The motor functions of an animal require precisely timed and coordinated sequences of movements. The cerebellum is crucial for performing these functions with precision. To investigate cerebellar computations involved in precise motor movements, behavioral paradigms such as delay eyelid conditioning have been used. Delay eyelid conditioning trains an animal to close its eye in response to a previously neutral stimulus. The timing of the eyelid closure responses suggests that the cerebellum is capable of keeping track of the elapsed time since the onset of the stimulus. This dissertation proposes a network mechanism for cerebellar timing based on biologically informed simulations of the cerebellum. In chapter 2, a simulation with over a million cells is described. This simulation approaches the observed cerebellar connectivity in several well studied mammals. Graphics processing units (GPUs) provide the computational power necessary to perform this simulation at a practical speed. This chapter describes simulation algorithms that efficiently utilize GPUs. In
chapter 3, the simulation is used to explore cerebellar timing mechanisms. The lateral inhibition among cerebellar Golgi cells is observed to be a potential mechanism for robust timing. Lateral Golgi inhibition enables the simulation to better replicate animal eyelid conditioning behavior for longer inter-stimulus intervals. In chapter 4, the emergent network mechanisms of lateral Golgi inhibition are analyzed by decomposing the network into its individual components. This component analysis demonstrates that nonreciprocal connectivity (where one Golgi cell inhibits another but does not receive inhibition in return) is useful for timing. Specifically, removing nonreciprocal connectivity greatly degrades the simulation's ability to keep track of time. This implies that the aforementioned component analyses are relevant to the emergent timing mechanisms of the network. Finally, in chapter 5, this dissertation discusses the relevance and limitations of the computational approach, biological predictions, and component analysis presented in previous chapters. / text
Identifer | oai:union.ndltd.org:UTEXAS/oai:repositories.lib.utexas.edu:2152/30351 |
Date | 12 August 2015 |
Creators | Li, Wenke |
Source Sets | University of Texas |
Language | English |
Detected Language | English |
Type | Thesis |
Format | application/pdf |
Page generated in 0.0015 seconds