Ahead of the (ROC) Curve: A Statistical Approach to Utilizing Ex-Gaussian Parameters of Reaction Time in Diagnosing ADHD Across Three Developmental Periods.

Abstract

INTRODUCTION\nPerformance on executive function (EF) tasks is only modestly predictive of a diagnosis of Attention Deficit Hyperactivity Disorder (ADHD), despite the common assumption that EF deficits are ubiquitous to the disorder. The current study sought to determine whether ex-Gaussian parameters of simple reaction time are better able to discriminate between children and adults with and without ADHD, compared with traditional measures of inhibitory control.\n\n\nMETHODS\nReceiver Operating Characteristic (ROC) analyses and the area under the curve (AUC) were used to examine the ability of performance on two commonly used tasks of inhibitory control (i.e. stop signal reaction time (SSRT) and go-no-go tasks) to predict ADHD status in preschool (N = 108), middle childhood (N = 309), and young adulthood (N = 133).\n\n\nRESULTS\nAcross all samples, SSRT, go-no-go percentage of failed inhibits, and standard deviation of reaction (SDRT) time to go trials, all successfully discriminated between individuals with and without ADHD. Ex-Gaussian decomposition of the RT distribution indicated that both larger tau and larger sigma drove findings for SDRT. Contrary to predictions, traditional measures of inhibitory control were equal if not better predictors of ADHD status than ex-Gaussian parameters.\n\n\nCONCLUSIONS\nFindings support ongoing work to quantify the separate contributions of cognitive subprocesses that drive task performance, which in turn is critical to developing and improving process-based approaches in clinical assessment.