Neural Changes and Cognitive Processes Underlying Novel Skill Acquisition / Neural Changes and Cognitive Processes Underlying Novel Skill Acquisition: Evidence from Motor Learning and Speech Production

At the boundary of motor learning and speech production, pronouncing novel speech sequences is an essential skill for communication. A form of adaptive behavior, it relies both on domain-general and task-specific cognitive mechanisms. Using a series of activation likelihood estimation coordinate-based meta-analyses, we examined cognitive processes involved in non-speech motor sequence learning and their neural correlates. These processes, observed as a function of task as well as of the properties of the sequences being learned, revealed a global network similar to the one found in speech production. This network varied considerably between individual tasks and baselines, which indicates that the choice of the experimental design can favor isolating certain cognitive processes and related brain regions over others. As a consequence, we were able to propose a refinement of existing neurocognitive models. We went onto focus on two cognitive processes at the core of skill acquisition, performance monitoring and internal model refinement, and performed a functional magnetic resonance imaging study in which participants learned to pronounce novel speech sequences of different difficulties due to their phonotactic legality. Sensorimotor-driven, representation-specific performance monitoring was observed in motor regions of the cerebellum as well as in the medial frontal cortex. Transmodal monitoring was observed in cognitive regions of the cerebellum, and the crus I in particular, in concert with temporo-parietal areas. These results highlight the presence of multiple - and possibly hierarchically interdependent - mechanisms at the service of optimizing performance during speech production. Finally, motor improvement, sustained by the basal ganglia, and internal model refinement, sustained by the cerebellum, were examined through the changes in resting state functional connectivity between these regions and the rest of the brain. The connectivity of both the cerebellum and the basal ganglia with frontal and medial frontal cortical areas changed with participants’ repetition-induced motor behavior improvement. Changes related to learning progress, arguably reflecting internal model updating, resulted in an increase in the connectivity between the right crus I in the cerebellum and the left precuneus in the parietal cortex. This change in connectivity could be explained by the cerebellum’s increasing capacity to predict the future state of the parietal cortical activity and synchronize with it. The parietal activity could then sustain the motor imagery that precedes movement articulation, further improving cerebellar prediction. Drawing from both motor learning and speech production literatures, the body of the work reported in this dissertation represents a significant step towards understanding novel speech sequence production as a skill relying on cognitive processes shared with other adaptive behaviors. Combining traditional approaches with resting state functional connectivity, we were able not only to observe these processes in distinct cerebral areas but also to pinpoint their coupling in functional networks.