How can an empty slate be filled with the instructions for skilled movement without an intrinsic model of the skill to be learned?

Sensory feedback and the calculation of prediction errors are essential for skill learning.  While synaptic tagging and Hebbian plasticity are critical for learning, so is the generation of functional noise. Processes that tag and strengthen neuronal connections must operate therefore in fine balance with processes of structural diversification that might lead to altered connectivity. We view neurons as multilingual signaling centers that communicate with their targets using multiple means in parallel to regulate these divergent processes.

We use conditional genetic gain and loss of function and gene expression tracer strategies in conjunction with opto-genetic approaches to study the contribution of individual types of cell signaling factors expressed by the same neuron in the regulation of structural plasticity in mice during self-paced reinforcement learning.  We focus our attention on determining structural correlates of motor learning in the basal ganglia and the spinal cord.

Our basic research appears to be of relevance to diseases of the basal ganglia and for spinal cord injury and recovery. Basal ganglia diseases like Parkinson’s Disease or Schizophrenia are characterized by aberrant motor learning and proprioceptive feedback is critical for the establishment of compensatory movement pathways in the spinal cord upon injury. In the course of our work we established a novel progressive, genetic model of Parkinson’s Disease and discovered a novel approach to reduce the formation of dyskinesia. We also found that injured motor neurons and proprioceptive neurons impinge on the physiology of glia in their immediate environment including NG2 cells whose turnover is critical for motor learning.