Transitions between pursuit eye movements and fixation in the monkey: dependence on context.

1. We compared the visuomotor processing underlying the onset and offset of pursuit by recording the eye movements of three monkeys as they smoothly tracked a target that was initially at rest, started to move suddenly at a constant velocity along the horizontal meridian, and then stopped. We presented this sequence of target motions in two different contexts. In the first context the target sometimes stopped after 500 ms, but on other interleaved trials the target either continued moving at a constant velocity, slowed down, speeded up, or reversed direction. In the second context the target always stopped, but the duration of the preceding constant velocity was randomized from 500 to 700 ms. 2. The dynamics of the eye velocity during the offset of pursuit were markedly different in the two experiments. When the target stopped only sometimes, the decrease in eye velocity at the offset of pursuit often overshot zero, producing a brief, small reversal in the direction of pursuit before eye speed settled to zero. When the target always stopped, the decrease in eye velocity at the offset of pursuit followed a more gradual transition toward zero with no overshoot. Thus the eye velocity profiles obtained in the first experiment contradict, whereas those obtained in the second experiment confirm, previous characterizations of the offset of pursuit as an exponential decay toward zero eye speed. 3. To investigate the basis of the different eye velocity profiles obtained in the two experiments, we probed the state of transmission along the visuomotor pathways for pursuit with the use of small perturbations in the motion of the target. We used perturbations consisting of 1 degree step changes in target position superimposed on the constant velocity motion of the target, on the basis of previous findings that such perturbations elicit saccades during fixation but smooth changes in eye speed during maintained pursuit. Single perturbations were imposed at regularly spaced intervals on separate interleaved trials during the onset, maintenance, and offset of pursuit. 4. Perturbations imposed during the onset and maintenance of pursuit had similar effects regardless of whether the target stopped only sometimes or always. In both experiments, perturbations that stepped the target in the direction opposite to the constant velocity of the target produced decreases in eye speed; perturbations in the same direction produced negligible or inconsistent changes in eye speed. The changes in eye speed caused by perturbations were largest for those perturbations introduced within the first 100 ms after the start of target motion, before the onset of the smooth eye movement, and became progressively smaller as target motion continued. The largest changes in eye speed were therefore caused by those perturbations imposed during periods of large retinal slip and by those perturbations whose direction opposed that slip. 5. Perturbations imposed during the offset of pursuit had different effects depending on whether the target stopped only sometimes or always. When the target stopped only sometimes, forward perturbations produced large increases in eye speed, whereas backward perturbations produced negligible or inconsistent changes in eye speed. Thus the visuomotor processing underlying the offset of pursuit in this experiment strongly resembled that underlying the onset of pursuit: in both cases, those perturbations in the direction opposing large retinal slip produced the largest effects. In contrast, when the target always stopped, neither forward nor backward perturbations imposed during the offset of pursuit produced large changes in eye speed. This indicates that the visuomotor processing underlying the offset of pursuit in this experiment was different from the processing underlying the onset of pursuit. 6. Perturbations also produced changes in the frequency of saccades, although these effects were less consistent than the changes in pursuit eye speed