Abstract
The antisaccade task entails decoupling the spatial relations between target and response and looking mirror-symmetrical to the location of a target. Extensive work documents that antisaccade reaction times are longer than prosaccades: a finding attributed to the top-down nature of inhibiting a stimulus-driven response. The present investigation sought to determine whether the top-down nature of antisaccading influences the extent to which saccade trajectories are modified online. Participants completed pro- and antisaccades to target stimuli that were visible (Experiment 1) or occluded (Experiment 2) throughout the response. To index trajectory modifications, we computed the proportion of variance (R2) explained by the spatial location of the eye at 10% increments of normalized movement time (i.e., 10%, 20%, … 80%, 90% of movement time) relative to the saccade's ultimate movement endpoint. The basis for this analysis is that robust R2 values indicate that the location of the eye at any point in the trajectory provides a strong prediction of the ultimate saccade endpoint – thus evidencing a primarily offline mode of control. In turn, modest R2 values indicate that the location of the eye does not reliably predict saccade endpoint - thus evidencing a more online mode of control. Results showed that endpoints for prosaccades were more accurate and less variable than antisaccades. What is more, prosaccades yielded lower R2 than antisaccades from 20% to 70% of movement time: a finding consistent across Experiments 1 and 2. That prosaccades were characterized by weak R2 values in combination with their accurate and stable endpoints suggests that stimulus-driven actions allow for online corrections to the saccade trajectory. In contrast, the robust R2 values and increased endpoint error and variability of antisaccades suggests that top-down decoupling of the spatial relations between target and response renders a primarily offline mode of control.
Natural Sciences and Engineering Research Council of Canada.