Taking the 'motor' out of motor learning: The role of cognitive brain networks and explicit processes in sensorimotor adaptation.

Abstract

The Human brain is capable of sophisticated learning and behavior even in the absence of conscious thought or attention -- a feature essential to our ability to navigate uncertain and variable environments. We are also capable of reasoning about the structure of the environment, and developing strategies in order to facilitate rapid learning. This kind of explicit learning has been linked to executive functions, and to the prefrontal cortex, but ultimately its neural substrates are poorly understood. In a series of fMRI experiments, we studied functional connectivity between cognitive and sensorimotor brain networks during various forms of sensorimotor adaptation, with a view towards characterizing networks associated with explicit learning processes. In the first of these experiments, we found that functional coupling between frontoparietal control regions and parts of the motor system were associated with improved performance. In the second experiment, we found that patterns of functional connectivity associated with explicit learning generalized across different motor learning tasks, while neural signatures of implicit learning did not, suggesting that explicit learning is supported by task-general processes. Moreover, we found that in all of these tasks, explicit learning was associated with increased functional coupling of the default mode network to frontoparietal regions, and parts of the motor system. We conclude by discussing an extensive body of literature linking features of explicit learning to control processes supported by core regions of the default mode network. Lastly, we argue that competing and vaguely defined conceptualizations of explicit learning make it difficult to relate neuroimaging findings to specific facets of learning, and that the sensorimotor learning field has done a poor job of synthesizing literature from other fields in which explicit and implicit processes have been extensively characterized, both behaviourally and neurally.