Revealing neural representations of movements and skill using multi voxel pattern analysis

One of the main functions of the human brain is to process information, such that we can interact efficiently with our environment by moving our body. Neuronal representations of information pertaining to the movement is fundamental for control. Using functional magnetic resonance imaging, researchers have studied brain areas that are responsible for motor control based on overall neuronal signal changes. It is assumed that the amount of overall activity indicates how much an area is involved in the control of movements. In this thesis, I start from the approach that the representation of critical variables describing the movements, rather than the overall activation, is the most relevant factor for a region to be important in the control of an action. Representations in three major fields of motor control were studied in this thesis. First, the integration of sensory and motor information was analysed via finger representations in the cerebellum and the neocortex. The findings suggest that sensory and motor representations of fingers overlap spatially in the neocortex but are interdigitated in the cerebellum, suggesting neuronal differences in how information are integrated in the brain structures. Then, neuronal reorganisations of representations were studied during motor learning. The results showed that the neural representation of sequences becomes more distinct with training, while the overall activity does not change. Lastly, I studied effector specific and effector independent representations of sequential motor behaviours by investigating the similarity of neuronal representations for left and right hand performance. Overall, this thesis demonstrates that the study of neural representations using multivariate methods in fMRI provides a new hypothesis-driven approach to the study of human motor control and learning of movements.