroscience at the University of Chicago, and his passion for neuroscience, collegiality, and sense of humor will be deeply missed.

Brent is a world-renowned computational neuroscientist, combining a detailed knowledge of experimental neuroscience with a full command of the tools of applied mathematics and statistical physics. He and his group have published ground-breaking work in the most high-profile journals in neuroscience detailing their pioneering contributions to our understanding of the mechanisms of neuronal variability and their consequences for information processing in the brain.

Brent’s time at Pitt has featured a parade of awards including a Sloan Fellowship (2009), the Chancellor’s Distinguished Research Award (2012), and a Vannevar Bush Faculty Fellowship (2017). He is currently a co-PI on several large-scale collaborative research grants from the NIH and the Simons Foundation to investigate cortical coding and dynamics. Brent has a stellar record as a student and postdoctoral mentor, with former trainees in tenure-track positions at institutions such as Columbia University, U. Notre Dame, and even Pitt, and his grants include support for a large group of postdocs and graduate students who will be joining Brent at Chicago. In this vein, Brent has also served for many years as co-Director of the graduate Program in Neural Computation, run jointly by Pitt and CMU, helping to secure funding for and recruit a steady stream of top-notch students working at both institutions.

Within our own department, Brent has made major contributions to our computational neuroscience courses and has served on numerous thesis committees. He has been a central member of our mathematical biology group and our Applied Math seminar. Brent is a gifted speaker, with a talent for clearly presenting complex ideas and for infusing humor into his presentations. He is also a dedicated colleague and team player, always ready to help out and to pull his weight when service is required – and to lead everyone over to the bar when it’s time to relax or celebrate.

So, let us (figuratively, for the moment) raise a toast to Brent, to thank him for his contributions to Pitt and to wish him well as he moves on to his new home.

Featured Research

Bard Ermentrout - Spatiotemporal Dynamics in Nonlocally Connected Networks

Ermentrout received a new NSF award in April 2020 to pursue this project

Ongoing activity in the nervous system and how it impacts sensory inputs is the subject of much recent experimental work in neuroscience. In particular, it is clear that the intrinsic interactions between neuronal circuits in absence of inputs can have a strong impact on how the system responds to incoming stimuli even at the large-scale cognitive level. Recordings of brain activity in humans and animals reveal that much of this activity is organized into waves.  Using a combination of techniques from applied mathematics, simulations of neuronal networks, and analysis of cortical data, Ermentrout will attempt to model these dynamics and provide some interpretation of how the waves might be useful in cognition.

Michael Neilan - Practical and Geometric Advancements in Divergence-Free Approximations for Incompressible Flow

Neilan recently received a new NSF award to pursue this project

Accurate computations of fluid flow models have a direct impact on the simulation and prediction of various applications, e.g., weather and climate, aircraft design, etc.  Typically, such computations are based on discretizations of partial differential equations which model the physics of the underlying system.  These approximations introduce errors into the simulations which need to be rigorously quantified in order to make reliable simulations and predictions.  This project will construct accurate, structure-preserving computational schemes that exactly enforce the underlying physical laws at the discrete level in fluid flow models.  This attribute leads to high fidelity schemes that are robust with respect to several model parameters.  A particular focus of the project is to develop and analyze computational algorithms that reduce geometric error and thus yield provably accurate answers.  In addition, the project aims to provide user-friendly methods, in particular, the project will incorporate modules into current computational and open source software to improve usability, portability, and outreach.