Cerebellar and basal ganglia contributions to motor learning

点击翻译按钮获取中文摘要

Learning and refinement of new movements is an essential component of our everyday lives. We use observation of previous movement outcomes and utilise our capacity for adaptable motor control to successfully interact with objects in our environment. This process is thought to involve brain areas concerned with action selection/vigour (basal ganglia) and motor control/learning (cerebellum), which converge at the level of the motor thalamus to select the correct action type and execute movements. Although we have an in-depth understanding of the brain areas involved, the importance of the timing of these signals remains unresolved. Furthermore, the consistency of these signals between different movements and learning periods is also unknown. In this programme of work, we will test three main hypotheses:1. Timing of cerebellar and basal ganglia input to motor thalamus becomes tightly correlated as a task becomes learned2. Across several different behaviours, basal ganglia provide a consistent signal to motor thalamus, correlating with the vigour of each behaviour3. Functional convergence of these inputs at the level of motor thalamus is important for the learning of a motor taskTo determine the population-level representations of motor learning in subcortical motor regions, we will use wide-field 2-photon population calcium imaging and electrophysiological recordings of neuronal activity in subcortical motor regions (basal ganglia / motor thalamus) during execution of a Go/NoGo task. Viral-based opto-/chemogenetic manipulation strategies will be used to investigate causality between activity and learning.We will also use dimensionality reduction methods and Bayesian decoders to explore causal links between neuronal activity patterns in motor regions and learning. By applying advanced computational methods and computer vision algorithms we will develop an in-depth understanding of how single cell and network dynamics in subcortical motor regions combine to learning complex tasks and generation of accurate movements.