J Abnorm Child Psychol DOI 10.1007/s10802-015-0016-9
Feedback May Harm: Role of Feedback in Probabilistic Decision Making of Adolescents with ADHD Yehuda Pollak 1,2 & Rachel Shoham 3,4
# Springer Science+Business Media New York 2015
Abstract Inept probabilistic decision making is commonly associated with ADHD. In experimental designs aimed to model probabilistic decision making in ADHD, feedback following each choice was, in the majority of studies, part of the paradigm. This study examined whether feedback processing plays a role in the maladaptive choice behavior of subjects with ADHD by comparing feedback and no-feedback conditions. Sixty adolescents (49 males), ages 13–18, with and without ADHD, performed a descriptive probabilistic choice task in which outcomes and probabilities were explicitly provided. Subjects performed the task either with or without feedback. Under the no-feedback condition, adolescents with ADHD and controls performed similarly, whereas under the feedback condition, subjects with ADHD chose the unfavorable outcomes more frequently and risked smaller sums than controls. These finding demonstrate the crucial role of feedback in the decision making of adolescents with ADHD. Keywords Attention Deficit Hyperactivity Disorder . Adolescents . Decision making . Risk taking . Feedback Poor decision making under uncertainty is commonly associated with ADHD (Barkley 2006). Supporting this claim are * Yehuda Pollak [email protected] 1
School of Education, The Hebrew University of Jerusalem, Mount Scopus, Jerusalem 91950, Israel
2
Neuropediatric Unit, Shaare Zedek Medical Center, Jerusalem, Israel
3
Department of Psychology, The Hebrew University of Jerusalem, Jerusalem, Israel
4
Talpiot College, Holon, Israel
studies reporting that individuals with ADHD are more likely to engage in risky behaviors such as dangerous driving, substance abuse and gambling (Barkley 2002a; Barkley and Fischer 2010; Faregh and Derevensky 2010; Nigg 2013; Rosenbloom and Wultz 2011; Wilens 2011). Probabilistic decision making in ADHD has become the focus of a substantial amount of research, using tasks in which subjects choose among different options, each associated with different probabilities of winning or losing. Several studies examined the performance of individuals with ADHD on the Iowa Gambling Test (IGT; Bechara et al. 1994). In the IGT, the participant draws cards from several decks associated with different probabilities of winning or losing; the probabilities are unknown to the participant at the beginning of the task. Importantly, successful performance depends upon learning the contingencies of each deck. Several studies using the original IGT or adapted version for children showed that subjects with ADHD were less successful on the task, with positive correlations between disadvantageous choices and measures of impulsivity and hyperactivity (Garon et al. 2006; Luman et al. 2008; Malloy-Diniz et al. 2007; Toplak et al. 2005; but see Ernst et al. 2003; Geurts et al. 2006; Masunami et al. 2009). These findings suggest that ADHD is associated with different processes of decision making under uncertainty. However, the IGT involves both learning the contingencies of each option and choosing between them. Since individuals with ADHD choose unfavorable and risky options more often than controls, their behavior might reflect inefficient learning processes rather than different preferences of probabilistic options. To avoid the learning issue, a different risk taking paradigm consisting of choices where outcomes and probabilities are explicitly described has been used to assess decision making in ADHD. Drechsler et al. (2008) used the Game of Dice Task (GDT), in which subjects implement decisions on a
J Abnorm Child Psychol
computerized dice game, during which they choose among four possible outcomes: two high risk choices, associated with high gains and a low probability to win, and two low risk choices, associated with small gains and a high probability to win. In the GDT, gains and probabilities are presented explicitly, and the game is played twice. Results showed similar performance of children with and without ADHD the first time the game was played. However, when playing a second time, children with ADHD made more high risk choices than controls. In a subsequent study, Drechsler et al. (2010) constructed a task that matches the principles of the GDT, and showed that children with ADHD more frequently chose less likely option but with larger rewards. Using a probabilistic discounting task, Scheres et al. (2006) measured the degree to which the subjective value of a large reward decreased as the probability of obtaining it decreased. Subjects had to choose between a certain amount of money and smaller amount of money that varied in likelihood. Amounts and likelihoods were presented to the subjects explicitly. The authors did not find differences between children with ADHD and controls on that explicit probabilistic task, and suggested that this unexpected result may be a function of methodological issues that hindered real effects. Another task probing probabilistic decision making used to assess risk taking in ADHD is the Cambridge Gambling Task (CGT; Rogers et al. 1999) which measures risk taking through betting behavior. Subjects are presented with explicit probabilities and asked to choose the more likely option and then to determine the magnitude of bet they are willing to risk given the probabilities. Using the CGT, DeVito et al. (2008) showed that subjects with ADHD more frequently chose the less likely option and on average risked similar wagers as did the control group. Poor choices were correlated with reports of disruptive behavior. In our study of probabilistic decision making in adolescents with ADHD using a modified version of the CGT (Kroyzer et al. 2014), we found that adolescents with ADHD, regardless of subtype, gender, comorbid psychiatric conditions and intelligence, chose the less likely option more often and risked smaller bets. No differences between groups were recorded in deliberation time, but shorter deliberation time was correlated with higher quality choices (i.e., choosing the more likely outcome), indicating that choice differences were not due to careless behavior. No differences between groups were found with regard to the rate of adjusting wagers to outcome probability, so it was clear that both groups understood the meaning of probabilities. Notably, as the chances to win in the gambling are always higher than the chances to lose, the higher the bet the more profitable the choice is likely to be, and therefore, subjects with ADHD’s cautious betting choices were less profitable. Further analysis revealed that the tendency of the ADHD group to choose smaller bets was greater when the preceding trial
ended in loss, rather than a win, suggesting that difference in choice behavior depends on different processing of serial feedback. In contrast to experience-based tasks in which learning is mandatory in order to make reasonable decisions, in description-based tasks, probabilities and outcomes are explicitly provided. In all the studies cited above in which subjects were given series of explicit choice trials, they were also provided with feedback after each trial. Basically, this feedback was chosen according to the probabilities assigned to a chosen option, and was independent of subsequent trials. One might expect that this feedback should not affect subjects’ subsequent choices. However, it has consistently been demonstrated that people alter their choice patterns when feedback is provided, even in description-based tasks (i.e., mixed description-experience tasks). For example, Brand et al. (2009) reported that feedback interacted with intelligence and strategy application to improve performance on the GDT. It was summarized that on a repeated gambling task, feedback is a critical component during the decision making process, even in the presence of fully specified descriptive information (Jessup et al. 2008). For individuals with ADHD, there is evidence showing impairments in basic aspects of feedback processing. For instance, several studies suggested that rewards are less salient in ADHD, resulting in reduced event-related potentials, cortical activation and heart rate changes (Crone et al. 2003; Herrmann et al. 2009; Rubia et al. 2009). Such difficulties in feedback processing might interfere with integrating the information provided by description and experience. Hence, in cases where feedback is not necessary for optimizing decision making, the presence of feedback might impair, rather than improve, decision making by individuals with ADHD. Co-morbid psychiatric conditions of ADHD have long been associated with compromised probabilistic decision making, as indicated by studies of decision making in populations with anxiety and behavioral disorders (Geurts et al. 2006; Hobson et al. 2011; Kim et al. 2006; Luman et al. 2010a). These findings follow more classical studies which applied Gray’s theory of brain function (Gray 1991) into child psychiatry. According to Gray, a behavioral activation system (BAS) underlies responding to conditioned stimuli for reward, and a behavioral inhibition system (BIS) responds to conditioned stimuli for punishment. Quay (1988), in a seminal paper, linked undersocialized aggressive conduct disorder to an excessive response of the BAS, ADHD to an underactive BIS, and anxiety disorder to an overactive BIS. Importantly, studies aimed to examine Quay’s hypotheses used tasks eliciting probabilistic decision making (i.e., Bdoor opening task^) (Daugherty et al. 1993; Matthys et al. 1998; O’Brien and Frick 1996; O’Brien et al. 1994). As anxiety and behavioral disorders affect probabilistic decision making and often co-exist in individuals with
J Abnorm Child Psychol
ADHD (Young 2008), their existence in the study sample was tracked and controlled for, and their relationship to decision making was evaluated. Given the uncertain function of feedback on risk taking in adolescents with ADHD, the purpose of this study was to test the hypothesis that flawed probabilistic decision making in ADHD is due to differences in feedback processing. This goal was achieved by manipulating the presence of feedback in a repeated description-based gambling task and controlling for co-morbid conditions. We hypothesized that if feedback is present subjects with ADHD would show less favorable probabilistic decision making and bet lower sums, whereas if feedback is absent no differences between groups would be observed.
Method The study was approved by the Shaare Zedek Medical Center Institutional Review Board for research on human subjects. Written informed consent was obtained from parents and assent was obtained from adolescents.
Participants Thirty one adolescents with ADHD, ages 13–18, were recruited through the outpatient ADHD clinic of the Neuropediatric Unit in Shaare Zedek Medical Center. A control group comprised of 31 subjects matched by age, gender, IQ, parents’ years of education and socioeconomic level (people per room) was recruited (see Table 1). Inclusion criterion for the ADHD group was a diagnosis made by a child neurologist or psychiatrist and confirmed by the Schedule for Affective Disorders and Schizophrenia for School-Age Children-Present and Lifetime Version (K-SADS-PL; Kaufman et al. 1997). History of ADHD was ruled out for controls using the KSADS-PL. Exclusion criteria were history of a neurological illness (i.e., epilepsy, cerebral palsy), head injury, pervasive developmental disorder, psychosis, or major depression, IQ< 70, and attending special education schools. One subject from the study group was excluded due to a diagnosis of major depression and one of the controls who had been diagnosed with ADHD. Twenty-three children from the experimental group were taking stimulants, and were instructed not to take it on the day of the experiment. No subject was taking nonstimulants. Based on Jessup et al. (2008) who found a large effect size for feedback in a decision making from description task (Cohen’s f=0.41), a power analysis for a between-subject ANCOVA was conducted. According to this analysis, a total sample size of 60 participants should reveal an effect of f=0.37 with an alpha level of 0.05 and power of 0.8.
Procedure The ADHD group subjects were diagnosed by an experienced child neurologist or psychiatrist according to DSM-IV criteria (American Psychiatric Association [APA] 1994). Both groups underwent the following assessment procedures: confirmation by a psychologist of (1) ADHD diagnosis or its absence, (2) co-morbid Axis I psychiatric disorders using the Schedule for Affective Disorders and Schizophrenia for School-Age Children-Present and Lifetime Version (K-SADS-PL; Kaufman et al. 1997). For the psychiatric interview only parent report was utilized. Severity of ADHD symptoms was assessed using the Disruptive Behavior Disorder Rating Scale (DBDRS; Pelham et al. 1992) without the conduct disorder scale. The conduct disorder scale was omitted due to the low rate of conduct disorder in the population treated at the unit, and in order to enhance parent cooperation. Intelligence was assessed using the short version of the Wechsler Intelligence Scale for Children – Revised (WISC-R; Sattler 1988). The newer Hebrew version of the WISC was not available in Israel at the time the study was conducted. To measure the level of anxiety, participants filled out the following questionnaires: The Revised Children’s Manifest Anxiety Scale (RCMAS; Reynolds and Richmond 1979) and the Screen for Child Anxiety Related Emotional Disorders (SCARED; Birmaher et al. 1997). At the end of the experiment, participants received a monetary prize, each point won valued at about 2 Eurocents. Families of subjects received $25 compensation for travel expenses, etc. Decision Making Task Participants performed a modified version of the CambridgeGambling Task (CGT; Kroyzer et al. 2014; Shalev et al. 2012). Subjects were told that they could earn about 20 NIS (about 4 Euros) by playing a gambling game. Next they were told that the computer has randomly hidden a yellow token inside one of ten red or blue boxes arrayed on the screen, representing one of the following odds: 9:1, 8:2, 7:3 and 6:4 (Fig. 1). The subject first guessed whether the token is hidden under a red or a blue box, and then decided how many points to gamble (options from 10 to100 in steps of 10 were presented on the screen simultaneously). Immediately after a bet was made, a feedback message appeared, reporting the results of the bet (Win/Lose) accompanied by a corresponding sound. This feedback was set to correspond to the ratio presented on the trial (e.g., if the red:blue ratio was 3:7, the probability that the token was under a red box was 0.3). If the subject chose the correct color, the bet placed was added to his score; if the choice was wrong, the bet was subtracted. The subject was instructed to treat the points as being valuable and to accumulate as many as possible during the test. The experiment began with a practice phase in which the experimenter ascertained
J Abnorm Child Psychol Table 1 Demographic and clinical characteristics by diagnostic group Mean age (SD) Gender Mean IQa (SD) Mean parent’s years of education (SD) Mean of socioeconomic, people per room (SD) Subtypes of ADHD diagnosis Inattentive type Hyperactive type Combined type Comorbid diagnosisb (n) Behavior disorders (ODD, CD) Anxiety disorder DBDRS Mean inattention score (SD) Mean hyperactivity score (SD) Mean ODD score (SD) Anxiety scales RCMAS SCARED
ADHD (n=30)
Controls (n=30)
Group comparison
14.9 (1.5) 20 males 105.1 (10.1) 15.1 (2.2) 1.0 (0.3)
15.2 (1.9) 19 males 108.9 (14.5) 14.5 (1.9) 1.1 (0.2)
t(58) =0.7 (p=0.50) χ2(1) =0.7 (p=0.78) t(58) =1.2 (p=0.24) t(56) =−1.2 (p=0.25) t(57) =0.2 (p=0.84)
7 7
0 1
χ2(1) =7.9 (p=0.005) χ2(1) =5.2 (p=0.023)
15.0 (4.9) 9.6 (6.6) 8.0 (5.9)
3.5 (2.9) 2.4 (3.7) 4.1 (3.8)
t(58) =−11.1 (p
Feedback May Harm: Role of Feedback in Probabilistic Decision Making of Adolescents with ADHD Yehuda Pollak 1,2 & Rachel Shoham 3,4
# Springer Science+Business Media New York 2015
Abstract Inept probabilistic decision making is commonly associated with ADHD. In experimental designs aimed to model probabilistic decision making in ADHD, feedback following each choice was, in the majority of studies, part of the paradigm. This study examined whether feedback processing plays a role in the maladaptive choice behavior of subjects with ADHD by comparing feedback and no-feedback conditions. Sixty adolescents (49 males), ages 13–18, with and without ADHD, performed a descriptive probabilistic choice task in which outcomes and probabilities were explicitly provided. Subjects performed the task either with or without feedback. Under the no-feedback condition, adolescents with ADHD and controls performed similarly, whereas under the feedback condition, subjects with ADHD chose the unfavorable outcomes more frequently and risked smaller sums than controls. These finding demonstrate the crucial role of feedback in the decision making of adolescents with ADHD. Keywords Attention Deficit Hyperactivity Disorder . Adolescents . Decision making . Risk taking . Feedback Poor decision making under uncertainty is commonly associated with ADHD (Barkley 2006). Supporting this claim are * Yehuda Pollak [email protected] 1
School of Education, The Hebrew University of Jerusalem, Mount Scopus, Jerusalem 91950, Israel
2
Neuropediatric Unit, Shaare Zedek Medical Center, Jerusalem, Israel
3
Department of Psychology, The Hebrew University of Jerusalem, Jerusalem, Israel
4
Talpiot College, Holon, Israel
studies reporting that individuals with ADHD are more likely to engage in risky behaviors such as dangerous driving, substance abuse and gambling (Barkley 2002a; Barkley and Fischer 2010; Faregh and Derevensky 2010; Nigg 2013; Rosenbloom and Wultz 2011; Wilens 2011). Probabilistic decision making in ADHD has become the focus of a substantial amount of research, using tasks in which subjects choose among different options, each associated with different probabilities of winning or losing. Several studies examined the performance of individuals with ADHD on the Iowa Gambling Test (IGT; Bechara et al. 1994). In the IGT, the participant draws cards from several decks associated with different probabilities of winning or losing; the probabilities are unknown to the participant at the beginning of the task. Importantly, successful performance depends upon learning the contingencies of each deck. Several studies using the original IGT or adapted version for children showed that subjects with ADHD were less successful on the task, with positive correlations between disadvantageous choices and measures of impulsivity and hyperactivity (Garon et al. 2006; Luman et al. 2008; Malloy-Diniz et al. 2007; Toplak et al. 2005; but see Ernst et al. 2003; Geurts et al. 2006; Masunami et al. 2009). These findings suggest that ADHD is associated with different processes of decision making under uncertainty. However, the IGT involves both learning the contingencies of each option and choosing between them. Since individuals with ADHD choose unfavorable and risky options more often than controls, their behavior might reflect inefficient learning processes rather than different preferences of probabilistic options. To avoid the learning issue, a different risk taking paradigm consisting of choices where outcomes and probabilities are explicitly described has been used to assess decision making in ADHD. Drechsler et al. (2008) used the Game of Dice Task (GDT), in which subjects implement decisions on a
J Abnorm Child Psychol
computerized dice game, during which they choose among four possible outcomes: two high risk choices, associated with high gains and a low probability to win, and two low risk choices, associated with small gains and a high probability to win. In the GDT, gains and probabilities are presented explicitly, and the game is played twice. Results showed similar performance of children with and without ADHD the first time the game was played. However, when playing a second time, children with ADHD made more high risk choices than controls. In a subsequent study, Drechsler et al. (2010) constructed a task that matches the principles of the GDT, and showed that children with ADHD more frequently chose less likely option but with larger rewards. Using a probabilistic discounting task, Scheres et al. (2006) measured the degree to which the subjective value of a large reward decreased as the probability of obtaining it decreased. Subjects had to choose between a certain amount of money and smaller amount of money that varied in likelihood. Amounts and likelihoods were presented to the subjects explicitly. The authors did not find differences between children with ADHD and controls on that explicit probabilistic task, and suggested that this unexpected result may be a function of methodological issues that hindered real effects. Another task probing probabilistic decision making used to assess risk taking in ADHD is the Cambridge Gambling Task (CGT; Rogers et al. 1999) which measures risk taking through betting behavior. Subjects are presented with explicit probabilities and asked to choose the more likely option and then to determine the magnitude of bet they are willing to risk given the probabilities. Using the CGT, DeVito et al. (2008) showed that subjects with ADHD more frequently chose the less likely option and on average risked similar wagers as did the control group. Poor choices were correlated with reports of disruptive behavior. In our study of probabilistic decision making in adolescents with ADHD using a modified version of the CGT (Kroyzer et al. 2014), we found that adolescents with ADHD, regardless of subtype, gender, comorbid psychiatric conditions and intelligence, chose the less likely option more often and risked smaller bets. No differences between groups were recorded in deliberation time, but shorter deliberation time was correlated with higher quality choices (i.e., choosing the more likely outcome), indicating that choice differences were not due to careless behavior. No differences between groups were found with regard to the rate of adjusting wagers to outcome probability, so it was clear that both groups understood the meaning of probabilities. Notably, as the chances to win in the gambling are always higher than the chances to lose, the higher the bet the more profitable the choice is likely to be, and therefore, subjects with ADHD’s cautious betting choices were less profitable. Further analysis revealed that the tendency of the ADHD group to choose smaller bets was greater when the preceding trial
ended in loss, rather than a win, suggesting that difference in choice behavior depends on different processing of serial feedback. In contrast to experience-based tasks in which learning is mandatory in order to make reasonable decisions, in description-based tasks, probabilities and outcomes are explicitly provided. In all the studies cited above in which subjects were given series of explicit choice trials, they were also provided with feedback after each trial. Basically, this feedback was chosen according to the probabilities assigned to a chosen option, and was independent of subsequent trials. One might expect that this feedback should not affect subjects’ subsequent choices. However, it has consistently been demonstrated that people alter their choice patterns when feedback is provided, even in description-based tasks (i.e., mixed description-experience tasks). For example, Brand et al. (2009) reported that feedback interacted with intelligence and strategy application to improve performance on the GDT. It was summarized that on a repeated gambling task, feedback is a critical component during the decision making process, even in the presence of fully specified descriptive information (Jessup et al. 2008). For individuals with ADHD, there is evidence showing impairments in basic aspects of feedback processing. For instance, several studies suggested that rewards are less salient in ADHD, resulting in reduced event-related potentials, cortical activation and heart rate changes (Crone et al. 2003; Herrmann et al. 2009; Rubia et al. 2009). Such difficulties in feedback processing might interfere with integrating the information provided by description and experience. Hence, in cases where feedback is not necessary for optimizing decision making, the presence of feedback might impair, rather than improve, decision making by individuals with ADHD. Co-morbid psychiatric conditions of ADHD have long been associated with compromised probabilistic decision making, as indicated by studies of decision making in populations with anxiety and behavioral disorders (Geurts et al. 2006; Hobson et al. 2011; Kim et al. 2006; Luman et al. 2010a). These findings follow more classical studies which applied Gray’s theory of brain function (Gray 1991) into child psychiatry. According to Gray, a behavioral activation system (BAS) underlies responding to conditioned stimuli for reward, and a behavioral inhibition system (BIS) responds to conditioned stimuli for punishment. Quay (1988), in a seminal paper, linked undersocialized aggressive conduct disorder to an excessive response of the BAS, ADHD to an underactive BIS, and anxiety disorder to an overactive BIS. Importantly, studies aimed to examine Quay’s hypotheses used tasks eliciting probabilistic decision making (i.e., Bdoor opening task^) (Daugherty et al. 1993; Matthys et al. 1998; O’Brien and Frick 1996; O’Brien et al. 1994). As anxiety and behavioral disorders affect probabilistic decision making and often co-exist in individuals with
J Abnorm Child Psychol
ADHD (Young 2008), their existence in the study sample was tracked and controlled for, and their relationship to decision making was evaluated. Given the uncertain function of feedback on risk taking in adolescents with ADHD, the purpose of this study was to test the hypothesis that flawed probabilistic decision making in ADHD is due to differences in feedback processing. This goal was achieved by manipulating the presence of feedback in a repeated description-based gambling task and controlling for co-morbid conditions. We hypothesized that if feedback is present subjects with ADHD would show less favorable probabilistic decision making and bet lower sums, whereas if feedback is absent no differences between groups would be observed.
Method The study was approved by the Shaare Zedek Medical Center Institutional Review Board for research on human subjects. Written informed consent was obtained from parents and assent was obtained from adolescents.
Participants Thirty one adolescents with ADHD, ages 13–18, were recruited through the outpatient ADHD clinic of the Neuropediatric Unit in Shaare Zedek Medical Center. A control group comprised of 31 subjects matched by age, gender, IQ, parents’ years of education and socioeconomic level (people per room) was recruited (see Table 1). Inclusion criterion for the ADHD group was a diagnosis made by a child neurologist or psychiatrist and confirmed by the Schedule for Affective Disorders and Schizophrenia for School-Age Children-Present and Lifetime Version (K-SADS-PL; Kaufman et al. 1997). History of ADHD was ruled out for controls using the KSADS-PL. Exclusion criteria were history of a neurological illness (i.e., epilepsy, cerebral palsy), head injury, pervasive developmental disorder, psychosis, or major depression, IQ< 70, and attending special education schools. One subject from the study group was excluded due to a diagnosis of major depression and one of the controls who had been diagnosed with ADHD. Twenty-three children from the experimental group were taking stimulants, and were instructed not to take it on the day of the experiment. No subject was taking nonstimulants. Based on Jessup et al. (2008) who found a large effect size for feedback in a decision making from description task (Cohen’s f=0.41), a power analysis for a between-subject ANCOVA was conducted. According to this analysis, a total sample size of 60 participants should reveal an effect of f=0.37 with an alpha level of 0.05 and power of 0.8.
Procedure The ADHD group subjects were diagnosed by an experienced child neurologist or psychiatrist according to DSM-IV criteria (American Psychiatric Association [APA] 1994). Both groups underwent the following assessment procedures: confirmation by a psychologist of (1) ADHD diagnosis or its absence, (2) co-morbid Axis I psychiatric disorders using the Schedule for Affective Disorders and Schizophrenia for School-Age Children-Present and Lifetime Version (K-SADS-PL; Kaufman et al. 1997). For the psychiatric interview only parent report was utilized. Severity of ADHD symptoms was assessed using the Disruptive Behavior Disorder Rating Scale (DBDRS; Pelham et al. 1992) without the conduct disorder scale. The conduct disorder scale was omitted due to the low rate of conduct disorder in the population treated at the unit, and in order to enhance parent cooperation. Intelligence was assessed using the short version of the Wechsler Intelligence Scale for Children – Revised (WISC-R; Sattler 1988). The newer Hebrew version of the WISC was not available in Israel at the time the study was conducted. To measure the level of anxiety, participants filled out the following questionnaires: The Revised Children’s Manifest Anxiety Scale (RCMAS; Reynolds and Richmond 1979) and the Screen for Child Anxiety Related Emotional Disorders (SCARED; Birmaher et al. 1997). At the end of the experiment, participants received a monetary prize, each point won valued at about 2 Eurocents. Families of subjects received $25 compensation for travel expenses, etc. Decision Making Task Participants performed a modified version of the CambridgeGambling Task (CGT; Kroyzer et al. 2014; Shalev et al. 2012). Subjects were told that they could earn about 20 NIS (about 4 Euros) by playing a gambling game. Next they were told that the computer has randomly hidden a yellow token inside one of ten red or blue boxes arrayed on the screen, representing one of the following odds: 9:1, 8:2, 7:3 and 6:4 (Fig. 1). The subject first guessed whether the token is hidden under a red or a blue box, and then decided how many points to gamble (options from 10 to100 in steps of 10 were presented on the screen simultaneously). Immediately after a bet was made, a feedback message appeared, reporting the results of the bet (Win/Lose) accompanied by a corresponding sound. This feedback was set to correspond to the ratio presented on the trial (e.g., if the red:blue ratio was 3:7, the probability that the token was under a red box was 0.3). If the subject chose the correct color, the bet placed was added to his score; if the choice was wrong, the bet was subtracted. The subject was instructed to treat the points as being valuable and to accumulate as many as possible during the test. The experiment began with a practice phase in which the experimenter ascertained
J Abnorm Child Psychol Table 1 Demographic and clinical characteristics by diagnostic group Mean age (SD) Gender Mean IQa (SD) Mean parent’s years of education (SD) Mean of socioeconomic, people per room (SD) Subtypes of ADHD diagnosis Inattentive type Hyperactive type Combined type Comorbid diagnosisb (n) Behavior disorders (ODD, CD) Anxiety disorder DBDRS Mean inattention score (SD) Mean hyperactivity score (SD) Mean ODD score (SD) Anxiety scales RCMAS SCARED
ADHD (n=30)
Controls (n=30)
Group comparison
14.9 (1.5) 20 males 105.1 (10.1) 15.1 (2.2) 1.0 (0.3)
15.2 (1.9) 19 males 108.9 (14.5) 14.5 (1.9) 1.1 (0.2)
t(58) =0.7 (p=0.50) χ2(1) =0.7 (p=0.78) t(58) =1.2 (p=0.24) t(56) =−1.2 (p=0.25) t(57) =0.2 (p=0.84)
7 7
0 1
χ2(1) =7.9 (p=0.005) χ2(1) =5.2 (p=0.023)
15.0 (4.9) 9.6 (6.6) 8.0 (5.9)
3.5 (2.9) 2.4 (3.7) 4.1 (3.8)
t(58) =−11.1 (p
