Movement has a purpose. The person who slaps her arm wants to stop the mosquito from biting her. The hungry man will lift the spoon to carry food to the mouth. But, what could happen if the hungry person is going to be bitten by a mosquito? How do we decide which movement is appropriate for a given situation? Somehow we need to evaluate the movement. We need to assign a value to each movement to tell the importance or usefulness of it. And when this is done, we need to select a movement that has the larger predicted value: Mosquito first, spoon later.
The basal ganglia, and the striatum, in particular, can assign value to movement. Even though the synaptical details are not fully known, it is clear that a dopaminergic signal somehow announces the probability of reward. Dopamine comes from two places. Substantia nigra pars compacta provides striatal dopamine, and the ventral tegmental area provides nucleus accumbens (ventral striatum) dopamine.
The striatum plays a role in forming motor habits. The quick slapping response of a person living close to mosquitoes is an example. In forming such a motor habit (quick slapping response) the striatal neurones would experience a change. To prove it, one would need to track the formation of the motor habit (i.e. the motor response), and at the same time, the changes in striatal neurones. In an optimal situation, one would like to track the same neurones as learning proceeds. As difficult as it sounds, it has been done. Tang and coworkers (Tang et al., 2007) used the motor habit of lifting the head to reach water of rats eager to drink. These authors study carefully the kinematic of movement (how fast, in which direction, with pauses etc.), and they were able to classify the type of movements as the rat learned. Of course the rat used more and more efficient movement without interruption as learning proceeded. At the same time, they implanted microwires in the dorsal striatum that allowed them to follow the activity of individual neurones during motor learning. These microwires allowed for recording of the same neurone in different days. More than that, the authors were able to follow individual neurones firing, and correlate them to a particular type of movement.
The results showed that the untrained animal uses many striatal neurones in lifting the head. As it learns and the movement becomes more efficient, fewer and fewer neurones were used.
The striatum can assign value to the movement. We know that. The child, without knowing neurophysiology, can reach for the largest piece of cake in the tray. Pasquereau and coworkers (Pasquereau et al., 2007) studied how the incentive value can shape the motor response. They used monkeys that learned to predict a juice rewarded while observing shapes in a screen, and move a joystick to select them. Each shape corresponded to a probability of reward; so after a triangle, juice always came; after a cross one in three times, and after a circle, never. Of course, the monkeys quickly learned to move the cursor to the triangle if it happen to appear in the screen and taste a yummy juice afterwards. In 44 recording sessions, the authors study the firing of 268 striatal projection neurones (medium spiny neurones) at two instances; as the cue appeared, and during the actual movement. Population perievent histograms showed that the neurones responded stronger if the cue has a large predictive value. Looking at both the striatum and the globus pallidus lateralis (internal portion of the globus pallidus, GPi), the authors made an interesting observation: During the actual movement, neurones responded stronger in the GPi, but not in the striatum. In addition, the authors examined the directional tuning curve, this is how well the neurones respond to different direction of movement (of the joystick). They found that, when the predictive value was low, neurones responded independent on the direction. But, when the triangle appeared (juice for sure!), neurones responded stronger to the triangle direction than to other directions. It is like the striatal neurones were telling: "move there, move there! There is the juice!" This sharp tunning of direction occurs in the striatatum during the appearance of the cue, but in the GPi during the actual movement. The authors concluded that the GPi computes the probability of reward "to assist the chosen action." It will be interesting to know the synaptic mechanism responsible for this directional tuning.
As we can see, the striatum can learn a habit, and it does according to the expected reward. The meaning of the reward, however, is not the same for everyone.
Tobler and coworkers (Tobler et al., 2007) examine motor learning triggered by a reward in rich and poor subjects. They trained people to predict the appearance of a coin in a screen. The researchers told the subjects that they will get the money of all correct responses. At the same time, the researchers examined the striatal activity by functional magntic resonance image (fMRI). Surprisingly, lower income subject learned faster. More than that, the striatal activity was inversely correlated to the individual assets. So we might want to say that คนจนเรียนเร็ว.
