Abstract
Motion in the visual world is often caused by the organism itself. Consider a primate's arm moving across his field of view, or the optic flow generated while an animal locomotes. As compared to other forms of naturally occurring motion, self-generated motion is inherently more predictable to the organism, and it is also more directly coupled with ongoing motor behavior. Perhaps it is processed differently by the brain.
We have been recording from direction-selective cells in the parietal lobe of an awake, behaving macaque monkey. While the monkey fixates he views a peripheral stimulus — a dot moving back and forth between two parallel bars, 22 apart. The bars are situated such that the dot reverses direction within the cell's receptive field, along the preferred-null axis.
The monkey does three different tasks in separate but interleaved blocks of trials. In the ACTIVE block the monkey himself induces the dot's motion reversals by pressing a lever. He is allowed to turn the dot only when it is within a few degrees of the bar. In the ACTIVE-DELAY block the monkey experiences a 200 msec delay between the time he presses the lever and time the dot turns. In the REACTION block the monkey no longer influences the turns. He has to act upon the lever immediately after noticing a computer-generated reversal. Some of the reaction-block reversals happen at identical locations to the ones in the active blocks. The dot's visible trajectory is nearly identical in all analyzed trials — what changes, block by block, is the phase of a required motor behavior, and whether the monkey actively controls or simply reacts to the moving stimulus.
Our recordings target the superior temporal sulcus and the intraparietal sulcus. In initial recordings from 61 neurons in MT/MST, the population has shown little difference across blocks. In agreement with past findings from our lab, neurons from MT/MST seem to faithfully follow the visual image independent of the hand movement.