Skip to main content
eScholarship
Open Access Publications from the University of California

UC San Diego

UC San Diego Electronic Theses and Dissertations bannerUC San Diego

Elucidating the role of retrosplenial cortex in history-based decision-making

Abstract

Using past experience to inform future choice is fundamental to decision-making and behavior. Integrating experience by comparing choices and outcomes across time is a critical task, and requires a neural mechanism by which information may be assessed and accumulated. This is a widespread phenomenon in the brain, involving many cortical and subcortical structures. In this dissertation I show that one cortical area, the retrosplenial cortex (RSC), is particularly enriched in neurons that encode behaviorally-relevant history information, and is necessary for decision-making that relies on reward-history. RSC neurons exhibit a diversity of time-constants over which this history information is integrated, and I have found that the timescales encoded in RSC match the temporal characteristics of the behavior better than other cortical areas. I developed a novel behavioral model in which decision is reached as the weighted sum of multiple exponential integrators using the observed diversity of time-constants. Acutely inactivating RSC results in the attenuation of this combinatorial behavioral strategy, and a decreased reliance on reward-history. From these results, I propose a conceptual model where reward-history information is encoded in neurons with a simple update rule, but the time-constants are heterogenous and vary across the population. The combination of diverse temporal information produces a behavioral strategy which is sensitive to both recent experience and long-term trends, a feature observed as the hyperbolic discounting of past experience. In Chapter 1 I introduce the concepts of reinforcement learning theory relevant for this dissertation, and survey how reward-based value information is encoded in the brain. Chapter 2 identifies RSC as particularly important for the integration of past reward experience into actionable value information. Chapter 3 further examines and tests the role of RSC in integrating information across a diversity of timescales, and proposes a model of independent temporal integration in the brain that underlies the hyperbolic discounting of past experience. Chapter 4 discusses the properties of RSC that support the integration and maintenance of diverse information, and contextualizes the results in the broader decision-making context

Main Content
For improved accessibility of PDF content, download the file to your device.
Current View