Browsing by Author "Hull, Court A"

The cerebellum plays a key role in motor coordination and learning. Classical models posit that cerebellar learning is instructed by teaching signals from climbing fibers that act according to supervised learning principles by reporting motor errors. However, the climbing fibers can also signal reward-related information in some behaviors, and cerebellar learning can be enhanced by certain internal behavioral states. To understand how the cerebellum guides behavior in different contexts, it is important to reveal what learning rules the cerebellar teaching signals obey and how different behavioral context modulates these teaching signals. My thesis research addresses these two questions through two main studies.To test whether climbing fibers can operate according to reinforcement learning principles, we have measured climbing fiber responses in Purkinje cells in the lobule simplex during a classical conditioning task using two-photon imaging. By performing a modified version of the reward conditioning task, we find that climbing fiber signals exhibit central hallmarks of reinforcement learning. By using an optogenetic approach, we find that the climbing fiber response to the unconditioned stimulus is necessary for establishing a learned climbing fiber response to the conditioned stimulus, and the climbing fiber teaching signals also contribute to generating reliable, well-timed, behavioral responses. These results suggest that the cerebellum can use reinforcement learning rules to guide behavior. To test how an internal behavior state that is known to promote learning modulates cerebellar processing, we have measured climbing fiber responses in Purkinje cells in vermis V/VI during self-paced and motorized locomotion. While head-fixed mice are running on the treadmill, wide-field or two-photon imaging is done to record Ca2+ activity in Purkinje cell dendrites, and an airpuff is delivered to the lower facial area as a sensory stimulus. We find that both the spontaneous and sensory-evoked Ca2+ activity in Purkinje cell dendrites in vermis V/VI are suppressed during locomotion, and this suppression is caused by both a decreased spike probability and event amplitude, suggesting both a decreased pre-synaptic climbing fiber excitation and a post-synaptic inhibition due to locomotion. These results provide insights into how a specific internal state modulates cerebellar processing and enhances learning.

Classical models of cerebellar learning posit that climbing fibers operate according to a supervised learning rule to instruct changes in motor output by signaling the occurrence of movement errors. This model is grounded largely in studies of behaviors that utilize hardwired neural pathways to link sensory input to motor output. Yet, cerebellar output is also associated with non-motor behaviors, and recently with modulating reward association pathways in the VTA. Here, I test whether the supervised learning model applies to more flexible learning regimes and how the cerebellum processes reward related signals. I have used both classical and operant condition paradigms in combination with calcium imaging. In the operant conditioning paradigm I find that climbing fibers are preferentially driven by and more time-locked to correctly executed movements and other task parameters that predict reward outcome in a manner consistent with an unsigned reinforcement learning rule. In the classical conditioning paradigm I find distinct climbing fiber responses in three lateral cerebellar regions that can each signal reward prediction, but not reward prediction errors per se. These instructional signals are well suited to guide cerebellar learning based on reward expectation and enable a cerebellar contribution to reward driven behaviors.

Examining the coordination of excitatory and inhibitory (E/I) activity within cortical circuits is a fundamental approach to understanding normal and aberrant circuit function. Alteration to this E/I coordination is one of the leading pathophysiological models for several neurological disorders, including epilepsy, schizophrenia, ADHD, and autism spectrum disorders (ASD) (Nelson et al. 2015, Mullins et al. 2016). The overarching goal of this research is to understand the mechanism that regulates excitatory and inhibitory coordination within the cerebellum. Using a range of techniques to monitor and manipulate cortical circuits; I examined how the neuromodulator, acetylcholine, alters E/I activity at the initial input stage of the cerebellar cortex, the granular layer. Additionally, in a collaboration with Dr. Hatten's group at Rockefeller, we investigated the post-migratory role of Astrotactin 2 (ASTN2), a risk gene for autism; and found a role in regulating the surface level expression of synaptic proteins. By investigating how basic circuit function is modified in a transient manner, i.e. neuromodulation, we can reveal mechanisms that facilitate context-dependent learning; moreover, by examining permanent modifications, e.g. ASTN2, to circuit function, we can begin to understand neural mechanisms that underlie both normal synaptic function and the pathophysiology of ASDs.

These studies revealed the following: 1) acetylcholine actively modifies two out of three main nodes - excitatory mossy fiber terminals and inhibitory Golgi cells - within the granule cell layer. This modulation resulted in a bidirectional change in the excitability of granule cells, suggesting that cholinergic circuits within the granule cell layer are well situated to alter information processing in a context-specific manner. 2) ASTN2 binds and regulates the surface expression of multiple synaptic proteins via endocytosis. A truncated form of ASTN2 (ASTN2-JDUP), resulting from copy number variation in its FNIII and MAC/Perforin domains, occurs in patients with neurodevelopmental disorders. Investigations into this JDUP variant revealed changes in the binding affinity to several different binding partners, including neuroligins 1-4. Additionally, conditional overexpression of JDUP or ASTN2 in Purkinje cells, revealed differential changes in postsynaptic glutamatergic and GABAergic activity. Overall, these ASTN2 results lay the foundation for future studies using an ASTN2 loss-of-function mouse model.

The cerebellum plays a central role in motor learning by establishing sensorimotor associations necessary for coordinated movement. To do so, the cerebellum’s input layer, the granule cell layer, has been proposed to integrate and transform sensorimotor input in a way that creates unique population ensembles that can be easily learned by downstream Purkinje cells. For my thesis work, I used calcium indicators in combination with a specific pharmacological block of inhibition to measure single and combined sensory responses in granule cells in awake, behaving mice to test if local synaptic inhibition sparsens and diversifies granule cell ensembles. I have also tested if this feature can be a component of cerebellar learning, and whether neuromodulation can influence inhibition in a manner consistent with regulating sensorimotor representations. I have found that local synaptic inhibition sparsens and thresholds sensory responses, as well as sculpts population responses by establishing intensity preferences and selectively suppressing inputs. I have also found that inhibition to granule cells is necessary for performance of a cerebellum-dependent sensorimotor task. Additionally, serotonin modulates inhibition to suppress granule cell output, suggesting that it could enable learning flexibility to accommodate changes in environment or internal states. Together, my data suggest that synaptic inhibition has an important role in forming sensorimotor associations in the cerebellum.