Probing cerebellar function and its role in autism spectrum disorders

Abstract

Over the past few decades, a view has emerged in which the cerebellum learns and adapts well-timed response to unexpected events. While the cerebellum is traditionally considered a motor structure, it has become clear that this description of cerebellar function also applies to the nonmotor domain. This conceptual shift is important in the face of evidence that some neuropsychiatric disorders, such as autism spectrum disorder (ASD), involve cerebellar dysfunction. Although there are few assays of cerebellum-dependent nonmotor function, the putatively canonical neural circuitry of the cerebellum makes it possible to identify disruptions in well-established motor tasks in these diseases that parallel the nonmotor disruptions.

In this thesis, I report my work studying delay eyeblink conditioning, an established form of cerebellum-dependent learning that requires forming an association between neutral and aversive stimuli and producing a well-timed learned response. The ultimate goal of this approach was to model the effects of ASD-related mutations on cerebellar function. Because laboratory mice are often the best animals for modeling genetic disorders, delay eyeblink conditioning in mice must be reliable. To this end, I developed a head-fixed training technique that produced conditioning on par with training in rabbits, the standard in the field. I then applied this technique to show a novel form of learning, backward blocking, for delay eyeblink conditioning. I discuss the implications of these findings for the basic understanding of cerebellar information processing at a neural circuits level. I then applied the delay eyeblink conditioning technique to five ASD-related mouse models: Shank3+/ΔC, MeCP2R308/Y, Cntnap2-/-, patDp(15q11-13)/+, and L7-Tsc1. I discovered that these mice had learning and/or response timing deficits that mapped to the components of the cerebellar cortex affected by the modeled deficits, as predicted by what is known about the roles of those components in delay eyeblink conditioning. In particular, response timing deficits were seen in the mutants in which the granule cell network was putatively affected (Shank3+/ΔC and MeCP2R308/Y), while learning deficits were primarily seen in the mutants in which Purkinje cells and the olivocerebellar loop were putatively affected (Cntnap2-/-, patDp(15q11-13)/+, and L7-Tsc1). Together with previous findings, this work brings evidence of cerebellar learning disruption in ASD-related mouse models to seven models. This work lays the foundation for understanding the whether and to what degree ASD-related mutations affect the cerebellum in a etiologically relevant way, and suggests experiments that could be done to address the contribution of cerebellar dysfunction to the cognitive, affective, and other nonmotor symptoms typically associated with the disease.