Attention-deficit hyperactivity disorder (AD/HD) is characterised by symptoms of severe inattention, overactivity and impulsiveness. With its prevalence of 3–5% in school-age-children, AD/HD is one of the most common disorders in child and adolescent psychiatry . According to Barkley's theory of AD/HD [2, 3], deficient behavioural inhibition is the core deficit of the disorder, and may lead to impairments of executive functions. Behavioural inhibition may be separated into three interrelated processes called 'inhibition of the initial prepotent response to an event', 'stopping of an ongoing response' and 'interference control'.
Several behavioural studies reported deficits of response-inhibition in children with AD/HD ([4–8]; for a review see ). However, impaired behavioural response inhibition is also observed in children with other disruptive disorders such as ODD/CD , which is the most prevalent comorbidity of AD/HD and poses significant additional clinical and public health problems. In addition, further deficits which are not likely to result from deficient inhibition are present in children with AD/HD, as evident from their poor performance in a variety of executive functions tasks such as the Continuous Performance Test (CPT) [10, 11], Wisconsin Card-Sorting-Task [12–14], Tower-of-Hanoi [13, 14] and Stroop-Test [12, 15]; for a review see [16, 17].
In a more neurophysiologically oriented theory covering both ADHD and ODD/CD, Quay [18, 19] following Gray  argued that the behavioural activation system (BAS, sensitive to reward) and the behavioural inhibition system (BIS, sensitive to punishment) may reflect distinct pathways for inhibition deficits. Children with AD/HD may suffer from an underactive BIS while their BAS seems to be unimpaired, whereas children with ODD/CD should have an overactive BAS that dominates their (unimpaired) BIS. Therefore, according to Quay's theory both AD/HD and ODD/CD groups should display deficits in inhibition, but for very different reasons. If comorbid AD/HD+ODD/CD is an additive combination of AD/HD and ODD/CD, this group should display the worst impairment in response inhibition because an overactive BAS may be combined with a weak BIS. Concerning response control, results from a recent neurophysiological study with the CPT-task are consistent with this prediction, and indicate that such deficits are indeed particularly pronounced in this comorbid group . Deficits in executive functioning in general, and inhibition deficits in particular are also explained by other neurophysiological theories focusing on either AD/HD or ODD/CD. For ODD/CD [22–24], it has been argued that deficits of the prefrontal cortex leads to reduced orienting and arousal, both of which predispose individuals to stimulation-seeking, disinhibition and attention deficits, and thereby to antisocial behaviour.
The ,Stop-signal paradigm' allows investigating well defined response inhibition processes directly. Generally, the subjects perform a simple or a two choice reaction task. In some of the trials, a Stop-signal follows the go-stimulus at a given delay and requires the inhibition of the ongoing response. The longer the Stop-signal-delay (SSD), the more difficult it becomes to inhibit the response. The „horse race“ model of the Stop-task, which assumes a race between the reaction to the primary task and the reaction to the Stop signal, further allows to estimate the "virtual" reaction time to the Stop-signal (SSRT) as a measure for response inhibition performance . In a meta-analysis of the Stop-task, Oosterlaan et al.  reported that behavioural studies showed consistently slower SSRT for children with ADHD, but also for children with CD compared to controls, Comparisons between AD/HD and CD as well as between AD/HD+CD and AD/HD revealed no differences. However, inferences based on performance data only may have limited validity, because differences in covert brain mechanisms may lead to similar overt performance [21, 27].
A more direct access to brain functions is provided by non-invasive methods such as functional magnetic resonance imaging (fMRI)  or event related potentials (ERP) . Briefly, in the blood-oxygenation-level-dependent (BOLD) fMRI, changes in cerebral blood-flow and metabolism related to neuronal activation are measured with high spatial but low temporal resolution reflecting the underlying hemodynamic process. ERPs are voltage topographies and fluctuations recorded on the scalp which reflect neural activation to an event such as the presentation of a stimulus or a response. A major advantage of the ERP technique is the high temporal resolution in the range of milliseconds which allows to measure brain-electrical correlates of information-processing in realtime. A number of studies therefore used electrophysiological or fMRI measures of response inhibition processes in AD/HD [7, 27, 30–32].
An ERP-study of Brandeis et al.  revealed that in AD/HD children, successful Stops differed from Stop-failures with topographic alterations in a microstate which reflected mainly processing of the go-stimulus, whereas normal controls differed at a slightly later stage of processing with increased global-field-power (GFP, the spatial standard deviation of voltages) in Stop-failures compared to correct Stops. Rubia et al.  reported that during their fMRI-study, decreased right-inferior-prefrontal activation in AD/HD occurred solely in the Stop-task, and thus hypothesized the "brake system of the brain"  to be located right-prefrontal. Pliszka et al.  reported for normal controls a negative wave 200 ms after onset of the Stop-signal (Stop-N200) over right inferior frontal electrodes which was reduced in ADHD-children. For both groups, this N200 after successful inhibitions was positively correlated with inhibition performance whereas correlations for Stop-N200 to Stop-failures were not that clear. Following Kok , the N200 to the Stop-signal could either reflect a 'red flag' or a subsequent "(action-) inhibitory process, emanating from structures in the prefrontal cortex" . A second finding was that at right-frontal electrode-sites 250–500 ms post Go-signal-onset the control-group displayed greater positivity to failed than successful Stop-trials whereas in the ADHD-group successful trials did not differ from failed ones. This preparatory activity in failed Stop-trials was more positive in controls than in ADHD patients. Further, Dimoska et al.  found, despite worse Go-task- and inhibition-performance in AD/HD compared to controls, different activation-patterns at an early stage of processing the Stop-signal. Again, a decreased N200 to the Stop-signal of successful Stops for AD/HD was found, whereas groups did not differ concerning Stop-N2 of Stop-failures. Following Pliszka et al. , the authors argued that this N200 would reflect activation of inhibitory processes. However, in contrast to Pliszka et al. their auditory evoked N200 was generally larger to failed than to successful Stops. Overtoom et al.  found slower SSRT and decreased inhibition performance for AD/HD compared to normal controls. Interestingly the study showed no N200-effects to the Stop-signal. This could be due to the use of an auditory Stop-signal, as Falkenstein et al.  found a Nogo-N2 which was smaller for auditory compared to visual stimuli despite similar performance in both modalities which could indicate that inhibition is related to a pre-motor level.
There is an ongoing debate whether the Nogo-N200 reflects inhibitory processes per se [33–37], or conflict monitoring [38–40] which may initiate inhibition. We did not intend to distinguish between these two models. Both of them predict that the Stop-N200 is related to inhibition performance: while the inhibition theory relates diminished Stop-N200 amplitudes directly to an impaired central inhibition mechanism, the conflict-signal theory suggests that impaired triggering of the inhibitory mechanisms is responsible.
Taken together, studies strongly suggest difficulties in response inhibition paralleled by neurophysiological deviances for children with AD/HD compared to normal controls, but to our knowledge there is no such evidence for ODD/CD and comorbid AD/HD+ODD/CD. Thus, the aim of this study was threefold, as we intended (1) to replicate the neurophysiological finding of Brandeis et al.  and of Pliszka et al.  concerning both early pre-Stop-signal processing and the later Stop-N200-differences between controls and children with AD/HD; (2) to clarify whether children with ODD/CD and especially those with comorbid AD/HD+ODD/CD also display an inhibitory-deficit as hypothesized according to Quay's model, i.e. a slower SSRT and slower Stop-failure reaction-times paralleled at the neuronal level by a reduced Stop-N200-amplitude; and (3) we wanted to test whether an additive model of AD/HD and ODD/CD explains response-inhibition performance of children with comorbid AD/HD+ODD/CD.