Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Conditional Learning Deficits in Children with ADHD can be Reduced Through Reward Optimization and Response-Specific Reinforcement

Conditional Learning Deficits in Children with ADHD can be Reduced Through Reward Optimization... When children with ADHD are presented with behavioral choices, they struggle more than Typically Developing [TD] children to take into account contextual information necessary for making adaptive choices. The challenge presented by this type of behavioral decision making can be operationalized as a Conditional Discrimination Learning [CDL] task. We previously showed that CDL is impaired in children with ADHD. The present study explores whether this impairment can be remediated by increasing reward for correct responding or by reinforcing correct conditional choice behavior with situationally specific outcomes (Differential Outcomes). An arbitrary Delayed Matching-To-Sample [aDMTS] procedure was used, in which children had to learn to select the correct response given the sample stimulus presented (CDL). We compared children with ADHD (N = 45) and TD children (N = 49) on a baseline aDMTS task and sequentially adapted the aDMTS task so that correct choice behavior was rewarded with a more potent reinforcer (reward manipulation) or with sample-specific (and hence response- specific) reinforcers (Differential Outcomes manipulation). At baseline, children with ADHD performed significantly worse than TD children. Both manipulations (reward optimization and Die ff rential Outcomes) improved performance in the ADHD group, resulting in a similar level of performance to the TD group. Increasing the reward value or the response-specificity of reinforcement enhances Conditional Discrimination Learning in children with ADHD. These behavioral techniques may be effective in promoting the learning of adaptive behavioral choices in children with ADHD. Keywords Attention Deficit Hyper activity Disorder · Differential Outcomes · Conditional Discrimination Learning · Reward · Treatment Background * Hasse De Meyer hasse.demeyer@kuleuven.be ADHD is marked by elevated levels of inattention, * Saskia van der Oord hyperactivity and impulsiveness that are inconsistent saskia.vanderoord@kuleuven.be with a child’s developmental stage (American Psychiatric Gail Tripp Association, 2013). As a result of these symptoms, children tripp@oist.jp with ADHD often fail to conform to the expected social, Tom Beckers cognitive and emotional requirements of their environment, tom.beckers@kuleuven.be increasing their risk of adverse outcomes later in Behaviour, Health and Psychopathology, KU Leuven, development (Wehmeier et al., 2010; Willcutt et al., 2005). Tiensestraat 102, 3000 Leuven, Belgium A critical skill for adaptive socio-emotional and cognitive Department of Psychology, Faculty of Behavioural Sciences, functioning is the ability to align one’s actions with the HELP University, Subang 2, Persiaran Cakerawala Seksyen, frequently changing expectations or requirements of U4, 40150 Shah Alam, Selangor, Malaysia the environment (e.g., being quiet and deferential in the Human Developmental Neurobiology Unit, Okinawa classroom, being energetic and assertive in the playground) Institute of Science and Technology Graduate University, (Urcuioli, 2005). Numerous researchers have observed that Okinawa 904-0495 1919-1 Tancha, Onna, Japan children with ADHD have difficulty adapting their behavior Leuven Brain Institute, KU Leuven, Leuven, Belgium to shifting environmental demands (Nigg & Casey, 2005; Developmental Psychology, University of Amsterdam, Sagvolden et al., 2005). Nieuwe Achtergracht 129, 1018 Amsterdam, Netherlands Vol.:(0123456789) 1 3 1166 Research on Child and Adolescent Psychopathology (2021) 49:1165–1178 A widely used paradigm for testing behavioral adaptation deficits linked to executive impairments (e.g., in working to environmental expectations in laboratory research is a memory) in children with ADHD (Dovis et al., 2012; Fosco Conditional Discrimination Learning [CDL] task (Martínez et  al.,  2015; Luman et  al.,  2005; Slusarek et  al.,  2001). et al., 2012; Mok et al., 2017). Conditional Discrimination Whether such an incentive-oriented manipulation would Learning can be tested in an arbitrary Matching-To-Sample also be effective in improving CDL performance is to be [aMTS] procedure in which associations have to be learned determined. between non-similar, non-related sample stimuli and choice Inspiration for a more cognitively oriented way to responses (Estévez et al., 2001; Trapold, 1970). On a given remediate impairments in adaptive choice behavior in trial participants are presented with a single sample stimulus, children with ADHD can be found in the associative e.g., S , followed by a choice between two responses R and learning literature. Providing Differential Outcomes [DO] 1 1 R , and they have to learn to select the correct response. or response-specific reinforcement is a frequently used Importantly, which choice response is correct is dependent technique to overcome learning and memory deficits in on the sample stimulus presented: R may be the correct clinical as well as non-clinical samples (Urcuioli, 2005). In choice for S , but for another stimulus S , R may be the DO, stimulus–response relationships are reinforced using 1 2 2 correct choice response. When a retention interval is response-unique, rather than general, outcomes; correctly inserted between the offset of the sample stimulus and the choosing R after presentation of S will result in outcome 1 1 appearance of the choice stimuli, the aMTS task becomes an O , while correctly choosing R after presentation of S 1 2 2 arbitrary Delayed Matching-To-Sample [aDMTS] task (Case will result in a different outcome O . This is in contrast et al., 2015; Skinner, 1950). Due to its conditional nature, to a non-Differential Outcomes [nDO] procedure, where this instrumental learning task models the capacity to adapt different outcomes are provided randomly (O or O ), 1 2 choice behavior to situational requirements or hierarchical irrespective of the response, or the standard Common reinforcement contingencies (Mok et al., 2009). Outcomes [CO] procedure where only one outcome is Despite its clear clinical relevance, there is very little used to signal correct responding across trials (Holden & research on CDL in ADHD (for exceptions, see De Meyer Overmier, 2014; Overmier & Linwick, 2001). The use of a et al., 2019; Gitten et al., 2006). The few available studies DO procedure allows for the formation of specific sample- show no evidence for a deficit in CDL learning in children outcome and response-outcome associations: besides an with ADHD compared to TD children when choice stimuli S-R association, an S-O association is formed that contains immediately follow the sample stimuli. However, in daily information on the specific outcome that can be earned for life a delay between environmental cues and behavioral correct responding to the sample stimulus (Urcuioli, 2005). choice is the rule rather than the exception (e.g., as Given that this specific outcome is also linked uniquely to a when a child is instructed to begin a new task after first specific choice response, the acquired S-O associations can completing another one). We recently showed that under help support correct choice behavior (i.e., the correct choice delay conditions (imposing a delay of 8 or 16 s between is not only supported by a direct S-R association but also by the sample stimulus and response choice), children with an indirect S-O-R associative chain) (Hochhalter & Joseph, ADHD show poorer learning than TD children on a CDL 2001; Mok & Overmier, 2007). Thus, when the sample task (De Meyer et al., 2019). In addition to being a good stimulus (S) is presented, it activates a prospective memory marker of children’s everyday ability to use environmental representation of a specific, to-be-earned outcome (S-O cues to adapt their behavior (Martínez et al., 2009), CDL association) that primes a specific choice response. This delay tasks can potentially be used as an indicator whether prospective memory representation can help to overcome the intervention strategies, aimed at improving the ability to memory load created by the imposition of a delay between adapt behavior to contextual demands have the potential to the disappearance of the sample stimulus and the appearance work. This knowledge can be used for further development of the choice stimuli (Overmier & Linwick, 2001). and testing of interventions in more ecologically valid In existing contingency management programs for designs and studies (e.g., micro trials; Staff et al., 2021). children with ADHD, explicitly linking specific rewards to One potential way to improve the degree to which certain behaviors (e.g., differently colored stickers or marbles children with ADHD adapt their behavior to environmental for various forms of situationally appropriate behaviors) as in expectations, and thus to increase task performance on a a DO procedure, is to our knowledge not specifically being CDL task, is by increasing the value of the associated taught in behavioral management programs (Staff et  al., reward. The use of a larger reinforcement (e.g., a large 2021). There is substantial evidence that DO has beneficial monetary reward, as compared to a small reward or effects on learning and memory in typically developing feedback only) has been shown capable of normalizing the children and adults (Plaza et al., 2018; Urcuioli, 2005) and on-task performance of children with ADHD compared to in other clinical groups (e.g., in Autism, Down’s, Prader- TD children, including the amelioration of performance Willi and Korsakoff syndromes, and in Alzheimer’s disease) 1 3 Research on Child and Adolescent Psychopathology (2021) 49:1165–1178 1167 (Esteban et al., 2014; Hochhalter & Joseph, 2001; Joseph which serves as the baseline for the current study, and et al., 1997; Vivas et al., 2018). Studies in Korsakoff and the associated delay which was used in the manipulations Prader-Willi patients show benefits of DO in delayed and non- reported here (see Fig.  1 – Phase 1). Two different delayed conditional discrimination learning tasks (Hochhalter procedures were sequentially evaluated for their effect on & Joseph, 2001; Joseph et al., 1997) and suggest that it has the children’s CDL performance: (1) in the first CDL task, potential for targeting forgetfulness in daily life in clinical €10 could be accumulated through correct responding (large patients (e.g., for remembering the intake of medication). reward condition) (see Fig. 1 – Phase 2); (2) in subsequent Given that Prader-Willi and Korsakoff syndromes share CDL tasks, we used a within-subjects manipulation to characteristics with ADHD (including the presence of compare the effects of DO and nDO on performance. In impairments in learning and memory) (Hochhalter & Joseph, DO, making the correct choice led to unique outcomes and 2001), we speculate that the use of DO may facilitate learning thus specific stimulus–response relations were learned. and attenuate the performance deficit that children with This differs from nDO where, non-specific outcomes were ADHD exhibit in conditional discrimination learning under provided after correct choice responses, irrespective of the delays (De Meyer et al., 2019; Martínez et al., 2009, 2013; response (within-subjects) (see Fig. 1 – Phase 3). Based on Overmier & Linwick, 2001). previous research (Dovis et al., 2012; Luman et al., 2005; Differential Outcomes procedures can be integrated Slusarek et al., 2001), we expected that both procedures (the in an aDMTS task in different ways, involving different provision of larger reinforcers as well as the provision of DO types of outcomes. In a standard implementation of DO, reinforcers for correct choice behavior) would improve CDL the outcomes used can be conditioned (or secondary) performance under delay conditions, in the two groups, with reinforcers (e.g., a token) and/or primary reinforcers a more pronounced effect for those with ADHD. (Estévez et  al., 2001; Martínez et  al., 2009, 2013). Between-subjects, we manipulated whether DO applied to The latter are hedonic reinforcers that are intrinsically secondary reinforcement only (with primary reinforcement motivating (e.g., food or water in animal research) being non-differential) or to secondary and primary (Estévez et  al., 2001; Martínez et  al., 2009, 2013). reinforcement (see Fig. 1 – Phase 3). This was achieved by Whereas often in DO studies response-specific secondary random assignment of the participants to either primary and response-specific primary reinforcers are used as and secondary DO or secondary DO only. We predicted outcomes (e.g., response-specific tokens that can later be that ADHD and TD groups would both benefit more from exchanged for token-specific candy), emerging evidence a DO procedure where primary and secondary reinforcers suggests that response-specificity at one level of outcome were response-unique than from a DO procedure where only (e.g., only at the level of the secondary reinforcer) may be secondary reinforcers were response-unique. We expected sufficient to achieve a beneficial DO effect. Then again, these effects would again be more pronounced in children children with ADHD are known to process reinforcement with ADHD, due to their altered reinforcement sensitivity. differently than TD children and may need more optimal reinforcement than TD children to perform well on tasks (Dovis et al., 2012; Luman et al., 2010). When studying Method DO in children with ADHD, it is therefore important to determine whether response-specific primary and Forty-six children with a prior diagnosis of ADHD (22 secondary reinforcement is superior to response-specific combined, 18 inattentive, and 6 hyperactive/impulsive) and secondary reinforcement only. 55 typically developing children participated in the study. In summary, the aim of the present study was to investigate Six participants (5 = TD, 1 = ADHD) were subsequently whether enhancing the value of reward (remediation 1) and/ excluded due to an error in administration; i.e., an incorrect or introducing differential outcomes (remediation 2) would delay was selected from the baseline aDMTS tasks for use ameliorate deficits in CDL performance, under conditions in the reward, nDO and DO tasks. The children, aged 8 of delay, observed in children with ADHD compared to TD to 12 years, were recruited through the clinical networks children. Within differential outcomes, we also assessed of the authors (ADHD group) and local schools (TD the importance of the degree of response-specificity of group). Study inclusion criteria were: (a) an estimated IQ reinforcement; i.e., is there a difference in the effectiveness score ≥ 80, based on the short form of the Dutch version of of response specific secondary reinforcer compared to the Wechsler Intelligence Scale for Children [WISC-III-NL] response specific primary and secondary reinforcement. (b) absence of any sensory, neurological or motor disorder Immediately before the start of the current study, all or a clinical diagnosis of Autism Spectrum Disorder (as children had performed a series of CDL tasks with increasing indicated by parents) (c) absence of a clinical diagnosis of delays to determine the delay at which their performance Conduct Disorder [CD] as assessed by the CD section of tapered off (De Meyer et al., 2019). It is that performance the Disruptive Behavior Disorders module of the Diagnostic 1 3 1168 Research on Child and Adolescent Psychopathology (2021) 49:1165–1178 Fig. 1 Study overview. Prior to the start of the current study, feedback only in the baseline condition) was administered. In phase the delay at which CDL performance declined was determined 3, aDMTS tasks were administered under nDO and DO conditions individually through administration of aDMTS tasks with increasing (with order counterbalanced across participants). Within the DO delays between sample and choice stimuli (Baseline, phase 1). manipulation, half of the participants were exposed to secondary In phase 2, an aDMTS task with the same delay but including a reinforcement only and the other half to secondary as well as primary monetary reward (possibility of obtaining 10 euros in addition to reinforcement, i.e., between-subject within the groups Interview Schedule for Children, Parent Version (PDISC; This composite score exhibits satisfactory validity and Shaffer et al.,  2000) and (d) not taking any medication other reliability (0.86 and 0.91) and is highly correlated with full- than stimulant medication (in the case of ADHD) which scale IQ (Sattler, 2001). participants were willing to withdraw 24 h prior to testing PDISC: The clinical assessment followed the algorithm (Greenhill, 1998). of the Diagnostic Interview Schedule for Children, Parent The diagnosis of ADHD was established by a certified Version (Shaffer et al., 2000). This interview, based on the psychiatrist or clinical psychologist and DSM-criteria were DSM-IV criteria, has adequate psychometric properties confirmed by the PDISC. Typically Developing children (test–retest reliability = 0.79) and is a reliable assessment were required to fall within the normal range on the tool to assess DSM symptoms of ADHD, ODD and CD th Inattentive and Hyperactivity/Impulsivity section (≤ 90.9 (Shaffer et al., 2000). The interviewers, licensed clinical percentile), Oppositional Defiant Disorder [ODD] section psychologists or Masters students in clinical psychology, nd nd (≤ 95.2 percentile) and CD section (≤ 95.2 percentile) were trained by the first author in administering the PDISC. of the Disruptive Behavior Disorder Rating Scale (DBDRS; DBDRS: The Dutch version of the Disruptive Behavior Dutch translation: Oosterlaan et al., 2008) as endorsed by Disorder Rating Scale (Oosterlaan et  al., 2008) contains parents. four DSM-IV-TR based scales assessing Inattention, Hyperactivity/Impulsivity, ODD and CD symptoms. The 42-item questionnaire is designed to be completed by parents Measures of children between six and sixteen years of age. Parents were asked to rate the behavior of their child on a 4-point Likert WISC-III-NL, short version: Vocabulary and Block Design, scale, ranging from 0 (not at all) to 3 (very much). Raw scores two subtests from the Dutch version of the WISC-III (Kort (ratings added across all symptoms) were transformed to norm th et al., 2005), were administered to estimate full-scale IQ. scores ranging between 10 (50 percentile, non-clinical) 1 3 Research on Child and Adolescent Psychopathology (2021) 49:1165–1178 1169 Fig. 2 Panel a – Baseline aDMTS procedure (retrieved from De Meyer incorrect responding is followed by a red cross (see Fig. 1 – Baseline). et  al., 2019). A randomly chosen sample stimulus is presented at the After 2  s, the next trial is presented.  Panel b – aDMTS task with top of a touchscreen. Upon touching the sample stimulus, the screen increased reward. Children are presented with the aDMTS task with is cleared. After a delay of 0, 8 or 16  s, three choice stimuli appear the relevant delay determined by their baseline performance. Correct and remain on the screen until the child responds by touching one responses yield smiley faces that accumulate towards a potential ten- of the stimuli. The child’s task is to learn to select the correct choice euro reward accompanied by specific reinforcement instructions “You stimulus for a given sample stimulus through trial and error. Correct have a higher chance at winning €10”. Incorrect responses yield a red responding yields a green or red smiley (randomly determined), cross (see Fig. 1 – Phase 2) th and 19 (99.9 percentile, clinical). Adequate psychometric third choice stimulus C was added on all trials. In order for properties are reported for a Flemish sample; internal participants to become acquainted with the task, a training consistencies for the Inattention (α = 0.90), Hyperactivity/ phase was presented prior to the first CDL task. Impulsivity (α = 0.87), ODD (α = 0.88) and CD subscales Across the task, there was a gradual increase in the retention (α = 0.66) (Oosterlaan et al., 2008) are moderate to high. interval between sample and choice stimuli (from 0 s through 8 s to 16 s) (see Fig. 1 – Baseline). Each retention interval involved Conditional Discrimination Learning Task: Baseline 24 trials (with the exception of the 0-s task, which included 36 Assessment trials). The appearance of the sample stimulus (S or S ) and 1 2 the position of the choice stimuli (C, C, C ) was determined 1 2 3 In phase 1, initial CDL performance was assessed through randomly for each trial (12 options) and a different set of stimuli repeated arbitrary Delayed Matching-To-Sample [aDMTS] was used for each delay. The correct sample-choice association tasks, as reported in De Meyer et al. (2019). In each aDMTS was determined in advance and not counterbalanced. With task, participants learnt arbitrary relationships between a new increasing delay between the sample and choice stimuli, a drop set of sample stimuli and choice stimuli (see Fig. 2 – Panel in learning performance is consistently observed, typically a); conditional upon the presentation of sample stimulus S , attributed to an increased memory load (Case et al., 2015). selection of C is the correct response and upon presentation The sample and choice stimuli for each CDL task were clearly of S , selection of C is the correct response. Children distinguishable, randomly chosen abstract figures from MS 2 2 learnt the correct (S -C and S -C ) associations through a Word 2008 presented in black on a white background square 1 1 2 2 feedback-based trial-and-error procedure; correct responses measuring 5 × 5 cm. Outcome stimuli were colored 10 × 10 cm were followed by a smiley face, incorrect responses were smiley figures. The task was presented on a 15-inch touchscreen. followed by a red cross. To increase the level of difficulty, a After performing the basic aDMTS tasks with increasing delays, the delay at which a participant failed to reach the criterion of 75% correct choices over the last 12 trials was A detailed description of the aDMTS task is provided in De Meyer used as the delay for that participant in the current study; final et al., 2019. 1 3 1170 Research on Child and Adolescent Psychopathology (2021) 49:1165–1178 performance at that delay during the basic aDMTS tasks is used accumulate towards non-differential primary reinforcers (both here as the children’s baseline. If participants achieved criterion types of smileys earn candy and toy rewards); For every blue for all tested delays, the 16-s delay was used for the current smiley you will earn blue points and for every yellow smiley you study and performance on that delay was used as the baseline will earn yellow points. At the end, you can exchange all the blue to which both of the remediation procedures were compared. points and all the yellow points together for toys and candy. The more points you earn, the higher the chance you have at earning Conditional Discrimination Learning Task: candy and toys (see Fig.  3 – Panel a). All task instructions Reinforcement Manipulations were explained to the participants and the researcher checked whether they understood all instructions. Participants were told For the first reinforcement manipulation (phase 2), (see Fig.  2 that they needed to obtain enough smileys in order to receive a – Panel b), the aDMTS task, with an individually determined reward. In effect, all children received identical rewards after the delay (see figure legend) was presented that included a 10 programmed 24 trials irrespective of their performance. For the euro monetary reward for correct choice behavior. At the task, the sample stimuli, choice stimuli and correct association beginning of the task, participants were informed about the were randomly determined by the computer program. change in reward outcome: ‘From now on, for every smiley In the nDO task, which could be presented before or after the you will earn a point. The more points you earn, the higher DO task, correct responses yielded a randomly colored smiley, the chance you have at winning ten euros. When you have i.e., either a black or white smiley, unrelated to the sample earned enough points, the game will end and you will see a presented (see Fig. 3 – Panel b). Children were told that smileys green screen’, which was assumed to maintain motivation over accumulated towards a reward (toys or candy, determined by the time (Dovis et al., 2012). The ten one-euro coins that the child experimenter), in a non-differential way; For every smiley you could earn were shown and placed in sight but out of reach; will earn a point. At the end, you can exchange all the points for they remained in view throughout the entire task. All children candy or toys. The more points you earn, the higher the chance received 10 euros at the end of testing, irrespective of their you have at candy or toys. To control for reward level across actual performance on the task. conditions (i.e., DO and nDO) and the potential influence on In phase 3, we evaluated the second reinforcement their motivation, upon completing the 24 trials in the nDO task, manipulation, that is if CDL performance could be improved children were rewarded with both toys and candy (explained through the use of Differential Outcomes as compared to non- as being a consequence of their outstanding performance) but Differential Outcomes. The order of nDO/DO and nature of DO unrelated to their actual performance (see Fig. 3 – Panel b). (primary and secondary vs secondary) were varied between In all tasks, CDL performance was measured as the participants, stratified for gender, age, and group. All children percentage of correct responses across all trials of the task performed both the nDO and DO tasks (within-subjects). The minus the first four (performance on the first four trials is nature of the DO task (primary and secondary DO or secondary determined by chance only). For each participant, this metric DO only) was manipulated between subjects. In the DO aDMTS was calculated for 1) the baseline task (i.e., the aDMTS task task, correctly choosing C after the presentation of S resulted from the baseline phase on which a participant failed to reach 1 1 in outcome O (a blue smiley), whereas correctly choosing the 75% criterion, with this delay used in all subsequent tasks; C after presentation of S resulted in a different outcome O 2) the aDMTS task from Phase 2 (increased reward); and 3) the 2 2 2 (a yellow smiley) (see Fig. 3 – Panel a). In order to test the two aDMTS tasks from Phase 3 (nDO and DO). influence of the degree of response-specificity of reinforcement (i.e., response specificity of secondary reinforcement only versus primary and secondary reinforcement), half of the children in Procedure each group (ADHD, TD) received primary and secondary DO, in that they were told that different smileys could later be Before participating in the study, separate information letters exchanged for different rewards: O (blue smiley) accumulated were given to parents and children, and informed consent towards candy and O (yellow smiley) towards a toy. At the was obtained from both parents. While parents completed the beginning of the task, participants were informed about the structured interview (ADHD group only) and questionnaires change in reward outcome: For every yellow smiley you will (DBDRS and demographic questionnaire), children performed earn yellow points and for every blue smiley you will earn blue the experimental tasks in a distraction-free room. All points. At the end, you can exchange all the yellow points for toys participants completed the basic aDMTS task immediately and all the blue points for candy. The more yellow points you before the tasks described here (baseline; for full results, see earn, the higher the chance you have at earning toys. The more De Meyer et al., 2019). After determining each participant’s blue points you earn, the higher the chance you have at earning baseline delay, used in all subsequent tasks, the aDMTS task candy. The other half of the children in each group received with monetary reward was conducted (Phase 2). All children secondary DO only. They were told that secondary reinforcers received the monetary reward immediately afterwards. 1 3 Research on Child and Adolescent Psychopathology (2021) 49:1165–1178 1171 Fig. 3 Panel a – DO aDMTS task. The basic aDMTS task was modi- aDMTS is presented in which correct responses yield a non-specific fied so that correct responses yield a sample-specific outcome (DO). outcome (nDO): Participants always receive non-differential sec- All participants receive differential secondary reinforcers (smileys of ondary reinforcers (randomly white or black smileys) that accumu- a sample-specific color). For half of the participants, those second- late towards a non-differential primary reinforcer (toys or candy, as ary reinforcers accumulate towards differential primary reinforcers instructed and determined by the experimenter) (see Fig.  1 – Phase (smileys of one color earn a toy reward, the other earns candy), for 3). In reality, children were rewarded with both toys and candy the other participants, secondary reinforcers accumulate towards non- (explained as being a consequence of their outstanding performance) differential primary reinforcers (both types of smileys earn candy and unrelated to their actual performance to control for reward level toy rewards) (see Fig. 1 – Phase 3). Panel b – nDO aDMTS task. An across conditions and the potential influence on their motivation Next, children performed the nDO and DO aDMTS tasks, The study was approved by the KU Leuven Social and in counterbalanced order (Phase 3). All tasks were separated Societal Ethics Committee (G-2015 01 156). The authors by a 10-min break. The experimenter remained in the room confirm that the study was conducted in line with the ethical throughout the testing procedure (± 100 min). All children standards of the institutional research committee and with the were able to complete the tasks and families were compensated 1975 declaration of Helsinki and its 2008 amendment. with an additional 10 euros for participating in the study. Results Review of the distribution of the outcome variables detected some extreme values (outliers) and high skewness and kurtosis for all outcome variables (unrelated to group), indicating The tasks were part of a larger test battery including tasks on work- ing memory and delay aversion (see De Meyer et al., 2019). 1 3 1172 Research on Child and Adolescent Psychopathology (2021) 49:1165–1178 Table 1 Demographic and ADHD TD Clinical Characteristics for the M(SD) M(SD) F/χ p ADHD and TD children Gender N 4.62 0.032* Male (N/%) 31 (68.89) 23 (46.94) Female (N/%) 14 (31.11) 26 (53.06) Age (years) 10.29 (0.99) 10.07 (1.21) 0.90 0.346 FSIQ 98.00 (11.72) 105.35 (9.84) 10.89 0.001** Dyscalculia (N/%) 1 (2.22) 0 Dyslexia (N/%) 4 (8.88) 0 ODD (PDISC – N/%) 44 (46.80) - Medication (N/%) 23 (51.11) 0 Maternal education 1.93 (0.82) 1.72 (0.74) 2.04 0.360 DBDRS (norm scores) Inattention 15.02 (2.01) 10.53 (1.00) 191.56 < 0.001*** Hyperactivity/Impulsivity 14.23 (2.44) 10.51 (0.98) 96.51 < 0.001*** ODD 12.42 (2.43) 10.71 (1.24) 18.57 < 0.001*** CD 11.42 (1.53) 10.96 (1.15) 2.68 0.105 PDISC (number of symptoms) Inattention 7.43 (1.65) - Hyperactivity/Impulsivity 6.00 (2.57) - ODD 3.23 (2.14) - CD 0.45 (0.79) - ADHD  Attention Deficit Hyperactivity Disorder, TD  Typically Developing, FSIQ  Full Scale IQ, PDISC  Diagnostic Interview Schedule for Children, Parent Version, DBDRS  Disruptive Behavior Rating Scale ODD Oppositional Defiant Disorder, CD Conduct Disorder *p < .05, **p < .01, ***p < .001 High (1) = University Education; Average (2) = Non-University Higher Education; Low (3) = Secondary Education (or less); 3 missing data points (1 = ADHD; 2 = TD) 2 missing data points for ADHD group non-normal distributions. Therefore, data was subjected to children with ADHD, F(1, 92) = 10.89, p = 0.001, ηp = 0.106. an arcsine transformation, as is recommended when outcome Neither IQ nor gender were included as covariates in the analysis variables are percentages (see also IBM Corp. 2019; Zar, 1984). as neither variable correlated with any of the outcome variables. After transformation, one extreme outlier was detected using To determine if adding a large reward or changing the boxplots and deleted from the dataset. For one other participant associative structure of the task improved CDL performance, the last 4 of 24 trials in the DO condition were missing. These two group x task repeated-measures ANOVAs were run, the missing values were replaced by the average score of the group first comparing performance on the baseline aDMTS task with (TD) for this variable (Field, 2013). The ADHD and TD groups the monetary reward aDMTS task across the ADHD and TD did not differ in mean age or family education level but did differ groups. The second one comparing performance on the DO and 2 = in gender distribution (χ 4.62, p = 0.032) with an uneven nDO aDMTS tasks across the two groups. Additionally, a group distribution of boys (N = 31) and girls (N = 14) in the ADHD x condition fixed-factors ANOVA was conducted to compare group compared with the control group (see Table 1 and the CDL performance between the two groups (ADHD/TD) when appendix for demographic characteristics). As often observed the DO manipulation involved secondary reinforcement only in studies of children with ADHD, there was a main effect of versus primary and secondary reinforcement. Post-hoc analyses, group for IQ with children in the TD group scoring higher than independent-samples t-tests and paired-samples t-tests were conducted to identify the source of the significant interaction effects. Effect sizes are reported for ease of interpretation; small 2 2 (ηp = 0.01; d = 0.2); moderate (ηp = 0.06; d = 0.5) and large Data analysis on transformed and untransformed data was applied (both parametric and non-parametric), leading to similar results. (ηp = 0.14; d = 0.8) (Cohen, 1988, 1992). Performance that is more extreme than three times the length of The distribution of the individually determined delays, th th the boxplot (between the 25 and 75 percentile), as marked with an selected on the basis of performance in the baseline tasks, was asterisk (*) on the boxplot. 1 3 Research on Child and Adolescent Psychopathology (2021) 49:1165–1178 1173 Fig. 4 CDL performance across tasks for children with ADHD and TD children. For ease of interpretation untransformed data is displayed. **p < 0.01 not significantly different between the groups, χ (2) = 5.75, ADHD compared to TD children (see Fig.  4, Table  2). Follow-up independent-samples t-tests indicated that p = 0.056. For the majority of children with ADHD and all TD children, a delay of 16 s was selected (ADHD: n = 40, children with ADHD differed significantly from the TD children at baseline t(79.18) = 3.29, p = 0.002, TD: n = 49); for the remaining children in the ADHD group, an 8-s (n = 3) or 0-s (n = 2) delay was selected. d = 0.69, but not following the addition of a monetary reward, t(92) = 1.56, p = 0.122, d = 0.32. Paired-samples The first 2 (group: ADHD vs TD) × 2 (condition: baseline vs monetary reward) repeated measures ANOVA t-tests showed that the difference in CDL performance between the baseline task and the monetary reward task yielded statistically significant main effects for condition, F(1, 92) = 6.97, p = 0.010, ηp = 0.070 and g roup, F(1, performance was significant for children with ADHD, t(44) = -2.85, p = 0.007, d = 0.43 but not for TD children, 92) = 9.45, p = 0.003, ηp = 0.093, and a significant condition by group interaction, F (1, 92) = 4.60, t(48) = -0.44, p = 0.666, d = 0.06. The effect of DO versus nDO was compared between p = 0.035, ηp = 0.048; adding a monetary reward had a larger impact on the CDL performance of children with groups in a 2 (condition; DO vs nDo) × 2 (group; ADHD vs Table 2 Percentage of Correct ADHD TD Responses (with Standard Deviations) and Univariate M (SD) M (SD) F p ηp ANOVA Results for Baseline, Baseline 72.53 (19.93) 85.51 (11.10) 11.10 0.001** 0.108 Reward, Differential Outcomes Reward 83.56 (12.23) 86.53 (11.69) 2.43 0.122 0.026 and Non-Differential Outcomes Tasks in the ADHD and TD Differential Outcomes 81.22 (16.89) 88.01 (7.72) 3.71 0.057 0.039 groups DO—secondary 80.42 (19.39) 88.94 (7.81) 2.15 0.149 0.043 DO—primary & secondary 82.14 (13.93) 86.96 (7.65) 1.53 0.223 0.035 Non-Differential Outcomes 76.33 (21.52) 87.86 (14.22) 12.71 0.001** 0.121 ADHD Attention Deficit Hyperactivity Disorder; TD Typically Developing ** p < .01 Based on untransformed data Based on transformed data We analyzed the data including only those children tested with a 16 s-delay. The main effects of condition and group remained signifi- cant. The previously significant interaction effect was no longer sig- nificant (p = 0.131). Degrees of freedom are for equal variances not assumed (signifi- cant Levene’s test). 1 3 1174 Research on Child and Adolescent Psychopathology (2021) 49:1165–1178 Fig. 5 DO performance across conditions for children with ADHD and TD children. For ease of interpretation untrans- formed data is displayed TD) repeated measures ANOVA. There was a signic fi ant main Discussion effect of group, F (1, 92) = 10.87, p = 0.001, ηp = 0.106, but not condition, F(1, 92) = 0.19, p = 0.665, ηp = 0.002. The group x Children with ADHD experience difficulty in using feedback condition interaction was significant, F(1, 92) = 5.37, p = 0.023, to adapt their behavior in the presence of delays (Conditional ηp = 0.055 (see Fig. 4, Table 2). Follow-up paired-samples Discrimination Learning). Here we evaluated the ability and independent-samples t-tests showed that performance of different reinforcement manipulations to improve did not differ significantly between nDO and DO for either performance on a CDL task. Specifically, we tested the group, t(48) = -1.46, p = 0.152, d = 0.23 (TD) and t(44) = 1.79, effects of increasing reinforcer size and the introduction p = 0.081, d = 0.26 (ADHD), however for children with ADHD, of Differential Outcomes (i.e. response-specific reward compared to TD children, the percentage of correct choices was outcomes). Within DO, we explored whether response- significantly lower under nDO, t (92) = 3.57, p = 0.001, d = 0.74, specific primary and secondary reinforcement was superior but not under DO, t(70.88) = 1.89, p = 0.063, d = 0.39. The to response-specific secondary reinforcement only. order in which the conditions were presented did not influence Contrary to our prediction that increasing reward size or the results when included as a covariate. value would have a positive effect on CDL performance under To explore whether a combination of response-specific conditions of delay in both groups, we found a significant secondary and primary reinforcement is more effective on CDL improvement in performance, i.e., a higher percentage of performance than response-specific secondary reinforcement correct responses compared with baseline conditions, in only, a 2 (condition; secondary vs primary and secondary) × 2 the ADHD group only. Under increased reward conditions, (group; ADHD vs TD) factorial ANOVA was run. Results the performance of the ADHD and TD group was no showed that there was no significant group, F (1, 90) = 3.56, longer significantly different, suggesting a normalization of p = 0.062, ηp = 0.038, or condition effect, F(1, 90) = 0.23, performance for the ADHD children. p = 0.636, ηp = 0.003, nor a group by condition interaction F(1, Similarly, adding a specific reward outcome to sample- 2 9 90) = 0.11, p = 0.740, ηp = 0.001 (see Fig. 5). Applying DO choice associations, i.e., DO, improved performance on the at secondary level only or at both primary and secondary level delayed CDL task in children with ADHD only. Although did not differentially impact performance in the two groups. a significant difference in performance was found between the groups when using non-differential outcomes, under DO reward conditions the difference between the ADHD group We reanalyzed the data including only those children tested with a 16  s-delay. The main effects of group remained significant. The previously significant interaction effect was no longer significant (p = 0.052). The order in which the conditions were presented did not influence Degrees of freedom are for equal variances not assumed. the results when included as a covariate. 1 3 Research on Child and Adolescent Psychopathology (2021) 49:1165–1178 1175 and the TD group was no longer significant. Again, this helps to overcome such aversion caused by delay, e.g., through suggests that adding DO may normalize CDL performance in increased attention towards the sample-choice association. children with ADHD. Further, we predicted that manipulating However, our previous study did not find an association between the nature of the reward within DO would affect performance. delay aversion and CDL performance under a delay (De Meyer The data do not support this hypothesis, i.e., primary and et al., 2019), rendering this explanation less likely. Alternatively, secondary DO performance did not differ from performance a more rewarding or response-unique outcome may simply under secondary DO only, for either group. serve to increase the motivation of children with ADHD to The observation of improved CDL performance through perform the CDL task as well as possible. However, given reward maximalization is in accordance with earlier findings reward intensity was equal across the DO and nDO conditions, that indicate beneficial effects of reward optimization on deficits a simple motivational account does not offer a convincing in Executive Functioning [EF] performance in ADHD (Dovis explanation for the effects of DO. Nevertheless, the results do et  al., 2012; Fosco et al., 2015; Slusarek et al., 2001). In a indicate an improvement in CDL performance through use of a DMTS task, a drop in accuracy under conditions of delay is large monetary reward for children with ADHD, removing the often attributed to deficits in short-term (Etkin & D’Amato, significant difference in performance between ADHD and TD 1969; Roberts & Grant, 1978) or working memory (Case groups that was observed at baseline. et al., 2015; Kempton et al., 1999), although the literature has Contrary to the results of earlier studies (e.g., Martínez et al., not addressed which specific memory aspect is involved in 2013; Mok & Overmier, 2007; Molina et al., 2015) and our own aDMTS. Our previous study (De Meyer et al., 2019), however, predictions, associating a specific outcome to a stimulus-choice suggested that neither short-term nor working memory was association (DO) did not significantly improve the performance related to performance on the aDMTS CDL task. The task of TD children. This may be due to a ceiling effect in the used to evaluate memory in that study (Corsi Block Tapping baseline performance of TD children, leaving limited room for Task; visual-spatial memory), may not have assessed memory change. Findings from previous research suggest a facilitating components required for CDL learning. The current study effect of DO on performance only when the task is sufficiently shows that adding a reward improves performance on a CDL challenging (e.g., a 4-cue task for adults) (Estévez et al., 2001; task under delay, although the specific mechanisms responsible Maki et al., 1995; Miller et al.,  2002). Despite care in task for this improvement have yet to be determined. development (balancing task dic ffi ulty for both groups through Improvement of CDL performance through a monetary pilot studies), DO would perhaps only facilitate performance in reward involves an increase in reward value, which might TD children in a more challenging task design. have served to remediate an underlying memory (short-term Additionally, it was hypothesized that the use of response- or working) deficit. The impact of DO on CDL performance specific secondary and primary reinforcement would might likewise be mediated by an effect on memory; the enhance performance as compared to response-specific nDO and DO conditions did not differ in reward size, but secondary reinforcement only. Unexpectedly, performance rather in the specificity of reward. Although the increase was similar in both DO conditions. It may be that presenting in performance from nDO to DO in children with ADHD a response-specific secondary reinforcer is already potent failed to reach significance (p = 0.081), DO did appear to enough to create a performance ceiling effect, thereby “normalize” performance in the ADHD group, i.e., they leaving little room for further improvement with response- performed more similarly to TD children on CDL. One specific primary reinforcement. It is also possible that the explanation for the effect of a DO procedure on aDMTS absence of a difference between these two forms of DO is performance is that it allows for the associative activation of related to the similarity in instructions between primary DO a prospective memory representation upon presentation of only and primary and secondary DO. The instructions given the sample stimulus that primes correct choice behavior. This to the children in both DO conditions were quite similar (see extra memory representation might have helped to counter Appendix), with children being promised candy and a toy possible short-term or working memory deficits that could in each (be it differential in one condition but not the other). impair aDMTS CDL performance in children with ADHD. The current study comes with some caveats. To begin with, An alternative explanation for increased performance on we did not succeed in collecting teacher ratings to confirm the the CDL task through increased reward or DO centers on their ADHD diagnosis for all participants due to practical constraints emotional-motivational effects (Sonuga-Barke, 2002). Within a (i.e., no response, children changing teachers, absence of CDL task, the delay between the sample and the choice stimuli contact information) and therefore cannot confirm the cross (and the associated reward) can trigger a negative emotional situational severity of symptoms, considered a core diagnostic state in children with ADHD, known as Delay Aversion (Antrop criterion in the diagnosis of ADHD. While acknowledging this et al., 2006), which can result in inattentive and hyperactive limitation, most of the children who entered the study had been symptoms and lead to decreased task performance (Marco et al., previously assessed and diagnosed through the KU Leuven 2009). The addition of a more salient reinforcer or DO possibly university hospital, by means of multi-method, multi-informant 1 3 1176 Research on Child and Adolescent Psychopathology (2021) 49:1165–1178 assessments where also cross-situational severity was taken only one token type to target various forms of situationally into account. Over the course of the study (including the appropriate behaviors (e.g., a sticker for sitting still during baseline testing reported in De Meyer et al., 2019), the aDMTS mathematics and for playing nicely with siblings) (Coelho task was administered four times. An influence of repeated et  al., 2015). Our results suggest that applying response- task administration on performance cannot be excluded, specific reinforcers may increase the learning of situation- although the significant group difference in nDO performance specific stimulus–response associations in children with (involving either the third or fourth aDMTS task, depending ADHD. This differential rewarding approach has already on counterbalancing) and the absence of a significant within- proven beneficial in a range of clinical populations (Esteban group difference between the baseline and nDO performance, et al., 2014; Hochhalter & Joseph, 2001; Overmier & Linwick, t(93) = -1.74, p = 0.086, argue against a simple task training 2001) and is often implemented in token economy programs effect. The current design did not allow us to control for the for children with Autism Spectrum Disorder (Fairbanks & contextual effect of being rewarded with a monetary reward Sugai, 2014; Neitzel, 2010), a neurodevelopmental disorder before the DO-nDO conditions, as the monetary reward with a significant overlap in clinical behavioral features and condition was always presented first, after which DO and nDO etiology with ADHD (Craig et al., 2015). were presented in counterbalanced order. Therefore, we cannot While the increase of reward value through monetary strictly rule out that delivering a large monetary reward ahead means also increases performance in children with ADHD, of the DO/nDO tasks differentially affected performance of the realistically, the addition of a high-value reward (e.g., 10 ADHD and TD groups. Another possible limitation relates to euro) is less feasible in real-life situations compared to the the stimuli used in the aDMTS task. Over the four aDMTS tasks relatively simple implementation of DO. In the present (baseline, reward, DO, nDO), the set of stimuli used was fixed study, DO was as effective as a high value of reward in and not counterbalanced. Therefore, performance differences increasing CDL performance in children with ADHD, and between aDMTS manipulations might, in principle, be due to the implementation of DO in token economy programs stimuli-specific characteristics. However, considerable care can be relatively easy (e.g., rewarding on-task behavior was taken to establish stimulus sets of equal difficulty. In with a red token and rewarding the raising of a hand before addition, it is possible that the task instructions influenced answering with a blue token). reward expectations differently for the DO and nDO conditions, In conclusion, the present research provides initial evidence favoring the DO condition. Although this effect was not evident that deficits in delayed conditional discrimination learning in in the performance of TD children (equal performance in DO vs ADHD on a DMTS task can be attenuated by enhanced reward nDO) we cannot rule out an ADHD specific differential effect. and DO manipulations. Our results have potential implications Finally, due the between-subjects manipulation of DO (primary for the refinement of behavioral interventions for children with and secondary versus secondary DO only), groups were rather ADHD. These findings can, for example inspire adaptations small, reducing power to detect significant differences between to existing token economies in ADHD, and further testing those two forms of DO. of these adapted “differential outcomes” token economies Despite the promising results, questions remain regarding versus “non-differential” token economies on proximal daily the mechanism underlying the DO phenomenon effect. Further life outcomes in micro-trails (Staff et  al., 2021). Further work is needed to disentangle what underlies the effectiveness research should also be directed at a better understanding of of DO. This is important to provide a better understanding of the mechanisms through which increased reward and DO exert its positive effects for children with ADHD. In future studies their beneficial effects on CDL performance. it would be important to test whether a DO effect can also be Supplementary Information The online version contains supplemen- achieved through other types of response-specific reinforcers tary material available at https://doi. or g/10. 1007/ s10802- 021- 00781-5 . (e.g., differential versus non-differential social reinforcers). Acknowledgements The authors would like to thank the participating Clinical Implications families for their contribution to the study. The findings of this study have a number of implications for Funding This work was supported by Research Grant G.0738.14 N of the Research Foundation Flanders (FWO). maximizing the impact of operant techniques in behavioral treatment for ADHD. In Behavioral Parent Training [BPT], a Declarations token economy is a widely used operant technique with the core aim of increasing adaptive and reducing inappropriate Conflict of Interest The authors declare that they have no conflicts of behavior in children with ADHD (Sullivan & O’Leary, interest. 1990). In a token economy, children are rewarded with Ethical Approval The study was approved by the KU Leuven Social specific tokens (e.g., marbles, stickers, etc.) for adaptive and Societal Ethics Committee (G-2015 01 156). The authors confirm behavior. A standard token economy, however, applies 1 3 Research on Child and Adolescent Psychopathology (2021) 49:1165–1178 1177 that the study was conducted in line with the ethical standards of the memory in children and in adults with Down syndrome. Research institutional research committee and with the 1975 declaration of Hel- in Developmental Disabilities, 35(6), 1384–1392. https://doi. or g/ sinki and its 2008 amendment.10. 1016/j. ridd. 2014. 03. 031 Estévez, A. F., Fuentes, L. J., & Marı́-Beffa, P., González, C., & Informed Consent Before participating in the study informed consent Alvarez, D. (2001). The Differential Outcome Effect as a Useful was obtained from both parents. Tool to Improve Conditional Discrimination Learning in Children. Learning and Motivation, 32(1), 48–64. https:// doi. org/ 10. 1006/ lmot. 2000. 1060 Open Access This article is licensed under a Creative Commons Attri- Etkin, M., & D’Amato, M. R. (1969). Delayed matching-to-sample bution 4.0 International License, which permits use, sharing, adapta- and short-term memory in the capuchin monkey. Journal tion, distribution and reproduction in any medium or format, as long of Comparative and Physiological Psychology, 69(3), 544– as you give appropriate credit to the original author(s) and the source, 549. https:// doi. org/ 10. 1037/ h0028 209 provide a link to the Creative Commons licence, and indicate if changes Fairbanks, S., & Sugai, G. (2014). Token Economy. Encyclopedia of were made. The images or other third party material in this article are Special Education, 3–4. https://doi. or g/10. 1002/ 97811 18660 584. included in the article’s Creative Commons licence, unless indicated ese24 03 otherwise in a credit line to the material. If material is not included in Field, A. P. (2013). Discovering statistics using IBM SPSS statistics. the article’s Creative Commons licence and your intended use is not Sage Publications. permitted by statutory regulation or exceeds the permitted use, you will Fosco, W. D., Hawk, L. W., Rosch, K. S., & Bubnik, M. G. (2015). need to obtain permission directly from the copyright holder. To view a Evaluating cognitive and motivational accounts of greater copy of this licence, visit http://cr eativ ecommons. or g/licen ses/ b y/4.0/ . reinforcement effects among children with attention-deficit/ hyperactivity disorder. Behavioral and Brain Functions, 11(1), 1–9. https:// doi. org/ 10. 1186/ s12993- 015- 0065-9 References Gitten, J. C., Winer, J. L., Festa, E. K., & Heindel, W. C. (2006). Conditional associative learning of spatial and object information American Psychiatric Association. (2013). Diagnostic and Statistical in children with attention deficit/hyperactivity disorder. Child Manual of Mental Disorders (5th ed.). https:// doi. org/ 10. 1176/ Neuropsychology, 12(1), 39–56.  https:// doi. or g/ 10. 1080/ appi. books. 97808 90425 596. 744053 09297 04050 02055 79 Antrop, I., Stock, P., Verté, S., Wiersema, J. R., Baeyens, D., & Greenhill, L. L. (1998). Childhood attention deficit hyperactivity Roeyers, H. (2006). ADHD and delay aversion: The influence of disorder: Pharmacological treatments. In P. E. Nathan & J. non-temporal stimulation on choice for delayed rewards. Journal Gorman (Eds.), A guide to treatments that work (pp. 42–64). New of Child Psychology and Psychiatry and Allied Disciplines, York: Oxford Universtity Press. 47(11), 1152–1158. https:// doi. org/ 10. 1111/j. 1469- 7610. 2006. Hochhalter, A. K., & Joseph, B. (2001). Differential Outcomes Training 01619.x Facilitates Memory in People with Korsakoff and Prader-Willi Case, J. P., Laude, J. R., & Zentall, T. R. (2015). Delayed matching Syndromes. Integrative Physiological and Behavioral Science, to sample in pigeons: Effects of delay of reinforcement and 36(3), 196–204. https:// doi. org/ 10. 1007/ BF027 34093 illuminated delays. Learning and Motivation, 49, 51–59. https:// Holden, J. M., & Overmier, J. B. (2014). Performance under differential doi. org/ 10. 1016/j. lmot. 2015. 01. 001 outcomes: Contributions of Reward-Specific Expectancies. Coelho, L. F., Barbosa, D. L. F., Rizzutti, S., Muszkat, M., Amodeo Learning and Motivation, 45(1), 1–14. https:// doi. org/ 10. 1016/j. Bueno, O. F., & Miranda, M. C. (2015). Use of cognitive lmot. 2013. 09. 001  behavioral therapy and token economy to alleviate dysfunctional IBM Corp. (2019). IBM SPSS Statistics for Macintosh, Version 26.0. behavior in children with attention-deficit hyperactivity disorder. Armonk, NY: IBM Corp. Frontiers in Psychiatry, 6, 1–9. https://doi. or g/10. 3389/ fpsyt. 2015. Joseph, B., Bruce Overmier, J., & Thompson, T. (1997). Food- and nonfood-related differential outcomes in equivalence learning by Cohen, J. (1988). Statistical power analysis for the behavioral sciences adults with Prader-Willi syndrome. American Journal on Mental (2nd ed.). Lawrence Erlbaum Associates. Retardation, 101(4), 374–386. Cohen, J. (1992). Statistical Power Analysis. Current Directions in Kempton, S., Vance, A., Maruff, P., Luk, E., Costin, J., & Pantelis, Psychological Science, 1(3), 98–101.  https:// doi. org/ 10. 1111/ C. (1999). Executive function and attention deficit hyperactivity 1467- 8721. ep107 68783 disorder: Stimulant medication and better executive function Craig, F., Lamanna, A. L., Margari, F., Matera, E., Simone, M., & performance in children. Psychological Medicine, 29(3), 527– Margari, L. (2015). Overlap Between Autism Spectrum Disorders 538. https:// doi. org/ 10. 1017/ S0033 29179 90083 38 and Attention Deficit Hyperactivity Disorder: Searching for Kort, W., Schittekatte, M., Bosmans, M., Compaan, E. L., Dekker, P. Distinctive/Common Clinical Features. Autism Research, 8(3), H., Vermeir, G., & Verhaeghe, P. (2005). WISC-III: handleiding 328–337. https:// doi. org/ 10. 1002/ aur. 1449 en verantwoording. Pearson. De Meyer, H., Beckers, T., Tripp, G., & van der Oord, S. (2019). Luman, M., Oosterlaan, J., & Sergeant, J. A. (2005). The impact Deficits in Conditional Discrimination Learning in Children with of reinforcement contingencies on AD/HD: A review and ADHD are Independent of Delay Aversion and Working Memory. theoretical appraisal. Clinical Psychology Review, 25(2), 183– Journal of Clinical Medicine, 8(9), 1381. https://doi. or g/10. 3390/ 213. https:// doi. org/ 10. 1016/j. cpr. 2004. 11. 001 jcm80 91381 Luman, M., Tripp, G., & Scheres, A. (2010). Identifying the Dovis, S., Van Der Oord, S., Wiers, R. W., & Prins, P. J. M. (2012). neurobiology of altered reinforcement sensitivity in ADHD: A Can motivation normalize working memory and task persistence review and research agenda. Neuroscience and Biobehavioral in children with attention-deficit/hyperactivity disorder? the Reviews, 34(5), 744–754. https:// doi. org/ 10. 1016/j. neubi orev. effects of money and computer-gaming. Journal of Abnormal 2009. 11. 021 Child Psychology, 40(5), 669–681.  https:// doi. or g/ 10. 1007/ Maki, P., Overmier, J. B., Delos, S., & Gutmann, J. A. (1995). s10802- 011- 9601-8 Expectancies as Factors Influencing Conditional Discrimination Esteban, L., Plaza, V., López-Crespo, G., Vivas, A. B., & Estévez, A. F. Performance of Children. The Psychological Record, 45(1), 45–71. (2014). Differential outcomes training improves face recognition 1 3 1178 Research on Child and Adolescent Psychopathology (2021) 49:1165–1178 Marco, R., Miranda, A., Schlotz, W., Melia, A., Mulligan, A., Müller, Experimental Psychology: Animal Behavior Processes, 4(3), U., & Sonuga-Barke, E. J. S. (2009). Delay and reward choice 219–236. https:// doi. org/ 10. 1037/ 0097- 7403.4. 3. 219 in ADHD: An experimental test of the role of delay aversion. Sagvolden, T., Johansen, E. B., Aase, H., & Russell, V. A. (2005). A Neuropsychology, 23(3), 367–380. https:// doi. or g/ 10. 1037/ dynamic developmental theory of attention-deficit/hyperactivity a0014 914 disorder (ADHD) predominantly hyperactive/impulsive and Martínez, L., Estévez, A. F., Fuentes, L. J., & Overmier, J. B. combined subtypes. Behavioral and Brain Sciences, 28(3), (2009). Improving conditional discrimination learning and 397–419. https:// doi. org/ 10. 1017/ S0140 525X0 50000 75 memory in five-year-old children: Differential outcomes effect Sattler, J. (2001). Assessment of Children: cognitive applications using different types of reinforcement. Quarterly Journal of (4th ed.). CA. Experimental Psychology, 62(8), 1617–1630. https://doi. or g/10. Shaffer, D., Fisher, P., Lucas, C. P., Dulcan, M. K., & Schwab- 1080/ 17470 21080 25578 27 Stone, M. E. (2000). NIMH Diagnostic Interview Schedule Martínez, L., Flores, P., González-Salinas, C., Fuentes, L. J., & for Children Version IV (NIMH DISC-IV): Description, Estévez, A. F. (2013). The effects of differential outcomes and Differences From Previous Versions, and Reliability of Some different types of consequential stimuli on 7-year-old children’s Common Diagnoses. Journal of the American Academy of Child discriminative learning and memory. Learning and Behavior, & Adolescent Psychiatry, 39(1), 28–38. https://doi. or g/10. 1097/ 41(3), 298–308. https:// doi. org/ 10. 3758/ s13420- 013- 0105-y00004 583- 20000 1000- 00014 Martínez, L., Marí-Beffa, P., Roldán-Tapia, D., Ramos-Lizana, Skinner, B. F. (1950). Are theories of learning necessary? Psychological J., Fuentes, L. J., & Estévez, A. F. (2012). Training with Review, 57(4), 193–216. https:// doi. org/ 10. 1037/ h0054 367 differential outcomes enhances discriminative learning and Slusarek, M., Velling, S., Bunk, D., & Eggers, C. (2001). visuospatial recognition memory in children born prematurely. Motivational effects on inhibitory control in children with Research in Developmental Disabilities, 33(1), 76–84. https:// ADHD. Journal of the American Academy of Child and doi. org/ 10. 1016/j. ridd. 2011. 08. 022 Adolescent Psychiatry, 40(3), 355–363. https://d oi.o rg/1 0.1 097/ Miller, O. T., Waugh, K. K. M., & Chambers, K. (2002). Differential 00004 583- 20010 3000- 00016 outcomes effect: Increased accuracy in adults learning kanji Sonuga-Barke, E. J. S. (2002). Psychological heterogeneity in with stimulus specific rewards. The  Psychological Record, AD/HD—a dual pathway model of behaviour and cognition. 52(3), 315–324. https:// doi. org/ 10. 1007/ BF033 95433 Behavioural Brain Research, 130(1–2), 29–36. https:// doi. org/ Mok, L. W., Estevez, A. F., & Overmier, J. B. (2017). Unique 10. 1016/ S0166- 4328(01) 00432-6 Outcome Expectations as a Training and Pedagogical Tool. Staff, A. I., Van den Hoofdakker, B. J., van der Oord, S., Hornstra, The Psychological Record, 60(2), 227–247. https:// doi. org/ 10. R., Hoekstra, P. J., Twisk, J. W. R., Oosterlaan, J., & Luman, 1007/ bf033 95705 M. (2021). Effectiveness of specific techniques in behavioral Mok, L. W., & Overmier, J. B. (2007). The differential outcomes teacher training for childhood ADHD: A randomized controlled effect in normal human adults using a concurrent-task within- microtrial. Journal of Clinical Child & Adolescent Psychology, subjects design and sensory outcomes. The  Psychological 1–17. https:// doi. org/ 10. 1080/ 15374 416. 2020. 18465 42 Record, 57(2), 187–200. Sullivan, M. A., & O’Leary, S. G. (1990). Maintenance following Mok, L. W., Thomas, K. M., Lungu, O. V., & Overmier, J. B. (2009). reward and cost token programs. Behavior Therapy, 21(1), 139– Neural correlates of cue-unique outcome expectations under 149. https:// doi. org/ 10. 1016/ S0005- 7894(05) 80195-9 differential outcomes training: An fMRI study. Brain Research, Trapold, M. A. (1970). Are expectancies based upon different 1265, 111–127. https:// doi. org/ 10. 1016/j. brain res. 2008. 12. 072 positive reinforcing events discriminably different? Learning Molina, M., Plaza, V., Fuentes, L. J., & Estévez, A. F. (2015). and Motivation, 1(2), 129–140. https:// doi. org/ 10. 1016/ 0023- The differential outcomes procedure enhances adherence to 9690(70) 90079-2 treatment: A simulated study with healthy adults. Frontiers in Urcuioli, P. J. (2005). Behavioral and associative effects of Psychology, 6(NOV), 1–7. https://doi. or g/10. 3389/ fpsy g.2015. differential outcomes in discrimination learning. Animal 01780 Learning & Behavior, 33(1), 1–21. https:// doi. or g/ 10. 3758/ Neitzel, J. (2010). Reinforcement for children and youth with autism bf031 96047 spectrum disorders: Online training module. Colombus: Vivas, A. B., Ypsilanti, A., Ladas, A. I., Kounti, F., Tsolaki, M., & OCALI. Estévez, A. F. (2018). Enhancement of Visuospatial Working Nigg, J. T., & Casey, B. J. (2005). An integrative theory of attention- Memory by the Differential Outcomes Procedure in Mild deficit/hyperactivity disorder based on the cognitive and Cognitive Impairment and Alzheimer’s Disease. Frontiers in affective neurosciences. Development and Psychopathology, Aging Neuroscience, 10(November), 1–7. https:// doi. or g/ 10. 17(3), 785–806. https:// doi. org/ 10. 1017/ S0954 57940 50503 763389/ fnagi. 2018. 00364 Oosterlaan, J., Baeyens, D., Scheres, A., Antrop, I., Roeyens, H., & Wehmeier, P. M., Schacht, A., & Barkley, R. A. (2010). Social and Sergeant, J. (2008). VvGK6-16: vragenlijst voor gedragsproblemen Emotional Impairment in Children and Adolescents with ADHD bij kinderen 6 tot en met 16 jaar. Pearson. and the Impact on Quality of Life. Journal of Adolescent Health, Overmier, J. B., & Linwick, D. (2001). Conditional Choice-Unique 46(3), 209–217. https://doi. or g/10. 1016/j. jadoh ealt h.2009. 09. 009 Outcomes Establish Expectancies That Mediate Choice Willcutt, E. G., Doyle, A. E., Nigg, J. T., Faraone, S. V., & Pennington, Behavior. Integrative Physiological and Behavioral Science, B. F. (2005). Validity of the Executive Function Theory of 36(3), 173–181. https:// doi. org/ 10. 1007/ BF027 34091 Attention-Deficit/Hyperactivity Disorder: A Meta-Analytic Plaza, V., Molina, M., Fuentes, L. J., & Estévez, A. F. (2018). Review. Biological Psychiatry, 57(11), 1336–1346. https:// doi. Learning and recall of medical treatment-related information in org/ 10. 1016/j. biops ych. 2005. 02. 006 older adults using the differential outcomes procedure. Frontiers Zar, J. H. (1984). Biostatistical Analysis. Journal of the American in Psychology, 9(FEB).  https:// doi. or g/ 10. 3389/ fpsy g. 2018. Statistical Association (2nd ed.). https://doi. or g/10. 2307/ 22854 23 Roberts, W. A., & Grant, D. S. (1978). An analysis of light-induced Publisher’s Note Springer Nature remains neutral with regard to retroactive inhibition in pigeon short-term memory. Journal of jurisdictional claims in published maps and institutional affiliations. 1 3 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Abnormal Child Psychology Springer Journals

Conditional Learning Deficits in Children with ADHD can be Reduced Through Reward Optimization and Response-Specific Reinforcement

Loading next page...
 
/lp/springer-journals/conditional-learning-deficits-in-children-with-adhd-can-be-reduced-m34iZQgmRO

References (104)

Publisher
Springer Journals
Copyright
Copyright © The Author(s) 2021
ISSN
0091-0627
eISSN
2730-7174
DOI
10.1007/s10802-021-00781-5
Publisher site
See Article on Publisher Site

Abstract

When children with ADHD are presented with behavioral choices, they struggle more than Typically Developing [TD] children to take into account contextual information necessary for making adaptive choices. The challenge presented by this type of behavioral decision making can be operationalized as a Conditional Discrimination Learning [CDL] task. We previously showed that CDL is impaired in children with ADHD. The present study explores whether this impairment can be remediated by increasing reward for correct responding or by reinforcing correct conditional choice behavior with situationally specific outcomes (Differential Outcomes). An arbitrary Delayed Matching-To-Sample [aDMTS] procedure was used, in which children had to learn to select the correct response given the sample stimulus presented (CDL). We compared children with ADHD (N = 45) and TD children (N = 49) on a baseline aDMTS task and sequentially adapted the aDMTS task so that correct choice behavior was rewarded with a more potent reinforcer (reward manipulation) or with sample-specific (and hence response- specific) reinforcers (Differential Outcomes manipulation). At baseline, children with ADHD performed significantly worse than TD children. Both manipulations (reward optimization and Die ff rential Outcomes) improved performance in the ADHD group, resulting in a similar level of performance to the TD group. Increasing the reward value or the response-specificity of reinforcement enhances Conditional Discrimination Learning in children with ADHD. These behavioral techniques may be effective in promoting the learning of adaptive behavioral choices in children with ADHD. Keywords Attention Deficit Hyper activity Disorder · Differential Outcomes · Conditional Discrimination Learning · Reward · Treatment Background * Hasse De Meyer hasse.demeyer@kuleuven.be ADHD is marked by elevated levels of inattention, * Saskia van der Oord hyperactivity and impulsiveness that are inconsistent saskia.vanderoord@kuleuven.be with a child’s developmental stage (American Psychiatric Gail Tripp Association, 2013). As a result of these symptoms, children tripp@oist.jp with ADHD often fail to conform to the expected social, Tom Beckers cognitive and emotional requirements of their environment, tom.beckers@kuleuven.be increasing their risk of adverse outcomes later in Behaviour, Health and Psychopathology, KU Leuven, development (Wehmeier et al., 2010; Willcutt et al., 2005). Tiensestraat 102, 3000 Leuven, Belgium A critical skill for adaptive socio-emotional and cognitive Department of Psychology, Faculty of Behavioural Sciences, functioning is the ability to align one’s actions with the HELP University, Subang 2, Persiaran Cakerawala Seksyen, frequently changing expectations or requirements of U4, 40150 Shah Alam, Selangor, Malaysia the environment (e.g., being quiet and deferential in the Human Developmental Neurobiology Unit, Okinawa classroom, being energetic and assertive in the playground) Institute of Science and Technology Graduate University, (Urcuioli, 2005). Numerous researchers have observed that Okinawa 904-0495 1919-1 Tancha, Onna, Japan children with ADHD have difficulty adapting their behavior Leuven Brain Institute, KU Leuven, Leuven, Belgium to shifting environmental demands (Nigg & Casey, 2005; Developmental Psychology, University of Amsterdam, Sagvolden et al., 2005). Nieuwe Achtergracht 129, 1018 Amsterdam, Netherlands Vol.:(0123456789) 1 3 1166 Research on Child and Adolescent Psychopathology (2021) 49:1165–1178 A widely used paradigm for testing behavioral adaptation deficits linked to executive impairments (e.g., in working to environmental expectations in laboratory research is a memory) in children with ADHD (Dovis et al., 2012; Fosco Conditional Discrimination Learning [CDL] task (Martínez et  al.,  2015; Luman et  al.,  2005; Slusarek et  al.,  2001). et al., 2012; Mok et al., 2017). Conditional Discrimination Whether such an incentive-oriented manipulation would Learning can be tested in an arbitrary Matching-To-Sample also be effective in improving CDL performance is to be [aMTS] procedure in which associations have to be learned determined. between non-similar, non-related sample stimuli and choice Inspiration for a more cognitively oriented way to responses (Estévez et al., 2001; Trapold, 1970). On a given remediate impairments in adaptive choice behavior in trial participants are presented with a single sample stimulus, children with ADHD can be found in the associative e.g., S , followed by a choice between two responses R and learning literature. Providing Differential Outcomes [DO] 1 1 R , and they have to learn to select the correct response. or response-specific reinforcement is a frequently used Importantly, which choice response is correct is dependent technique to overcome learning and memory deficits in on the sample stimulus presented: R may be the correct clinical as well as non-clinical samples (Urcuioli, 2005). In choice for S , but for another stimulus S , R may be the DO, stimulus–response relationships are reinforced using 1 2 2 correct choice response. When a retention interval is response-unique, rather than general, outcomes; correctly inserted between the offset of the sample stimulus and the choosing R after presentation of S will result in outcome 1 1 appearance of the choice stimuli, the aMTS task becomes an O , while correctly choosing R after presentation of S 1 2 2 arbitrary Delayed Matching-To-Sample [aDMTS] task (Case will result in a different outcome O . This is in contrast et al., 2015; Skinner, 1950). Due to its conditional nature, to a non-Differential Outcomes [nDO] procedure, where this instrumental learning task models the capacity to adapt different outcomes are provided randomly (O or O ), 1 2 choice behavior to situational requirements or hierarchical irrespective of the response, or the standard Common reinforcement contingencies (Mok et al., 2009). Outcomes [CO] procedure where only one outcome is Despite its clear clinical relevance, there is very little used to signal correct responding across trials (Holden & research on CDL in ADHD (for exceptions, see De Meyer Overmier, 2014; Overmier & Linwick, 2001). The use of a et al., 2019; Gitten et al., 2006). The few available studies DO procedure allows for the formation of specific sample- show no evidence for a deficit in CDL learning in children outcome and response-outcome associations: besides an with ADHD compared to TD children when choice stimuli S-R association, an S-O association is formed that contains immediately follow the sample stimuli. However, in daily information on the specific outcome that can be earned for life a delay between environmental cues and behavioral correct responding to the sample stimulus (Urcuioli, 2005). choice is the rule rather than the exception (e.g., as Given that this specific outcome is also linked uniquely to a when a child is instructed to begin a new task after first specific choice response, the acquired S-O associations can completing another one). We recently showed that under help support correct choice behavior (i.e., the correct choice delay conditions (imposing a delay of 8 or 16 s between is not only supported by a direct S-R association but also by the sample stimulus and response choice), children with an indirect S-O-R associative chain) (Hochhalter & Joseph, ADHD show poorer learning than TD children on a CDL 2001; Mok & Overmier, 2007). Thus, when the sample task (De Meyer et al., 2019). In addition to being a good stimulus (S) is presented, it activates a prospective memory marker of children’s everyday ability to use environmental representation of a specific, to-be-earned outcome (S-O cues to adapt their behavior (Martínez et al., 2009), CDL association) that primes a specific choice response. This delay tasks can potentially be used as an indicator whether prospective memory representation can help to overcome the intervention strategies, aimed at improving the ability to memory load created by the imposition of a delay between adapt behavior to contextual demands have the potential to the disappearance of the sample stimulus and the appearance work. This knowledge can be used for further development of the choice stimuli (Overmier & Linwick, 2001). and testing of interventions in more ecologically valid In existing contingency management programs for designs and studies (e.g., micro trials; Staff et al., 2021). children with ADHD, explicitly linking specific rewards to One potential way to improve the degree to which certain behaviors (e.g., differently colored stickers or marbles children with ADHD adapt their behavior to environmental for various forms of situationally appropriate behaviors) as in expectations, and thus to increase task performance on a a DO procedure, is to our knowledge not specifically being CDL task, is by increasing the value of the associated taught in behavioral management programs (Staff et  al., reward. The use of a larger reinforcement (e.g., a large 2021). There is substantial evidence that DO has beneficial monetary reward, as compared to a small reward or effects on learning and memory in typically developing feedback only) has been shown capable of normalizing the children and adults (Plaza et al., 2018; Urcuioli, 2005) and on-task performance of children with ADHD compared to in other clinical groups (e.g., in Autism, Down’s, Prader- TD children, including the amelioration of performance Willi and Korsakoff syndromes, and in Alzheimer’s disease) 1 3 Research on Child and Adolescent Psychopathology (2021) 49:1165–1178 1167 (Esteban et al., 2014; Hochhalter & Joseph, 2001; Joseph which serves as the baseline for the current study, and et al., 1997; Vivas et al., 2018). Studies in Korsakoff and the associated delay which was used in the manipulations Prader-Willi patients show benefits of DO in delayed and non- reported here (see Fig.  1 – Phase 1). Two different delayed conditional discrimination learning tasks (Hochhalter procedures were sequentially evaluated for their effect on & Joseph, 2001; Joseph et al., 1997) and suggest that it has the children’s CDL performance: (1) in the first CDL task, potential for targeting forgetfulness in daily life in clinical €10 could be accumulated through correct responding (large patients (e.g., for remembering the intake of medication). reward condition) (see Fig. 1 – Phase 2); (2) in subsequent Given that Prader-Willi and Korsakoff syndromes share CDL tasks, we used a within-subjects manipulation to characteristics with ADHD (including the presence of compare the effects of DO and nDO on performance. In impairments in learning and memory) (Hochhalter & Joseph, DO, making the correct choice led to unique outcomes and 2001), we speculate that the use of DO may facilitate learning thus specific stimulus–response relations were learned. and attenuate the performance deficit that children with This differs from nDO where, non-specific outcomes were ADHD exhibit in conditional discrimination learning under provided after correct choice responses, irrespective of the delays (De Meyer et al., 2019; Martínez et al., 2009, 2013; response (within-subjects) (see Fig. 1 – Phase 3). Based on Overmier & Linwick, 2001). previous research (Dovis et al., 2012; Luman et al., 2005; Differential Outcomes procedures can be integrated Slusarek et al., 2001), we expected that both procedures (the in an aDMTS task in different ways, involving different provision of larger reinforcers as well as the provision of DO types of outcomes. In a standard implementation of DO, reinforcers for correct choice behavior) would improve CDL the outcomes used can be conditioned (or secondary) performance under delay conditions, in the two groups, with reinforcers (e.g., a token) and/or primary reinforcers a more pronounced effect for those with ADHD. (Estévez et  al., 2001; Martínez et  al., 2009, 2013). Between-subjects, we manipulated whether DO applied to The latter are hedonic reinforcers that are intrinsically secondary reinforcement only (with primary reinforcement motivating (e.g., food or water in animal research) being non-differential) or to secondary and primary (Estévez et  al., 2001; Martínez et  al., 2009, 2013). reinforcement (see Fig. 1 – Phase 3). This was achieved by Whereas often in DO studies response-specific secondary random assignment of the participants to either primary and response-specific primary reinforcers are used as and secondary DO or secondary DO only. We predicted outcomes (e.g., response-specific tokens that can later be that ADHD and TD groups would both benefit more from exchanged for token-specific candy), emerging evidence a DO procedure where primary and secondary reinforcers suggests that response-specificity at one level of outcome were response-unique than from a DO procedure where only (e.g., only at the level of the secondary reinforcer) may be secondary reinforcers were response-unique. We expected sufficient to achieve a beneficial DO effect. Then again, these effects would again be more pronounced in children children with ADHD are known to process reinforcement with ADHD, due to their altered reinforcement sensitivity. differently than TD children and may need more optimal reinforcement than TD children to perform well on tasks (Dovis et al., 2012; Luman et al., 2010). When studying Method DO in children with ADHD, it is therefore important to determine whether response-specific primary and Forty-six children with a prior diagnosis of ADHD (22 secondary reinforcement is superior to response-specific combined, 18 inattentive, and 6 hyperactive/impulsive) and secondary reinforcement only. 55 typically developing children participated in the study. In summary, the aim of the present study was to investigate Six participants (5 = TD, 1 = ADHD) were subsequently whether enhancing the value of reward (remediation 1) and/ excluded due to an error in administration; i.e., an incorrect or introducing differential outcomes (remediation 2) would delay was selected from the baseline aDMTS tasks for use ameliorate deficits in CDL performance, under conditions in the reward, nDO and DO tasks. The children, aged 8 of delay, observed in children with ADHD compared to TD to 12 years, were recruited through the clinical networks children. Within differential outcomes, we also assessed of the authors (ADHD group) and local schools (TD the importance of the degree of response-specificity of group). Study inclusion criteria were: (a) an estimated IQ reinforcement; i.e., is there a difference in the effectiveness score ≥ 80, based on the short form of the Dutch version of of response specific secondary reinforcer compared to the Wechsler Intelligence Scale for Children [WISC-III-NL] response specific primary and secondary reinforcement. (b) absence of any sensory, neurological or motor disorder Immediately before the start of the current study, all or a clinical diagnosis of Autism Spectrum Disorder (as children had performed a series of CDL tasks with increasing indicated by parents) (c) absence of a clinical diagnosis of delays to determine the delay at which their performance Conduct Disorder [CD] as assessed by the CD section of tapered off (De Meyer et al., 2019). It is that performance the Disruptive Behavior Disorders module of the Diagnostic 1 3 1168 Research on Child and Adolescent Psychopathology (2021) 49:1165–1178 Fig. 1 Study overview. Prior to the start of the current study, feedback only in the baseline condition) was administered. In phase the delay at which CDL performance declined was determined 3, aDMTS tasks were administered under nDO and DO conditions individually through administration of aDMTS tasks with increasing (with order counterbalanced across participants). Within the DO delays between sample and choice stimuli (Baseline, phase 1). manipulation, half of the participants were exposed to secondary In phase 2, an aDMTS task with the same delay but including a reinforcement only and the other half to secondary as well as primary monetary reward (possibility of obtaining 10 euros in addition to reinforcement, i.e., between-subject within the groups Interview Schedule for Children, Parent Version (PDISC; This composite score exhibits satisfactory validity and Shaffer et al.,  2000) and (d) not taking any medication other reliability (0.86 and 0.91) and is highly correlated with full- than stimulant medication (in the case of ADHD) which scale IQ (Sattler, 2001). participants were willing to withdraw 24 h prior to testing PDISC: The clinical assessment followed the algorithm (Greenhill, 1998). of the Diagnostic Interview Schedule for Children, Parent The diagnosis of ADHD was established by a certified Version (Shaffer et al., 2000). This interview, based on the psychiatrist or clinical psychologist and DSM-criteria were DSM-IV criteria, has adequate psychometric properties confirmed by the PDISC. Typically Developing children (test–retest reliability = 0.79) and is a reliable assessment were required to fall within the normal range on the tool to assess DSM symptoms of ADHD, ODD and CD th Inattentive and Hyperactivity/Impulsivity section (≤ 90.9 (Shaffer et al., 2000). The interviewers, licensed clinical percentile), Oppositional Defiant Disorder [ODD] section psychologists or Masters students in clinical psychology, nd nd (≤ 95.2 percentile) and CD section (≤ 95.2 percentile) were trained by the first author in administering the PDISC. of the Disruptive Behavior Disorder Rating Scale (DBDRS; DBDRS: The Dutch version of the Disruptive Behavior Dutch translation: Oosterlaan et al., 2008) as endorsed by Disorder Rating Scale (Oosterlaan et  al., 2008) contains parents. four DSM-IV-TR based scales assessing Inattention, Hyperactivity/Impulsivity, ODD and CD symptoms. The 42-item questionnaire is designed to be completed by parents Measures of children between six and sixteen years of age. Parents were asked to rate the behavior of their child on a 4-point Likert WISC-III-NL, short version: Vocabulary and Block Design, scale, ranging from 0 (not at all) to 3 (very much). Raw scores two subtests from the Dutch version of the WISC-III (Kort (ratings added across all symptoms) were transformed to norm th et al., 2005), were administered to estimate full-scale IQ. scores ranging between 10 (50 percentile, non-clinical) 1 3 Research on Child and Adolescent Psychopathology (2021) 49:1165–1178 1169 Fig. 2 Panel a – Baseline aDMTS procedure (retrieved from De Meyer incorrect responding is followed by a red cross (see Fig. 1 – Baseline). et  al., 2019). A randomly chosen sample stimulus is presented at the After 2  s, the next trial is presented.  Panel b – aDMTS task with top of a touchscreen. Upon touching the sample stimulus, the screen increased reward. Children are presented with the aDMTS task with is cleared. After a delay of 0, 8 or 16  s, three choice stimuli appear the relevant delay determined by their baseline performance. Correct and remain on the screen until the child responds by touching one responses yield smiley faces that accumulate towards a potential ten- of the stimuli. The child’s task is to learn to select the correct choice euro reward accompanied by specific reinforcement instructions “You stimulus for a given sample stimulus through trial and error. Correct have a higher chance at winning €10”. Incorrect responses yield a red responding yields a green or red smiley (randomly determined), cross (see Fig. 1 – Phase 2) th and 19 (99.9 percentile, clinical). Adequate psychometric third choice stimulus C was added on all trials. In order for properties are reported for a Flemish sample; internal participants to become acquainted with the task, a training consistencies for the Inattention (α = 0.90), Hyperactivity/ phase was presented prior to the first CDL task. Impulsivity (α = 0.87), ODD (α = 0.88) and CD subscales Across the task, there was a gradual increase in the retention (α = 0.66) (Oosterlaan et al., 2008) are moderate to high. interval between sample and choice stimuli (from 0 s through 8 s to 16 s) (see Fig. 1 – Baseline). Each retention interval involved Conditional Discrimination Learning Task: Baseline 24 trials (with the exception of the 0-s task, which included 36 Assessment trials). The appearance of the sample stimulus (S or S ) and 1 2 the position of the choice stimuli (C, C, C ) was determined 1 2 3 In phase 1, initial CDL performance was assessed through randomly for each trial (12 options) and a different set of stimuli repeated arbitrary Delayed Matching-To-Sample [aDMTS] was used for each delay. The correct sample-choice association tasks, as reported in De Meyer et al. (2019). In each aDMTS was determined in advance and not counterbalanced. With task, participants learnt arbitrary relationships between a new increasing delay between the sample and choice stimuli, a drop set of sample stimuli and choice stimuli (see Fig. 2 – Panel in learning performance is consistently observed, typically a); conditional upon the presentation of sample stimulus S , attributed to an increased memory load (Case et al., 2015). selection of C is the correct response and upon presentation The sample and choice stimuli for each CDL task were clearly of S , selection of C is the correct response. Children distinguishable, randomly chosen abstract figures from MS 2 2 learnt the correct (S -C and S -C ) associations through a Word 2008 presented in black on a white background square 1 1 2 2 feedback-based trial-and-error procedure; correct responses measuring 5 × 5 cm. Outcome stimuli were colored 10 × 10 cm were followed by a smiley face, incorrect responses were smiley figures. The task was presented on a 15-inch touchscreen. followed by a red cross. To increase the level of difficulty, a After performing the basic aDMTS tasks with increasing delays, the delay at which a participant failed to reach the criterion of 75% correct choices over the last 12 trials was A detailed description of the aDMTS task is provided in De Meyer used as the delay for that participant in the current study; final et al., 2019. 1 3 1170 Research on Child and Adolescent Psychopathology (2021) 49:1165–1178 performance at that delay during the basic aDMTS tasks is used accumulate towards non-differential primary reinforcers (both here as the children’s baseline. If participants achieved criterion types of smileys earn candy and toy rewards); For every blue for all tested delays, the 16-s delay was used for the current smiley you will earn blue points and for every yellow smiley you study and performance on that delay was used as the baseline will earn yellow points. At the end, you can exchange all the blue to which both of the remediation procedures were compared. points and all the yellow points together for toys and candy. The more points you earn, the higher the chance you have at earning Conditional Discrimination Learning Task: candy and toys (see Fig.  3 – Panel a). All task instructions Reinforcement Manipulations were explained to the participants and the researcher checked whether they understood all instructions. Participants were told For the first reinforcement manipulation (phase 2), (see Fig.  2 that they needed to obtain enough smileys in order to receive a – Panel b), the aDMTS task, with an individually determined reward. In effect, all children received identical rewards after the delay (see figure legend) was presented that included a 10 programmed 24 trials irrespective of their performance. For the euro monetary reward for correct choice behavior. At the task, the sample stimuli, choice stimuli and correct association beginning of the task, participants were informed about the were randomly determined by the computer program. change in reward outcome: ‘From now on, for every smiley In the nDO task, which could be presented before or after the you will earn a point. The more points you earn, the higher DO task, correct responses yielded a randomly colored smiley, the chance you have at winning ten euros. When you have i.e., either a black or white smiley, unrelated to the sample earned enough points, the game will end and you will see a presented (see Fig. 3 – Panel b). Children were told that smileys green screen’, which was assumed to maintain motivation over accumulated towards a reward (toys or candy, determined by the time (Dovis et al., 2012). The ten one-euro coins that the child experimenter), in a non-differential way; For every smiley you could earn were shown and placed in sight but out of reach; will earn a point. At the end, you can exchange all the points for they remained in view throughout the entire task. All children candy or toys. The more points you earn, the higher the chance received 10 euros at the end of testing, irrespective of their you have at candy or toys. To control for reward level across actual performance on the task. conditions (i.e., DO and nDO) and the potential influence on In phase 3, we evaluated the second reinforcement their motivation, upon completing the 24 trials in the nDO task, manipulation, that is if CDL performance could be improved children were rewarded with both toys and candy (explained through the use of Differential Outcomes as compared to non- as being a consequence of their outstanding performance) but Differential Outcomes. The order of nDO/DO and nature of DO unrelated to their actual performance (see Fig. 3 – Panel b). (primary and secondary vs secondary) were varied between In all tasks, CDL performance was measured as the participants, stratified for gender, age, and group. All children percentage of correct responses across all trials of the task performed both the nDO and DO tasks (within-subjects). The minus the first four (performance on the first four trials is nature of the DO task (primary and secondary DO or secondary determined by chance only). For each participant, this metric DO only) was manipulated between subjects. In the DO aDMTS was calculated for 1) the baseline task (i.e., the aDMTS task task, correctly choosing C after the presentation of S resulted from the baseline phase on which a participant failed to reach 1 1 in outcome O (a blue smiley), whereas correctly choosing the 75% criterion, with this delay used in all subsequent tasks; C after presentation of S resulted in a different outcome O 2) the aDMTS task from Phase 2 (increased reward); and 3) the 2 2 2 (a yellow smiley) (see Fig. 3 – Panel a). In order to test the two aDMTS tasks from Phase 3 (nDO and DO). influence of the degree of response-specificity of reinforcement (i.e., response specificity of secondary reinforcement only versus primary and secondary reinforcement), half of the children in Procedure each group (ADHD, TD) received primary and secondary DO, in that they were told that different smileys could later be Before participating in the study, separate information letters exchanged for different rewards: O (blue smiley) accumulated were given to parents and children, and informed consent towards candy and O (yellow smiley) towards a toy. At the was obtained from both parents. While parents completed the beginning of the task, participants were informed about the structured interview (ADHD group only) and questionnaires change in reward outcome: For every yellow smiley you will (DBDRS and demographic questionnaire), children performed earn yellow points and for every blue smiley you will earn blue the experimental tasks in a distraction-free room. All points. At the end, you can exchange all the yellow points for toys participants completed the basic aDMTS task immediately and all the blue points for candy. The more yellow points you before the tasks described here (baseline; for full results, see earn, the higher the chance you have at earning toys. The more De Meyer et al., 2019). After determining each participant’s blue points you earn, the higher the chance you have at earning baseline delay, used in all subsequent tasks, the aDMTS task candy. The other half of the children in each group received with monetary reward was conducted (Phase 2). All children secondary DO only. They were told that secondary reinforcers received the monetary reward immediately afterwards. 1 3 Research on Child and Adolescent Psychopathology (2021) 49:1165–1178 1171 Fig. 3 Panel a – DO aDMTS task. The basic aDMTS task was modi- aDMTS is presented in which correct responses yield a non-specific fied so that correct responses yield a sample-specific outcome (DO). outcome (nDO): Participants always receive non-differential sec- All participants receive differential secondary reinforcers (smileys of ondary reinforcers (randomly white or black smileys) that accumu- a sample-specific color). For half of the participants, those second- late towards a non-differential primary reinforcer (toys or candy, as ary reinforcers accumulate towards differential primary reinforcers instructed and determined by the experimenter) (see Fig.  1 – Phase (smileys of one color earn a toy reward, the other earns candy), for 3). In reality, children were rewarded with both toys and candy the other participants, secondary reinforcers accumulate towards non- (explained as being a consequence of their outstanding performance) differential primary reinforcers (both types of smileys earn candy and unrelated to their actual performance to control for reward level toy rewards) (see Fig. 1 – Phase 3). Panel b – nDO aDMTS task. An across conditions and the potential influence on their motivation Next, children performed the nDO and DO aDMTS tasks, The study was approved by the KU Leuven Social and in counterbalanced order (Phase 3). All tasks were separated Societal Ethics Committee (G-2015 01 156). The authors by a 10-min break. The experimenter remained in the room confirm that the study was conducted in line with the ethical throughout the testing procedure (± 100 min). All children standards of the institutional research committee and with the were able to complete the tasks and families were compensated 1975 declaration of Helsinki and its 2008 amendment. with an additional 10 euros for participating in the study. Results Review of the distribution of the outcome variables detected some extreme values (outliers) and high skewness and kurtosis for all outcome variables (unrelated to group), indicating The tasks were part of a larger test battery including tasks on work- ing memory and delay aversion (see De Meyer et al., 2019). 1 3 1172 Research on Child and Adolescent Psychopathology (2021) 49:1165–1178 Table 1 Demographic and ADHD TD Clinical Characteristics for the M(SD) M(SD) F/χ p ADHD and TD children Gender N 4.62 0.032* Male (N/%) 31 (68.89) 23 (46.94) Female (N/%) 14 (31.11) 26 (53.06) Age (years) 10.29 (0.99) 10.07 (1.21) 0.90 0.346 FSIQ 98.00 (11.72) 105.35 (9.84) 10.89 0.001** Dyscalculia (N/%) 1 (2.22) 0 Dyslexia (N/%) 4 (8.88) 0 ODD (PDISC – N/%) 44 (46.80) - Medication (N/%) 23 (51.11) 0 Maternal education 1.93 (0.82) 1.72 (0.74) 2.04 0.360 DBDRS (norm scores) Inattention 15.02 (2.01) 10.53 (1.00) 191.56 < 0.001*** Hyperactivity/Impulsivity 14.23 (2.44) 10.51 (0.98) 96.51 < 0.001*** ODD 12.42 (2.43) 10.71 (1.24) 18.57 < 0.001*** CD 11.42 (1.53) 10.96 (1.15) 2.68 0.105 PDISC (number of symptoms) Inattention 7.43 (1.65) - Hyperactivity/Impulsivity 6.00 (2.57) - ODD 3.23 (2.14) - CD 0.45 (0.79) - ADHD  Attention Deficit Hyperactivity Disorder, TD  Typically Developing, FSIQ  Full Scale IQ, PDISC  Diagnostic Interview Schedule for Children, Parent Version, DBDRS  Disruptive Behavior Rating Scale ODD Oppositional Defiant Disorder, CD Conduct Disorder *p < .05, **p < .01, ***p < .001 High (1) = University Education; Average (2) = Non-University Higher Education; Low (3) = Secondary Education (or less); 3 missing data points (1 = ADHD; 2 = TD) 2 missing data points for ADHD group non-normal distributions. Therefore, data was subjected to children with ADHD, F(1, 92) = 10.89, p = 0.001, ηp = 0.106. an arcsine transformation, as is recommended when outcome Neither IQ nor gender were included as covariates in the analysis variables are percentages (see also IBM Corp. 2019; Zar, 1984). as neither variable correlated with any of the outcome variables. After transformation, one extreme outlier was detected using To determine if adding a large reward or changing the boxplots and deleted from the dataset. For one other participant associative structure of the task improved CDL performance, the last 4 of 24 trials in the DO condition were missing. These two group x task repeated-measures ANOVAs were run, the missing values were replaced by the average score of the group first comparing performance on the baseline aDMTS task with (TD) for this variable (Field, 2013). The ADHD and TD groups the monetary reward aDMTS task across the ADHD and TD did not differ in mean age or family education level but did differ groups. The second one comparing performance on the DO and 2 = in gender distribution (χ 4.62, p = 0.032) with an uneven nDO aDMTS tasks across the two groups. Additionally, a group distribution of boys (N = 31) and girls (N = 14) in the ADHD x condition fixed-factors ANOVA was conducted to compare group compared with the control group (see Table 1 and the CDL performance between the two groups (ADHD/TD) when appendix for demographic characteristics). As often observed the DO manipulation involved secondary reinforcement only in studies of children with ADHD, there was a main effect of versus primary and secondary reinforcement. Post-hoc analyses, group for IQ with children in the TD group scoring higher than independent-samples t-tests and paired-samples t-tests were conducted to identify the source of the significant interaction effects. Effect sizes are reported for ease of interpretation; small 2 2 (ηp = 0.01; d = 0.2); moderate (ηp = 0.06; d = 0.5) and large Data analysis on transformed and untransformed data was applied (both parametric and non-parametric), leading to similar results. (ηp = 0.14; d = 0.8) (Cohen, 1988, 1992). Performance that is more extreme than three times the length of The distribution of the individually determined delays, th th the boxplot (between the 25 and 75 percentile), as marked with an selected on the basis of performance in the baseline tasks, was asterisk (*) on the boxplot. 1 3 Research on Child and Adolescent Psychopathology (2021) 49:1165–1178 1173 Fig. 4 CDL performance across tasks for children with ADHD and TD children. For ease of interpretation untransformed data is displayed. **p < 0.01 not significantly different between the groups, χ (2) = 5.75, ADHD compared to TD children (see Fig.  4, Table  2). Follow-up independent-samples t-tests indicated that p = 0.056. For the majority of children with ADHD and all TD children, a delay of 16 s was selected (ADHD: n = 40, children with ADHD differed significantly from the TD children at baseline t(79.18) = 3.29, p = 0.002, TD: n = 49); for the remaining children in the ADHD group, an 8-s (n = 3) or 0-s (n = 2) delay was selected. d = 0.69, but not following the addition of a monetary reward, t(92) = 1.56, p = 0.122, d = 0.32. Paired-samples The first 2 (group: ADHD vs TD) × 2 (condition: baseline vs monetary reward) repeated measures ANOVA t-tests showed that the difference in CDL performance between the baseline task and the monetary reward task yielded statistically significant main effects for condition, F(1, 92) = 6.97, p = 0.010, ηp = 0.070 and g roup, F(1, performance was significant for children with ADHD, t(44) = -2.85, p = 0.007, d = 0.43 but not for TD children, 92) = 9.45, p = 0.003, ηp = 0.093, and a significant condition by group interaction, F (1, 92) = 4.60, t(48) = -0.44, p = 0.666, d = 0.06. The effect of DO versus nDO was compared between p = 0.035, ηp = 0.048; adding a monetary reward had a larger impact on the CDL performance of children with groups in a 2 (condition; DO vs nDo) × 2 (group; ADHD vs Table 2 Percentage of Correct ADHD TD Responses (with Standard Deviations) and Univariate M (SD) M (SD) F p ηp ANOVA Results for Baseline, Baseline 72.53 (19.93) 85.51 (11.10) 11.10 0.001** 0.108 Reward, Differential Outcomes Reward 83.56 (12.23) 86.53 (11.69) 2.43 0.122 0.026 and Non-Differential Outcomes Tasks in the ADHD and TD Differential Outcomes 81.22 (16.89) 88.01 (7.72) 3.71 0.057 0.039 groups DO—secondary 80.42 (19.39) 88.94 (7.81) 2.15 0.149 0.043 DO—primary & secondary 82.14 (13.93) 86.96 (7.65) 1.53 0.223 0.035 Non-Differential Outcomes 76.33 (21.52) 87.86 (14.22) 12.71 0.001** 0.121 ADHD Attention Deficit Hyperactivity Disorder; TD Typically Developing ** p < .01 Based on untransformed data Based on transformed data We analyzed the data including only those children tested with a 16 s-delay. The main effects of condition and group remained signifi- cant. The previously significant interaction effect was no longer sig- nificant (p = 0.131). Degrees of freedom are for equal variances not assumed (signifi- cant Levene’s test). 1 3 1174 Research on Child and Adolescent Psychopathology (2021) 49:1165–1178 Fig. 5 DO performance across conditions for children with ADHD and TD children. For ease of interpretation untrans- formed data is displayed TD) repeated measures ANOVA. There was a signic fi ant main Discussion effect of group, F (1, 92) = 10.87, p = 0.001, ηp = 0.106, but not condition, F(1, 92) = 0.19, p = 0.665, ηp = 0.002. The group x Children with ADHD experience difficulty in using feedback condition interaction was significant, F(1, 92) = 5.37, p = 0.023, to adapt their behavior in the presence of delays (Conditional ηp = 0.055 (see Fig. 4, Table 2). Follow-up paired-samples Discrimination Learning). Here we evaluated the ability and independent-samples t-tests showed that performance of different reinforcement manipulations to improve did not differ significantly between nDO and DO for either performance on a CDL task. Specifically, we tested the group, t(48) = -1.46, p = 0.152, d = 0.23 (TD) and t(44) = 1.79, effects of increasing reinforcer size and the introduction p = 0.081, d = 0.26 (ADHD), however for children with ADHD, of Differential Outcomes (i.e. response-specific reward compared to TD children, the percentage of correct choices was outcomes). Within DO, we explored whether response- significantly lower under nDO, t (92) = 3.57, p = 0.001, d = 0.74, specific primary and secondary reinforcement was superior but not under DO, t(70.88) = 1.89, p = 0.063, d = 0.39. The to response-specific secondary reinforcement only. order in which the conditions were presented did not influence Contrary to our prediction that increasing reward size or the results when included as a covariate. value would have a positive effect on CDL performance under To explore whether a combination of response-specific conditions of delay in both groups, we found a significant secondary and primary reinforcement is more effective on CDL improvement in performance, i.e., a higher percentage of performance than response-specific secondary reinforcement correct responses compared with baseline conditions, in only, a 2 (condition; secondary vs primary and secondary) × 2 the ADHD group only. Under increased reward conditions, (group; ADHD vs TD) factorial ANOVA was run. Results the performance of the ADHD and TD group was no showed that there was no significant group, F (1, 90) = 3.56, longer significantly different, suggesting a normalization of p = 0.062, ηp = 0.038, or condition effect, F(1, 90) = 0.23, performance for the ADHD children. p = 0.636, ηp = 0.003, nor a group by condition interaction F(1, Similarly, adding a specific reward outcome to sample- 2 9 90) = 0.11, p = 0.740, ηp = 0.001 (see Fig. 5). Applying DO choice associations, i.e., DO, improved performance on the at secondary level only or at both primary and secondary level delayed CDL task in children with ADHD only. Although did not differentially impact performance in the two groups. a significant difference in performance was found between the groups when using non-differential outcomes, under DO reward conditions the difference between the ADHD group We reanalyzed the data including only those children tested with a 16  s-delay. The main effects of group remained significant. The previously significant interaction effect was no longer significant (p = 0.052). The order in which the conditions were presented did not influence Degrees of freedom are for equal variances not assumed. the results when included as a covariate. 1 3 Research on Child and Adolescent Psychopathology (2021) 49:1165–1178 1175 and the TD group was no longer significant. Again, this helps to overcome such aversion caused by delay, e.g., through suggests that adding DO may normalize CDL performance in increased attention towards the sample-choice association. children with ADHD. Further, we predicted that manipulating However, our previous study did not find an association between the nature of the reward within DO would affect performance. delay aversion and CDL performance under a delay (De Meyer The data do not support this hypothesis, i.e., primary and et al., 2019), rendering this explanation less likely. Alternatively, secondary DO performance did not differ from performance a more rewarding or response-unique outcome may simply under secondary DO only, for either group. serve to increase the motivation of children with ADHD to The observation of improved CDL performance through perform the CDL task as well as possible. However, given reward maximalization is in accordance with earlier findings reward intensity was equal across the DO and nDO conditions, that indicate beneficial effects of reward optimization on deficits a simple motivational account does not offer a convincing in Executive Functioning [EF] performance in ADHD (Dovis explanation for the effects of DO. Nevertheless, the results do et  al., 2012; Fosco et al., 2015; Slusarek et al., 2001). In a indicate an improvement in CDL performance through use of a DMTS task, a drop in accuracy under conditions of delay is large monetary reward for children with ADHD, removing the often attributed to deficits in short-term (Etkin & D’Amato, significant difference in performance between ADHD and TD 1969; Roberts & Grant, 1978) or working memory (Case groups that was observed at baseline. et al., 2015; Kempton et al., 1999), although the literature has Contrary to the results of earlier studies (e.g., Martínez et al., not addressed which specific memory aspect is involved in 2013; Mok & Overmier, 2007; Molina et al., 2015) and our own aDMTS. Our previous study (De Meyer et al., 2019), however, predictions, associating a specific outcome to a stimulus-choice suggested that neither short-term nor working memory was association (DO) did not significantly improve the performance related to performance on the aDMTS CDL task. The task of TD children. This may be due to a ceiling effect in the used to evaluate memory in that study (Corsi Block Tapping baseline performance of TD children, leaving limited room for Task; visual-spatial memory), may not have assessed memory change. Findings from previous research suggest a facilitating components required for CDL learning. The current study effect of DO on performance only when the task is sufficiently shows that adding a reward improves performance on a CDL challenging (e.g., a 4-cue task for adults) (Estévez et al., 2001; task under delay, although the specific mechanisms responsible Maki et al., 1995; Miller et al.,  2002). Despite care in task for this improvement have yet to be determined. development (balancing task dic ffi ulty for both groups through Improvement of CDL performance through a monetary pilot studies), DO would perhaps only facilitate performance in reward involves an increase in reward value, which might TD children in a more challenging task design. have served to remediate an underlying memory (short-term Additionally, it was hypothesized that the use of response- or working) deficit. The impact of DO on CDL performance specific secondary and primary reinforcement would might likewise be mediated by an effect on memory; the enhance performance as compared to response-specific nDO and DO conditions did not differ in reward size, but secondary reinforcement only. Unexpectedly, performance rather in the specificity of reward. Although the increase was similar in both DO conditions. It may be that presenting in performance from nDO to DO in children with ADHD a response-specific secondary reinforcer is already potent failed to reach significance (p = 0.081), DO did appear to enough to create a performance ceiling effect, thereby “normalize” performance in the ADHD group, i.e., they leaving little room for further improvement with response- performed more similarly to TD children on CDL. One specific primary reinforcement. It is also possible that the explanation for the effect of a DO procedure on aDMTS absence of a difference between these two forms of DO is performance is that it allows for the associative activation of related to the similarity in instructions between primary DO a prospective memory representation upon presentation of only and primary and secondary DO. The instructions given the sample stimulus that primes correct choice behavior. This to the children in both DO conditions were quite similar (see extra memory representation might have helped to counter Appendix), with children being promised candy and a toy possible short-term or working memory deficits that could in each (be it differential in one condition but not the other). impair aDMTS CDL performance in children with ADHD. The current study comes with some caveats. To begin with, An alternative explanation for increased performance on we did not succeed in collecting teacher ratings to confirm the the CDL task through increased reward or DO centers on their ADHD diagnosis for all participants due to practical constraints emotional-motivational effects (Sonuga-Barke, 2002). Within a (i.e., no response, children changing teachers, absence of CDL task, the delay between the sample and the choice stimuli contact information) and therefore cannot confirm the cross (and the associated reward) can trigger a negative emotional situational severity of symptoms, considered a core diagnostic state in children with ADHD, known as Delay Aversion (Antrop criterion in the diagnosis of ADHD. While acknowledging this et al., 2006), which can result in inattentive and hyperactive limitation, most of the children who entered the study had been symptoms and lead to decreased task performance (Marco et al., previously assessed and diagnosed through the KU Leuven 2009). The addition of a more salient reinforcer or DO possibly university hospital, by means of multi-method, multi-informant 1 3 1176 Research on Child and Adolescent Psychopathology (2021) 49:1165–1178 assessments where also cross-situational severity was taken only one token type to target various forms of situationally into account. Over the course of the study (including the appropriate behaviors (e.g., a sticker for sitting still during baseline testing reported in De Meyer et al., 2019), the aDMTS mathematics and for playing nicely with siblings) (Coelho task was administered four times. An influence of repeated et  al., 2015). Our results suggest that applying response- task administration on performance cannot be excluded, specific reinforcers may increase the learning of situation- although the significant group difference in nDO performance specific stimulus–response associations in children with (involving either the third or fourth aDMTS task, depending ADHD. This differential rewarding approach has already on counterbalancing) and the absence of a significant within- proven beneficial in a range of clinical populations (Esteban group difference between the baseline and nDO performance, et al., 2014; Hochhalter & Joseph, 2001; Overmier & Linwick, t(93) = -1.74, p = 0.086, argue against a simple task training 2001) and is often implemented in token economy programs effect. The current design did not allow us to control for the for children with Autism Spectrum Disorder (Fairbanks & contextual effect of being rewarded with a monetary reward Sugai, 2014; Neitzel, 2010), a neurodevelopmental disorder before the DO-nDO conditions, as the monetary reward with a significant overlap in clinical behavioral features and condition was always presented first, after which DO and nDO etiology with ADHD (Craig et al., 2015). were presented in counterbalanced order. Therefore, we cannot While the increase of reward value through monetary strictly rule out that delivering a large monetary reward ahead means also increases performance in children with ADHD, of the DO/nDO tasks differentially affected performance of the realistically, the addition of a high-value reward (e.g., 10 ADHD and TD groups. Another possible limitation relates to euro) is less feasible in real-life situations compared to the the stimuli used in the aDMTS task. Over the four aDMTS tasks relatively simple implementation of DO. In the present (baseline, reward, DO, nDO), the set of stimuli used was fixed study, DO was as effective as a high value of reward in and not counterbalanced. Therefore, performance differences increasing CDL performance in children with ADHD, and between aDMTS manipulations might, in principle, be due to the implementation of DO in token economy programs stimuli-specific characteristics. However, considerable care can be relatively easy (e.g., rewarding on-task behavior was taken to establish stimulus sets of equal difficulty. In with a red token and rewarding the raising of a hand before addition, it is possible that the task instructions influenced answering with a blue token). reward expectations differently for the DO and nDO conditions, In conclusion, the present research provides initial evidence favoring the DO condition. Although this effect was not evident that deficits in delayed conditional discrimination learning in in the performance of TD children (equal performance in DO vs ADHD on a DMTS task can be attenuated by enhanced reward nDO) we cannot rule out an ADHD specific differential effect. and DO manipulations. Our results have potential implications Finally, due the between-subjects manipulation of DO (primary for the refinement of behavioral interventions for children with and secondary versus secondary DO only), groups were rather ADHD. These findings can, for example inspire adaptations small, reducing power to detect significant differences between to existing token economies in ADHD, and further testing those two forms of DO. of these adapted “differential outcomes” token economies Despite the promising results, questions remain regarding versus “non-differential” token economies on proximal daily the mechanism underlying the DO phenomenon effect. Further life outcomes in micro-trails (Staff et  al., 2021). Further work is needed to disentangle what underlies the effectiveness research should also be directed at a better understanding of of DO. This is important to provide a better understanding of the mechanisms through which increased reward and DO exert its positive effects for children with ADHD. In future studies their beneficial effects on CDL performance. it would be important to test whether a DO effect can also be Supplementary Information The online version contains supplemen- achieved through other types of response-specific reinforcers tary material available at https://doi. or g/10. 1007/ s10802- 021- 00781-5 . (e.g., differential versus non-differential social reinforcers). Acknowledgements The authors would like to thank the participating Clinical Implications families for their contribution to the study. The findings of this study have a number of implications for Funding This work was supported by Research Grant G.0738.14 N of the Research Foundation Flanders (FWO). maximizing the impact of operant techniques in behavioral treatment for ADHD. In Behavioral Parent Training [BPT], a Declarations token economy is a widely used operant technique with the core aim of increasing adaptive and reducing inappropriate Conflict of Interest The authors declare that they have no conflicts of behavior in children with ADHD (Sullivan & O’Leary, interest. 1990). In a token economy, children are rewarded with Ethical Approval The study was approved by the KU Leuven Social specific tokens (e.g., marbles, stickers, etc.) for adaptive and Societal Ethics Committee (G-2015 01 156). The authors confirm behavior. A standard token economy, however, applies 1 3 Research on Child and Adolescent Psychopathology (2021) 49:1165–1178 1177 that the study was conducted in line with the ethical standards of the memory in children and in adults with Down syndrome. Research institutional research committee and with the 1975 declaration of Hel- in Developmental Disabilities, 35(6), 1384–1392. https://doi. or g/ sinki and its 2008 amendment.10. 1016/j. ridd. 2014. 03. 031 Estévez, A. F., Fuentes, L. J., & Marı́-Beffa, P., González, C., & Informed Consent Before participating in the study informed consent Alvarez, D. (2001). The Differential Outcome Effect as a Useful was obtained from both parents. Tool to Improve Conditional Discrimination Learning in Children. Learning and Motivation, 32(1), 48–64. https:// doi. org/ 10. 1006/ lmot. 2000. 1060 Open Access This article is licensed under a Creative Commons Attri- Etkin, M., & D’Amato, M. R. (1969). Delayed matching-to-sample bution 4.0 International License, which permits use, sharing, adapta- and short-term memory in the capuchin monkey. Journal tion, distribution and reproduction in any medium or format, as long of Comparative and Physiological Psychology, 69(3), 544– as you give appropriate credit to the original author(s) and the source, 549. https:// doi. org/ 10. 1037/ h0028 209 provide a link to the Creative Commons licence, and indicate if changes Fairbanks, S., & Sugai, G. (2014). Token Economy. Encyclopedia of were made. The images or other third party material in this article are Special Education, 3–4. https://doi. or g/10. 1002/ 97811 18660 584. included in the article’s Creative Commons licence, unless indicated ese24 03 otherwise in a credit line to the material. If material is not included in Field, A. P. (2013). Discovering statistics using IBM SPSS statistics. the article’s Creative Commons licence and your intended use is not Sage Publications. permitted by statutory regulation or exceeds the permitted use, you will Fosco, W. D., Hawk, L. W., Rosch, K. S., & Bubnik, M. G. (2015). need to obtain permission directly from the copyright holder. To view a Evaluating cognitive and motivational accounts of greater copy of this licence, visit http://cr eativ ecommons. or g/licen ses/ b y/4.0/ . reinforcement effects among children with attention-deficit/ hyperactivity disorder. Behavioral and Brain Functions, 11(1), 1–9. https:// doi. org/ 10. 1186/ s12993- 015- 0065-9 References Gitten, J. C., Winer, J. L., Festa, E. K., & Heindel, W. C. (2006). Conditional associative learning of spatial and object information American Psychiatric Association. (2013). Diagnostic and Statistical in children with attention deficit/hyperactivity disorder. Child Manual of Mental Disorders (5th ed.). https:// doi. org/ 10. 1176/ Neuropsychology, 12(1), 39–56.  https:// doi. or g/ 10. 1080/ appi. books. 97808 90425 596. 744053 09297 04050 02055 79 Antrop, I., Stock, P., Verté, S., Wiersema, J. R., Baeyens, D., & Greenhill, L. L. (1998). Childhood attention deficit hyperactivity Roeyers, H. (2006). ADHD and delay aversion: The influence of disorder: Pharmacological treatments. In P. E. Nathan & J. non-temporal stimulation on choice for delayed rewards. Journal Gorman (Eds.), A guide to treatments that work (pp. 42–64). New of Child Psychology and Psychiatry and Allied Disciplines, York: Oxford Universtity Press. 47(11), 1152–1158. https:// doi. org/ 10. 1111/j. 1469- 7610. 2006. Hochhalter, A. K., & Joseph, B. (2001). Differential Outcomes Training 01619.x Facilitates Memory in People with Korsakoff and Prader-Willi Case, J. P., Laude, J. R., & Zentall, T. R. (2015). Delayed matching Syndromes. Integrative Physiological and Behavioral Science, to sample in pigeons: Effects of delay of reinforcement and 36(3), 196–204. https:// doi. org/ 10. 1007/ BF027 34093 illuminated delays. Learning and Motivation, 49, 51–59. https:// Holden, J. M., & Overmier, J. B. (2014). Performance under differential doi. org/ 10. 1016/j. lmot. 2015. 01. 001 outcomes: Contributions of Reward-Specific Expectancies. Coelho, L. F., Barbosa, D. L. F., Rizzutti, S., Muszkat, M., Amodeo Learning and Motivation, 45(1), 1–14. https:// doi. org/ 10. 1016/j. Bueno, O. F., & Miranda, M. C. (2015). Use of cognitive lmot. 2013. 09. 001  behavioral therapy and token economy to alleviate dysfunctional IBM Corp. (2019). IBM SPSS Statistics for Macintosh, Version 26.0. behavior in children with attention-deficit hyperactivity disorder. Armonk, NY: IBM Corp. Frontiers in Psychiatry, 6, 1–9. https://doi. or g/10. 3389/ fpsyt. 2015. Joseph, B., Bruce Overmier, J., & Thompson, T. (1997). Food- and nonfood-related differential outcomes in equivalence learning by Cohen, J. (1988). Statistical power analysis for the behavioral sciences adults with Prader-Willi syndrome. American Journal on Mental (2nd ed.). Lawrence Erlbaum Associates. Retardation, 101(4), 374–386. Cohen, J. (1992). Statistical Power Analysis. Current Directions in Kempton, S., Vance, A., Maruff, P., Luk, E., Costin, J., & Pantelis, Psychological Science, 1(3), 98–101.  https:// doi. org/ 10. 1111/ C. (1999). Executive function and attention deficit hyperactivity 1467- 8721. ep107 68783 disorder: Stimulant medication and better executive function Craig, F., Lamanna, A. L., Margari, F., Matera, E., Simone, M., & performance in children. Psychological Medicine, 29(3), 527– Margari, L. (2015). Overlap Between Autism Spectrum Disorders 538. https:// doi. org/ 10. 1017/ S0033 29179 90083 38 and Attention Deficit Hyperactivity Disorder: Searching for Kort, W., Schittekatte, M., Bosmans, M., Compaan, E. L., Dekker, P. Distinctive/Common Clinical Features. Autism Research, 8(3), H., Vermeir, G., & Verhaeghe, P. (2005). WISC-III: handleiding 328–337. https:// doi. org/ 10. 1002/ aur. 1449 en verantwoording. Pearson. De Meyer, H., Beckers, T., Tripp, G., & van der Oord, S. (2019). Luman, M., Oosterlaan, J., & Sergeant, J. A. (2005). The impact Deficits in Conditional Discrimination Learning in Children with of reinforcement contingencies on AD/HD: A review and ADHD are Independent of Delay Aversion and Working Memory. theoretical appraisal. Clinical Psychology Review, 25(2), 183– Journal of Clinical Medicine, 8(9), 1381. https://doi. or g/10. 3390/ 213. https:// doi. org/ 10. 1016/j. cpr. 2004. 11. 001 jcm80 91381 Luman, M., Tripp, G., & Scheres, A. (2010). Identifying the Dovis, S., Van Der Oord, S., Wiers, R. W., & Prins, P. J. M. (2012). neurobiology of altered reinforcement sensitivity in ADHD: A Can motivation normalize working memory and task persistence review and research agenda. Neuroscience and Biobehavioral in children with attention-deficit/hyperactivity disorder? the Reviews, 34(5), 744–754. https:// doi. org/ 10. 1016/j. neubi orev. effects of money and computer-gaming. Journal of Abnormal 2009. 11. 021 Child Psychology, 40(5), 669–681.  https:// doi. or g/ 10. 1007/ Maki, P., Overmier, J. B., Delos, S., & Gutmann, J. A. (1995). s10802- 011- 9601-8 Expectancies as Factors Influencing Conditional Discrimination Esteban, L., Plaza, V., López-Crespo, G., Vivas, A. B., & Estévez, A. F. Performance of Children. The Psychological Record, 45(1), 45–71. (2014). Differential outcomes training improves face recognition 1 3 1178 Research on Child and Adolescent Psychopathology (2021) 49:1165–1178 Marco, R., Miranda, A., Schlotz, W., Melia, A., Mulligan, A., Müller, Experimental Psychology: Animal Behavior Processes, 4(3), U., & Sonuga-Barke, E. J. S. (2009). Delay and reward choice 219–236. https:// doi. org/ 10. 1037/ 0097- 7403.4. 3. 219 in ADHD: An experimental test of the role of delay aversion. Sagvolden, T., Johansen, E. B., Aase, H., & Russell, V. A. (2005). A Neuropsychology, 23(3), 367–380. https:// doi. or g/ 10. 1037/ dynamic developmental theory of attention-deficit/hyperactivity a0014 914 disorder (ADHD) predominantly hyperactive/impulsive and Martínez, L., Estévez, A. F., Fuentes, L. J., & Overmier, J. B. combined subtypes. Behavioral and Brain Sciences, 28(3), (2009). Improving conditional discrimination learning and 397–419. https:// doi. org/ 10. 1017/ S0140 525X0 50000 75 memory in five-year-old children: Differential outcomes effect Sattler, J. (2001). Assessment of Children: cognitive applications using different types of reinforcement. Quarterly Journal of (4th ed.). CA. Experimental Psychology, 62(8), 1617–1630. https://doi. or g/10. Shaffer, D., Fisher, P., Lucas, C. P., Dulcan, M. K., & Schwab- 1080/ 17470 21080 25578 27 Stone, M. E. (2000). NIMH Diagnostic Interview Schedule Martínez, L., Flores, P., González-Salinas, C., Fuentes, L. J., & for Children Version IV (NIMH DISC-IV): Description, Estévez, A. F. (2013). The effects of differential outcomes and Differences From Previous Versions, and Reliability of Some different types of consequential stimuli on 7-year-old children’s Common Diagnoses. Journal of the American Academy of Child discriminative learning and memory. Learning and Behavior, & Adolescent Psychiatry, 39(1), 28–38. https://doi. or g/10. 1097/ 41(3), 298–308. https:// doi. org/ 10. 3758/ s13420- 013- 0105-y00004 583- 20000 1000- 00014 Martínez, L., Marí-Beffa, P., Roldán-Tapia, D., Ramos-Lizana, Skinner, B. F. (1950). Are theories of learning necessary? Psychological J., Fuentes, L. J., & Estévez, A. F. (2012). Training with Review, 57(4), 193–216. https:// doi. org/ 10. 1037/ h0054 367 differential outcomes enhances discriminative learning and Slusarek, M., Velling, S., Bunk, D., & Eggers, C. (2001). visuospatial recognition memory in children born prematurely. Motivational effects on inhibitory control in children with Research in Developmental Disabilities, 33(1), 76–84. https:// ADHD. Journal of the American Academy of Child and doi. org/ 10. 1016/j. ridd. 2011. 08. 022 Adolescent Psychiatry, 40(3), 355–363. https://d oi.o rg/1 0.1 097/ Miller, O. T., Waugh, K. K. M., & Chambers, K. (2002). Differential 00004 583- 20010 3000- 00016 outcomes effect: Increased accuracy in adults learning kanji Sonuga-Barke, E. J. S. (2002). Psychological heterogeneity in with stimulus specific rewards. The  Psychological Record, AD/HD—a dual pathway model of behaviour and cognition. 52(3), 315–324. https:// doi. org/ 10. 1007/ BF033 95433 Behavioural Brain Research, 130(1–2), 29–36. https:// doi. org/ Mok, L. W., Estevez, A. F., & Overmier, J. B. (2017). Unique 10. 1016/ S0166- 4328(01) 00432-6 Outcome Expectations as a Training and Pedagogical Tool. Staff, A. I., Van den Hoofdakker, B. J., van der Oord, S., Hornstra, The Psychological Record, 60(2), 227–247. https:// doi. org/ 10. R., Hoekstra, P. J., Twisk, J. W. R., Oosterlaan, J., & Luman, 1007/ bf033 95705 M. (2021). Effectiveness of specific techniques in behavioral Mok, L. W., & Overmier, J. B. (2007). The differential outcomes teacher training for childhood ADHD: A randomized controlled effect in normal human adults using a concurrent-task within- microtrial. Journal of Clinical Child & Adolescent Psychology, subjects design and sensory outcomes. The  Psychological 1–17. https:// doi. org/ 10. 1080/ 15374 416. 2020. 18465 42 Record, 57(2), 187–200. Sullivan, M. A., & O’Leary, S. G. (1990). Maintenance following Mok, L. W., Thomas, K. M., Lungu, O. V., & Overmier, J. B. (2009). reward and cost token programs. Behavior Therapy, 21(1), 139– Neural correlates of cue-unique outcome expectations under 149. https:// doi. org/ 10. 1016/ S0005- 7894(05) 80195-9 differential outcomes training: An fMRI study. Brain Research, Trapold, M. A. (1970). Are expectancies based upon different 1265, 111–127. https:// doi. org/ 10. 1016/j. brain res. 2008. 12. 072 positive reinforcing events discriminably different? Learning Molina, M., Plaza, V., Fuentes, L. J., & Estévez, A. F. (2015). and Motivation, 1(2), 129–140. https:// doi. org/ 10. 1016/ 0023- The differential outcomes procedure enhances adherence to 9690(70) 90079-2 treatment: A simulated study with healthy adults. Frontiers in Urcuioli, P. J. (2005). Behavioral and associative effects of Psychology, 6(NOV), 1–7. https://doi. or g/10. 3389/ fpsy g.2015. differential outcomes in discrimination learning. Animal 01780 Learning & Behavior, 33(1), 1–21. https:// doi. or g/ 10. 3758/ Neitzel, J. (2010). Reinforcement for children and youth with autism bf031 96047 spectrum disorders: Online training module. Colombus: Vivas, A. B., Ypsilanti, A., Ladas, A. I., Kounti, F., Tsolaki, M., & OCALI. Estévez, A. F. (2018). Enhancement of Visuospatial Working Nigg, J. T., & Casey, B. J. (2005). An integrative theory of attention- Memory by the Differential Outcomes Procedure in Mild deficit/hyperactivity disorder based on the cognitive and Cognitive Impairment and Alzheimer’s Disease. Frontiers in affective neurosciences. Development and Psychopathology, Aging Neuroscience, 10(November), 1–7. https:// doi. or g/ 10. 17(3), 785–806. https:// doi. org/ 10. 1017/ S0954 57940 50503 763389/ fnagi. 2018. 00364 Oosterlaan, J., Baeyens, D., Scheres, A., Antrop, I., Roeyens, H., & Wehmeier, P. M., Schacht, A., & Barkley, R. A. (2010). Social and Sergeant, J. (2008). VvGK6-16: vragenlijst voor gedragsproblemen Emotional Impairment in Children and Adolescents with ADHD bij kinderen 6 tot en met 16 jaar. Pearson. and the Impact on Quality of Life. Journal of Adolescent Health, Overmier, J. B., & Linwick, D. (2001). Conditional Choice-Unique 46(3), 209–217. https://doi. or g/10. 1016/j. jadoh ealt h.2009. 09. 009 Outcomes Establish Expectancies That Mediate Choice Willcutt, E. G., Doyle, A. E., Nigg, J. T., Faraone, S. V., & Pennington, Behavior. Integrative Physiological and Behavioral Science, B. F. (2005). Validity of the Executive Function Theory of 36(3), 173–181. https:// doi. org/ 10. 1007/ BF027 34091 Attention-Deficit/Hyperactivity Disorder: A Meta-Analytic Plaza, V., Molina, M., Fuentes, L. J., & Estévez, A. F. (2018). Review. Biological Psychiatry, 57(11), 1336–1346. https:// doi. Learning and recall of medical treatment-related information in org/ 10. 1016/j. biops ych. 2005. 02. 006 older adults using the differential outcomes procedure. Frontiers Zar, J. H. (1984). Biostatistical Analysis. Journal of the American in Psychology, 9(FEB).  https:// doi. or g/ 10. 3389/ fpsy g. 2018. Statistical Association (2nd ed.). https://doi. or g/10. 2307/ 22854 23 Roberts, W. A., & Grant, D. S. (1978). An analysis of light-induced Publisher’s Note Springer Nature remains neutral with regard to retroactive inhibition in pigeon short-term memory. Journal of jurisdictional claims in published maps and institutional affiliations. 1 3

Journal

Journal of Abnormal Child PsychologySpringer Journals

Published: Apr 1, 2021

Keywords: Attention Deficit Hyperactivity Disorder; Differential Outcomes; Conditional Discrimination Learning; Reward; Treatment

There are no references for this article.