Background This study examined the effects of atomoxetine (ATX) and OROS methylphenidate (MPH) on laboratory measures of inhibitory control and attention in youth with attention-deficit/hyperactivity disorder (ADHD). It was hypothesized that performance would be improved by both treatments, but response profiles would differ because the medications work via different mechanisms. Methods 102 youth (77 male; mean age = 10.5 +/− 2.7 years) with ADHD received ATX (1.4 +/− 0.5 mg/kg) and MPH (52.4 +/− 16.6 mg) in a randomized, double-blind, crossover design. Medication was titrated in 4–6-week blocks separated by a 2-week placebo washout. Inhibitory control and attention measures were obtained at baseline, following washout, and at the end of each treatment using Conners’ Continuous Performance Test II (CPT-II), which provided age-adjusted T-scores for reaction time (RT), reaction time variability (RT variability), and errors. Repeated measures analyses of variance were performed, with Time (pre-medication, post-medication) and Treatment type (ATX, MPH) entered as within-subject factors. Data from the two treatment blocks were checked for order effects and combined if order effects were not present. Clinical trial registration Clinicaltrials.gov:NCT00183391 Results Main effects for Time on RT (p = 0.03), RTSD (p = 0.001) and omission errors (p = 0.01) were significant. A significant Drug × Time interaction indicated that MPH improved RT, RTSD and omission errors more than ATX (p <0.05). Changes in performance with treatment did not correlate with changes in ADHD symptoms. Conclusions MPH has greater effects than ATX on CPT measures of sustained attention in youth with ADHD. However, the dissociation of cognitive and behavioral change with treatment indicates that CPT measures cannot be considered proxies for symptomatic improvement. Further research on the dissociation of cognitive and behavioral endpoints for ADHD is indicated.