The Neurobiology of Action Lab at the Champalimaud Neuroscience Programme is focused on the study of actions and how we do things. One main goal of action learning is to understand how brain networks and circuits are modified in order to allow for our motor system to acquire new motor skills.
We tackle this question by dividing it in three main topics: Action generation, focused on what are the mechanisms that allow us to generate a novel action; Action shaping, focused on the mechanisms that underlie changes during action performance leading to performance improvements and organisation of actions; Action goals, focused on the mechanisms that underlie changes in why actions are performed, if they are performed intentionally and how we create habits or routines.
Imagine that you are learning how to play tennis. Similarly to the process of learning most motor skills, this will involve a degree of trial and error. The first goal will always be to return the ball to the opposite side, while the force and accuracy at which this is done can remain variable within the first attempts. The motor system overcomes the problem of learning a complex task by identifying and reducing variability specifically on the motor elements that are essential for achieving the end goal. In our most recent study we found evidence for this, demonstrating that when learning to execute a complex action, some of the movement elements are more critical than others. Since in a challenging task it is difficult to master and optimize all elements of behavior, we observed that animals are able to select and refine specific motor patterns that lead to the end goal.
This behavioral observation led us to the subsequent question of how these processes are performed in the brain. Since there is a growing body of evidence suggesting that the circuits between the cortex (outer layer of cerebrum, composed of folded grey matter) and the basal ganglia (group of structures in the base of the brain, best known for their role facilitating voluntary movements) are involved in motor skill learning, action generation and selection, we then investigated the mechanisms that could subserve the observed effects.
By recording the electrical activity of neurons within the motor cortex (involved in the planning, control and execution of voluntary movements) and striatum (the major input of the basal ganglia, involved movement and also a part of the reward system) of mice, we were able to observe the emergence of a neuronal activity pattern that resembled the previously observed behavioral effect. The neurons displayed initial high variability of their firing rates that decreased over training, and importantly, these neuronal modulations seemed to track the more critical aspects of the movement by becoming increasingly linked to these specific elements of the action. Using genetic techniques we were also able to demonstrate that these processes depend on the capacity of these brain areas to plastically change the connections between neurons.
These results provide an important contribution to understanding how the brain chooses which action to perform and how to perform it. Understanding how the motor system selects and refines specific aspects of behavior, and how this information is encoded within the brain networks associated with motor skill learning is a crucial piece of information that can help us to get a better grasp of the complex process of how we perform motor actions.
See the original study at: Santos F et al. (2015) Corticostriatal dynamics encode the refinement of specific behavioral variability during skill learning. eLife 2015;10.7554/eLife.09423
Fernando Santos is a former PhD student at the Neurobiology of Action Lab, Champalimaud Neuroscience Programme.
Edited by: Márcia Aranha (page editor), Clara Ferreira (editor-in-chief)
Loading Likes...
You must be logged in to post a comment.