In this light it is plausible that Cajal’s depiction of the nuclear organization and settling position of neurons in the developing brain represents a supraspinal complement to Romanes’s focused analysis of motor pool organization. If so, neuronal settling position could turn out to be a critical determinant

of connectivity and circuit assembly throughout the vertebrate CNS. We are indebted to George Romanes for many enlightening discussions on spinal cord anatomy and function, and for his gracious mTOR inhibitor hospitality. We also thank Gwen and Muriel Romanes for kindly providing the photograph shown in Figure 2, Kendall Doerr for an objective view on neural architecture, and Trevor Drew for advice on motor cortical maps. T.M.J. is supported by grants from NINDS and ProjectALS and is an HHMI Investigator. “
“The sensorimotor control system has exceptional abilities to perform skillful action. For example as an opposing ice hockey player skates in on the net and shoots, within a split second the goalie reaches out, catches the puck, and prevents the goal. However, there are several issues that make this a difficult computational problem for the

brain to solve. The first is uncertainty: although the goalie sees the puck coming toward the goal, he cannot be certain when and where the player will aim or where the puck will actually go. Second, once the goalie estimates the puck’s likely trajectory, Selleck Cabozantinib he must determine which of the over 200 joints and 600 muscles he will use in order to move his body Tryptophan synthase or stick to block the puck—this is the problem of redundancy.

Third, both his sensory feedback, such as the puck’s visual location, and his motor outputs are corrupted by noise. This noise in combination with the variable environment, such as the unevenness of the ice surface, leads to variability in both perception and action. Fourth, both the sensory feedback processing and motor outputs are subject to delays, with visual perception of the puck location, for example, already around 100 ms out of date. The fifth issue is nonstationarity—the system’s physical properties do not remain constant. Throughout a game the goalie must correct for weaker muscles as he fatigues, and changes in the ice surface. Finally, the entire neuromuscular system is nonlinear: for example the output of the muscle (force) is dependent on the descending activation command in a complexly nonlinear manner based on the muscle state. We will discuss each of these issues and then describe five computational mechanisms that the sensorimotor control system uses to solve the complex problem of motor control, which it does with so much skill. Our primary focus will be at the computational and behavioral level because, at present, rather little is known about how these computations are implemented. Our hope is that the neurophysiological community will see ways in which different neural areas and circuits might be mapped onto these computations.