| To what extent do humans exert control over low-level components of the oculomotor system? Previous studies of the fixation-offset effect (FOE)---a decrease in the latency of saccades to a peripheral target produced by concomitant removal of a fixated stimulus---suggest that this effect is stimulus-driven (i.e., dependent only on the presence or absence of a fixated stimulus upon target presentation). However, two eye-tracking studies reported here indicate that humans can modulate the magnitude of the prosaccade FOE through purely strategic processes engaged when information is provided regarding the probability of target presentation. In contrast, the additivity of target probability and fixation offset/overlap factors in a subsequent antisaccade experiment suggests that different control processes are engaged for preparation of antisaccades than those engaged for prosaccades. From these results, I proposed that performers in the present antisaccade paradigm strategically maintained a high level of fixation-related activity in the oculomotor system while preparing the amplitude component of the antisaccade vector; evidence for this hypothesis was inferred from the results of a previous precueing study (Abrams & Jonides, 1988), and it is consistent with a theoretical model proposed here. This model accounts quantitatively for the saccade latencies reported in the present studies and for those reported in Machado and Rafal (2000b), in some cases with striking precision. The success of this parsimonious model suggests that it reflects the neural mechanisms underlying performance in pro- and anti-saccade tasks. Finally, an fMRI study of prosaccade performance is presented that revealed for the first time that humans can strategically modulate superior colliculus activity in the service of saccade preparation. |
