Contribution of cerebellar sensorimotor representations to behaviour

Abstract

The cerebellum is important for behavioural control. It is not necessary for the generation of behaviour, but it is critical for its fluency and coordination. Its computations are often explained in terms of an estimation process, which allows to predict, and therefore refine, behavioural dynamics. At its simplest, this estimation process allows to integrate sensory and motor information, in order to refine actions (based on incoming sensory information), and predict sensations (based on actions). However, it is still unclear (i) how cerebellar state estimation bidirectionally interacts with brain-wide neuronal dynamics, and (ii) how this estimation process may be adapted to different behavioural contexts.
Here we addressed these two questions using computational and experimental techniques. The computational work investigated cerebellar computations, and how they contribute to whole-brain dynamics - here approximated by a simple set of differential equations. In essence, we sustain that behavioural coordination is a fundamental cerebellar function, explaining its pervasive role in behavioural control, and described a general model of cerebellar state estimation
that enables to coordinate different behavioural variables or domains.
The experimental work investigated how the cerebellar cortex represents sensorimotor information, and how these representations change as a function of the state of the network, which could reflect different behavioural contexts. In particular, we recorded both population activity in the lateral cerebellar cortex and whisking behaviour in mice. To change the state of the network, we manipulated the level of inhibition, by reducing the activity of a population of
inhibitory neurons, the Golgi cells. Our results suggests that Golgi cell inhibition may be an important mechanism through which sensorimotor representations in the cerebellum are tuned, which in turn has an impact on downstream behavioural control.
In summary, we combined computational and experimental techniques to gain insights into how neuronal dynamics in the cerebellum may be linked to and inform theories of cerebellar computations and functioning, and how these dynamics may contribute to behavioural control.

Date of Award	23 Jan 2024
Original language	English
Awarding Institution	University of Bristol
Supervisor	Paul T Chadderton (Supervisor) & Conor Houghton (Supervisor)