Hi! I am a bioengineer interested in how populations of neurons interact to learn a new motor skill and allow us to navigate the world.
I am currently a postdoc at the University of Colorado with Abby Person and Diego Restrepo studying what computational logic the early cerebellum engages to facilitate endpoint reach precision.
Prior to this, I received my PhD at Carnegie Mellon University, where I worked with Aryn Gittis and Steve Chase studying movement behavior and kinematics to disentangle the circuit basis of motor skill learning.
Summary: Here, we examined how downstream pallidal targets of indirect pathway spiny neurons mediate motor and reinforcement. Surprisingly, our results suggest that the indirect pathway plays a more prominent role in negative reinforcement than in motor control.
Summary: We developed a behavioral task to track body and paw kinematics in mice as they learn to stay atop an accelerating wheel. We found that learning was accompanied by stereotypes progressions of paw kinematics that correlated with early, intermediate, and late states of performance.
Summary: Here, we trained subjects on a task then asked them to come back the following day. After a 24-hour break, subjects recalled very little of what they learned before despite being given the same instructions and being in the same environment. Astonishingly, after only experiencing a single error like they saw before, we observed near total recall of the learned task.
Summary: High Fat Diet (HFD) did not disrupt the regulation of total daily caloric intake, even when up to 90% of total calories came from the HFD. However, HFDs increased daily caloric intake when provided ad libitum and were readily consumed by mice outside of their normal feeding cycle. Ad libitum HFDs appear to induce overconsumption beyond the mechanisms that regulate daily caloric intake.
Summary: Both simulation and behavioral results show that velocity-based learning decays at a slower rate than position-based learning, even when learning is significantly biased towards the latter at the end of training. Collectively, these results suggest that motion-state learning based on movement velocity is more stable than that based on limb position.
Summary: The Feeding Experimentation Device (FED) is a low-cost, open-source, home cage-compatible feeding system that automatically quantifies feeding with high accuracy and temporal resolution. Learn more about FED and stay up-to-date here!