Abstract
Cancellation tasks are widely used measures of attention but are conventionally scored by metrics such as mean accuracy or concentration scores, disregarding the temporal structure of responding. Here, we derive two reaction-time (RT) biomarkers from our tablet-based UCancellation task in 388 young adults: intra-individual RT variability (RTV), a marker of sustained attention, and switch cost, an index of cognitive flexibility from transitions between two different task's target categories. Because consecutive targets are separated by a variable number of scanned distractors, both are confounded by search effort, which we remove with a scanning-load (p) correction framework. RTV behaved as a scored individual-difference measure: correction improved reliability and convergent validity against TOVA low-frequency RTV (n = 373), and a cross-target composite was the most reliable index. Switch cost behaved as a within-task phenomenon whose sign reversed with load—large and positive when targets were adjacent, reversing as scanning load increased—consistent with inhibition of return. Together, these metrics show that RT structure within a brief cancellation task carries attentional information that standard metrics alone does not capture.