Individuals with Attention-Deficit-Hyperactivity Disorder (ADHD) are more likely than their peers to make poor real-life decisions, as these individuals are described as impulsive , engage in more risky activities than controls [2, 3], and tend to exhibit a preference for immediate rather than delayed rewards [4–7]. Much of the recent work over the last 20 years in ADHD has focused on the cognitive features of ADHD, defined as executive functions [8–10]. However, recent theories of ADHD have included both executive processes and motivational style characterized by delay aversion as two important pathways in ADHD [11, 12]. The purpose of this study was to elaborate and extend this conceptualization by examining performance on a risky-choice decision-making card task, also known as the Iowa Gambling Task, in a sample of adolescents with ADHD and comparison controls. We refer to this task as the card task for simplicity. In addition, we were also interested in examining associations between performance on the card task and measures of intellectual ability and working memory.
The most recent development in the field has been an emphasis on multiple pathways for explaining impairment in ADHD [13–16, 12]. Sonuga-Barke's  most recent model of ADHD argues for a dual pathway model of ADHD, highlighting both the executive and motivational, delay aversion aspects of ADHD. Specifically, the executive pathway involves a dysregulation of thought and action that is primarily characterized by a core deficit in inhibitory control . Alternatively, the motivational pathway is hypothesized to mediate a link between behavioural symptoms, task engagement, and a biologically embedded alteration in reward mechanisms . One important change reflected in this dual-pathway conceptualization is that the executive and motivational pathways are not regarded as competing theories, rather that deficits in both processes are thought to give rise to the manifestation of ADHD . Sonuga-Barke  argues that these two pathways likely give rise to an ADHD diagnosis of the Combined subtype, and that the executive pathway is more likely associated with more severe and generalized cognitive impairment. This idea of dual pathways maps well onto clinical and research characterizations of ADHD, as the Inattentive subtype has been associated with executive dysfunction  and the Hyperactive/Impulsive subtype has been associated with the impulsiveness feature of ADHD . Sonuga-Barke, Dalen, & Remington  reported that executive function and delay aversion made significant, independent contributions to ADHD symptoms in a sample of children. Similar findings have also been reported by Crone, Jennings, and van der Molen .
In addition to separability at a neuropsychological level, it has been hypothesized that the executive and motivational pathways are rooted in conceptually similar, but functionally segregated brain circuits . It has been hypothesized that the executive pathway receives inputs from the dorsolateral prefrontal cortex to the dorsal portion of the neostriatum, as well as reciprocal connections from subcortical regions, including the dorso-medial sections of the thalamus. Alternatively, the motivational pathway centers on the reward circuits of the ventral striatum, specifically the nucleus accumbens, with connections from frontal regions, including the anterior cingulate and orbitofrontal cortex, and the amygdala. Importantly, this dual pathway model offers a theoretical account of interactions between cortical and subcortical pathways in the regulation of executive processes and motivation. The notion of separable pathways has also been theorized at a neurobiological level as well. Sagvolden et al.'s  dynamic developmental model of ADHD suggests that altered dopamine branches give rise to the different ADHD symptomotology, specifically, a hypofunctioning mesolimbic dopamine branch gives rise to delay aversion, a hypofunctioning mesocortical dopamine branch gives risk to poor executive functions, and a hypofunctioning nigrostriatal dopamine branch gives rise to other motor symptoms. This neurobiological account is consistent with the idea of separable pathways in ADHD, specifically the executive and delay aversion pathways. It is important to understand and test the relative contribution of these different pathways to the clinical manifestation of ADHD at both behavioural and neurobiological levels of analysis.
An important behavioural measure, called the Iowa Gambling Task (or card task), was designed to simulate the uncertainties of real-life decision-making, necessitating the weighting of rewards and penalties. In this task, participants are asked to select cards from four decks, which unbeknownst to these participants vary on expected outcome [two decks are composed of quick high gains and high losses (disadvantageous decks) or low gains and low losses (advantageous decks)] and in frequency of penalties (two decks have frequent, smaller penalties, while the other two decks have infrequent, large penalties). Expected outcome and frequency of penalties are crossed, creating four different conditions with these four decks. What has been particularly striking and important about these studies on Iowa Gambling Task performance with patients who have ventromedial cortex lesions is not simply the link between brain function and higher cognitive processes, but also the possibility that this form of reasoning may be fairly modular and localized, and separable from other cognitive abilities, such as intelligence .
The card task was originally used to study patients with ventromedial cortex lesions . The ventromedial cortex, which may include the orbitofrontal cortex, but some have argued for subtle distinctions between these areas [23, 24]. For present purposes, the orbitofrontal cortex and ventromedial areas have generally been implicated in the emotional experience associated with gains and losses in decision processes, which is to be differentiated from other frontal processes and regions, including the dorsolateral prefrontal cortex and anterior cingulate .
Patients with ventromedial prefrontal cortex lesions displayed less optimal performance than normal controls, making considerably more selections from the disadvantageous decks than from the advantageous decks . This link between the orbitofrontal cortex and impulsivity has been described previously by others, such as, Newman et al. , who reported on brain lesion studies with rats which suggested an association between disinhibitory syndromes in humans (including, psychopathy, addictions, and ADHD) and lesions in the orbitofrontal cortex. Other clinical samples have reportedly displayed less optimal performance on the card task, including high school students with multiple suspensions , heroin addicts , and individuals with antisocial behaviour and psychopathic tendencies . Sex differences have also been reported on card task performance, with females tending to make more selections from disadvantageous decks than males .
Performance on the card task has also been examined in a small group of adults with and without ADHD  and in a sample of adolescents with disruptive behaviour disorders, including ADHD and conduct problems . In the study which included adolescents with disruptive behaviour disorders , adolescents with disruptive behaviour disorders displayed less optimal performance on this task than controls. In the study using adults with ADHD , no behavioural differences were observed on decision-making performance. However, PET scans revealed that brain activation in the ADHD group was less extended than in the control group. Specifically, control participants recruited hippocampal and insular regions more than adults with ADHD, and the adults with ADHD engaged the caudal part of the right anterior cingulated more than the controls. These results, therefore, suggest that this task activates different processes and brain regions. The behavioural significance of these differential activations is unclear at this time, and suggests that more work needs to be done on elaborating our understanding of the motivational pathway . The card task and its links with patients with ventromedial lesions makes it an important task to further study and understand the motivational pathway in ADHD.
As recent models of ADHD suggest two potential pathways supporting the motivational and executive processes in ADHD , we must also investigate the possible associations and/or dissociations between these processes. One study investigated card task performance and working memory in normal controls, patients with lesions in the ventromedial prefrontal cortex, and patients with lesions of the dorsolatoral/mesial region of the prefrontal cortex . Bechara et al.  reported a cognitive and anatomical double dissociation between card task performance and working memory. Others have also reported a dissociation between impulsivity and working memory in the orbitofrontal and dorsolateral regions of the prefrontal cortex . However, Ernst et al.  found that IQ was a significant predictor of card task performance in their sample of adults with substance abuse disorders and adolescents with behaviour disorders. Hinson, Jameson, & Whitney  reported that increasing working memory load resulted in poorer performance on the card task. Therefore, mixed results have been reported on the behavioural associations between card task performance and working memory. These findings suggest that lesions in the ventromedial cortex may not uniquely or independently explain risky decision-making, and that contributions from other regions, such as the frontostriatal dopamine system , and other processes, such as working memory [35, 36], may also play a role in decision-making involving gains and losses. Therefore, we also examined the relationship between card task performance and intellectual ability, and working memory in our sample of adolescents with ADHD.
The purpose of the present study was to examine how adolescents with ADHD weight risks and benefits in the card task compared to comparison controls. Impaired performance on the card task would importantly implicate ventromedial prefrontal regions for study in ADHD, which would importantly extend the motivational pathway of the dual pathway model of ADHD [11, 12] by suggesting a link between subcortical structures, like the nucleus accumbens, and cortical structures, like the ventromedial cortex. In addition, associations between card task performance, intellectual ability, and working memory were also examined. A dissociation between performance on the card task and intellectual ability and working memory (our executive tasks) was predicted, which would be consistent with dual pathway models of ADHD.