Skip to main content

Faculty Profile: Ann Kennedy, PhD, assistant professor of Physiology

Ann Kennedy, PhD, is an assistant professor of Physiology. She is a theoretical neuroscientist who uses dynamical systems, statistical modeling and machine learning to study the structure of animal behavior and the neural mechanisms of behavior control. The goal of her research is to develop new theories and models to better understand how neural structure controls function and shapes behavior across the animal kingdom, as well as to improve the understanding of how neural circuit function is altered in various neurological disorders.  

What are your research interests? 

We’re interested in how different parts of the brain interact to shape our behavior, and how the structure of neural networks is related to their computational function. Our lab develops machine learning tools for automated tracking and analysis of animal behavior, and builds models to relate that behavior to activity in the brain. We collaborate with labs recording from brain areas involved in behavior and motor control, including the hypothalamus, basal ganglia, cerebellum and motor cortex, to help them identify structure in their high-dimensional neural recordings and determine how that structure relates to what those brain areas are computing.

What is the ultimate goal of your research?

Behavioral control is a massively distributed thing — hundreds of thousands of neurons across dozens of interconnected brain areas all contribute their own small piece to steering us through the day. Our brains form representations of our motivations — to eat, interact, escape predators — and somehow are able to balance all of these competing drives in a way that is robust and highly flexible. We want to understand what computational principles these massive, distributed networks are applying that allows them to be so successful. We also hope that by better understanding the computational mechanisms at work in these circuits, we’ll be able to shed light onto how their function is altered in neurological disorders such as autism, PTSD and Parkinson’s.

How did you become interested in this area of research?

I learned to write code at a young age (my mother was an assembly coder who developed operating systems back in the 70s) and I’ve always been interested in the question of what makes certain problems easy for us but hard for computers. During my PhD, I studied how the structure of the neural microcircuit in cerebellum determines how that system can learn and how learning generalizes. From there, I pursued a postdoc with Dr. David Anderson’s group at Caltech to study the control of more naturalistic, innate behaviors by nuclei of the hypothalamus. While we have a decent handle on the basic computational principles at work in cerebellum, making sense of bigger structures like the densely interconnected nuclei of hypothalamus is much tougher. Technologies for recording from behaving brains have evolved a lot in the past decade, which means now there is a tremendous demand for models and theoretical frameworks that can incorporate this new kind of data.

What types of collaborations are you engaged in across campus (and beyond)?

Collaboration is very important to our group, and we work broadly with experimental labs to study complex behavior and its control by the brain. At Northwestern, we are collaborating now with the labs of Jim Surmeier, Jones Parker, and Greg Schwartz to develop improved computer vision systems for automated tracking and behavior recognition in mice. We are also working with Lee Miller to study the structure of complex motor behaviors. Beyond Northwestern, one new project I’m excited about is a team effort with Drs. Weizhe Hong, Zoe Donaldson, Michael Yartsev, Peyman Golshani and Daniel Aharoni to study the formation and structure of social memories across different species, including mice, prairie voles and Egyptian fruit bats.

How is your research funded?

We are currently funded by NIH grants, including an National Institute of Mental Health Pathway to Independence (K99/R00) award, and a U01 from the National Institute of Neurological Disorders and Stroke. Our research is also supported by startup funds from Feinberg’s Department of Physiology.

Where has your work been published?

Our lab is still getting off the ground, but towards the end of my postdoc I published papers in Nature, Nature Neuroscience, Neuron and through the CVPR machine learning conference.