Date of Award

Spring 1-1-2016

Document Type


Degree Name

Doctor of Philosophy (PhD)



First Advisor

Randall C. O'Reilly

Second Advisor

Tim Curran

Third Advisor

R. McKell Carter

Fourth Advisor

Anu Sharma

Fifth Advisor

Erik Willcutt


How does the brain selectively retrieve information from long term memory? What neural mechanisms are critical for this process, and how are these mechanisms brought into service in a task dependent way? What are the implications for the representations that are processed through these mechanisms, and can we use our understanding of them to better utilize encoding and retrieval of information in long term memory? These are some of the fundamental questions being addressed in this dissertation. Through the use of neural network models of the hippocampus and surrounding cortex this dissertation proposes a framework for understanding how time frequency signatures measured at the scalp can be used to track long term memory processes, and make quantitative predictions about how information in long term memory is altered by these processes. The fundamental thesis of this dissertation is that neural oscillations in the theta (3-8 Hz), alpha (8-12 Hz), and beta (12-30 Hz) frequency bands can be tied to specific functional mechanisms supporting long term memory, and that these oscillatory signatures can be tracked in human scalp EEG recordings to predict behavioral changes in the retrieval of items from memory. Specifically that oscillatory power in the theta band positively correlates with the how much information the hippocampus is reactivating for a given retrieval event, power in the alpha band positively correlates with how much information is being inhibited from being retrieved, and beta power negatively correlates with how much non-hippocampal dependent information is being retrieved. This thesis is supported by three behavioral experiments, two EEG experiments and two explorations with a computational neural network model of the hippocampus and surrounding cortex.