Adaptive interactions between memory systems for structure learning Dhairyya Singh

Format:

Book

Thesis/Dissertation

Author/Creator:

Singh, Dhairyya, author.

Contributor:

University of Pennsylvania. Psychology., degree granting institution.

Language:

English

Subjects (All):

Psychology.

Neurosciences.

Cognitive psychology.

0621.

0317.

0800.

0633.

Local Subjects:

Psychology.

Neurosciences.

Cognitive psychology.

0621.

0317.

0800.

0633.

Genre:

Academic theses

Physical Description:

1 online resource (255 pages)

Contained In:

Dissertations Abstracts International 87-12B

Place of Publication:

Ann Arbor : ProQuest Dissertations and Theses, 2026

Language Note:

English

Summary:

How does the brain extract the structure of the world from ongoing experience, and how do memory systems interact to support this ability? Discovering shared structure requires integrating across experiences, placing it in tension with episodic memory, which requires keeping experiences separate. Longstanding theories have proposed that these demands are resolved through a division of labor between the hippocampus, which rapidly encodes individual episodes, and the neocortex, which slowly extracts shared structure over extended timescales. However, much evidence indicates that the hippocampus itself supports rapid structure learning, and theoretical work has incorporated this capacity into accounts of hippocampal function. This raises new questions about how the hippocampus coordinates its complementary functions and how its learning is consolidated into long-term knowledge. This dissertation uses behavioral experiments and biologically grounded neural network modeling to investigate the adaptive interactions within and between brain systems that support structure learning at multiple scales. In Chapter 2, I review evidence for heterogeneous representational and functional specialization within the hippocampus, arguing that the evidence supports distinct hippocampal subcircuits being implicated in separable roles in episodic memory and rapid structure learning. In Chapter 3, I show that structure learning has direct consequences for memory representations: in a hippocampally-dependent statistical learning task, object memories are systematically distorted by the statistical structure linking them, providing a characterization of how rapid structure learning can shape individual object memory representations. In Chapter 4, I introduce a mixture-of-experts framework in which an mPFC-like gating network learns to adaptively recruit complementary hippocampal pathways according to task demands, providing a mechanistic account of how prefrontal-hippocampal interactions can arbitrate between episodic and structure learning. In Chapter 5, I present a novel model of autonomous hippocampal-neocortical interactions during sleep, showing how replay-driven consolidation integrates new hippocampal learning into existing cortical knowledge while protecting prior knowledge from interference. Together, this work characterizes how adaptive interactions within the hippocampus, between the hippocampus and prefrontal cortex, and between hippocampus and neocortex enable the brain to learn, deploy, and consolidate structured knowledge

Notes:

Source: Dissertations Abstracts International, Volume: 87-12, Section: B.

Advisors: Schapiro, Anna C. Committee members: Bhatia, Sudeep; Thompson-Schill, Sharon L.

Ph.D. University of Pennsylvania 2026

Vendor supplied data

Local Notes:

School code: 0175

ISBN:

9798247981275

Access Restriction:

Restricted for use by site license