In the previous meeting of the journal club, we discussed how there are several parts of an optimal motor control system:
First, to control a plant (i.e., to control an arm, leg, eye, or vocal tract) we need to know certain minimal information about the current status of the plant, i.e., we need to know the current state of that plant.
Second, that because the state of the plant is, in general, not directly available, we need an observer, or internal model, of the plant that integrates information from all possible sources (e.g. the muscle commands sent to the plant, visual, auditory, and somatosensory feedback from the plant) and furnishes an estimate of the current state of the plant.
Finally, we need a controller that, based on the current state of the plant, sends the right muscle commands to the plant so it ultimately reaches some goal (e.g, we send the right muscle commands to our arm so that our hand reaches the cup we want) in some optimal way (e.g., minimum time, minimum effort, etc…)
Each of these parts of the control system must be learned, but how? Tomorrow, at Sensorimotor Journal Club, we will begin to discuss a paper suggesting that each part (the state representation, the observer, and the controller) requires a different type of learning (unsupervised, supervised, and reinforcement learning), and that different regions of the CNS (the cortex, cerebellum, and basal ganglia) are responsible for these different types of learning:
Doya, K. (2000). Complementary roles of basal ganglia and cerebellum in learning and motor control. Current Opinion in Neurobiology, 10(6), 732-739. (link to pdf of article)
The classical notion that the basal ganglia and the cerebellum are dedicated to motor control has been challenged by the accumulation of evidence revealing their involvement in non-motor, cognitive functions. From a computational viewpoint, it has been suggested that the cerebellum, the basal ganglia, and the cerebral cortex are specialized for different types of learning: namely, supervised learning, reinforcement learning and unsupervised learning, respectively. This idea of learning-oriented specialization is helpful in understanding the complementary roles of the basal ganglia and the cerebellum in motor control and cognitive functions.