Experience-Dependent Network Modification in the Medial Temporal Lobe

• Main Author: Thome, Alexander
• Other Authors: Barnes, Carol A.
• Language: en
• Published: The University of Arizona. 2012
• Subjects: Memory; Single-Unit; Neuroscience; Correlations; Medial Temporal Lobe
• Online Access: http://hdl.handle.net/10150/223358

Theoretical models of information storage in the brain have suggested that neurons may undergo an experience-dependent tuning or sharpening of their representations in order to maximize the amount of information that can be stored. Changes in the tuning profiles of neurons have been demonstrated to occur when animals must learn perceptual discriminations, however, whether similar changes occur in the absence of behavioral demands is unclear. To address these questions, the activity of simultaneously recorded medial temporal lobe (MTL) neurons was studied in relation to a passive visual recognition memory task. The structure of this task was such that it allowed for a comparison between novelty related responses as well as tuning properties of individual neurons. A total of 565 well isolated single neurons were recorded. The first contribution of this dissertation is the finding of a dissociation between different medial temporal lobe regions such that neurons in temporal area F (TF), but not perirhinal cortex (PRC) or the hippocampus, show an experience-dependent change in their stimulus selectivity. This finding indicates that tuning of stimulus representations may be an effective mechanism for maximizing information storage in some brain regions. The absence of stimulus tuning in higher level association regions (i.e. TF and PRC) suggests that tuning in these regions may be disadvantageous due to the need to construct unified representations across sensory modalities. A complimentary question to the question of network storage capacity is how networks avoid saturation in the connections between neurons. The second contribution of this dissertation is the finding that there exists a decrease in the magnitude of the short time scale correlations between pairs of neurons; suggesting that networks reduce the number of connections between neurons as a stimulus becomes familiar. Gamma oscillations have been proposed to be the mechanism by which groups of neurons coordinate their activity. However, network coordination has only been indirectly measured. The final contribution of this dissertation is the finding that the magnitude of gamma oscillations is strongly correlated with enhanced magnitude of correlations between neurons.