Temporal and probabilistic discounting of rewards in children and adolescents: Effects of age and ADHD symptoms

https://doi.org/10.1016/j.neuropsychologia.2005.10.012Get rights and content

Abstract

This study investigated whether age and ADHD symptoms affected choice preferences in children and adolescents when they chose between (1) small immediate rewards and larger delayed rewards and (2) small certain rewards and larger probabilistic uncertain rewards.

A temporal discounting (TD) task and a probabilistic discounting (PD) task were used to measure the degree to which the subjective value of a large reward decreased as one had to wait longer for it (TD), and as the probability of obtaining it decreased (PD). Rewards used were small amounts of money. In the TD task, the large reward (10 cents) was delayed by between 0 and 30 s, and the immediate reward varied in magnitude (0–10 cents). In the PD task, receipt of the large reward (10 cents) varied in likelihood, with probabilities of 0, 0.25, 0.5, 0.75, and 1.0 used, and the certain reward varied in magnitude (0–10 cents).

Age and diagnostic group did not affect the degree of PD of rewards: All participants made choices so that total gains were maximized. As predicted, young children, aged 6–11 years (n = 25) demonstrated steeper TD of rewards than adolescents, aged 12–17 years (n = 21). This effect remained significant even when choosing the immediate reward did not shorten overall task duration. This, together with the lack of interaction between TD task version and age, suggests that steeper discounting in young children is driven by reward immediacy and not by delay aversion. Contrary to our predictions, participants with ADHD (n = 22) did not demonstrate steeper TD of rewards than controls (n = 24).

These results raise the possibility that strong preferences for small immediate rewards in ADHD, as found in previous research, depend on factors such as total maximum gain and the use of fixed versus varied delay durations. The decrease in TD as observed in adolescents compared to children may be related to developmental changes in the (dorsolateral) prefrontal cortex. Future research needs to investigate these possibilities.

Introduction

The cognitive, motivational, and neural pathways underlying attention deficit/hyperactivity disorder (ADHD) remain largely unspecified. One promising candidate mechanism consists of the mesocorticolimbic dopamine system (Johansen, Aase, Meyer, & Sagvolden, 2002; Sonuga-Barke, 2002; Viggiano, Vallone, & Sadile, 2004). This neural circuitry originates in the ventral tegmental area and projects to the striatum, nucleus accumbens, limbic areas, and frontal cortex, and underlies the attribution of salience, a key component in reward and motivation (e.g., Schultz, Dayan, & Montague, 1997). Part of this system, the striatum, has been shown to be active during reward anticipation both in studies of animals and healthy adults (Knutson, Fong, Bennett, Adams, & Hommer, 2003; Pagnoni, Zink, Montague, & Berns, 2002), and to play a role in preferences for small immediate rewards over large delayed rewards (McClure, Laibson, Loewenstein, & Cohen, 2004). Hypofunctioning of the mesolimbic reward circuitry has been proposed to contribute to the development of symptoms of ADHD, and to be associated with altered reinforcement mechanisms in ADHD (e.g., Castellanos & Tannock, 2002; Johansen et al., 2002; Sagvolden, Aase, Zeiner, & Berger, 1998). Indeed, various researchers have hypothesized that a core problem of children with ADHD is an abnormality relating to reward processes (Blum et al., 2000; Castellanos & Tannock, 2002; Douglas, 1999; Douglas & Parry, 1983; Ernst et al., 2003; Haenlein & Caul, 1987; Iaboni, Douglas, & Ditto, 1997; Sagvolden et al., 1998, Sonuga-Barke, 2002, Sonuga-Barke, 2003; Tripp and Alsop, 1999, Tripp and Alsop, 2001; Wender, 1972).

The nature of a supposed deficiency in reward processing in ADHD remains largely unclear for three main reasons. First, there are many different aspects to reward such as magnitude, immediacy, and probability (Williams & Taylor, 2004), and the contribution of each of these aspects to reward sensitivity in ADHD has not been comprehensively studied. Second, little research has focused on the neural basis of reward processing in ADHD (for a recent contribution in this area, see van Meel, Oosterlaan, Heslenfeld, & Sergeant, 2005). Finally, studies of reward in ADHD have used a range of tasks and task manipulations, and findings have been inconsistent (see, for a review, Luman, Oosterlaan, & Sergeant, 2005). The overarching goals of this study were to systematically examine the effects of variations in reward magnitude, immediacy, probability and participant age on preferences for reward in ADHD and matched comparison subjects, beginning with the task that has yielded the most consistent results. Although research on the neural basis of these reward aspects is also needed, the goal of this initial study was purely to look at choice preferences, without measuring brain activation.

Investigators have repeatedly demonstrated that children with ADHD have a greater preference for small immediate compared to larger delayed rewards (Barkley, Edwards, Laneri, Fletcher, & Metevia, 2001; Kuntsi, Oosterlaan, & Stevenson, 2001; Schweitzer & Sulzer-Azaroff, 1995; Solanto et al., 2001; Sonuga-Barke, Taylor, Sembi, & Smith, 1992; Tripp & Alsop, 2001; for a review, see Luman et al., 2005). It should be noted that the specific paradigm and task parameters used in these studies varied. In two studies (Solanto et al., 2001, Sonuga-Barke et al., 1992), the paradigm that was used was identical. Participants were presented with a choice-delay task (CDT) between a small immediate reward (e.g., 5 cents now) and a larger delayed reward (e.g., 10 cents after 30 s) on repeated trials. One limitation of such a fixed-choice design is that it does not give insight into how trade-offs are made between reward magnitude and delay duration, because these variables are not systematically varied. Hence, it is unclear how reward preferences in ADHD may vary as a function of delay durations and reward magnitudes.

By contrast, studies of impulsivity in adults have employed more sophisticated paradigms in which reward magnitude and delay duration are varied in order to obtain a temporal discounting function (e.g., Coffey, Gudleski, Saladin, & Brady, 2003; Crean, de Wit, & Richards, 2000; Green, Meyerson, Lichtman, Rosen, & Fry, 1996; Kirby, Petry, & Bickel, 1999; Kollins, 2003; Petry, 2001, Petry, 2002; Petry & Casarella, 1999; Reynolds, Richards, Horn, & Karraker, 2004; Richards, Zhang, Mitchell, & de Wit, 1999). Temporal discounting (TD) refers to the decrease of subjective reward value as a function of increasing delay (Critchfield & Kollins, 2001; Monterosso & Ainslie, 1999). In one study on adolescents with ADHD and controls (mean age for each group was 15 years), a TD paradigm that allowed for discounting functions was used, with hypothetical choices, i.e., participants were asked to predict their choices over intervals as long as a year and with sums of money that were not really being offered (such as US$ 100 or 1000) (Barkley et al., 2001). This study showed that adolescents with ADHD chose the smaller immediate reward more often than controls when the large delayed reward was US$ 100, but not when the large delayed reward was US$ 1000. While the results of such a hypothetical exercise were informative, they also raise questions of applicability to decisions of daily life.

Thus we decided to employ a TD paradigm in which delay durations and reward magnitudes were systematically varied so that the subjective value of the large delayed reward could be plotted as a function of delay. These TD functions capture the trade-off between reward magnitude and delay, and can be expressed as a single parameter (area under the discounting curve) per individual (Myerson, Green, & Warusawitharana, 2001). Real choices were used in which participants experienced the pre-reward delay on trials on which they chose the delayed reward which consisted of small sums of money that were paid at the end of the task. We predicted that subjects with ADHD would exhibit steeper discounting of delayed rewards than comparison subjects.

There is debate as to whether relatively strong preferences for small immediate rewards reflect hypersensitivity to the immediacy of the reward (Tripp & Alsop, 2001), or, alternatively, a preference for shorter delays (Sonuga-Barke et al., 1992). In order to distinguish between these two explanations for the expected steeper discounting in ADHD, two task versions were used: one in which preferences for immediate rewards led to a reduction in trial duration, and therefore, task duration; one in which choosing the immediate reward did not result in shorter trial durations (see Section 2). We expected that, if reward preferences in ADHD were driven by the reward immediacy, then subjects with ADHD would demonstrate steeper discounting than controls in both versions. If, however, their preferences were driven by aversion to delay, we expected them to show steeper discounting than controls on the version that would lead to shorter task durations, but not on the version in which post-reward delays were inevitable.

We also examined reward preferences in ADHD in a probabilistic discounting paradigm. Probabilistic discounting (PD) refers to the decrease of the subjective value of a reward due to decreasing probability. In PD paradigms, participants are typically presented with choices between a small certain reward and a large probabilistic uncertain reward. While magnitude of the certain reward usually varies, the magnitude of the probabilistic reward is kept constant while its probability level varies. We included a PD paradigm for two reasons: (1) ADHD, and impulsivity in particular, have been associated with a tendency toward greater risk-taking (APA, 1994; Barkley, Murphy, DuPaul, & Bush, 2002; Richards et al., 1999). Therefore, we expected to see less discounting of probabilistic rewards in ADHD; (2) comparisons of temporal and probabilistic discounting have suggested that TD and PD are positively correlated (Myerson, Green, Hanson, Holt, & Estle, 2003; Reynolds, Karraker, Horn, & Richards, 2003; Richards et al., 1999). This means that individuals who demonstrate steeper TD (stronger preferences for small immediate rewards) also demonstrate steeper PD (stronger preferences for small certain rewards). However, the relationship between TD and PD has not been examined in ADHD. Based on the literature that individuals with ADHD demonstrate strong preferences for small immediate rewards and the clinical correlate of predilection towards risk-taking behavior, it can be hypothesized that ADHD should be characterized by steeper TD and weaker PD, i.e., by a negative correlation (Myerson et al., 2003).

A final goal of this study was to investigate age effects on TD and PD of rewards. The ability to wait for a large/preferred delayed reward in preschoolers predicts cognitive and social competence in the teenage years (Mischel, Shoda, & Rodriguez, 1989). However, the development of TD of rewards has not been studied. Similarly, with respect to PD, although the ability to judge probabilities appears to be established by age 6 years (Schlottmann, 2000), little is known about whether younger children are more or less tempted than adolescents to make riskier choices. We sought to address the question of whether younger children would exhibit riskier or more risk-averse choice strategies than adolescents, both of which would result in a lower total gain.

In summary, the four goals of this study were: (1) to investigate whether children and adolescents with ADHD demonstrate steeper TD of rewards than controls; (2) to investigate whether children and adolescents with ADHD differ from controls in the extent to which they would discount less certain rewards; (3) to investigate whether TD and PD correlate positively in controls and negatively in ADHD; (4) to study the effects of age on TD and PD of small monetary rewards.

Section snippets

Recruitment

Participants were recruited from the New York University Child Study Center ADHD Clinical Service, and through notices on our website, University newsletters, and through community resources including local independent schools and not-for-profit organizations. Participants with ADHD and healthy controls were separately recruited, i.e., we specifically looked for participants with symptoms of ADHD and participants who were symptom-free. This was explicit in the ads.

Inclusion criteria

To be assigned to the ADHD

Temporal discounting

The ADHD group did not differ in rate of discounting from the control group for either of the TD tasks (F(1,42) = .69, n.s.; η2 = .02 without post-reward delays; F(1,40) = .07, n.s.; η2 = .00 with post-reward delays) (see Fig. 2, Fig. 3). Younger participants discounted delayed rewards significantly more strongly than adolescents on both versions of the TD task (F(1,42) = 4.3; p < .05; η2 = .09 without post-reward delays; F(1,40) = 6.9, p < .01; η2 = .15 with post-reward delays). Diagnostic and age groups did not

Discussion

Contrary to expectations, we did not find differences between children and adolescents with ADHD and controls in TD of real monetary rewards. We observed a significantly positive correlation between TD and PD in the ADHD group but not in controls. Children (ages 6–11 years) discounted delayed rewards more steeply than adolescents (ages 12–17 years) regardless of diagnosis. There were no age or diagnostic effects on PD.

The lack of steeper TD in ADHD compared to controls raises a number of

Acknowledgements

This work was partially supported by NIMH grant MH0660393. We gratefully acknowledge B.J. Casey, Keith Ditkowski, Monique Ernst, Richard Gallagher, and Mary Solanto for their help with and useful discussions about task development; Harriet de Wit for useful advise; Lori Evans, Vilma Gabbay, Glenn Hirsch, Mark Krushelnycky, Steven Kurtz, Melvin Oatis, and Maurizio Zamebenedetti for assistance in recruiting participants; most of all, the participants and their parents.

References (66)

  • B. Reynolds et al.

    Delay and probability discounting as related to different stages of adolescent smoking and non-smoking

    Behavioural Processes

    (2003)
  • B. Reynolds et al.

    Delay discounting and probability discounting as related to cigarette smoking status in adults

    Behavioural Processes

    (2004)
  • T. Sagvolden et al.

    Altered reinforcement mechanisms in attention-deficit/hyperactivity disorder

    Behavioural Brain Research

    (1998)
  • E.J. Sonuga-Barke

    Psychological heterogeneity in AD/HD—A dual pathway model of behaviour and cognition

    Behavioural Brain Research

    (2002)
  • E.J. Sonuga-Barke

    The dual pathway model of AD/HD: An elaboration of neuro-developmental characteristics

    Neuroscience & Biobehavioral Reviews

    (2003)
  • C.S. van Meel et al.

    Motivational effects on motor timing in attention-deficit/hyperactivity disorder

    Journal of the American Academy of Child & Adolescent Psychiatry

    (2005)
  • T.E. Wilens

    Attention-deficit/hyperactivity disorder and the substance use disorders: The nature of the relationship, subtypes at risk, and treatment issues

    Psychiatric Clinics of North America

    (2004)
  • L.J. Woodward et al.

    Driving outcomes of young people with attentional difficulties in adolescence

    Journal of the American Academy of Child & Adolescent Psychiatry

    (2000)
  • American Psychiatric Association (1994). Diagnostic and Statistical Manual of Mental Disorders. Washington, DC:...
  • R.A. Barkley

    Behavioral inhibition, sustained attention, and executive functions: Constructing a unifying theory of ADHD

    Psychological Bulletin

    (1997)
  • R.A. Barkley et al.

    Executive functioning, temporal discounting, and sense of time in adolescents with attention deficit hyperactivity disorder (ADHD) and oppositional defiant disorder (ODD)

    Journal of Abnormal Child Psychology

    (2001)
  • R.A. Barkley et al.

    Driving-related risks and outcomes of attention deficit hyperactivity disorder in adolescents and young adults: A 3- to 5-year follow-up survey

    Pediatrics

    (1993)
  • R.A. Barkley et al.

    Driving in young adults with attention deficit hyperactivity disorder: Knowledge, performance, adverse outcomes, and the role of executive functioning

    Journal of the International Neuropsychological Society

    (2002)
  • K. Blum et al.

    Reward deficiency syndrome: A biogenetic model for the diagnosis and treatment of impulsive, addictive, and compulsive behaviors

    Journal of Psychoactive Drugs

    (2000)
  • F.X. Castellanos et al.

    Neuroscience of attention-deficit hyperactivity disorder: The search for endophenotypes

    Nature Reviews Neuroscience

    (2002)
  • S.F. Coffey et al.

    Impulsivity and rapid discounting of delayed hypothetical rewards in cocaine-dependent individuals

    Experimental and Clinical Psychopharmacology

    (2003)
  • J.P. Crean et al.

    Reward discounting as a measure of impulsive behavior in a psychiatric outpatient population

    Experimental and Clinical Psychopharmacology

    (2000)
  • T.S. Critchfield et al.

    Temporal discounting: Basic research and the analysis of socially important behavior

    Journal of Applied Behavior Analysis

    (2001)
  • E.A. Crone et al.

    Developmental changes in real life decision making: Performance on a gambling task previously shown to depend on the ventromedial prefrontal cortex

    Developmental Neuropsychology

    (2004)
  • V.I. Douglas

    Cognitive control processes in attention-deficit/hyperactivity disorder

  • V.I. Douglas et al.

    Effects of reward on delayed reaction time task performance of hyperactive children

    Journal of Abnormal Child Psychology

    (1983)
  • V.I. Douglas et al.

    Effects of reward and nonreward on frustration and attention in attention deficit disorder

    Journal of Abnormal Child Psychology

    (1994)
  • M. Ernst et al.

    Decision making in adolescents with behavior disorders and adults with substance abuse

    The American Journal of Psychiatry

    (2003)
  • Cited by (250)

    View all citing articles on Scopus
    View full text