Sensitivity derivatives for flexible sensorimotor learning.

To learn effectively, an adaptive controller needs to know its sensitivity derivatives--the variables that quantify how system performance depends on the commands from the controller. In the case of biological sensorimotor control, no one has explained how those derivatives themselves might be learned, and some authors suggest they are not learned at all but are known innately. Here we show that this knowledge cannot be solely innate, given the adaptive flexibility of neural systems. And we show how it could be learned using forms of information transport that are available in the brain. The mechanism, which we call implicit supervision, helps explain the flexibility and speed of sensorimotor learning and our ability to cope with high-dimensional work spaces and tools.

Task-dependent constraints in motor control: pinhole goggles make the head move like an eye.

In the 19th century, Donders observed that only one three-dimensional eye orientation is used for each gaze direction. Listing's law further specifies that the full set of eye orientation vectors forms a plane, whereas the equivalent Donders' law for the head, the Fick strategy, specifies a twisted two-dimensional range. Surprisingly, despite considerable research and speculation, the biological reasons for choosing one such range over another remain obscure. In the current study, human subjects performed head-free gaze shifts between visual targets while wearing pinhole goggles. During fixations, the head orientation range still obeyed Donders' law, but in most subjects, it immediately changed from the twisted Fick-like range to a flattened Listing-like range. Further controls showed that this was not attributable to loss of binocular vision or increased range of head motion, nor was it attributable to blocked peripheral vision; when subjects pointed a helmet-mounted laser toward targets (a task with goggle-like motor demands but normal vision), the head followed Listing's law even more closely. Donders' law of the head only broke down (in favor of a "minimum-rotation strategy") when head motion was dissociated from gaze. These behaviors could not be modeled using current "Donders' operators" but were readily simulated nonholonomically, i.e., by modulating head velocity commands as a function of position and task. We conclude that the gaze control system uses such velocity rules to shape Donders' law on a moment-to-moment basis, not primarily to satisfy perceptual or anatomic demands, but rather for motor optimization; the Fick strategy optimizes the role of the head as a platform for eye movement, whereas Listing's law optimizes rapid control of the eye (or head) as a gaze pointer.

Non-commutativity in the brain.

In non-commutative algebra, order makes a difference to multiplication, so that a x b not equal to b x a. This feature is necessary for computing rotary motion, because order makes a difference to the combined effect of two rotations. It has therefore been proposed that there are non-commutative operators in the brain circuits that deal with rotations, including motor circuits that steer the eyes, head and limbs, and sensory circuits that handle spatial information. This idea is controversial: studies of eye and head control have revealed behaviours that are consistent with non-commutativity in the brain, but none that clearly rules out all commutative models. Here we demonstrate non-commutative computation in the vestibulo-ocular reflex. We show that subjects rotated in darkness can hold their gaze points stable in space, correctly computing different final eye-position commands when put through the same two rotations in different orders, in a way that is unattainable by any commutative system.