Research
Circuitry Underlying Decision-Making
The Constantinople lab is interested in the mechanisms by which multi-regional neural circuits implement algorithmic computations for value-based decision-making. We use high-throughput training of rats to combine sophisticated behavioral paradigms with large-scale electrophysiology, cell-type specific recordings, measurements of neuromodulators, and circuit perturbation techniques including pharmacology and optogenetics. The high-throughput approach enables statistically powerful and rigorous characterization of behavior, reduces the effort cost of behavioral training, and facilitates the study of complex behaviors that have long training times. Well-validated behavioral models are used to generate hypotheses about the algorithmic consequences of neural perturbations, and specify the computations that might be implemented by neural circuits.
Most of the lab studies a temporal wagering task that we developed. In the task, rats reveal how much they value different amounts of water rewards based on how long they are willing to wait for them. The task has a hidden block structure where different reward blocks offer different distributions of rewards. Well-trained rats infer the reward block when deciding how long to wait. Our goal is a brain-wide mapping of how neural circuits implement computations that rats use to solve this task, including inferring the block, updating beliefs about the block, and deciding how long to wait.