Prior information and oculomotor initiation: the effect of cues in gaps.

Eye movements reflect not only an important output of various neural control systems, but also often reflect cognitive processing. For example, saccades are frequently used as a behavioural index of attentional processing. A second important eye movement type, smooth pursuit (SP), has received much less attention in this regard. These two types of eye movement were classically thought of as being separate, but recent results have suggested a closer linkage of their control mechanisms and perhaps their interactions with cognitive processes. Prior information, in the form of cues, alters saccade latency leading to characteristic cueing effects. When the period between the appearance of the cue and the appearance of the saccade target is sufficiently long, the latency of saccades to targets appearing at cued locations is increased. This "inhibition of return" is enhanced by a second type of stimulus manipulation, the early removal of the fixation target a few hundred milliseconds before the target appears (the gap paradigm). In the current experiments, the effect of cues, and interactions between cues and long gaps were investigated. In the main pursuit experiment, and in a separate saccade experiment, subjects were presented with interleaved runs of tasks with and without long gaps (gap duration = 1 s), and with and without cues. In tasks without cues, SP latency was reduced by long gaps (mean reduction 8 ms); unexpectedly, saccade latency for non-cue tasks was increased by long gaps (mean increase 41 ms). In a control experiment with only non-cue tasks, in which SP and saccade gap and non-gap tasks were run together, SP latency was again reduced in gap tasks, while saccade latency was increased, but by much less than in the first experiment. Analysis of individual subjects' data showed that while gaps increased saccade latency in two subjects who had participated in the main experiment (in which cues and gaps had been combined), in two naive subjects long gaps did not affect saccade latency. In the main pursuit experiment, cues had both spatially specific and non-spatially specific (warning) effects on pursuit latency. In non-gap conditions, latency was greater when contralateral cues were presented 250 ms prior to the appearance of the pursuit target, compared to ipsilateral cues, a pattern of effect consistent with inhibition of return. However, this was reversed when cues appeared during a gap--contralateral cues increased while ipsilateral cues decreased latency. For saccades, as expected, in both gap and non-gap conditions, cue effects were consistent with inhibition of return (latency was lower with contralateral cues), and the inhibition of return effect was larger in gap, compared to non-gap conditions. The results suggest that, in appropriate contexts (or as a result of appropriate training), there are distinct inhibitory mechanisms that operate on saccades but not pursuit. What appears to be an inhibition of return effect on pursuit latency when static cues are presented in pursuit tasks, may be better understood as the product of a modulation of mechanisms active in pursuit initiation, perhaps related to motion processing. In contrast to some recent evidence suggesting a close anatomical and functional linkage between pursuit and saccade initiation, the results are consistent with the involvement of a wider range of mechanisms, or a greater degree of flexibility, in programming the initiation of these two oculomotor behaviours.