Evidence from cluster analysis for differentiation of antisaccade performance groups based on speed/accuracy trade-offs.

Anti and pro-saccade performance in single or mixed contexts was explored in a large sample of young adults (n=281). ANOVAs were first conducted to evaluate trial type, context and gender effects. A cluster analysis was then used to determine whether subgroups could be identified based on saccadic performance variables. Increased antisaccade errors were observed among females and during mixed-saccade runs. Cluster analysis identified two groups: 1) increased errors clustered with faster latencies and 2) decreased errors clustered with slower latencies. These data offer justification for examining subgroups based on saccadic performance and may help elucidate mechanisms underlying response variability within and between different populations.