Abstract
Previous work by our group has shown that antisaccades exhibit reduced online corrections relative to prosaccades (Heath et al., 2010: Exp Brain Res). In that previous work, the onset of a target stimulus served as the movement imperative cue, and as a result, prosaccade reaction times (RT) were reliably faster than antisaccades. However, it is possible that the reported difference in online corrections may relate to the temporal limitations of the constituent planning processes underlying the antisaccade task (i.e., response suppression and vector inversion). To address this issue, the present study had participants complete stimulus-driven and memory-guided pro- and antisaccades. Notably, the memory-guided condition was employed to equate pro- and antisaccade RTs. To index online control, we computed the proportion of variance (R2 values) explained the spatial location of the eye at normalized deciles of movement time relative to the response's ultimate movement endpoint. The basis for this analysis is that smaller R2 values at any point in the trajectory indicate that the position of the eye does not reliably predict the response's ultimate endpoint and is therefore indicative of online trajectory corrections. In turn, larger R2 values are interpreted to reflect fewer online corrections. RTs for stimulus-driven prosaccades were faster than matched condition antisaccades; however, this advantage was nullified in the memory-guided task. Moreover, antisaccade R2 values from the early to late stages of the response were larger than their prosaccades counterparts and this finding was consistent across stimulus-driven and memory-guided conditions. Thus, between-task differences in online control cannot be attributed to premovement costs in response planning; rather, evidence indicates that the top-down nature of the antisaccade task engenders a presetting of the oculomotor system and renders a mode of control that is not optimized to support online corrections.