Individual Differences During an Inhibitory Control Training for 7- to 9-Year-Old Children: A Multilevel Approach
Carline J. M. van Heijningen Leiden University
Master program: Developmental Psychology (Research) Student number: 1208616
Supervisors: B. Westhoff and dr. A. C. K. van Duijvenvoorde External supervisors: dr. N. A. J. Steinbeis and dr. M. A. Schel
Abstract
Inhibitory control (IC) enables children to adapt to their environment and its norms and facilitates the pursuit of long-term goals. Therefore, improvement of children’s IC skills may be beneficial. The current study aimed to investigate the effectiveness of a five-week IC training program for children in improving their IC performance. Additionally, the role of individual differences in IC during training were examined.
The current study followed a pre-test – training – post-test format. Twenty-two 7- to 9-year-old children were randomly assigned to either an IC training or a sham training
condition. Both computerized, adaptive training programs were based on the stop-signal paradigm. Children who received IC training were instructed to inhibit their response when a stop-signal appeared, whereas children who received sham training were instructed to respond to go-signals and stop-signals.
Although children’s IC performance did not improve from pre- to post-test after IC training, children did show an improved IC performance during IC training on an individual level. We found that children’s IC performance improved most when they completed more home training sessions. These findings underline the importance of examining individual differences and learning curves during training to investigate whether a training might work better for some individuals than for others.
Keywords: inhibitory control training, stop-signal paradigm, individual differences, learning curve, children
Individual Differences During an Inhibitory Control Training for 7- to 9-Year-Old Children
Improving children’s ability to inhibit impulses can be beneficial for various aspects of children’s everyday life, such as their academic performance and social-emotional
development. For example, children are better able to pay attention to a teacher’s instruction, when they can suppress their own internal urges (e.g., to talk, daydream, or stand up) and ignore distractions in the classroom. Moreover, sharing with other children while resisting the temptation to act selfishly will help children to develop and maintain social relationships. This ability to inhibit impulses is often labeled as inhibitory control (IC; or inhibition), which can be defined as the ability “to control one’s attention, behavior, thoughts, and/or emotions to override a strong internal predisposition or external lure” (Diamond, 2013, p. 137). As demonstrated by these examples, IC enables children to adapt to their environment and its norms and facilitates the pursuit of long-term goals (Diamond, 2013; Duckworth, 2011).
IC is considered as the foundation of executive functions (Diamond, 2013; Miyake et al., 2000). Executive functions (EFs), which also include working memory and cognitive flexibility, are a broad concept that encompasses multiple cognitive processes that help individuals to think and act adaptively and goal-directed with an important underlying role of the prefrontal cortex (Diamond, 2013; Miller & Cohen, 2001; Miyake & Friedman, 2012; Hofmann, Schmeichel, & Baddeley, 2012). EFs are suggested to be essential for social-emotional development and successful outcomes in life, such as mental and physical health and academic achievement (for a review, see Diamond, 2013). For example, individual differences in IC during childhood appear to be an important predictor of several outcomes later in life (Moffitt et al., 2011). More specifically, poor childhood IC was related to ‘poor’ decision-making during adolescence (e.g., dropping out of school without any educational degrees). Additionally, individuals with poor childhood IC were at the age of 32 more likely
to develop substance abuse problems and other health-related problems, to show financial (management) difficulties, and to be convicted of a crime (Moffitt et al., 2011).
Although poor IC can have a negative impact on outcomes later in life, it is suggested that IC can be improved by training (Hofmann et al., 2012). Improvement of children’s IC skills may be beneficial for their learning and success within educational contexts, might prevent poor decisions during adolescence, and promote well-being and other outcomes during adulthood (Diamond & Lee, 2011; Jolles & Crone, 2012; Goswami, 2006; Moffitt et al., 2011). However, it is also suggested that individual differences in IC should be taken into account during training(Shah, Buschkuehl, Jaeggi, & Jonides, 2012). Although all children show to some degree IC difficulties due to the ongoing development of cognitive functions along with the associated process of brain maturation (Casey, Tottenham, Liston, & Durston, 2005; Jolles & Crone, 2012; Goswami, 2006; Melby-Lervåg & Hulme, 2013; Wass, Scerif, & Johnson, 2012), some children show relatively poor IC. Examining these individual
differences in IC as well as differences in children’s learning curve during IC training increases our understanding of their learning processes (Jolles & Crone, 2012; Shah et al., 2012). This might help us to develop more effective training programs based on children’s individual needs.
The Effectiveness of Cognitive Training Programs
Given the relevance of the different EFs and the suggestion that EFs are trainable (Hofmann et al., 2012), cognitive training has been an emerging area of research. However, studies examining the trainability of EF and the effectiveness of cognitive training programs differ to a great extent in for example targeted population, experimental paradigms to train EF, intensity, and duration. These factors complicate the comparison of training programs regarding their effectiveness (Karbach & Kray, 2009). Recently, however, it has been argued that training programs might be especially useful for children due to their ongoing
development and associated increased neural plasticity (Berkman, Graham, & Fisher, 2012; Karbach & Kray, 2009; Wass, Scerif, & Johnson, 2012).
Previous studies investigating the effectiveness of cognitive training programs show that children generally show an improved performance during training as well as on trained tasks and/or tasks that measure a similar construct as the trained task(s) from pre- to post-test. This has been shown for several cognitive training studies on different domains, such as working memory (WM) training (e.g., Bergman Nutley et al., 2011; Holmes, Gathercole, & Dunning, 2009; Jaeggi, Buschkuehl, Jonides, & Shah, 2011; Jolles, Van Buchem, Rombouts, & Crone, 2012; Klingberg et al., 2005; Klingberg, Forssberg, & Westerberg, 2002; Thorell, Lindqvist, Bergman Nutley, Bohlin, & Klingberg, 2009; Van der Molen, Van Luit, Van der Molen, Klugkist, & Jongmans, 2010), non-verbal reasoning training (Bergman Nutley et al., 2011), and task-switching training (e.g., Karbach & Kray, 2009; Kray, Karbach, Haenig, & Freitag, 2012). Although there is lack of evidence of long-term improvement on these domains or transfer to other, non-trained domains (Lervåg & Hulme, 2013; Melby-Lervåg, Redick, & Hulme, 2016; Shipstead, Redick, & Engle, 2012), EF training programs are potentially suitable interventions for typically developing children as well as children with poor EF performance.
The Effectiveness of IC Training Programs
Compared to studies regarding the trainability of other EF, only few studies have investigated the mechanisms of IC training. IC performance is mostly trained by training programs aimed at response IC with experimental paradigms, such as go/no-go paradigm and stop-signal paradigm (Bari & Robbins, 2013; Zanolie & Crone, 2018). Findings from studies with adult subjects that investigated the effectiveness of IC training programs appear to be inconsistent (e.g., Berkman, Kahn, & Merchant, 2014; Spierer, Chavan, & Manuel, 2013), whereas studies in children and/or adolescents are scarce.
A study that investigated the effectiveness of a training program for children combined IC and WM in a training due to the close relationship between both EFs (Diamond, 2013; Johnstone, Roodenrys, Philips, Watt, & Mantz, 2010). This high intensity, adaptive training program was not effective in improving children’s IC performance from pre- to post-test, even though their performance during training improved (Johnstone et al., 2010). Another study examining the effectiveness of IC training showed that across the IC training sessions, preschool children generally improved their performance on two of the three IC tasks which were trained (e.g., on go/no-go task and flanker task, but not on the stop-signal task). Despite their improved performance during training, again children did not improve their IC
performance on a trained and a non-trained IC task from pre- to post-test compared to children in active or passive control conditions (Thorell et al., 2009). In light of these findings, future studies are needed to investigate the extent of the trainability of IC.
The Importance of Examining Individual Differences
Many studies have investigated whether IC or other cognitive training programs led to improved performance on trained and non-trained tasks. To our knowledge, few studies have looked more closely at children’s performance during EF training to unravel the effect(s) of individual differences. However, in light of the inconsistent findings regarding the
(in)effectiveness of IC training programs, individual differences might play a crucial role in whether or not an individual is able to benefit from EF training (Shah et al., 2012).
A recent study examined the effectiveness of an adaptive, four-week WM training program and the role of individual differences (Jaeggi et al., 2011). The authors found that children significantly improved their performance on the trained task during the training. Next, they looked at the extent to which children improved during training and divided the group in children with either ‘small’ or ‘large’ training gains. Children who showed large gains during training performed significantly better on non-trained reasoning tasks compared
to children who showed small gains and children who received a sham training (i.e., control condition). This study indicates that individual differences during training might explain whether or not training demonstrates a transfer effect and underlines the importance of investigating individual differences in the ability to profit from a training and the rate of this learning process (Jaeggi et al., 2011; Willis & Schaie, 2009).
Additionally, individual differences in initial EF performance might play an important role in the effectiveness of EF training. Some studies suggested that individuals who show relatively poor EF performance are most likely to profit from an EF training (Diamond, 2013; Hofmann et al., 2012; Karbach & Kray, 2009). For example, a study examining the
effectiveness of an adaptive WM training found that children who showed lower initial WM performance and greater improvement during WM training performed better on tasks related to the trained task after training (Karbach, Strobach, & Schubert, 2015). Furthermore, a factor that may affect children’s performance during IC training is children’s level of impulsivity. Impulsivity and IC are interrelated, since poor IC manifests in for example impulsive behavior. Therefore, impulsivity can be seen as a consequence of poor IC skills (Bari & Robbins, 2013; Dalley, Everitt, & Robbins, 2011). A recent study found a strong positive correlation between two measures that both examine children’s IC skills in children. More specifically, lower stop-signal reaction time (SSRT; i.e., better IC skills) during a stop-signal task (SST) as a measure of response IC was related to less impulsive decision-making as measured by a temporal discounting task (Steinbeis, Haushofer, Fehr, & Singer, 2014).
To summarize, children’s initial EF performance as well as children’s learning curve during training might be important factors in the effectiveness of EF training. Due to the ongoing development in children, there is a large variability in individual performance (Titz & Karbach, 2014). When more attention is paid to individual differences instead of group differences (Karbach & Unger, 2014), EF training programs would be suitable as potential
interventions for typically developing children as well as children with poor EF performance. In the current study, we examined whether a training works better for some individuals than for others; as such, we hope to gain a more thorough understanding of mechanisms of change (Berkman, Graham, & Fisher, 2008; Shah et al., 2012; Willis & Schaie, 2009). This might help us to tailor interventions to children’s individual needs (Berkman, Graham, & Fisher, 2008; Karbach & Unger, 2014).
The Current Study
For the current study, we designed a five-week adaptive response IC training program based on previous research (e.g., Berkman, Kahn, & Merchant, 2014; Diamond, 2012; Diamond & Lee, 2011; Klingberg, 2010; Klingberg et al., 2005; Thorell et al., 2009). Additionally, we incorporated an overarching story and a game element to enhance and maintain children’s motivation. We aimed to investigate the effectiveness of this IC training program for 7- to 9-year-old children in improving their IC performance. Children were assigned to either an IC training condition or a sham training condition. Children in the IC condition performed a stop-signal task, which was focused on improving their IC
performance. On the contrary, children in the sham condition performed a highly similar stop-signal task, which did not include IC demands but only focused on enhancing children’s response speed. During the pre- and post-test, children’s IC performance and impulsive decision-making were assessed.
In line with previous research (e.g., Jaeggi et al., 2011; Karbach & Kray, 2009; Thorell et al., 2009), we expected that children in the IC training condition would significantly
improve their IC performance on a related IC task (e.g., show a significant decrease in stop-signal reaction time; SSRT) from pre- to post-test compared to children in the sham condition (hypothesis 1). Since the stop-signal paradigm does not only incorporate IC performance but also response speed demand (Band, Van der Molen, & Logan, 2003; Verbruggen & Logan,
2008), we expected that children in the IC as well as sham training condition would
significantly improve their response speed on an IC task (i.e., show a significant decrease in reaction time) from pre- to post-test (hypothesis 2).
An additional aim of the study was to investigate the role of individual differences in impulsivity on children’s IC performance during a five-week training program. In line with the first hypothesis (e.g., Thorell et al., 2009), we also expected that children in the IC training condition would significantly improve their IC performance across training sessions (hypothesis 3a). Regarding children’s initial level of impulsivity, we expected that children who showed higher levels of initial impulsive decision-making (impulsivity) would have a higher starting point (i.e., a higher SSRT, indicating poor IC performance) compared to children who showed lower levels of impulsive decision-making (hypothesis 3b). Lastly, we expected that children who can be seen as relatively impulsive would profit more from IC training than children who can be seen as relatively less impulsive (hypothesis 3c; Diamond, 2013; Jolles & Crone, 2012; Karbach & Kray, 2009; Karbach, Strobach, & Schubert, 2015).
Method Participants
Participants were 22 native Dutch-speaking 7- to 9-year-old children (13 boys; Mage = 8.55;
SDage = .57). Participants were recruited via teachers at five primary schools. Primary schools
as well as parent(s) or legal guardian(s) received an information letter and provided written informed consent based on their willingness to participate. Children provided verbal consent and were compensated for their participation (gift with a value of €10,-) after post-test administration. No exclusion criteria were specified. The study was approved by the Psychology Ethics Committee of Leiden University.
Design and Procedure
The current study followed a pre-test – training – post-test format. Within each school, children were paired based on gender and these pairs were randomly assigned to either an IC training condition or a sham training condition. All children were asked to perform four 15-minute training sessions per week (one supervised session at school and three unsupervised sessions at home) for a period of five weeks (see Table 1). Supervised training sessions at school were administered individually.
The study consisted of a two (time points; within-subjects factor) by two (conditions; between-subjects factor) mixed factorial design. Pre- and post-test (time points) were
administered individually (outside the classroom in a separate room) and lasted approximately 55 minutes each. The order in which the tasks were administered during the pre- and post-test differed and was counterbalanced across participants (see Table 1for an overview of
administered tests). This study was part of a broader research project and therefore, additional materials were administered to measure children’s prosocial decision-making, risk-taking behavior, self-efficacy, and motivation.
Due to motivational or technical difficulties with the (installation of the) training program, some children completed the training in less than five weeks (n = 3), were not able to train at home (n = 8), or did not complete all home training sessions (n = 8). Due to technical difficulties, data from seven children were missing from the first and/or second supervised training session(s). However, these participants were not excluded. Missing data was handled in further analyses.
Table 1. Overview of the procedure and administered tests
Time point Tasks
Pre-test Stop-Signal Anticipation Task Temporal Discounting Task Corsi Block Tapping Task
Raven's Standard Progressive Matrices Five-week training
program
Stop-Signal Task (SST) Each week:
- One supervised training session (at school; auditory SST) and a motivation questionnaire
- Three unsupervised training sessions (at home; visual SST) Post-test Stop-Signal Anticipation Task
Temporal Discounting Task Corsi Block Tapping Task
Raven's Standard Progressive Matrices
Materials
Training program. Children received either an IC (experimental) training or a sham
training. The computerized training program was programmed and presented in Presentation® software (Version 18.0, Neurobehavioral Systems, Inc., Berkeley, CA, www.neurobs.com). The training programs were based on the stop-signal paradigm (Logan, 1994) and adjusted for 7- to 9-year old children by incorporating an overarching story and adding a game component.
During a stop-signal task (SST), participants are instructed to respond fast on all trials, but simultaneously they are required to inhibit their response on a certain amount of trials (Band, Van der Molen, & Logan, 2003, p. 107; Verbruggen & Logan, 2008). Each training session, children were asked to perform a SST. They were instructed to press the spacebar as fast as possible when a stimulus (i.e., go-signal) appeared on the screen (see Figure 1a). However, on approximately 25 percent of the trials, a go-signal was followed by the presentation of a stop-signal (see Figure 1b). Children in the IC training condition were instructed to inhibit their response when a stop-signal appeared. Stop-signals were presented in both conditions and were presented in two modalities: auditory (during training sessions at school) and visual (during the training sessions at home). After the instructions, the instructor asked standardized control questions to ensure children understood the task correctly. The practice block included ten trials (i.e., seven à eight go-trials and two à three stop-trials). The training block included 120 experimental trials (i.e., 90 go-trials and 30 stop-trials) and lasted approximately 10 minutes.
Previous studies highlighted the need for adaptive training designs to ensure that participants will be challenged (Berkman, Kahn, & Merchant, 2014; Thorell et al., 2009; Diamond & Lee, 2011). Therefore, our training program was adaptive in terms of complexity to maintain a demanding level for each participant during the training task. Task difficulty was continuously adapted (i.e., after each trial) by increasing or decreasing the stop-signal delay (SSD) with 50 milliseconds, so that participants can successfully inhibit their response in approximately 50% of the stop-trials (Berkman, Kahn, & Merchant, 2014; Williams, Ponesse, Schachar, Logan, & Tannock, 1999). The SSD represents the time between the onset of a go-signal and the onset of a stop-signal and was originally set at 250 milliseconds after the go-signal (see Figure 2; Logan & Cowan, 1984; Verbruggen & Logan, 2008). The SSD was increased when children successfully inhibited their response, whereas
the SSD was decreased when children were not able to inhibit their response (Berkman et al., 2014). When the SSD was increased, it became more difficult to inhibit a go-response and vice versa (Bari & Robbins, 2013). In light of the adaptiveness of the training program, the SSD at the end of a training session was used at the beginning of the next training session.
The SST yielded two outcome measures. First, reaction time (RT) was a measure of response speed, which was calculated as the average RT across all go-trials in which children responded. Secondly, the stop-signal reaction time (SSRT) was a measure of IC performance, which represents the time that individuals require to stop their response when a stop-signal is presented after a go-signal (see Figure 2; Band et al., 2003; Logan, 1994). The SSRT was calculated according to the mean method (Logan & Cowan, 1984). A short SSRT indicates efficient IC performance, since individuals are able to successfully inhibit their response, even though the stop-signal is delayed. A relatively longer SSRT indicates poor IC performance. Therefore, individuals with a longer SSRT can be seen as relatively more impulsive than individuals with a shorter SSRT (Kim & Lee, 2011; Verbruggen & Logan, 2008). a. Go-trial (N = 90)
fixation screen stimulus feedback screen blank screen 1000-1500 ms 1000 ms 1000 ms 1000-2000 ms b. Stop-trial (N = 30)
fixation screen stimulus: go- and stop-signal feedback screen blank screen 1000-1500 ms 1000 ms 1000 ms 1000-2000 ms Figure 1. A visual representation of (a) go- and (b) stop-trials during the SST training program
Figure 2. A visual representation of a stimulus during the SST. Reprinted from “Release the BEESTS: Bayesian estimation of ex-Gaussian stop-signal reaction time distributions,” by D. Matzke, J. Love, T. V. Wiecki, S. D. Brown, G. D. Logan, & E. J. Wagenmakers, 2013, Frontiers in Psychology, 4, p. 2. Copyright 2013 by American Psychological Association.
The training program included some motivational elements. For example, children were told an overarching story, in which a monkey wanted to organize a party and he needed their help to collect items for his party. Children collected different items (e.g., bananas, balloons, animal friends) during each training session (at school as well as at home) on the route to the party. The different items were presented as stimuli during go-trials. During stop-trials in the auditory SST, go-signals were accompanied by a tone as a stop-signal (e.g., a balloon that ‘popped’ as a stop-signal). During stop-trials in the visual SST, go-signals changed into stop-signals (e.g., a yellow banana as go-stimulus turned into a rotten, brown banana as stop-stimulus; see Figure 1b). The story about the monkey and the description of the go- and stop-signals were incorporated in the standardized instruction (see Appendix A) at the beginning of each training session and in the control questions (e.g., “How can you help the monkey collect items?”). Children’s progress on the route to the monkey’s party was
shown at the end of each training session (see Appendix B). To maintain high levels of motivation, children received feedback on their performance during each training session. Coins were presented as feedback (see Appendix B) when children either successfully
inhibited their response (in the IC training condition) or when they successfully responded on a go-trial (i.e., faster than the previous go-trial; in both training conditions). Another feedback screen was presented to indicate when children responded too fast on a trial (i.e., before a go-stimulus was presented; see Appendix B).
The sham training was highly similar to the IC training (i.e., SST with a similar
adaptive design). However, the sham training focused on improving children’s response speed and did not include IC demands. Children in this training condition were instructed to respond as fast as possible to go-stimuli as well as stop-stimuli.
Stop Signal Anticipation Task. Children were asked to perform a stop-signal
anticipation task (SSAT) as a measure of their IC performance (Coxon, Stinear, & Byblow, 2007; Zandbelt & Vink, 2010), which is based on the stop-signal paradigm as described above. During the SSAT, children were instructed to respond when a moving object (i.e., squirrel) reached a target line (i.e., the nut), which represented a go-signal (see Figure 3). In approximately 20 percent of the trials, children were required to inhibit their response when the moving object automatically stopped before the target line, which represented the stop-signal (see Figure 3; Coxon et al., 2007; Zandbelt & Vink, 2010). During the first trial, the stop-signal was presented 250 milliseconds before the go-signal (SSD; see Figure 3). Task difficulty was continuously adapted (i.e., after each stop-trial) by increasing or decreasing the SSD with 50 milliseconds. When the SSD was decreased, it became more difficult to inhibit a go-response and vice versa (Bari & Robbins, 2013). The task consisted of ten practice trials and three experimental blocks. Each experimental block included 100 trials and lasted approximately 5 minutes.
Contrary to the SST, the SSAT includes cues which indicate the likelihood that a stop-signal will be presented. Since five different probabilities are included in the SSAT (see Figure 4), five proactive measures of inhibition (different RTs and SSRTs) can be obtained from this task. Our main focus lies on response inhibition and therefore, a general SSRT measure was obtained as a measure of IC performance. The SSRT was calculated according to the mean method (Logan & Cowan, 1984) and was used during subsequent analyses.
Figure 3. The presentation of a trial during the SSAT
Figure 4. A visual representation of the five different cues, which indicate the likelihood of the presentation of a stop-signal (from left to right: 0%, 17%, 20%, 25%, and 33%)
Temporal Discounting Task. A computerized temporal discounting (or delay of
gratification) task (Ainslie, 1975; Mazur, 1987) was used to measure children’s impulsive decision-making. During this task, children were asked to choose between two rewards in monetary units (MUs): a smaller, immediate reward (e.g., two, four, or six MUs) or a larger, delayed reward (e.g., eight MUs after four, eight, 28, or 56 nights). Children were told that the options were hypothetical as they would not receive these rewards themselves. All available
800 ms
550 ms SSD
250 ms
200 ms
pairs of choices (three immediate options timesfour delayed options) resulted in twelve trials. Based on children’s choices for either immediate options or delayed options on these twelve trials, four indifference points were estimated for each individual. An indifference point can be described as the subjective value when participants do not differentiate between the smaller, immediate reward and the larger, delayed reward (Myerson, Green, &
Warusawitharana, 2001). Next, the area under the curve (AUC) was determined according to the procedure of Myerson and colleagues (2001). The AUC was used as a measure of
impulsive decision-making. When participants prefer smaller, immediate rewards and assign a low subjective value to larger, delayed rewards, this will result in a small AUC (steeper discounting). Therefore, individuals with smaller AUC values can be seen as relatively more impulsive than individuals with higher AUC values (Myerson, Green, Hanson, Holt, & Estle, 2003; Myerson et al., 2001; Scheres et al., 2006).
Additional Materials. Two additional materials were administered to establish
whether the two training conditions were comparable at baseline (pre-test). First, Raven’s Standard Progressive Matrices (Raven, Raven, & Court, 2003; Raven, Raven, & Court, 2004) were administered to obtain a general indication of children’s intelligence level. Secondly, a computerized version of the Corsi Block Tapping Test (Corsi, 1972) was used to measure children’s working memory performance.
Statistical Analyses
To examine the effectiveness of the IC training in improving children’s IC performance (hypothesis 1), a factorial/ mixed analysis of variance (ANOVA) was performed to take into account between-subjects as well as within-subjects variables. Children’s IC performance (SSRT during the SST) was used as dependent variable, time (pre- and post-test) as within-subjects variable, and condition (IC training and sham training) as between-within-subjects variable. An additional factorial/ mixed ANOVA was performed to examine the effectiveness of the
response speed demand in the sham training condition as well as in the IC training condition (hypothesis 2), in which children’s response speed (RT on go-trials) was used as dependent variable.
To examine individual growth trajectories of children’s IC performance across training sessions (at school) and the effect of individual differences in impulsivity on IC performance (hypothesis 3), we performed a multilevel analysis. A multilevel analysis takes into account the dependency between observations (multiple observations within individuals) in
hierarchical, longitudinal data. Additionally, a multilevel analysis can handle missing and unbalanced data and time-varying predictor variables (i.e., training sessions over time; Singer & Willett, 2003). A multilevel (mixed models) analysis was performed in R, using the R package lme4 (Bates, Mächler, Bolker, & Walker, 2014). P-values were calculated by Satterthwaite’s approximation as used in R package lmerTest (Kuznetsova, Brockhoff, & Christensen, 2017).
After data exploration, we started the model-building process (see Table 2). Children’s IC performance as measured by the SSRT during the SST was used as dependent variable. We anticipated on large variability in the number of completed unsupervised training sessions (at home). Therefore, the number of training sessions at school (Session; with a maximum of five sessions) was used as within-subject, time-varying predictor variable. First, a null model (random intercepts only; Model 1) was performed in order to calculate the intra-class
correlation (ICC) as an indication of whether a multilevel analysis was needed (Meyers, Gamst, & Guarino, 2013). To examine whether there were differences in intercepts and/or slopes, we first added a fixed, linear effect of Session (Model 2). Based on visual inspection of the data, we added a quadratic effect of Session to the model (Model 3). Afterwards, the random effect of Session was added (Model 4). Model 4 was considered as a basis before adding children’s initial (pre-test) level of impulsive decision-making (AUC, centered) as
between-subjects, time-invariant predictor. First, we added the main effect of AUC as well as the interaction effect between Session and AUC (Model 5). Next, we added the interaction effect between the quadratic effect of Session and AUC (Model 6). Additionally, we wanted to control for children’s total number of completed home training sessions (Total) by
including the main effect as well as the interaction between total number of completed home training sessions and Session (Model 7). Lastly, we removed all non-significant terms to obtain a model that provided the best fit to the data (Model 8). The fit indices Akaike Information Criterion (AIC; Akaike, 1974) and Bayesian Information Criterion (BIC; Schwarz, 1978) were used to compare the models, in which lower AIC and BIC indices represented a better fit to the data. Likelihood Ratio Tests (LRT) were used to compare (nested) models. All models were performed using Full Maximum Likelihood estimation (FML) to examine fixed effects as well as random effects (Singer & Willett, 2003).
Table 2. Overview of the subsequently fitted models Model Model-building steps
1 Null model (random intercepts only) 2 Model 1 plus Session (fixed effect of time)
3 Model 2 plus Session_quadratic (fixed, quadratic effect of time) 4 Model 3 plus Session (random effect of time)
5 Model 4 plus impulsive decision-making (AUC) and Session * AUC 6 Model 5 plus Session_quadratic * AUC
7 Model 6 plus total completed sessions (Total) and Session * Total 8 Final model: Removing all non-significant terms
Results Initial Group Comparisons
Both training groups did not significantly differ on children’s gender (χ2(1) = 1.88, p = .665)
and age in years (t(20) = -.61, p = .550). Furthermore, children’s general intelligence level as measured by Raven’s Standard Progressive Matrices (t(20) = 1.15, p = .264), initial working memory performance as measured by Corsi Block Tapping Test (forward span: t(15.06) = -.41, p = .685; backward span: t(20) = -.50, p = .621), and initial IC performance (t(20) = .62, p = .543) did not significantly differ between conditions prior to the training (see Table 3).
Table 3. Descriptive statistics of initial group comparison measures
n Age (years) RSPM WM forward WM backward IC Condition M SD M SD M SD M SD M SD IC training 11 8.09 .70 33.54 5.89 4.55 1.29 4.45 1.13 214.35 27.13 Sham training 11 7.91 .70 36.18 4.81 4.36 .67 4.18 1.40 220.64 20.01 RSPM = Raven’s Standard Progressive Matrices (total correct); WM forward = WM forward span performance; WM backward = WM backward span performance; IC = IC performance (SSRT)
Effectiveness of the IC Training
To examine the effectiveness of the IC training in improving children’s IC performance (hypothesis 1) and their response speed (hypothesis 2), two factorial/ mixed analyses of variance (ANOVA) were performed. Descriptive statistics are reported in Table 4.
Firstly, children’s IC performance was examined. The main effect of time was not significant (F(1) = .18, p = .673), which indicated that children’s IC performance did not improve from pre- to post-test regardless of condition. The main effect of condition was also not significant (F(1) = .03, p = .864), which indicated that children’s IC performance did not significantly differ between conditions. Furthermore, the interaction effect between time and condition was not significant (F(1) = .48, p = .495), which indicated that children’s IC
performance from pre- to post-test did not differ between conditions (see Table 4 and Figure 5). To conclude, contrary to our expectations, the IC training was not effective in improving children’s IC performance.
Secondly, children’s response speed was examined. The main effect of time was not significant (F(1) = .19, p = .668), which indicated that children’s response speed did not increase from pre- to post-test regardless of condition. In other words, children did not
significantly respond faster after training. The main effect of condition was significant (F(1) = 6.48, p = .019), which indicated that children in the IC condition showed a significantly higher response speed (i.e., a lower go RT) than children in the sham condition (see Table 4 and Figure 6). Furthermore, the interaction effect between time and condition was not
significant (F(1) = .77, p = .389), which indicated that children’s response speed from pre- to post-test did not differ between conditions (see Figure 6). Contrary to our expectations, these results indicated that although children’s response speed differed across conditions, response speed did not increase from pre- to post-test in both conditions.
Table 4. Children’s IC performance and response speed on pre- and post-test across conditions
Pre-test Post-test
Condition M SD M SD
IC performance (SSRT) IC training 214.35 27.13 216.18 35.31 Sham training 220.64 20.01 212.95 19.00 Response speed (go RT) IC training 808.13 39.76 799.96 29.46 Sham training 773.37 32.75 776.13 15.57
Figure 5. Children’s IC performance (SSRT) from pre- to post-test across conditions
Figure 6. Children’s response speed (go RT) from pre- to post-test across conditions
0 25 50 75 100 125 150 175 200 225 250 275 300 Pre-test Post-test IC pe rf orm an c e (S S RT i n m s ) Sham training IC training 500 550 600 650 700 750 800 850 900 950 1000 Pre-test Post-test Res po ns e s pe ed ( RT i n m s ) Sham training IC training
Individual Differences across Training Sessions
Data exploration. Before we fitted the multilevel models (see Table 2), the
relationship between children’s IC performance and training sessions was explored (see Table 5 and Figure 7). We observed a strong decrease in SSRT from session 0 to session 1 and this relationship seemed to flatten out from session 1 to session 4. This led to the conclusion that the overall relationship between IC performance was not strictly linear. Therefore, we included a quadratic effect of session in our model (Model 3).
Additionally, a negative, strong Pearson’s correlation (r = -.669) between subject-specific intercepts and subject-subject-specific slopes was computed and showed that children with a higher intercept showed a larger decrease of their IC performance across training sessions compared to children with a lower intercept.
Model-building process. Based on the null model (random intercepts only; Model 1),
the ICC was calculated (σ02 / (σ02 + σe2) = 584.6 / (584.6 + 6688.8) = 0.080375). The ICC
indicated that 8.0% of the total variance of children’s IC performance can be attributed to between-subject differences (Meyers, Gamst, & Guarino, 2013). This low ICC implied that a multilevel is not needed. Nevertheless, we decided the use of a multilevel analysis due to clear dependency between multiple observations within individuals in the design of the study.
Next, the models were fitted as previously described (see Table 6 for an overview of parameter estimates and fit information). A quadratic effect (Model 3) significantly improved the model as indicated by a LRT (χ2(1) = 13.24, p < .001) and lower AIC and BIC values
compared to Model 2. A random effect of session (Model 4) did not significantly improve the model as indicated by a LRT (χ2(2) = .17, p = .920) and elevated AIC and BIC values
compared to Model 3 (see Table 5). Therefore, the final model did not include a random effect of session, which indicated that the effect of session (slopes) was similar for all children. We hypothesized that trajectories of IC performance over time would differ by
children’s initial level of impulsive decision-making (hypothesis 3b and c). However,
impulsive decision-making (AUC) and the interaction effects between AUC and session were non-significant predictors (Model 5 and 6) and led to higher AIC and BIC values of these models, which indicated a poorer fit of the model. Therefore, these non-significant predictors, AUC, session * AUC, and session_quad * AUC, were removed from the final model (see Table 6).
Final model. The intercept of the final model was significant (t(37.29) = 12.68, p <
.001), which indicated that the IC score after one training session (session 0) was 449.05 for an average child within the IC training condition who did not complete any training sessions at home. Furthermore, the estimates of session (-100.39) as well as the quadratic effect of session (21.44) were significant (respectively, t(39.59) = -3.95, p < .001 and t(39) = 4.55, p < .001) This indicated that the IC scores (SSRT) showed a decrease of 100.39 across training sessions for an average child within the IC training condition. However, this predicted
decrease of 100.39 was diminished by 21.44 for each session (see Figure 8). The total number of completed training sessions at home (total; M = 5.91, SD = 7.10) was a significant
covariate in the model (t(36.75) = 2.08, p = .044). The significant interaction effect between sessions and total number of completed home training sessions (-3.04, t(42) = -3.29, p = .002) indicated that children who completed more training sessions at home showed greater
improvement of IC performance across training sessions.
Table 5. Descriptive statistics of children's IC performance (SSRT) across training sessions (at school)
Training session n M SD Minimum Maximum
Session 0 8 492.53 82.22 385.42 673.30
Session 1 8 347.98 50.73 267.72 403.01
Session 2 11 340.89 59.22 223.67 414.88
Session 3 11 336.75 82.81 224.45 524.37
Figure 7. Children’s individual trajectories (n = 11) and average trajectory of IC performance (SSRT) across training sessions (at school) within the IC training condition
Figure 8. Predicted values based on the linear and quadratic effect of session (final model) 0 50 100 150 200 250 300 350 400 450 500 550 600 650 700 750 800
Session 0 Session 1 Session 2 Session 3 Session 4
IC pe rf orm an c e (S S RT i n m s ) 1 2 3 4 5 6 7 8 9 10 11 Mean 0 50 100 150 200 250 300 350 400 450 500
Session 0 Session 1 Session 2 Session 3 Session 4
P red ic ted IC pe rf orm an c e (S S RT i n m s )
Table 6. Overview of parameter estimates and fit information of the fitted models
Parameter Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Model 8
Intercept 365.42 427.04*** 473.28*** 475.14*** 472.32*** 477.52*** 446.99*** 449.05*** Session -28.03*** -111.55*** -112.80*** -113.31*** -119.42*** -106.56*** -100.39*** Session_quad 20.38*** 20.58*** 20.56*** 21.87*** 22.81*** 21.44*** AUC 32.02 -4.76 6.88 Session * AUC 16.54 77.96 77.34 Session_quad * AUC -14.75 -17.50 Total 6.35* 5.85* Session * Total -3.02** -3.04** (σe2) 6688.8 5095.0 3643.9 3467.05 3481.45 3384.3 2842.0 2866.2 (σ02) 584.6 597.4 851.3 987.32 872.44 1015.5 558.3 778.9 (σ12) 78.43 53.55 76.3 .00 (ρ01) -0.40 -0.58 -0.64 1.00 Log Likelihood -287.1 -281 -274.3 -274.3 -273 -272.7 -267.5 -268.9 Deviance 574.3 561.9 548.7 548.5 546.0 545.4 535.0 537.9 Dfresiduals 46 45 44 42 40 39 37 42 AIC 580.3 569.9 558.7 562.5 564.0 565.4 559.0 551.8 BIC 585.9 577.5 568.1 575.8 581.0 584.4 581.7 565.0
Note. σe2 = unexplained within-subject variance; σ02 = unexplained between-subject variance in intercepts; σ12 = unexplained between-subject variance in slopes; ρ01 = correlation between subject-specific intercepts and slopes (Singer & Willett, 2003)
Discussion
The current study aimed to investigate the effectiveness of an IC training for 7- to 9-year-old children. Our main finding is that our IC training was ineffective, since it neither improved children’s IC performance nor their response speed from pre- to post-test. Additionally, we aimed to investigate individual trajectories during IC training as well as the role of individual differences in impulsivity on children’s IC performance during training. During IC training, children improved their IC performance, although their improvement seemed to decrease over time. Children who completed more training sessions at home showed greater improvement of IC performance across training sessions at school.
Our main finding is in line with previous studies demonstrating the ineffectiveness of IC training programs for children (Johnstone et al., 2010; Thorell et al., 2009). However, this finding was contrary to our expectations in light of the strengths of our training program. A possible explanation for these null findings regarding the effectiveness of IC training
programs might be the design of IC tasks (Thorell et al., 2009). These tasks require inhibition in approximately one fourth of the trials and, therefore, they do not focus exclusively on IC. As a result, the time spent on specifically training IC is relatively short (Thorell et al., 2009). However, the stop-signal paradigm is a paradigm that is commonly used to train IC
performance (Bari & Robbins, 2013; Zanolie & Crone, 2018). Another suggested alternative explanation is that the trainability of IC differs from the trainability of other EFs, such as WM and task-switching (e.g., Klingberg, 2010). In other words, IC might be less easily improved by means of a cognitive training (Thorell et al., 2009). However, an alternative explanation for our null findings is that our sample size is too small to detect training effects due to limited power. Hence, the (null) results should be interpreted with caution and future studies with larger sample sizes are needed to investigate the extent of the trainability of IC in children.
Although children’s IC performance did not improve from pre- to post-test on group level, children in the IC training condition did improve their IC performance during training on an individual level, which is in line with previous research (e.g., Johnstone et al., 2010; Thorell et al., 2009). However, this finding might have been mainly driven by a strong observed improvement in IC performance during the second training session compared to the first session. Prior studies suggested that this improvement might not represent actual
improvement (e.g., Jolles & Crone, 2012; Thorell et al., 2009; Van der Molen et al., 2010), but instead children’s increased understanding of the task after the first session (Thorell et al., 2009). More specifically, the decrease in children’s improvement across training sessions might be explained by children’s acquaintance with the task and their adopted strategy to complete each training session (Jolles & Crone, 2012).
Contrary to our expectations, based on the close relationship between IC and impulsivity, individual differences in impulsivity did not affect children’s IC performance across training sessions. In other words, children who can be seen as relatively impulsive did not profit more from IC training than children who can be seen as less impulsive. Again, these findings should be interpreted with caution due to the small sample size. Although individual differences in impulsivity did not affect children’s IC performance during training, we did find that children who completed more training sessions at home showed greater
improvement of IC performance across training sessions at school. This finding is in line with previous research demonstrating that repeated practice leads to improvement (e.g., Diamond, 2012; Ericson, Nandagopal, & Roring, 2009; Klingberg et al., 2005).
Limitations and Strengths
Although the stop-signal paradigm is a valid and reliable paradigm to measure children’s response IC performance (Band et al., 2003), it is questionable whether this paradigm is ecologically valid. Children are instructed to inhibit their response when an
external cue (i.e., a stop-signal) is presented. However, in everyday (social) contexts children are required to intentionally inhibit unwanted behavior (e.g., Schel, Scheres, & Crone, 2014), such as suppressing the urge to talk or stand up in the classroom during a teacher’s
instruction. Thus, IC that is required during the stop-signal paradigm differs from intentional IC, since intentional IC is motivated by internal considerations rather than explicit external cues (Filevich, Kühn, & Haggard, 2012; Haggard, 2008; Schel et al., 2014). In addition to response IC paradigms, an intentional IC paradigm can be included, such as the marble paradigm. During this paradigm, participants are instructed, on several specific trials, to decide for themselves whether they want to inhibit their response or not (Kühn, Haggard, & Brass, 2009). By including an additional intentional IC paradigm to train IC, a more complete picture of the trainability of IC can be obtained.
Despite we did not observe effectiveness of our IC training, possibly due to our small sample size, the current study shed light on the trainability of IC and children’s ability to learn during training. Based on previous literature (e.g., Berkman, Kahn, & Merchant, 2014;
Diamond, 2012; Diamond & Lee, 2011; Diamond & Ling, 2016; Morceau & Conway, 2014), we designed an adaptive response IC training program, which included several key
components. Most importantly, our training program was adaptive in terms of complexity to maintain a demanding, challenging level for each participant during training. An adaptive design such as this is most likely to lead to improvement (e.g., Diamond, 2012; Diamond & Ling, 2016). Furthermore, the cover story and game element were also aimed at enhancing and maintaining children’s motivation, since motivation is suggested to be a crucial aspect for learning (e.g., Diamond, 2012; Wass et al., 2012). Thus, our training program included
various motivational components, which are essential to train IC in children.
Other strengths of the current study are the random assignment of children to conditions and the use of an active sham training as control condition, which enabled us to
adequately examine the effectiveness of the IC training (e.g., Wass et al., 2012; Jolles & Crone, 2012). More specifically, the sham training was highly similar to the IC training with respect to the overarching story, visual presentation, adaptiveness, intensity, and duration of the training program. Due to this high similarity between training programs, children were equally well challenged in both conditions. Consequently, expectancy effects in addition to test-retest effects were taken into account (Wass et al., 2012; Jolles & Crone, 2012).
Future Studies
Given the relevance of childhood IC skills for various outcomes later in life (Moffitt et al., 2011), future studies are needed to investigate the extent of the trainability of IC from a behavioral as well as a neural perspective. In light of the large variability in children’s (neural) development and performance, individual differences in IC and differences in learning trajectories should be taken into account as these differences might explain whether an IC training works better for some individuals than for others. Insight in children’s ability to learn during IC training(Jolles & Crone, 2012; Shah et al., 2012) helps us to tailor IC training programs to children’s individual needs (Berkman, Graham, & Fisher, 2008; Karbach & Unger, 2014). In turn, these training programs might improve children’s IC skills, which helps children to adjust themselves to the changing needs of society.
References
Ainslie, G. (1975). Specious reward: A behavioral theory of impulsiveness and impulse control. Psychological Bulletin, 82(4), 463–496. doi:10.1037/h0076860
Akaike, H. (1974). A new look at the statistical model identification. IEEE Transactions on Automatic Control, 19(6), 716–723. doi:10.1109/TAC.1974.1100705
Band, G. P. H., Van Der Molen, M. W., & Logan, G. D. (2003). Horse-race model simulations of the stop-signal procedure. Acta Psychologica, 112(2), 105–142. doi:10.1016/S0001-6918(02)00079-3
Bari, A., & Robbins, T. W. (2013). Inhibition and impulsivity: Behavioral and neural basis of response control. Progress in Neurobiology, 108, 44–79.
doi:10.1016/j.pneurobio.2013.06.005
Bates, D., Mächler, M., Bolker, B., & Walker, S. (2014). Fitting linear mixed-effects models using lme4. Journal of Statistical Software, 67(1), 1–48. doi:10.18637/jss.v067.i01 Bergman Nutley, S., Söderqvist, S., Bryde, S., Thorell, L. B., Humphreys, K., & Klingberg,
T. (2011). Gains in fluid intelligence after training non‐verbal reasoning in 4‐year‐old children: A controlled, randomized study. Developmental Science, 14(3), 591–601. doi:10.1111/j.1467-7687.2010.01022.x
Berkman, E. T., Graham, A. M., & Fisher, P. A. (2012). Training self‐control: A domain‐ general translational neuroscience approach. Child Development Perspectives, 6(4), 374–384. doi:10.1111/j.1750-8606.2012.00248.x
Berkman, E. T., Kahn, L. E., & Merchant, J. S. (2014). Training-induced changes in inhibitory control network activity. Journal of Neuroscience, 34(1), 149–157. doi:10.1523/JNEUROSCI.3564-13.2014
Casey, B. J., Tottenham, N., Liston, C., & Durston, S. (2005). Imaging the developing brain: What have we learned about cognitive development? Trends in Cognitive Sciences, 9(3), 104–110. doi:10.1016/j.tics.2005.01.011
Corsi, P. M. (1972). Human memory and the medial temporal region of the brain. Dissertation Abstracts International, 34, 819B.
Coxon, J. P., Stinear, C. M., & Byblow, W. D. (2007). Selective inhibition of movement. Journal of Neurophysiology, 97(3), 2480–2489. doi:10.1152/jn.01284.2006 Dalley, J. W., Everitt, B. J., & Robbins, T. W. (2011). Impulsivity, compulsivity, and
top-down cognitive control. Neuron, 69(4), 680–694. doi:10.1016/j.neuron.2011.01.020 Diamond, A. (2012). Activities and programs that improve children’s executive functions.
Current Directions in Psychological Sciences, 21(5), 335–341. doi:10.1177/0963721412453722
Diamond, A. (2013). Executive functions. Annual Review of Psychology, 64(1), 135–168. doi:10.1146/annurev-psych-113011-143750
Diamond, A., & Lee, K. (2011). Interventions shown to aid executive function development in children 4 to 12 years old. Science, 333(6045), 959–964.
doi:10.1126/science.1204529
Diamond, A., & Ling, D. S. (2016). Conclusions about interventions, programs, and approaches for improving executive functions that appear justified and those that, despite much hype, do not. Developmental Cognitive Neuroscience, 18, 34–48. doi:10.1016/j.dcn.2015.11.005
Duckworth, A. L. (2011). The significance of self-control. Proceedings of the National Academy of Sciences, 108(7), 2639–2640. doi:10.1073/pnas.1019725108
Ericsson, K. A., Nandagopal, K., & Roring, R. W. (2009). Toward a science of exceptional achievement. Annals of the New York Academy of Sciences, 1172(1), 199–217. doi:10.1196/annals.1393.001
Filevich, E., Kühn, S., & Haggard, P. (2012). Intentional inhibition in human action: The power of ‘no’. Neuroscience and Biobehavioral Reviews, 36(4), 1107–1118. doi:10.1016/j.neubiorev.2012.01.006.
Goswami, U. (2006). Neuroscience and education: From research to practice? Nature Reviews Neuroscience, 7(5), 406–413. doi:10.1038/nrn1907
Haggard, P. (2008). Human volition: Towards a neuroscience of will. Nature Reviews Neuroscience, 9(12), 934–946. doi:10.1038/nrn2497
Hofmann, W., Schmeichel, B. J., & Baddeley, A. D. (2012). Executive functions and self-regulation. Trends in Cognitive Sciences, 16(3), 174–180.
doi:10.1016/j.tics.2012.01.006
Holmes, J., Gathercole, S. E., & Dunning, D. L. (2009). Adaptive training leads to sustained enhancement of poor working memory in children. Developmental Science, 12(4), F9– F15. doi:10.1111/j.1467-7687.2009.00848.x
Jaeggi, S. M., Buschkuehl, M., Jonides, J., & Shah, P. (2011). Short- and long-term benefits of cognitive training. Proceedings of the National Academy of Sciences, 108(25), 10081–10086. doi:10.1073/pnas.1103228108
Johnstone, S. J., Roodenrys, S., Phillips, E., Watt, A. J., & Mantz, S. (2010). A pilot study of combined working memory and inhibition training for children with AD/HD. ADHD Attention Deficit and Hyperactivity Disorders, 2(1), 31–42. doi:10.1007/s12402-009-0017-z
Jolles, D. D., & Crone, E. A. (2012). Training the developing brain: A neurocognitive
Jolles, D. D., van Buchem, M. A., Rombouts, S. A. R. B., & Crone, E. A. (2012). Practice effects in the developing brain: A pilot study. Developmental Cognitive Neuroscience, 2S, S180–S191. doi:10.1016/j.dcn.2011.09.001
Karbach, J., & Kray, J. (2009). How useful is executive control training? Age differences in near and far transfer of task‐switching training. Developmental Science, 12(6), 978– 990. doi:10.1111/j.1467-7687.2009.00846.x
Karbach, J., Strobach, T., & Schubert, T. (2015). Adaptive working-memory training benefits reading, but not mathematics in middle childhood. Child Neuropsychology, 21(3), 285–301. doi:10.1080/09297049.2014.899336
Karbach, J., & Unger, K. (2014). Executive control training from middle childhood to adolescence. Frontiers in Psychology, 5, 390. doi:10.3389/fpsyg.2014.00390 Kim, S., & Lee, D. (2011). Prefrontal cortex and impulsive decision making. Biological
Psychiatry, 69(12), 1140–1146. doi:10.1016/j.biopsych.2010.07.005
Klingberg, T. (2010). Training and plasticity of working memory. Trends in Cognitive Sciences, 14(7), 317–324. doi:10.1016/j.tics.2010.05.002
Klingberg, T., Fernell, E., Olesen, P. J., Johnson, M., Gustafsson, P., Dahlström, K., ... & Westerberg, H. (2005). Computerized training of working memory in children with ADHD-A randomized, controlled trial. Journal of the American Academy of Child & Adolescent Psychiatry, 44(2), 177–186. doi:10.1097/00004583-200502000-00010 Klingberg, T., Forssberg, H., & Westerberg, H. (2002). Training of working memory in
children with ADHD. Journal of Clinical and Experimental Neuropsychology, 24(6), 781–791. doi:10.1076/jcen.24.6.781.8395
Kray, J., Karbach, J., Haenig, S., & Freitag, C. (2012). Can task-switching training enhance executive control functioning in children with attention deficit/-hyperactivity disorder? Frontiers in Human Neuroscience, 5, 180. doi:10.3389/fnhum.2011.00180
Kühn, S., Haggard, P., & Brass, M. (2009). Intentional inhibition: How the “veto-area” exerts control. Human Brain Mapping, 30(9), 2834–2843. doi:10.1002/hbm.20711
Kuznetsova, A., Brockhoff, P. B., & Christensen, R. H. B. (2017). lmerTest package: Tests in linear mixed effects models. Journal of Statistical Software, 82(13), 1–26.
doi:10.18637/jss.v082.i13
Logan, G. D. (1994). On the ability to inhibit thought and action: A user’s guide to the stop signal paradigm. In D. Dagenbach, & T. H. Carr (Eds.), Inhibitory processes in attention, memory, and language (pp. 189–239). San Diego, CA: Academic Press. Logan, G.D., & Cowan, W.B. (1984). On the ability to inhibit thought and action: A theory of
an act of control. Psychological Review, 91(3), 295–327. doi:10.1037/0033-295X.91.3.295
Matzke, D., Love, J., Wiecki, T. V., Brown, S. D., Logan, G. D., & Wagenmakers, E. J. (2013). Release the BEESTS: Bayesian estimation of ex-Gaussian stop-signal reaction time distributions. Frontiers in Psychology, 4, 918. doi:10.3389/fpsyg.2013.00918 Mazur, J. E. (1987). An adjusting procedure for studying delayed reinforcement. In M. L.
Commons, J. E. Mazur, J. A. Nevin, & H. Rachlin (Eds.), Quantitative Analysis of Behavior: Vol 5. The Effects of Delay and Intervening Events on Reinforcement Value (pp. 55–73). Hillsdale, NJ: Lawrence Erlbaum Associates, Inc.
Melby-Lervåg, M., & Hulme, C. (2013). Is working memory training effective? A meta-analytic review. Developmental Psychology, 49(2), 270–291. doi:10.1037/a0028228 Melby-Lervåg, M., Redick, T. S., & Hulme, C. (2016). Working memory training does not
improve performance on measures of intelligence or other measures of “far transfer”: Evidence from a meta-analytic review. Perspectives on Psychological Science, 11(4), 512–534. doi:10.1177/1745691616635612
Meyers, L. S., Gamst, G., & Guarino, A. J. (2013). Applied Multivariate Research: Design and Interpretation (2nd ed.). Thousand Oaks, CA: Sage Publications.
Miller, E. K., & Cohen, J. D. (2001). An integrative theory of prefrontal cortex function. Annual Review of Neuroscience, 24(1), 167–202. doi:10.1146/annurev.neuro.24.1.167 Miyake, A., & Friedman, N. P. (2012). The nature and organization of individual differences
in executive functions: Four general conclusions. Current Directions in Psychological Science, 21(1), 8–14. doi:10.1177/0963721411429458
Miyake, A., Friedman, N. P., Emerson, M. J., Witzki, A. H., Howerter, A., & Wager, T. D. (2000). The unity and diversity of executive functions and their contributions to complex “frontal lobe” tasks: A latent variable analysis. Cognitive Psychology, 41(1), 49–100. doi:10.1006/cogp.1999.0734
Moffitt, T. E., Arseneault, L., Belsky, D., Dickson, N., Hancox, R. J., Harrington, H., ... & Caspi, A. (2011). A gradient of childhood self-control predicts health, wealth, and public safety. Proceedings of the National Academy of Sciences, 108(7), 2693–2698. doi:10.1073/pnas.1010076108
Moreau, D., & Conway, A. R. A. (2014). The case for an ecological approach to cognitive training. Trends in Cognitive Science, 18(7), 334–336. doi:10.1016/j.tics.2014.03.009 Myerson, J., Green, L., Hanson, J. S., Holt, D. D., & Estle, S. J. (2003). Discounting delayed
and probabilistic rewards: Processes and traits. Journal of Economic Psychology, 24(5), 619–635. doi:10.1016/S0167-4870(03)00005-9
Myerson, J., Green, L., & Warusawitharana, M. (2001). Area under the curve as a measure of discounting. Journal of the Experimental Analysis of Behavior, 76(2), 235–243. doi:10.1901/jeab.2001.76-235
Raven, J., Raven, J. C., & Court, J. H. (2003). Manual for Raven's progressive matrices and vocabulary scales. Section 1: General overview. San Antonio, TX: Harcourt
Assessment.
Raven, J., Raven, J. C., & Court, J. H. (2004). Manual for Raven's progressive matrices and vocabulary scales. Section 3: The standard progressive matrices. San Antonio, TX: Harcourt Assessment.
Schel, M. A., Scheres, A., & Crone, E. A. (2014). New perspectives on self-control
development: Highlighting the role of intentional inhibition. Neuropsychologia, 65, 236–246. doi:10.1016/j.neuropsychologia.2014.08.022
Scheres, A., Dijkstra, M., Ainslie, E., Balkan, J., Reynolds, B., Sonuga-Barke, E., & Castellanos, F. X. (2006). Temporal and probabilistic discounting of rewards in children and adolescents: Effects of age and ADHD symptoms. Neuropsychologia, 44(11), 2092–2103. doi:10.1016/j.neuropsychologia.2005.10.012
Schwarz, G. (1978). Estimating the dimension of a model. The Annals of Statistics, 6(2), 461– 464. doi:10.1214/aos/1176344136
Shah, P., Buschkuehl, M., Jaeggi, S., & Jonides, J. (2012). Cognitive training for ADHD: The importance of individual differences. Journal of Applied Research in Memory and Cognition, 1(3), 204–205. doi:10.1016/j.jarmac.2012.07.001
Shipstead, Z., Redick, T. S., & Engle, R. W. (2012). Is working memory training effective? Psychological Bulletin, 138(4), 628–654. doi:10.1037/a0027473
Singer, J. D. & Willett, J. B. (2003). Applied Longitudinal Data Analysis: Modeling Change and Event Occurrence. New York, NY: Oxford University Press.
Spierer, L., Chavan, C. F., & Manuel, A. L. (2013). Training-induced behavioral and brain plasticity in inhibitory control. Frontiers in Human Neuroscience, 7, 427.
Steinbeis, N., Haushofer, J., Fehr, E., & Singer, T. (2014). Development of behavioral control and associated vmPFC–DLPFC connectivity explains children's increased resistance to temptation in intertemporal choice. Cerebral Cortex, 26(1), 32–42.
doi:10.1093/cercor/bhu167
Thorell, L. B., Lindqvist, S., Bergman Nutley, S., Bohlin, G., & Klingberg, T. (2009). Training and transfer effects of executive functions in preschool children.
Developmental Science, 12(1), 106–113. doi:10.1111/j.1467-7687.2008.00745.x Titz, C., & Karbach, J. (2014). Working memory and executive functions: Effects of training
on academic achievement. Psychological Research, 78(6), 852–868. doi:10.1007/s00426-013-0537-1
Van der Molen, M. J., Van Luit, J. E. H., Van der Molen, M. W., Klugkist, I., & Jongmans, M. J. (2010). Effectiveness of a computerised working memory training in adolescents with mild to borderline intellectual disabilities. Journal of Intellectual Disability Research, 54(5), 433–447. doi:10.1111/j.1365-2788.2010.01285.x
Verbruggen, F., & Logan, G. D. (2008). Response inhibition in the stop-signal paradigm. Trends in Cognitive Sciences, 12(11), 418–424. doi:10.1016/j.tics.2008.07.005 Wass, S. V., Scerif, G., & Johnson, M. H. (2012). Training attentional control and working
memory – Is younger, better? Developmental Review, 32(4), 360–387. doi:10.1016/j.dr.2012.07.001
Williams, B. R., Ponesse, J. S., Schachar, R. J., Logan, G. D., & Tannock, R. (1999). Development of inhibitory control across the life span. Developmental Psychology, 35(1), 205–213. doi:10.1037/0012-1649.35.1.205
Willis, S. L., & Schaie, K. W. (2009). Cognitive training and plasticity: Theoretical perspective and methodological consequences. Restorative Neurology and Neuroscience, 27(5), 375–389. doi:10.3233/RNN-2009-0527
Zandbelt, B. B., & Vink, M. (2010). On the role of the striatum in response inhibition. PloS One, 5(11), e13848. doi:10.1371/journal.pone.0013848
Zanolie, K. & Crone, E. A. (2018). Development of cognitive control across childhood and adolescence. In J. Wixted (Ed.), The Stevens' handbook of experimental psychology and cognitive neuroscience, 5th edition: Volume 3, Developmental & Social
Psychology (pp. 1–24). New York, NY: Wiley & Sons, Inc. doi:10.1002/9781119170174.epcn405
Appendix A
Instructions within the Training Program Example of Instruction within the Inhibitory Control Training
Visual modality. This time, Pedro wants to collect some balloons for decorating his
house! Every time you see a balloon on the screen, you must press the spacebar as soon as you can! Sometimes the balloon bursts. Pedro definitely does NOT want those balloons, so you must not press the space bar when that happens. If you are fast enough, you will see the coin with the stars. You also see the coin if you do NOT press when the balloons have burst. So try to be as fast as you can when helping Pedro, but only pick the balloons he wants.
Auditory modality. This time, Pedro wants to collect vehicles to help his friends get
to the party! Every time you see a vehicle on the screen, you must press the spacebar as soon as you can! Sometimes the vehicles are noisy. Pedro definitely does NOT want those vehicles, so you must not press the space bar when that happens. If you are fast enough, you will see the coin with the stars. You also see the coin if you do NOT press when the vehicles are noisy. So try to be as fast as you can when helping Pedro, but only pick the vehicles he wants.
Examples of Instructions within the Sham Training
Visual modality. This time, Pedro wants to collect some balloons for decorating his
house! Every time you see a balloon on the screen, you must press the spacebar as soon as you can! Don’t worry if the balloons sometimes burst – Pedro wants every balloon there is. If you are fast enough, you will see the coin with the stars. So try to be as fast as you can when helping Pedro.
Auditory modality. This time, Pedro wants to collect vehicles to help his friends get
to the party! Every time you see a vehicle on the screen, you must press the spacebar as soon as you can! Don’t worry if the vehicle sometimes makes a sound – Pedro wants all the
vehicles there are. If you are fast enough, you will see the coin with the stars. So try to be as fast as you can when helping Pedro.
Appendix B
The Presentation of the Computerized Training Program
During training sessions, a positive feedback screen (see Figure B1) was presented after children either successfully inhibited their response on a stop-trial (in the IC condition) or when they successfully responded on a go-trial (e.g., faster than the previous go-trial; in both conditions). Another feedback screen (see Figure B1) was presented to indicate when children responded too fast on a trial (i.e., before a go-stimulus was presented). After each training session, children’s progress on the route to the monkey’s party was presented (see Figure B2).
Figure B1. Feedback screen which was shown based on children’s performance (left) and when children responded too fast (right)
