Investigating the site of human saccadic adaptation with express and targeting saccades

Abstract

To focus on various objects of interest within the visual environment, primates employ rapid eye movements called saccades. When the accuracy of these movements becomes impaired, the brain can adjust their amplitude by a process known as saccadic adaptation. To investigate the locus of this plasticity in the human brain, we behaviorally adapted two types of saccade thought to be generated through different neuronal pathways. Targeting saccades, which are made to sequentially illuminated targets and have long latencies, are thought to involve higher cortical processing whereas express saccades, which have very short latencies, apparently do not. If adaptation transfers between these two types of saccade, one may conclude that the plasticity must exist at a locus common to the two pathways generating these saccades. We directly reduced the gain of either targeting or express saccades by intrasaccadically moving the target one-third of its amplitude back toward the initial fixation location and then examined whether the gain was also reduced in the other type of saccade. When targeting saccades were adapted directly, all subjects showed significant reductions in the gain of these saccades. In 75% of the 32 experimental target conditions across all subjects, there were also significant reductions in the gain of express saccades, thus providing evidence of adaptation transfer. In 71% of these conditions (i.e., 53% of all target conditions) there was no significant difference between the reductions in gain of the two types of saccade, suggesting that adaptation transfer was complete (100%). Similar results were obtained when express saccades were adapted directly: significant reductions in gain occurred in 91% of express saccades and in 100% of targeting saccades. In 86% of the target conditions, across subjects, in which both express and targeting saccades showed significant reductions in gain, the two types of saccade did not differ significantly in the amount of gain reduction. This suggests that adaptation transfer was complete for 78% of all target conditions. Therefore, we conclude that saccadic adaptation transfers robustly between targeting and express saccades. These results suggest that adaptation in humans occurs after the pathways generating these two types of saccade converge, probably at or downstream from the superior colliculus.