Novel strategy for the analysis of CPT data provides new insight into the effects of methylphenidate on attentional states in children with ADHD

Abstract

Continuous performance tasks (CPTs) provide a method for studying some components of attention, but do not take into account that attention fluctuates from moment to moment. To address this issue, CPT performance was classified into one of four states (on-task, impulsive, distracted, or randomly responding) every 30 seconds, based on commission and omission error rates. We evaluated this method on 60 boys (10.6 +/- 1.1 years) with attention-deficit hyperactivity disorder (ADHD)-Combined subtype, tested before and after a dose of methylphenidate (MPH, 0.4 mg/kg), and 8 unmedicated healthy control boys (11.3 +/- 2.0 years of age). Healthy controls were on-task during 82.4% of the 30-second epochs, and made an average of 5.4 attention shifts. In contrast, children with ADHD were only on-task during 42.6% of the epochs (p = 0.0006), and they made an average of 12.8 attention shifts (p = 0.00004). These state measures provided more robust indicators of the difference between children with ADHD and controls than did traditional CPT measures of error rates, latency, and variability. The new state measures were also more significantly affected by MPH. MPH produced a 77% increase in the percent of time children with ADHD spent on-task (p < 10(12)). Conversely, MPH reduced time spent in the distracted, impulsive, and random response states by 79%, 44.5%, and 69.2%, respectively (all p values < 0.0002). Unlike errors of omission and commission, which are highly correlated (r = 0.722, n = 60, p < 10(-11)), the percent of epochs spent in impulsive, distracted, and random response states were uncorrelated, and loaded onto discrete independent factors on principal component analysis. The level of activity during the CPT correlated with the degree of distraction, but not with the degree of impulsivity. Children with ADHD could be subtyped according to the nature of their attention performance problems, and these subtypes differed in levels of hyperactivity and degrees of response to MPH.