Orbitofrontal reward sensitivity and impulsivity in adult attention deficit hyperactivity disorder

Abstract

Impulsivity symptoms of adult attention deficit hyperactivity disorder (ADHD) such as increased risk taking have been linked with impaired reward processing. Previous studies have focused on reward anticipation or on rewarded executive functioning tasks and have described a striatal hyporesponsiveness and orbitofrontal alterations in adult and adolescent ADHD. Passive reward delivery and its link to behavioral impulsivity are less well understood. To study this crucial aspect of reward processing we used functional magnetic resonance imaging (fMRI) combined with electrodermal assessment in male and female adult ADHD patients (N = 28) and matched healthy control participants (N = 28) during delivery of monetary and non-monetary rewards. Further, two behavioral tasks assessed risky decision making (game of dice task) and delay discounting. Results indicated that both groups activated ventral and dorsal striatum and the medial orbitofrontal cortex (mOFC) in response to high-incentive (i.e. monetary) rewards. A similar, albeit less strong activation pattern was found for low-incentive (i.e. non-monetary) rewards. Group differences emerged when comparing high and low incentive rewards directly: activation in the mOFC coded for the motivational change in reward delivery in healthy controls, but not ADHD patients. Additionally, this dysfunctional mOFC activity in patients correlated with risky decision making and delay discounting and was paralleled by physiological arousal. Together, these results suggest that the mOFC codes reward value and type in healthy individuals whereas this function is deficient in ADHD. The brain–behavior correlations suggest that this deficit might be related to behavioral impulsivity. Reward value processing difficulties in ADHD should be considered when assessing reward anticipation and emotional learning in research and applied settings.

Introduction

Between 15 and 65% of children with attention deficit hyperactivity disorder (ADHD) continue to show symptoms during adulthood (prevalence rate: 2.5%) (Faraone and Biederman, 2005, Simon et al., 2009). Whereas hyperactivity symptoms remit over time, attentional deficits and impulsivity persist (Wender et al., 2001). In adult ADHD, impulsivity manifests in poor occupational performance (Mannuzza et al., 1997), drug abuse (Elkins et al., 2007), sexual risk taking (Flory et al., 2006), intimate partner violence (Fang et al., 2010), risky driving (Fischer et al., 2007), and other disadvantageous behaviors (Biederman et al., 1994, Eakin et al., 2004, Weafer et al., 2011). In terms of gender ratio, ADHD in adulthood is more evenly distributed in men and women.

Impulsivity in ADHD has been explained primarily by deficits in executive functioning and inhibition (e.g. Barkley, 1997). Premature responses, e.g., might be related to timing disturbances (Rubia et al., 2009a). Another line of research has highlighted the role of emotional and motivational aspects for impulsivity in ADHD (e.g. Sonuga-Barke, 2005). One crucial motivational aspect of impulsive behavior is an altered reward sensitivity (see e.g. Luman et al., 2010 for review). Accordingly, ADHD patients prefer immediate over delayed rewards due to a steeper gradient for delay-of-gratification during learning (Sagvolden et al., 2005). Neurobiologically, this might reflect dysfunctions in tonic or phasic dopamine levels in response to reward (Tripp and Wickens, 2008).

Behavioral studies have indeed demonstrated altered responses to reinforcement in ADHD (Aase and Sagvolden, 2006, Douglas and Parry, 1994, Frank et al., 2007, Luman et al., 2009). Also a specific preference for immediate over delayed reward has been described in childhood and adolescent ADHD (Bitsakou et al., 2009, Paloyelis et al., 2010, Solanto et al., 2001) accompanied by an orbitofrontal hypoactivation (Rubia et al., 2009a). A brain imaging study in adult ADHD patients demonstrated a neural dissociation between decisions for immediate and delayed reward (Plichta et al., 2009) but no behavioral preference. ADHD patients exhibited a diminished response to immediate reward in the ventral striatum (VS) and an increased response to delayed reward in the dorsal striatum.

A related line of fMRI studies has focused on reward anticipation as studied in the Monetary Incentive Delay (MID) task (Knutson et al., 2001). In this task, distinct cues inform participants about the possibility to win or to avoid losing money by responding quickly to a target. During the presentation of these cues, i.e. during the reward anticipation phase, adolescents (Scheres et al., 2007) and adults (Carmona et al., 2011, Hoogman et al., 2011, Strohle et al., 2008) with ADHD showed decreased activation in the VS compared to controls. Strohle et al. (2008) also analyzed neural responses during the reward delivery phase and found increased activations in the dorsal striatum and lateral orbitofrontal cortex (OFC) in male ADHD patients.

Altered OFC functioning in ADHD has also been reported in studies investigating the effect of reward on executive functioning (Cubillo et al., 2011, Dibbets et al., 2009, Rubia et al., 2009b). However, in these study designs a specific attribution of OFC deficits to either anticipatory or consumatory processes is difficult since the focus is the effect of reward on inhibition or sustained attention rather than reward processing alone. Moreover, some uncertainties remain with regard to the direction of the OFC deficit (some studies found a hypoactivation whereas others a hyperactivation), its specificity to type of reward (positive feedback vs. monetary reward), its location within the OFC (medial, antero-lateral, postero-lateral) as well as its generalizability to female ADHD patients (cf. Valera et al., 2010) and interaction with comorbid disorders (e.g. conduct disorder, Cubillo et al., 2011, Rubia et al., 2009c).

Taken together, there is clear evidence for a ventral striatal deficit (hypoactivation) during the anticipation of reward which has been replicated consistently in ADHD patients of different ages and sexes. In contrast, the role of the OFC and alterations in neural response to reward delivery in ADHD are less well understood. Yet, both issues are of great importance for several reasons: First, from a theoretical standpoint, the interpretation of altered reward anticipation implicitly relies on assumptions about a normal reward receipt, e.g. a normal subjective valuation of this reward. In other words, differences in reward valuation would effectuate similar differences during anticipation. Second, electrophysiological studies, which due to excellent temporal resolution can reliably characterize both anticipation and delivery of rewards, suggest altered processing of the latter, i.e. altered responses to feedback and monetary outcomes, in ADHD (Groen et al., 2008, Holroyd et al., 2008, van Meel et al., 2005, van Meel et al., 2011). Third, together with observed abnormalities in the OFC of ADHD patients this altered processing of actual reward could refer to altered reward coding in ADHD (Kahnt et al., 2010, Kringelbach and Rolls, 2004, Sescousse et al., 2010). Fourth, a better understanding of reward delivery processing and the role of the OFC might help to clarify discrepant findings about the association of VS activity and impulsivity: findings in ADHD patients suggest VS hypoactivation as a neural correlate of impulsivity (Scheres et al., 2007, Strohle et al., 2008) whereas findings in healthy controls suggest that impulsivity is represented by VS hyperactivation (Hariri et al., 2006). Therefore, there has been a discussion about a missing link between VS hypoactivation and impulsivity in ADHD (Carmona et al., 2011, Hoogman et al., 2011, Strohle et al., 2008). Fifth, due to close interactions with the VS as well as impulsivity in ADHD the OFC should be considered as an important candidate (Konrad et al., 2010). Overall, a study with particular focus on reward delivery and the OFC, where alterations from normal functioning can be expected, could help to clarify these issues.

One task that proved particularly useful for the study of ‘pure’ reward delivery (that is, reward delivery processing in the absence of performance or learning aspects) is the card guessing task (Delgado et al., 2004). In this task, participants guess the value of a card and receive feedback about the actual outcome. In high-incentive blocks, they receive monetary feedback (gain/loss of money), in low-incentive blocks they receive non-monetary feedback (correct, incorrect). Thus this task allows the independent modeling of outcome and motivational value/incentive. Increased striatal and lateral as well as medial OFC activations under high incentive conditions have been observed consistently in this task (Delgado et al., 2000, Delgado et al., 2004, May et al., 2004).

Using this task we aimed to extend previous reports of altered reward processing in male and female adults with ADHD by focusing on striatal and orbitofrontal regions. We expected altered reward sensitivity – defined as differential neural responses to reward delivery – in ADHD compared to healthy controls (Luman et al., 2010). This could be evident in altered striatal or orbitofrontal responses to rewarding feedback on either low or high incentive level (pure feedback or monetary outcome, respectively). The direction of these group differences (i.e. hyper or hypo responding in ADHD on either incentive level) is hard to predict. In fact, the electrophysiological results introduced above suggest a more complex pattern where group differences in outcome processing depend on incentive level: ADHD patients showed normal responses to low incentive reinforcers but differed from controls in responding to high incentive outcomes (van Meel et al., 2005, van Meel et al., 2011). In the present design, this might manifest in an incentive × outcome interaction between groups.

Two types of complementary data were acquired to reinforce potential fMRI findings. First, we recorded electrodermal activity concurrently to the card guessing task since reward processing should affect autonomic arousal. Second, we administered two behavioral impulsivity tasks to test whether reward delivery processing indeed correlates with behavioral impulsivity as suggested by current theory (Luman et al., 2010). Furthermore we addressed some common limitations of this kind of research (predominantly small sample sizes and only male participants) by recruiting a large sample that would allow us to assess the role of gender and comorbidity (Biederman et al., 2004, de Zwaan et al., 2011).