Fear and anxiety disorders affect many individuals and hinders their ability to accurately assess threat (Bar-Haim et al., 2007; Mogg and Bradley, 1998). This type of fear learning has been successfully modeled in the laboratory using Pavlovian fear conditioning and has provided much understanding of the underlying neural mechanisms. Classic theories of animal learning define fear generalization as responding to novel stimuli in the absence of explicit pairing with reinforcement, but as a result of previous training with a similar stimulus (Mackintosh, 1974; Hull, 1947).
Auditory fear conditioning studies show there are significant fear responses to novel auditory stimuli and although associations between tones and shock can be made without and intact hippocampus, lesions of the dorsal hippocampus are able to abolish tone generalization (Duvarci, Bauer, and Pare, 2009; Quinn, Wied, and Fanselow, 2009; Anagnostaras, Maren, and Fanselow, 2009). In addition, the hippocampus is one of few structures in the brain that undergo neurogenesis, where new granule cells (GCs) are continually added to the dentate gyrus throughout a lifetime (Altman and Das, 1965; Gage, 2000). Therefore, at any given point, the dentate is composed of a heterogeneous population of immature and mature GCs. Adult-generated immature GCs have been shown to exhibit distinct neuronal properties and are preferentially activated during learning and memory tasks (Kee et al., 2007; Synder et al., 2012; Deng, Aimone, and Gage, 2010). More importantly, mice that lack post-natal neurogenesis show enhanced fear generalization to novel auditory stimuli (Cushman et al., 2012). Together, these findings indicate that the tendency to generalize may be regulated by processing within the hippocampus and that adult-born GCs are involved.
My dissertation examines the effect of manipulating hippocampal neurogenesis on auditory fear generalization. This was achieved through use of genetic alterations in mice along with x-irradiation techniques to regulate hippocampal neurogenesis. The effects of these manipulations on conditioned fear behavior were assessed with auditory fear conditioning protocols. In chapter 2, experiments described how we independently tested whether mature, immature, both or neither granule cell population would result in generalization to a novel cue after training with either a pure tone or white noise. Results showed that learning about an auditory cue that predicts shock does not require an intact dentate, however changing the population of GCs that are present produced an effect in amount of generalized freezing to a novel cue at test. In addition, mice with mature GCs silenced showed a normal generalization gradient compared to a control group in which the gradient was biased towards high frequency tones.
Chapter 3 covers an experiment that used only irradiation to investigate how depleting neurogenesis affects generalization to two pure tones after undergoing discrimination training. Mice learned to discriminate between a high and low frequency tone, but groups that had the high frequency as the CS+ performed better than those with the low frequency as the CS+. During the generalization test, irradiated mice showed little generalization but only if trained with the high frequency tone. Lastly, chapter 4 describes the last study in which a genetic manipulation of neurogenesis is used to measure effects on tone generalization using a high and low frequency auditory stimulus. Mice again showed a generalization bias at test that was dependent on the cue that was used for conditioning.
The overarching pattern of results showed that different populations of dentate GCs may lead to an increased ability to differentiate between stimuli that signal shock and those that do not. This effect, however, is substantially affected by the characteristics of the auditory stimulus that are used during conditioning.