It’s a matter of timing: Delay of reward and resistance to
peer influence in adolescents with mild to borderline
intellectual disabilities.
Name Maaike Vertregt
Student number 5958415
Date September 7, 2014
Supervisor Mw. dr. A. M. L. Collot d’Escury-Koenigs Second assessor Mw. dr. E. Salemink
1 Abstract
This study investigated the ability to delay a reward and resistance to peer influence in adolescents with and without mild to borderline intellectual disability (MBID). Furthermore, we assessed cool executive functioning (EF) and studied its predictive value with respect to delay of reward. To this end, 89 adolescents performed a go/no go task as a measure of cool EF and a newly developed delay-of-reward task with and without peer influence. Results indicated that adolescents with MBID are impaired in the ability to delay a reward and in cool EF. However, the effect of peers and the relation between cool EF and delay of reward remain unclear. Exploratory analyses revealed that the effect of waiting may be different for
adolescents with MBID than for typically developing adolescents. Suggestions for future studies are discussed.
Keywords: ADOLESCENTS, DELAY OF REWARD, GO/NO GO TASK, INTELLECTUAL DISABILITY, PEER INFLUENCE
2 Introduction
Adolescents with mild to borderline intellectual disability (MBID) experience more health problems (Chen, Lawlor, Duggan, Hardy, & Eaton, 2006), they encounter more problems at school (Fernell & Ek, 2010), they are overrepresented in the criminal justice system (Collot d’Escury, 2007) and when using drugs, they abuse drugs more often compared to typically developing (TD) adolescents (Chapman & Wu, 2012). Thus, suffering from MBID seems to be a risk factor for a variety of negative outcomes in life.
Vice versa, a determinant predicting positive outcomes in life is the ability to delay a reward: children who are able to delay a reward turn out to be more healthy, perform better at school, are cognitively more competent and use less drugs (Mischel, Shoda, & Rodriguez, 1989; Moffit et al., 2010). The ability to delay a reward is ascribed to executive functions (EF) such as self-regulation and inhibition (Olson, Hooper, Collins, & Luciana, 2007; Weatherly & Ferraro, 2011). EF refers to inhibiting inappropriate responses, self-regulation and planning of behavior (Goswami, 2008, p. 295). MBID is reported to coincide with a lag in EF (Willner, Bailey, Parry, & Dymond, 2010a; Carretti, Belacchi, & Cornoldi, 2010;
Bexkens, Ruzzano, Collot d’Escury-Koenigs, Van der Molen, & Huizenga, 2014). Consequently a hampered ability to delay a reward is to be expected in adolescents with MBID. However, research shows inconsistent results (Bexkens et al., 2014). Hence, the aim of this study is to assess the ability to delay a reward in adolescents with MBID.
Adolescent behavior typically takes place in a social environment with peers, especially where various challenging high reward behaviors (drinking alcohol, using drugs and having sex) are concerned. Adolescents are more susceptible to peer influence (Crone & Dahl, 2012) due to a highly active dopamine system (Wahlstrom, White, & Luciana, 2010) that facilitates approach tendencies to rewards (Enter, Colzato, & Roelofs, 2012). This hyperactivity of the reward system also seems to deteriorate the ability to delay a reward when peers are present (O’Brien, Albert, Chein, & Steinberg, 2011). Furthermore, adolescents with MBID tend to be more susceptible to peer influence than typically developing adolescents (Bexkens et al.,
3
2013a). Therefore, the influence of peers on delay of reward is taken into account in this study.
A theoretical framework to explain the ability to delay a reward was offered by
Metcalfe and Mischel (1999). They proposed a hot/cool distinction in information processing: hot information processing refers to reflexive, fast and emotional information processing and cool information processing refers to reflective, slow and cognitive information processing. When the cool system is powerful enough to control the hot system, people are able to delay a reward. More recently, Zelazo and Müller (2002) introduced hot and cool EF. Although hot EF and hot information processing are overlapping constructs they also differ: hot EF refers to top down processes whereas hot information processing refers to bottom up information processing (Zelazo & Carlson, 2012). In this study we will use the hot/cool system and hot/cool EF interchangeably, since both hot EF and hot information processing favor preference for a small immediate reward and both cool EF and cool information processing favor preference for a larger delayed reward.
As mentioned, empirical findings with respect to the ability to delay a reward in individuals with intellectual disabilities show mixed results: three studies indicated that individuals with intellectual disability were less able to delay a reward compared to TD developing individuals (Eisenhouwer, Baker, & Blacher, 2007; McIntyre, Blacher, & Baker, 2006; Willner, Bailey, Parry, & Dymond, 2010b) but two other studies did not find such a difference (Rourke & Quinlan, 1973; Koolhof, Loeber, Wei, Pardini, & Collot d’Escury, 2007, see for delay-of-reward task Bexkens et al., 2014). These mixed results may have been caused by several methodological issues (Bexkens et al., 2014). Sometimes not only delay, but also probability was measured (Willner et al., 2010b; Koolhof et al., 2007), possibly resulting in impure measures of the ability to delay a reward. In one study, the reward was presented the next day (Rourke & Quinlan, 1973). Distrust in the experimenter may have been a confound in this study (Kid, Palmeri & Aslin, 2013). Furthermore, participants’ ages varied between 8 and 14 year. Since the ability to delay a reward increases from childhood to
4
adolescence (Scheres et al., 2006; Olson et al., 2007), age differences may have increased variability of the outcomes, thereby concealing existing effects.
In sum, assessments of the ability to delay a reward in individuals with intellectual disability have been confounded by impure tasks, timing of the reward and a large age range. Therefore this study will assess the ability to delay a reward with a newly developed delay-of-reward task without a probability component, the delay-of-rewards will be handed immediately and adolescents of a restricted age range will be studied. Since peers influence adolescent behavior, we will compare adolescents with MBID to TD adolescents in a delay-of-reward task with and without peers, resulting in a two by two design. Furthermore, cool EF as
assessed by a go/no go task and the relation of cool EF to the ability to delay a reward will be studied.
Key questions
First, we want to replicate the finding that the cool system is impaired in adolescents with MBID. The second question is whether adolescents with MBID have more difficulty in delaying a reward than TD adolescents. Third, we want to replicate the finding that
adolescents are less able to delay a reward when peers are present (O’Brien et al., 2011). The fourth research question is whether adolescents with MBID are more susceptible to peer influence than TD adolescents in their ability to resist the temptation of a reward. Finally, we want to replicate the findings that cool system functioning predicts the ability to delay a reward (Olson et al., 2007).
The first hypothesis is that the cool system is impaired in adolescents with MBID compared to TD adolescents. Since the cool system in adolescents with MBID is expected to control the hot system less well, the second hypothesis is that adolescents with MBID have more difficulty delaying a reward than TD adolescents. Motivational cues may activate the hot system. Therefore the third hypothesis is that all adolescents will delay rewards less often when peers are present compared to when peers are absent. Adolescents with MBID seem to
5
be more susceptible to peer influence than TD adolescents (Bexkens et al., 2013a). Therefore, the fourth hypothesis is that adolescent with MBID perform worse on the delay-of-reward task with peer influence compared to TD adolescents. The fifth hypothesis is that better cool EF predicts longer waiting times on a delay-of-reward task.
Methods Participants
Based on a power analysis using G*Power 3 (Faul, Erdfelder, Lang, & Buchner, 2007), 89 adolescents aged 13 to 15 years were recruited (Mage = 14.12, SD = 0.65, 51 males). This
age range was chosen because peer influence is strongest around 14 years (Steinberg & Monahan, 2007) and because adolescents in this age range have more problems to delay a reward (Steinberg et al., 2009).
Forty seven adolescents with MBID (Mage = 14.25, SD = 0.65, 28 males) were recruited
from special vocational education (‘praktijkonderwijs’). Four schools for special vocational education in different towns participated. Special vocational education features strict admittance criteria consisting of IQ-scores between 60 and 80 assessed by the Regional Referral Committee (RVC)(praktijkonderwijs, n.d.). In adolescents at special vocational education, behavioral problems do not predominate. Adolescents with severe, persistent behavioral problems are referred to other specialized education (‘cluster 4 education’).
Forty two adolescents without MBID were recruited (Mage = 13.99, SD = 0.63, 23
males). The adolescents attended classes for pre-vocational1 education (VMBO-t, n = 11), senior general secondary education (HAVO, n = 9), pre-university education
(VWO/gymnasium, n = 14) or a combination of the latter two (HAVO/VWO, n = 8), all in the same school. Although no intelligence scores were available for this group, admittance
criteria for VMBO-t are intelligence scores between 97 and 104 (as measured by the Dutch
1 Note that pre-vocational education refers to regular education and special vocational education refers to
6
Intelligence Test for Educational Level, NIO) or equivalent CITO-scores. The NIO and CITO scores correlate with r = .78 (van Dijk & Tellegen, 2004). Therefore we were confident that this group was not intellectually impaired. At general secondary education and pre-university education admittance criteria are even more stringent.
Four participant characteristics were assessed: age, socio-economic status, the value they attached to money (the variable ‘money value’) (see table 1) and gender. Socio-economic status of adolescents with MBID is likely to be lower than of TD adolescents (Emerson & Hatton, 2007). Perceived lower SES may make people choose smaller immediate rewards (Liu, Feng, Suo, Lee, & Li, 2012) therefore this characteristic may present a possible confound. A second possible confound is ‘money value’: adolescents who value money highly, may be motivated more to wait for money or less able to resist quick money.
Age, gender and ‘money value’ did not differ significantly between groups. In the analysis of SES, four participants were excluded since no data were available: three
adolescents from the MBID-group did not report an existing address and one SES-value was missing in the group. SES was lower for participants in the MBID group than the TD-group. This difference was significant, t(78.49) = 4.8, p < .001, but SES did not predict delay of reward times or amount of money earned. The correlation between SES and ‘money value’ was not significant, neither in the total sample nor in subgroups (MBID-TD).
Measures Go/no go task
A go/no go task was used as a measure of cool EF. The go/no go task measures cognitive control (Eigsti et al., 2006). First a fixation dot was presented for 500 ms in the center of the screen after which a stimulus was presented for 500 ms. Stimuli consisted of the letters “A” to “F” in the go trials (80% of the trials) and an “X” in the no/go trials (20% of the trials). Then the screen was empty for 1000 ms after which the new trial started (see figure 1). Participants were required to respond as quickly and as accurate as possible to a non X by
7
pressing the spacebar with the index finger of the dominant hand. The dependent variable was the amount of responses to an “X” (i.e. commission errors). Two practice sessions ensured that the participant understood the task. The practice sessions consisted of 20 trials each. In the practice trials a sound indicated an error: a response to an “X” or a non-response to a non “X” . The task itself consisted of two times 120 trials with a short break (at least 8 seconds) in between. The task started self-paced. Total duration of the task was approximately 10
minutes.
Delay-of-reward task
The delay-of-reward task consisted of a newly developed computer task2 with a solo and a peer version. Methodological problems of prior research were omitted: the task measured delay of reward without using probabilities and rewards were given immediately after testing. To increase motivational salience, real rewards were used. Real waiting times were used since they make it more difficult to delay a reward (Scheres et al., 2006). Waiting times of 10-13 seconds did not differentiate between groups (Scheres et al.; Koolhof et al., 2007) therefore longer waiting times were used in the original version and prolonged after a pilot study (see appendix 1 for description of the pilot study). Two practice trials were used to ensure that the participant understood the task. A tray filled with coins was on the table, visible for the participant. In the solo condition, a 1-cent and a 10-cent picture appeared on either side of the screen (randomized within participants) with a red rectangle beneath (see figure 2, left picture). When the rectangle turned green, the coin could be collected by clicking on the corresponding button: ‘z’ for the left coin, ‘/’ for the right coin. The buttons were marked by white or yellow stickers. The 1-cent was available after 2 seconds (see figure 2, right picture), the 10-cent after 10 to 53.5 seconds in random order. The total score was visible on the computer screen. Between all trials a fixation cross appeared in the middle of the screen for 500 ms. There were 30 trials in both the solo and the peer condition.
8 Peer influence
The delay-of-reward task in the peer condition was adapted from the study by Bexkens et al. (2013a). Pictures of three same sex peers were presented on the screen. Audio files presented same sex adolescent voices encouraging the participant every 8 seconds to take the 1 cent, for example: “Take the 1 cent!” or “Don’t wait so long”.
Money value
To assess the extent to which participants appreciated money, a Likert scale was used. All participants were asked the following question (the ‘money value’ question): ”What grade suits you with respect to the extent to which you appreciate money on a scale of 1 to 10 when 1 means that money means nothing to you and 10 means that money is super important to you?” A Likert scale has been shown reliable and valid to use with adolescents with MBID (Hartley & MacLean, 2006).
Socio-Economic Status (SES)
An indication of Socio-Economic Status (SES) was assessed by asking the four numbers of the postal code area of the home address of the participant (statusscores, n.d.). This yields a rough estimate of SES of the participant (Deonondan, Campbell, Ostbye, Tummon, & Robertson, 2000).
Procedure
Approval was obtained of the ethics committee of the University of Amsterdam and schools were informed about the study. Participating schools sent information to the parents. At schools for adolescents with MBID (‘praktijkonderwijs’), passive consent was used: parents and students could refuse to participate. At the regular high school, active consent was
9
used3: participants had to hand in an informed consent form signed by the parents. After consent was obtained (either active or passive), participants were tested. All tasks were performed on DELL latitude E5510 laptops with a 15.6 inch screen. The test took place in a quiet room in school in one individual session of 30 to 60 minutes depending on waiting times in the delay-of-reward task. At vocational education participants were tested during all types of lessons, at regular education participants were only tested during homework hour. Before the tasks started, participants were asked for their birth date and for the four digits of the postal code of their home address. They were not informed about the amount of trials they were going to perform in either task. Then participants started with the delay-of-reward task, either the solo version or the peer version, counterbalanced between participants. After the first delay-of-reward task, the participants performed the go/no go task, followed by the other version of the delay-of-reward task (resp. peer or solo). The tasks were explained using a protocol by a test assistant (the researcher or a collaborator) who sat quietly in the room during the test. Participants were informed they were not allowed to communicate with the test assistant during the tasks. After the tasks, participants were asked how important money was for them (the ‘money value’ question) and they were asked not to tell their classmates how much they earned. Finally, the participants received the money they earned and a small present.
Results
First, missing data in the delay-of-reward task were marked. Missing data represented participants who waited too long to press a key in one trial in the delay-of-reward task. Due to a technical problem, four data points of the go/no go task were missing in a practice trial. They were marked as missing values. The data were checked for outliers: no extreme values were found for both the delay-of-reward task and the go/no go task, therefore: all participants (N = 89) were included in the analyses.
10
Prior to the analyses the effect of task order on delay times was studied with a one way ANOVA. There was no significant effect of task order on delay of reward, F(1, 87) = 0.62, p = .44, partial η2 = .07, therefore groups were collapsed over the two task orders.
The dependent variables, delay of reward time and amount of commission errors, were not normally distributed. The amount of commission errors deviated only slightly from normality due to some high scores in the group of MBID-adolescents. This resulted in a right tailed distribution. The dependent variable ‘delay of reward’ deviated from normality in both groups. This was due to a ceiling effect of the delay-of-reward task, especially in the group of TD adolescents. Since the dependent variables were not normally distributed, in the analyses that were performed to generalize to the larger population, bias corrected and accelerated (BCa) 95% confidence intervals were computed using a bootstrapping procedure.
Confirmatory analyses
In line with the expectations, adolescents with MBID made more commission errors in the go/no go task than TD adolescents: an independent t-test showed that on average,
adolescents with MBID pressed the “X” more often (M = 16.81, S.E. = 1.06) than TD adolescents (M = 11.71, S.E. = 0.79). This difference, 5.09, BCa 95% CI [2.34, 7.84], was significant t (87) = 3.68, p < .001. It represents a large effect size, d = 0.78.
A two group (MBID-TD) repeated measures (solo vs peer) mixed model ANOVA was performed to test the main effects of peer, of group and the interaction effect on delay of reward. Since the data of the delay-of-reward task were not normally distributed, it would be preferable to use a non-parametric test or to use a bootstrapping procedure. Unfortunately, a non-parametric counterpart of the mixed model ANOVA does not exist, transforming the data was unsuccessful and the bootstrapping procedure for the mixed model ANOVA was not available. Therefore, the parametric ANOVA was performed first to test for the interaction effect (note that the result should be interpreted cautiously). After that, the main effect for peers (within effect) was tested using a paired samples t-test with a bootstrapping procedure
11
and the main effect for group (between effect) was tested with an independent t-test with a bootstrapping procedure. Since three analyses were performed, hypotheses were tested at α = 0.02.
Contrary to the expectations, the mixed model ANOVA (see figure 3) showed no significant interaction effect: the effect of peer influence was the same for MBID and TD adolescents, F(1, 87) = .67, p = .41, partial η2 = .01. Also contrary to the expectations, the dependent samples t-test showed no main effect of peers. On average, participants waited longer when peers were present (Mpeerdelay =23.51 sec, S.E. = 1.15) as compared to when peers
were absent (Msolodelay = 21.67 s, S.E. = 1.26). This difference, 1.83 s, BCa 95% CI [3.74,
-0.05], was not significant, t(88) = -1.96, p = .05, d = -0.16. In line with the expectations, the independent samples t-test showed a significant main effect for group: MBID adolescents waited shorter for a reward (Mdelay = 19.00 s, S.E. = 1.56) than TD adolescents (Mdelay = 26.38
s, S.E. = 1.28). This difference, -7.39 s, BCa 95% CI [-11.28, -3.41], was significant, t(84.68) = 3.62, p = .001, d = 0.76. It represents a large effect size.
Contrary to the expectations, a regression analysis indicated that the amount of commission errors on the go/no go task did not predict performance on the delay-of-reward task in the total sample and in the group of TD adolescents (see table 2). The explained variance in these models was zero, indicating that there was no effect. Analysis in adolescents with MBID indicated a trend opposite to the expectations: adolescents with MBID who made more commission errors in the go/no go task tended to wait longer in the delay-of-reward task—not shorter—see figure 4 for the results. The amount of explained variance in this model was 7%. This represents a medium effect. However, when the analyses were repeated without the participant with MBID with the highest score excluded, the effect became non-significant.
12 Exploratory analyses
Since three comparisons were performed on the same data—both for the delay-of-reward data and the go/no go data—, Bonferroni corrected alpha levels were set at α = .02 for the exploratory analyses.
Learning curves
To explore learning curves in the go/no go task, a mixed model ANOVA was used. It was studied what type of learning curve the adolescents showed. This was tested with a polynomial trend (i.e. a linear trend) for time. Main interest was whether learning curves differed for adolescents with MBID and TD adolescents, therefore we aimed to find an interaction effect between time and group. To this end, percentage correct responses were computed for six blocks of trials including the two blocks of 20 practice trials (1-20, 21-40, 41-100, 101-160, 161-220, 221-280). Analyses revealed that learning curves for both MBID and TD adolescents were quadratic and not linear. The quadratic trend was significant, F(1,87) = 9.76, p = .002, partial η2 = .10. The learning curve for adolescents with MBID and TD adolescents was the same: there was no significant interaction effect between learning curve and group (MBID/TD), see figure 5 for the results. Since the data were not linear, the analyses were repeated from block two to block six. Now the relation was linear, F(1,87) = 19.22, p < .001, partial η2 = .18. There was no interaction effect between time and group: the learning curve was the same for adolescents with MBID and TD adolescents.
Next, learning curves in the delay-of-reward task were studied. First the average delay time in every ten trials (in chronological order) was computed, resulting in six blocks. We studied the type of learning curve and whether learning was the same for adolescents with MBID and TD adolescents. The mixed model ANOVA revealed that the learning curve was cubic and not linear. The cubic trend was significant, F(1,87) = 15.61, p < .001, partial η2 = .15. In the second block there was a trend towards an interaction effect: adolescents with MBID tended to wait shorter worse over time, whereas TD adolescents did not, F(2,174) =
13
2.43, p = .09, partial η2 = .05 (see figure 6). Contrasts revealed that this interaction effect appeared between block 5 to block 6, F (1, 87) = 5.61, p = .02, partial η2 = .06. This finding must be interpreted cautiously, since assumptions were violated.
Effect of waiting times
The final exploratory analysis studied the effect of short and long waiting times on performance on the subsequent trial. Waiting times of 10 to 13.5 s were defined as short waiting times, waiting times of 40 to 53.5 s were defined as long waiting times. The
percentage 10 cent gained after long waiting trials (in which 10 cent was collected) and after short waiting trials was computed. Participants without any variation in reaction (either 100% 10 cent or 100% 1 cent) were excluded from the analyses resulting in 69 participants for these analysis (NMBID = 42 and NTD= 27). Then a mixed model ANOVA was performed with
percentage 10 cent gained after long and short waiting times in the previous trial as within factor and group (MBID-TD) as between factor. There was a slight trend towards an interaction: when they had to wait long, performance of adolescents with MBID seemed to deteriorate more compared to TD adolescents, F(1, 67) = 2.64, p = .11, partial η2 = .04, see table 4 and figure 7 for the results. However, this result must be interpreted cautiously since the dependent variable, percentage of trials after respectively short and long waiting times, was not normally distributed.
Discussion
We studied cool system functioning and the ability to delay a reward in adolescents with MBID. We expected that a better developed cool system would exert more power over the hot system thus enabling TD adolescents to delay a reward longer. We anticipated that the cool system would be less developed in adolescents with MBID thus deteriorating their ability to delay a reward. Since motivational cues—especially peer presence—activate the hot
14
system, we predicted that all adolescents would be less able to delay a reward when peers were present and we expected this effect to be stronger for adolescents with MBID. Our findings partly supported the theory: the cool system seemed to be impaired in adolescents with MBID and they delayed rewards less well than TD adolescents. These findings were robust and showed a large effect. This was so, even though the delay-of-reward task showed a ceiling effect, especially in the group of TD adolescents. However, the relation between cool system functioning and the ability to delay a reward was not as expected: adolescents with a better functioning cool system did not delay rewards longer. Although their cool system was better developed, it did not exert more power over the hot system. Also unexpected, peer presence did not seem to activate the hot system: adolescents delayed a reward to the same extent with and without peers.
We were not able to replicate the finding of Olson and colleagues (2007) that cool system functioning as assessed by a go/no go task predicted performance on a delay-of-reward task. This null finding was also reported by Eigsti and colleagues (2006). They
suggested that delay does not only depend on EF, but also on motivation to obtain the delayed rewards. This idea is also expressed in ‘resource depletion’(Huizenga, Van der Molen,
Bexkens, & Van den Wildenberg, 2013): motivation drops when resources of EF are depleted. It is possible that in our study the ambiguity about the length of the task—participants did not know in advance how many trials the task contained—depleted resources of adolescents with MBID more than of TD adolescents. This would explain why performance dropped radically in the final block of the delay task. Future studies could compare the ability to delay a reward in tasks with and without ambiguity to test this idea.
Another factor that may influence the ability to delay a reward is future orientation. This was studied by Peters and Büchel (2010). Before participants performed a delay of reward task, the experimenters collected personal information of the participants about what activities they planned when they would receive the reward. People who were reminded of this personal information, delayed rewards longer than people who were not reminded of this information.
15
It is possible that adolescents with MBID think less about future gains thus deteriorating their ability to delay a reward. It is for future studies to evaluate this possibility.
Peers did not influence adolescents in the delay-of-reward task. This contradicts the findings of the two other studies that examined the influence of peers on a delay-of-reward task (O’Brien et al., 2011; Weigard, Chein, Albert, Smith, & Steinberg, 2014). An
explanation is that the peer version of the delay task in the current study may be susceptible to improvement. First, it is possible that the 1 cent was too little to increase the susceptibility to rewards: the studies that found peer influence on delay of reward used $ 200,- as the
minimum—hypothetical—reward (O’Brien et al.; Weigard et al.). Perhaps we should have used 5 cents as an alternative for the 10 cent. Second, adolescents reported that the voices they heard every 8 seconds shortened perceived waiting times, so perhaps the comments should be given only once. Furthermore, the comments may have focused at social exclusion, instead of just encouraging adolescents to take the 1 cent, since social exclusion enhances peer influence (Falk et al., 2014). Finally, in the current study, adolescents were not made to believe they were being watched by an existing peer while doing the task. This was the case in the studies that found influence of peers in delay of reward (O’Brien et al.; Weigard et al.). This possibly resulted in a manipulation that was too weak. However, the manipulation used in the current study was effective in a risk taking task with TD and MBID adolescents of the same age (Bexkens, Huizenga, Neville, Bredman, Collot d’Escury-Koenigs, & Van der Molen, 2013). The third null finding—that the effect of peers was the same for adolescents with MBID and TD adolescents—seems to be a logical consequence of the lack of peer influence.
Apart from the perhaps suboptimal operationalization of peer influence, we want to mention two other limitations of this study. First, although we studied the influence of SES, we measured it only roughly. For a better assessment of SES, individual characteristics such as parental occupation and education should be included (Deonondan et al., 2000). Second, we had no information on comorbid disorders and on medication of the participating
16
adolescents. With this information perhaps we could have defined subgroups within the MBID group who were and were not able to delay a reward.
The finding that adolescents with MBID show deficits in cool EF—as measured by the go/no go task— is consistent with the literature (Bexkens et al., 2014). With respect to delay of reward the literature is, as mentioned, inconsistent. Our study indicates that this
inconsistency may be related to the time frames that were used: in our pilot study, adolescents with MBID seemed to be able to delay a reward when waiting times were on average 7 seconds shorter than in the final task. Furthermore, exploring the effect of waiting times on subsequent trial, we found that long waiting times may have a more detrimental effect on the ability to delay a reward in adolescents with MBID than in TD adolescents. This possibly explains why the studies by McIntyre and colleagues (2006) and Eisenhouwer and colleagues (2007) yielded large effect sizes—they used long waiting times of 3 minutes—and why the study by Koolhof and colleagues (2007) did not find and effect—they used waiting times of 10 seconds—.
Our conclusion is that it is a matter of timing whether adolescents with MBID are able to delay a reward. This implies for instance that we cannot expect from adolescents with MBID to wait very long in the class room before they surrender to the reward of peer chatting, that they will surrender to the reward of stealing more often than TD adolescents when they do not have enough money.
In sum, adolescents with MBID show cool EF deficits and an impairment to delay a reward. However, the circumstances under which they are impaired in delaying a reward, the effects of timing, and the relation to cool EF remain unclear until now. Studies to
motivational aspects in delay of reward, the impact of depletion and future orientation may increase knowledge. Furthermore, the influence of peers in the ability to delay a reward needs clarification. Adolescence is a time of rapid growth and flexibility (Crone & Dahl, 2012; Blakemore & Choudhury, 2006), therefore improvement may be possible especially in
17
opportunity to design interventions that enable adolescents with MBID to delay rewards, giving them better outcomes in life.
18 References
Bexkens, A., Huizenga, H. M., Neville, D. A., Bredman, J., Collot d’Escury-Koenigs, A. M. L., & Van der Molen, M. W. (2013). Peer-influence effects on risk-taking in male adolescents with mild-to-borderline intellectual disabilities and/r behavior disorder. Manuscript in preparation.
Bexkens, A., Ruzzano, L., Collot d’Escury-Koenigs, A. M. L., Van der Molen, M. W., & Huizenga, H. M. (2014). Inhibition deficits in individuals with intellectual disability: a meta-regression analysis. Journal of Intellectual Disability Research, 58, 3-16.
Bexkens, A., Huizenga, H. M., Neville, D. A., Bredman, J., Collot d’Escury-Koenigs, A. M. L., & Van der Molen, M. W. (2013a). Peer-influence effects on risk-taking in male adolescents with mild-to-borderline intellectual disabilities and/or behavior disorder. Manuscript in preparation.
Blakemore, S. J., & Choudhury, S. (2006). Development of the adolescent brain: implications for executive function and social cognition. Journal of Child Psychology and
Psychiatry, 47, 296-312/
Carretti, B., Belacchi, C., & Cornoldi, C. (2010). Difficulties in working memory updating in individuals with intellectual disability. Journal of Intellectual Disability Research, 54, 337-345.
Chapman, S. L. C., Wu, L. T. (2012). Substance abuse among individuals with intellectual disabilities. Research in Developmental Disabilities, 33, 1147-1156.
Chen, C. Y., Lawlor, J. P., Duggan, A. K., Hardy, J. B., Eaton, W. W. (2006). Mild cognitive impairment in early life and mental health problems in adulthood. The American Journal of Public Health, 96, p. 1772.
Cohen, J. (1992). A power primer. Psychological Bulletin, 112, 155-159.
Collot d’Escury, A. M. (2007). Lopen jongeren met een lichte verstandelijke beperking meer kans om in aanraking te komen met justitie? Kind en Adolescent, 28, 128-137/
19
engagement and goal flexibility. Nature reviews. Neuroscience, 13, 636-650.
Deonondan, R., Campbell, K., Ostbye, T., Tummon, I., & Robertson, J. (2000). A comparison of methods for measuring socio-economic status by occupation or postal area. Chronic Diseases in Canada, 21 – 3.
Eigsti, I. M., Zayas, V., Mischel, W., Shoda, Y., Ayduk, O., Dadlani, M. B., Davidson, M. C., Aber, J. L., & Casey, J. (2006). Predicting cognitive control from preschool to late adolescence and young adulthood. Psychological Science, 17, 478-484.
Eisenhouwer, A. S., Baker, B., & Blacher, J. (2007). Early student-teacher relationships of children with and without intellectual disability: Contributions of behavioral, social, and self-regulatory competence. Journal of School Psychology, 45, 363-383.
Emerson, E., & Hatton, C. (2007). Contribution of socioeconomic position to health inequalities of Britisch children and adolescents with intellectual disabilities. American Journal on Mental Retardation, 112, 140-150.
Enter, D., Colzato, L. S., & Roelofs, K. (2012). Dopamine transporter polymorphisms affect social approach-avoidance tendencies. Genes, Brain and Behavior, 11, 671-676.
Falk, E. B., Cascio, C. N., O’Donnell, M. B., Carp, J., Tinney, F. J., Bingham, C. R., Shope, J. T., Ouimet, M. C., Pradhan, A. K., & Simons-Morton, B. G. (2014). Neural responses to exclusion predict susceptibility to social influence. Journal of Adolescent Health, 54, s22-31.
Faul, F., Erdfelder, E., Lang, A. G., & Buchner, A. (2007). G*Power 3: A flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behavior Research Methods, 39, 175-191.
Fernell, E., & Ek, U. (2010). Borderline intellectual functioning in children and adolescents – insufficiently recognized difficulties. Acta Paediatrica, 99, 748-753.
Goswami, U. (2008). Cognitive Development. The Learning Brain. New York: Psychology Press.
20
type scales for people with intellectual disability. Journal of Intellectual Disability Research, 50, 813-827.
Huizenga, H. M., Van der Molen, M. W., Bexkens, A., & Van den Wildenberg, W. P. M. (2013). Formal models of “resource depletion”. The Behavioral and Brain Sciences, 36, 694-695.
Kidd, C., Palmeri, H., & Aslin, R. N. (2013). Rational snacking: Young children’s decision- making on the marshmallow task is moderated by beliefs about environmental
reliability. Cognition, 126, 109-114.
Koolhof, R., Loeber, R., Wei, E. H., Pardini, D., & Collot d’Escury, A. (2007). Inhibition deficits of serious delinquent boys of low intelligence. Criminal Behaviour and Mental Health, 17, 274-292.
Liu, L., Feng, T., Suo, T., Lee, K., & Li, H. B. (2012). Adapting to the destitute situations: Poverty cues lead to short-term choice (poverty cues and immediate gratification). PloS ONE, 7, e33950.
McIntyre, L. L., Blacher, J., & Baker, B. L. (2006). The transition to school: adaptation in Young children with and without intellectual disability. Journal of Intellectual Disability Research, 50, 349-361.
Metcalfe J., & Mischel, W. (1999). A hot/cool-system analysis of delay of gratification: Dynamics of willpower. Psychological review, 106, 3-19.
Mischel, W., Shoda, Y., & Rodriguez, M. L. (1989). Delay of gratification in children. Science, 244, 933-938.
Moffit, T. E., Arsenault, L., Belsky, D., Dickson, N., Hancox, R. J., Harrington, H., Houts, R., Poulton, R., Roberts, B. W., Ross, S., Sears, M. R., Murray, W., & Caspi, A. (2011). A gradient of childhood self-control predicts health, wealth, and public safety.
Proceedings of the National Academy of Sciences of the United States, 108, p. 2693. O’Brien, L., Albert, D., Chein, J., & Steinberg, L. (2011). Adolescents prefer more immediate
21 747-753.
Olson, E., A., Hooper, C. J., Collins, P., & Luciana, M. (2007). Adolescents’ performance on delay and probability discounting tasks: Contributions of age, intelligence, executive functioning, and self-reported externalizing behavior. Personality and Individual Differences, 43, 1886-1897.
Peters, J., & Büchel, C. (2010). Episodic future thinking reduces reward delay discounting through an enhancement of prefrontal-mediotemporal interactions. Neuron, 66, 138- 148.
Praktijkonderwijs (n.d.). Retrieved at March 13, 2014 from
http://www.steunpuntdyslexie.nl/dyslexie-op-school/voortgezet-onderwijs/praktijkonderwijs/
Rourke, Ph. G., & Quinlan, D. M. (1973). Psychological characteristics of problem children at the borderline of mental retardation. Journal of Consulting and Clinical Psychology, 40, 59-68.
Scheres, A., Dijkstra, M., Ainslie, E., Balkan, J., Reynolds, B., Sonuga-Barke, E., &
Castellanos, F. X. (2006). Temporal and probabilistic discounting of rewards in children and adolescents: Effect of age and ADHD symptoms. Neuropsychologia, 44, 2092- 2103.
Statusscores (n.d.). Retrieved at June 29, 2014, from
http://www.scp.nl/Onderzoek/Lopend_onderzoek/A_Z_alle_lopende_onderzoeken/Sta
tusscores
Steinberg, L., & Monahan, K. C. (2007). Age differences in resistance to peer influence. Developmental Psychology, 43, 1531-1543.
Steinberg, L., Graham, S., O’Brien, L., Woolard, J., Cauffman, E., & Banich, M. (2009). Age differences in future orientation and delay discounting. Child Development, 80, 28-44. Van Dijk, H. & Tellegen, P. J. (2004). NIO. Nederlandse Intelligentietest voor
22
Wahlstrom, D., White, T., & Luciana, M. (2010). Neurobehavioral evidence for changes in dopamine system activity during adolescence. Neuroscience and Biobehavioral Reviews, 34, 631-648.
Weatherly, J. N., & Ferraro, F. R. (2011). Executive functioning and delay discounting of four different outcomes in university students. Personality and Individual Differences, 51, 183-187.
Weigard, A., Chein, J., Albert, D., Smith, A., & Steinberg, L. (2014). Effects of anonymous peer observation on adolescents’ preference for immediate rewards. Developmental Science, 17, 71- 78.
Willner, P., Bailey, R., Parry, R., & Dymond, S. (2010a). Evaluation of executive functioning in people with intellectual disabilities. Journal of Intellectual Disability Research, 54, 366-379.
Willner, P., Bailey, R., Parry, R., & Dymond, S. (2010b). Performance in temporal
discounting tasks by people with intellectual disabilities reveals difficulties in decision- making and impulse control. American Journal on Intellectual and Developmental Disabilities, 115, 157-171.
Zelazo, P. D., & Carlson, S. M. (2012). Hot and cool executive function in childhood and adolescence: Development and plasticity. Child Development Perspectives,6 , 354-360. Zelazo, P. D., & Müller, U. (2002). Executive functions in typical and atypical development.
In U. Goswami (Ed.), Handbook of Childhood Cognitive Development (pp. 445–469). Oxford, UK: Blackwell.
23 Table 1
Participant characteristics in each group.
MBID TD (n = 47) (n = 42) ______________ ______________ M S.D. M S.D. Age in months 170.98 7.84 167.83 7.55 Money value 6.66 2.43 6.86 1.34 SES*) 0.16 1.07 1.25 0.78
Note. MBID = Mild to Borderline Intellectual Disability, SES = Social Economic Status, *) for SES n = 44 (MBID) and n = 41 (TD).
24 Table 2
Linear model of predictor ‘amount of commission errors’ on delay of reward times in the total group and in the subgroups, with 95% bias corrected and accelerated confidence intervals reported in parentheses. Confidence intervals and standard errors based on 1000 bootstrap samples b S.E. B β p Total sample (n=89) Constant 22065.28 2409.53 (17195.45, 26810.75) .000 CE 28.88 143.05 (-268.09, 301.10) 0.02 .86 MBID (n=47) Constant 12635.03 3550.34 (6134.42, 20006.46) .002 CE 378.64 173.18 (14.83, 681.15) 0.26 .03† TD (n=42) Constant 26317.25 3026.19 (19549.98, 32672.96) .001 CE 5.16 239.12 (-468.93, 409.31) 0.00 .99
Note. R2 = 0.00 in the model of the total sample and the TD model, R2 = 0.07 for the MBID-model, CE = amount of commission errors in the go/no go task. † = borderline significant
25 Table 3
Average amount of money in cents (standard deviation between parentheses) earned in the pilot study for the original version with short waiting times (3 – 64.5 seconds) and for the final version with long waiting times (10 – 53.5seconds).
MBID TD
(n = 6) (n = 0)
Original task (short waiting times) 491.67 (136.76) -
(n = 4) (n = 2)
Final task (longer waiting times) 262.25 (109.13) 311.50 (109.13)
Table 4
Average percentage of 10 cents collected after short and long waiting times (standard deviations between parentheses)
MBID TD
(n = 41) (n = 29)
Perc 10 after short waiting times 79.89 (18.98) 90.85 (16.53) Perc 10 after long waiting times 63.93 (36.75) 86.19 (19.53)
26 Figure 1. Sequence of events in the go/no go task.
Figure 2. Computer screens of the delay-of-reward of reward task, Solo condition. Left: no coins available. Right: 1 cent is available and 10 cent not.
27
Figure 3. Average waiting times of the solo and peer version of the delay-of-reward task for both adolescents with MBID and TD adolescents.
Figure 4. Scatterplot with amount of commission errors on the x-axis and the average total delay per trial on the y-axis for adolescents with MBID and TD adolescents.
28
Figure 5. Average percentage correct responses in the go/no go task for MBID and TD adolescents in six blocks of trials (1-20, 21-40, 41-100, 101-160, 161-220 and 221-280)
Figure 6. Average waiting times in milliseconds (ms) in six blocks of ten trials in the delay-of-reward task in chronological order for MBID and TD adolescents. Note: between block 3 and 4 the go/no go task was performed.
29
Figure 7. Effect of waiting times on consecutive items. On the x-axis short waiting (1) and long waiting (2). On the y-axis the percentage 10 cent collected after resp. short and long waiting times in the previous trial.
30 Appendix 1
Pilot study
A pilot study was performed to test the newly developed delay-of-reward task. The main goal was to test the task on floor and ceiling effects.
Methods Participants
For the pilot study, a separate sample was used. This sample was recruited in the same schools as the main study, see participants section of the main study for more information on type of education. Twelve participants were recruited: 10 adolescents with MBID (Mage =
14.23, SD = 0.51, 4 males) and 2 TD adolescents (Mage = 13.92, SD = 0.24, 1 male). Only two
TD adolescents were recruited for the pilot study since it was difficult to find time to test TD adolescents.
Measure
Delay-of-reward task
For a full description of the delay-of-reward task, see the measures section of the main study. The original version of the delay-of-reward task was identical to the final version but for the waiting times: in the original version waiting times varied from 3 to 46.5 seconds for the 10 cent. For the pilot study, the dependent variable in the delay-of-reward task was the total amount of money earned.
Procedure
See procedure section of the main study for the procedure. In the pilot study, two aspects deviated from the main study: participants were not asked to keep secret how much they earned and the go/no go task that was performed in between the two versions of the task (peer and solo) consisted of less items (120 in total).
31 Results
When 6 adolescents with MBID had performed the delay-of-reward task, a ceiling effect appeared (see for test results table 3). The average amount of money that was earned by adolescents with MBID was in the upper decile ( they earned 91.05% of the maximum
amount of money). Since it was expected that TD adolescents would be able to wait longer and thus earn more, these results indicated a ceiling effect, therefore waiting times for the 10 cents were prolonged to 10 to 53.5 seconds. Another problem was that adolescents
extensively discussed the amount of money they earned in class. This resulted in excited adolescents who were eager to participate in the study, thus reducing the independency of the data. Therefore the protocol was changed and participants were asked not to inform
classmates about the amount of money they earned. After waiting times were prolonged and the protocol was changed, 6 more adolescents were tested (4 with MBID and 2 TD). Results showed a reduction of the ceiling effect (see table 3): adolescents with MBID now earned 49% of the maximum amount of money.
Discussion
In the pilot study, the delay-of-reward task was tested on floor and ceiling effects. With the prolonged waiting times, it seemed to measure delay of reward without floor or ceiling effects. Peer influence was not studied, since the expected effect was too small to be detected in this small sample. Furthermore, the operationalization of peer influence had been effective before (Bexkens et al., 2013).
