Abstract
An integral aspect of life is the ability to learn associations, particularly associations between experiences and negative outcomes which could impact survival. Many hypotheses have been put forth regarding how such associations are formed in the brain, and many experiments have followed, pointing to coincident activity between neurons resulting in strengthened synapses which store the association. While these hypotheses and results explain shorter timescale forms of associative learning, there are many paradigms in which associations are made between events on a much longer timescale. In this dissertation, we examine a new hypothesis explaining a longer timescale associative learning paradigm. In Chapter 1, we describe how experiments have shown that animals can create associations between two stimuli separated by many hours, yet there are no sufficient mechanistic explanations of how this happens in the brain. One mechanism underlying learning, known as reactivation, has potential for spanning such a long timescale. Findings in many brain regions, beginning with the hippocampus, have shown that there are short timescale periods of activity (outside of stimulus experience) that are strongly reminiscent of stimulus response activity. This similarity is what gives the events the name of “reactivation”. Such events have the potential of scaling a gap of many hours between events in such a way that classic mechanisms underlying learning can combine with the reactivation events to help create the association in the brain.
In Chapter 2, we characterize newly-found spontaneous bursts of activity in the anterior insular cortex, or gustatory cortex (GC), and investigate their relationship to induced activity in the region as a result of novel taste exposure. We further characterize these events in the context of a conditioned taste aversion (CTA) paradigm, a long-delay associative learning paradigm, in which we can utilize different neural states (such as rest, sickness rest, and post-CTA learning rest). The culmination of the findings is supporting evidence that these events are possible reactivations/pre-activations of taste responses and may underly associative learning in the long delay between taste experience and sickness in the CTA paradigm.
In Chapter 3, we provide theoretical evidence that a random network with clustered organization can give rise to spontaneous bursts of activity that strongly correlate to non-spontaneous / induced activity through a neural network model and simulations. This evidence is then compared with statistics of hippocampal spontaneous activity in relation to induced activity and placed in the context of reactivation research. The pre-experience nature of these events makes them comparable to preplay/pre-activation events in the hippocampus.
In Chapter 4, we summarize and review the results of the prior chapters and discuss future directions. We focus the discussion on analyses and experiments that would strengthen the argument that GC reactivation could underly learning and memory, and particularly associative learning underlying CTA.