Our Big Question
We want to know how the brain learns from the past to make predictions about the future. To make real inroads on this big question, we target a special feat at which the brains of humans, mice, and other animals excel:
predicting the sounds of one's own actions.
This ability to predict the acoustic consequences of our movements is vital for learning and maintaining complex behaviors such as speech and, more fundamentally, is foundational to our sense of acoustic self-awareness - anticipating how our actions will influence ourselves, others, and the world around us.
Using rodents as a model organism, we study two distinct and complimentary behavioral paradigms:
natural, ethological behaviors
engineered, lab-based behaviors.
In our ethological experiments, we study unrestrained rodents engaged in natural sound-generating behaviors including walking, foraging, and vocalizing.
In our engineered experiments, we design simple sound-generating behaviors in the lab that retain core computational principles of natural behaviors, but are performed in augmented reality (e.g. operating a sound-producing lever).
In each of these paradigms, we combine rigorous behavioral quantification, high-throughput neurophysiology and circuit perturbations. Studying the brain and behavior in both reduced and unrestrained preparations allows us to gain deep insight into the cellular and circuit mechanisms through which the brain makes predictions and also provides important insights into how these same prediction-related circuits operate in the wild.
Specific Research Projects
Local circuits for building and using internal models
Global networks for making predictions and learning from mistakes
The neurobiology of cautious walking
The social-vocal lives of Mongolian gerbils