Overarching Methodological Framework

To overcome these limitations, Kaunhoven and Dorjee (2017) proposed an integrative neurodevelopmental framework. Here, theory-driven research hypotheses are tested with data from a concurrent collection of self-report, other-report, behavioral, psychophysiological and neuroscientific measures; Kaunhoven and Dorjee (2017) particularly highlighted the benefits of using event-related potentials (ERPs). ERPs—averaged brain wave responses to stimuli—provide a non-invasive, cost-effective method with high temporal resolution for investigating the neurocognitive processes underlying self-regulation. The current project follows this approach, combining self- and other-report measures with behavioral performance in a computerized task, an AX type Continuous Performance T (AX-CPT; van Leeuwen et al., 1998), and event-related potentials (ERPs). The AX-CPT is a 2 stimuli Go/NoGo paradigm assessing several cognitive and underlying neural processes relevant to EF (Braver & Barch, 2002). Three types of stimuli appear—cue, target, and distractor—in the following trial types: Go trials (cue followed by target) expecting a button press, NoGo trials (cue followed by distractor) requiring inhibition of button presses, and Ignore trials (distractor followed by target) which can be disregarded. Behavioral performance is indexed by accuracy and reaction time. Brain processes modulated by the three conditions can also be assessed by four ERP components. Two of these ERP components are locked to the cue: the Cue-P3 and the contingent negative variation (CNV).

The Cue-P3

The Cue-P3 is a positive peak between 300 and 600 ms after stimulus onset and can index cue utilization (Hämmerer et al., 2010). Superior cue utilization in pre-adolescents is expressed by less positive amplitudes, but bound to the following two conditions: first, that this reduction in amplitude is accompanied by reduced Cue-P3 latency, and second, that behavioral response to the cue is withheld (Hämmerer et al., 2010; Tsai et al., 2009).

The CNV

The CNV is a negative component appearing between 1100 and 1600 ms after cue onset and signals resource optimization (Kononowicz & Penney, 2016). Advantageous levels of resource optimization during pre-adolescence are suggested by more negative CNV amplitudes (Hämmerer et al., 2010). Two further ERP components are locked to targets: the NoGo-N2 and the NoGo-P3.

The NoGo-N2

The NoGo-N2 is a negative deflection occurring 200–300 ms after the target and is related to conflict processing (Groom & Cragg, 2015). Improved conflict processing in pre-adolescents is demonstrated by less negative NoGo-N2 amplitudes (meta-analyses: Hoyniak, 2017; Hoyniak & Petersen, 2019).

The NoGo-P3

The NoGo-P3 is a positive peak between 300 and 500 ms post target indexing response inhibition (Groom & Cragg, 2015). More proficient response inhibition is indicated by more positive NoGo-P3 amplitudes (Jonkman, 2006) and shorter NoGo-P3 latency (Liu et al., 2011).

Study Aims and Hypotheses

In these studies, the overall assumption is tested that DMW, DMT, and long-term attendance of an MBI during pre-adolescence are linked with improved well-being, and that this benefit is achieved via enhancement of EF and emotion regulation. Study 1 focuses on associations of DMW, Study 2 employed a longitudinal controlled design comparing the effects of long-term mindfulness practice with an active control treatment (i.e., a positive psychology intervention).

Study 1

Part of the sample (n = 20) constituted the control group in a study by Wimmer and Dorjee (2020). This presented a cross-sectional comparison of mindfulness-experienced preadolescents with mindfulness-inexperienced preadolescents regarding EF, emotion regulation, and well-being. The current study investigates a distinct research question and uses statistical analyses different from the ones reported by Wimmer and Dorjee (2020). In the present study it was hypothesized that higher DMW would be associated with better EF, emotion regulation, and well-being. It was also expected that the positive association between DMW and well-being would be mediated by EF and emotion regulation.

Study 2

Study 2 aimed to examine whether long-term mindfulness training is linked with increased DMT. It also aimed to investigate whether changes resulting from mindfulness practice in pre-adolescents show the same pattern of associations with EF, emotion regulation, and well-being as DMW (Study 1). Study 2 tracked a group of pre-adolescents engaging in continuous mindfulness practice in three assessment sessions (i.e. T1 = June 2017 for the mindfulness group, T2 = December 2017 for the mindfulness group, T3 = June 2018 for the mindfulness group) spread over 1 year. At time 1 (T1), the mindfulness training group had practiced mindfulness on average for 2.11 years. An active control group received a positive psychology intervention based on the “Three Good Things” exercise (Seligman et al., 2005) and was also assessed at three time points (i.e. T1 = March 2017 for the control group, T2 = July 2017 for the control group, T3 = December 2017 for the control group)—immediately before and after the 17-week intervention period, plus at 5-month follow-up. A cross-sectional comparison of both groups at T1 is reported in Wimmer and Dorjee (2020). This study extends the previous one since it compares both groups across all three time points by means of distinct statistical analyses.

Positive psychology interventions including “Three Good Things” have proven to increase well-being in school-aged children (Carter et al., 2016; McCabe et al., 2011; Seligman et al., 2009). Here, assumed pathways to enhanced well-being include a more positive attributional style or a memory search bias towards positive events (Carter et al., 2016). As distinct from mindfulness though, improvements of EF and emotion regulation have not been put forward as underlying mechanisms. Hence, use of this comparison group controlled for mindfulness-unrelated improvement of well-being.

To aid comparisons with DMW, Study 2 applied the same measures as those in Study 1. An open question regarding the implementation of MBIs in schools addresses moderating effects of practice, in particular dosage (In the context of this study dosage is defined as self-reported frequency of mindfulness practice). Empirical evidence on dosage effects has so far been inconsistent (e.g., Dunning et al., 2019). One of the recommendations for further research has been that evaluations should also take into the account the impact of relational mindfulness practice characteristics such as practice enjoyment (Jensen, 2014). For instance, it is important to understand whether dosage is positively related to outcomes only when pupils like practicing mindfulness. This can inform adaptations of mindfulness-based programs for children and their implementation based on individual differences. Therefore, indicators of practice frequency and enjoyment were assessed in addition to measures used in Study 1.

It was hypothesized that:

1.

Mindfulness group (MG) would show greater improvement in EF than the positive psychology active control group (CG) from T1 to T2 and from T2 to T3 (due to continued mindfulness practice),

2.

MG would show greater gains on emotion regulation assessments than CG from T1 to T2 and from T2 to T3 (due to continued mindfulness practice);

3.

MG scores on DMT would show greater increase than CG mindfulness scores from T1 to T2 and from T2 to T3 (due to continued mindfulness practice);

4.

Both MG and CG would show improvements in well-being from T1 to T2 (because this was the only period during which both groups received training expected to foster well-being). Between T2 and T3, further gains were expected for MG, whereas CG were hypothesized to maintain well-being levels from T2 (due to continued mindfulness practice and previous findings from Carter et al. (2016), indicating sustained gains in well-being after a positive psychology intervention);

5.

It was also expected that increases in well-being from T1 to T2 in MG would be mediated by gains in EF and emotion regulation between T1 and T2. A similar mediation pattern was expected to occur from T2 to T3.

In addition, it was explored whether outcomes would be affected by frequency and enjoyment of mindfulness practice in MG.

Materials and Methods

The Ethics Committee in the School of Psychology at Bangor University approved both studies prior to their start.

Participants

For study 1, parents and children from years 3 to 6 in three schools in North Wales, which expressed interest in the project, were invited to participate in the study and 62 pupils were recruited; see Table 1 for sample characteristics. None of the participants had received training in mindfulness previously, none had a history of brain injury or brain operation in the past or suffered from epilepsy. Self-report questionnaires were completed by all 62 children. 60 children performed the computer task and 56 of them volunteered for EEG recording during the task; data from 43 pupils was useable for the final ERP analyses. An informant-based questionnaire was returned by 44 parents.

Table 1 Study 1: descriptive statistics/frequencies of demographic variables, questionnaire-based and behavioral dependent measures

For study 2, pupils (N = 53) were recruited from two primary schools in North Wales, 33 children from a school with an established mindfulness curriculum, and 20 children from a school without any previous experience with mindfulness. This initial sample constituted the sample of Wimmer and Dorjee (2020). First, the former school was invited to participate in research on their longstanding implementation of mindfulness; the headmaster accepted this invitation. Subsequently, all primary schools within a distance of 30 miles from Bangor University that had not engaged in mindfulness training were sent invitation letters. The latter school was the only one expressing interest. Next, parents and pupils were invited through information letters handed out at the two schools. None of the participants had a brain injury or brain operation in the past or suffered from epilepsy. The flow of participants through the study is illustrated in Fig. 1. The final sample consisted of N = 43, with n = 28 in MG and n = 15 in the positive psychology CG. Sample characteristics are displayed in Table 2. Although children in both groups were from the same year in school, the MG children were older than CG children at T1, T2, and T3. Both groups also differed in their primary language: While the primary language in MG was exclusively English, Welsh dominated as primary language in CG. However, both groups were exposed to English in everyday life. Furthermore, children in CG didn’t have difficulty understanding the assessments. Groups did not differ in gender or handedness.

Fig. 1

Flow of participants through study 2

Table 2 Study 2: sample characteristics

In both studies, parents provided written informed consent for themselves (for the informant-based assessments) and their children to participate in the respective study. In addition, children were asked for verbal consent on the day of assessment. Data was collected only if both parents and children consented to participation. Pupils received a small gift, such as a pencil, after each assessment.

Measures

The following instruments were employed in Study 1 and Study 2 (see Table 3 for internal consistencies observed in Study 2).

Table 3 Study 2: descriptive statistics and reliabilities of questionnaire-based and behavioral dependent measuresSelf-report Measures

Participants’ DMW was assessed using the Child and Adolescent Mindfulness Measure (CAMM; Greco et al., 2011; please note that in Study 2, CAMM was used to measure DMT). Internal consistency in Study 1 was ωRT = 0.67.Footnote 1 This 10-item self-report instrument assesses mindfulness skills in youths from 10 years of age. 53 children (85%) in Study 1 were younger than 10 years (age range 7.50–11.50 years). Nevertheless CAMM was used with the whole sample because there is no comprehensive published mindfulness measure for children below 10 years. It should be acknowledged that the Mindful Attention Awareness Scale for Children (MAAS-C; Lawlor et al., 2014) has been validated for application with children from the age of 9 years; however, as distinct from CAMM, MAAS-C is restricted to assessing present-moment awareness, so does not cover the non-judgmental aspect of mindfulness. Furthermore, even though most of the current participants were younger than 10 years, they seemed to understand the items correctly. Immediately before the questionnaire was handed out to participants, they were asked to turn to the examiner if they did not or not fully understand any of the items. Immediately after each child had filled in the CAMM, the examiner again asked the participant how they had got along and if anything had been unclear to them. None of the pupils raised any issues.

The 10-item Positive and Negative Affect Schedule for Children (PANAS-C short version; Ebesutani et al., 2012) was administered to measure well-being. In the children’s short version, which has a target group of 6- to 18-year-olds, higher subjective well-being is indicated by a relatively high positive affect score and/or a relatively low negative affect score. In Study 1 ωRT coefficients were 0.80 and 0.81, respectively. Several studies investigating mindfulness in (pre-)adolescents (Bluth & Blanton, 2014; Lawlor et al., 2014; Schonert-Reichl & Lawlor, 2010) used the PANAS as a proxy measure of subjective/emotional well-being. Thus the use of this measure made the current study comparable with previous research.

As a second measure of child well-being the 5-item Satisfaction with Life Scale for Children (SWLS-C; Gadermann et al., 2010) was employed. Study 1 showed an internal consistency of ωRT = 0.70.

Informant-Based Measure

Children’s emotion regulation was assessed with the parent version of the 24-item Emotion Regulation Checklist (ERC; Shields & Cicchetti, 1997). This consists of the subscales negativity/lability and emotion regulation (Adrian et al., 2011). The age group targeted is six to 18 years. Internal consistencies in Study 1 were ωRT = 0.89 for negativity/lability and ωRT = 0.86 for emotion regulation.

Experimental Task

EF was assessed using a visual version of the AX-CPT (adapted from Brocki & Bohlin, 2004) as outlined above. Participants were presented with five different simple black and white stimuli, one of which is a cue, two are targets, and two are distractors. Each trial is composed of 2 consecutive stimuli, each with a display time of 1000 ms and an inter-stimulus interval varying randomly between 1810 and 2043 ms. The following trial types were implemented: 40 (33.33%) Go trials, 40 (33.33%) NoGo trials, and 40 (33.33%) Ignore trials. Trials were presented in a random order except that the same cue or target type did not appear twice in a row. Behavioral responses were assessed using the following indicators: Mean reaction time (RT) and SD for correct Go trials, i.e., hits; error percentages—impulsive responses to 1st stimulus defined as an unexpected button press in response to the cue; target omissions; disinhibited responses defined as an unexpected button press in response to the 2nd stimulus in a NoGo trial; and inattentive impulses defined as an unexpected button press in response to the 2nd stimulus in an Ignore trial.

The following variables were assessed in Study 2 only: Here, the mindfulness training group responded to three questions related to their mindfulness practice. The first two questions were about practice frequency. On a 6-point Likert scale ranging from 0 = ‘never’, ‘about once a month’, ‘a few times in a month’, ‘about once a week’, ‘a few times in a week’, to 5 = ‘every day’ pupils rated how often they practiced mindfulness at school (question one) and at home (question two). The third question assessed enjoyment of practicing mindfulness expressed on a 4-point scale with response options 0 = ‘I don’t like it at all’, 1 = ‘I mostly don’t like it’, 2 = ‘I mostly like it’, 3 = ‘I like it very much’.

Children in the positive psychology intervention group kept logs indicating the days on which they had completed the “Three Good Things” diary.

Interventions Implemented in Study 2Mindfulness Training

The MG received continuous training in the Paws b curriculum developed by Sarah Silverton, Tabitha Sawyer, and Rhian Roxburgh in collaboration with the Mindfulness in Schools Project (https://mindfulnessinschools.org/). This curriculum is intended for children aged 7–11 years. Paws b consists of classroom-based activities with a total duration of approximately 360 min. Depending on preferences activities can either be delivered over twelve single, 30-min sessions, or grouped into pairs of six double, 60-min sessions. Each of the twelve lessons is designed to impart a specific mindfulness skill: Lesson one offers fundamental information about the human brain and how mindfulness training can impact on the brain. Lesson two conveys how humans can focus their minds and make adaptive choices using mindfulness. In lesson three, attendees experience how a mindful attitude can help them widen and constrict the focus of their attention. Lesson four addresses the application of mindfulness in daily life. In lesson five, children learn about the dynamic nature shaping human mind and body. Lesson six invites children to experience how mindful attention can settle their mind and body. Lesson seven deals with ways of coping with challenging situations. In lesson eight, participants learn how mindfulness can help them enhance their well-being. Lesson nine addresses the power of thoughts and mental habits. Lesson 10 informs about the relationship of thoughts with emotions, behavior, and bodily reactions, and how this connection can be regulated using mindfulness. In lesson eleven, pupils learn how mindfulness is related to taking care of themselves and others. Lesson twelve deals with embracing moments of joy and happiness.

Schoolteachers in the participating school were formally trained in mindfulness and in Paws b instruction. Teachers attended an 8-week secular mindfulness course, such as mindfulness-based stress reduction (MBSR). This course was led by a mindfulness teacher with more than 10 years of teaching experience. After this, regularly practicing mindfulness for at least 2 months was required to be admitted to an intensive 3-day Teach Paws b course. Teachers also promised to continuously implement mindfulness in their lessons. In the partaking school, Paws b lessons were incorporated into regular lessons throughout the school year; in addition, pupils were provided with the option to practice mindfulness during voluntary lunchtime sessions.

Positive Psychology Intervention “Three Good Things”

The positive psychology intervention was based on the “Three Good Things” exercise (Seligman et al., 2005). Within this exercise one writes down three good things that happened over a number of days (historically one or more weeks).

Children were asked to keep four versions of the “Three Good Things” diary for a duration of 2 weeks (i.e., 10 school days), each. When all diaries had been completed once (after 8 weeks), pupils started again with version 1 and continued until version 3. Thus, pupils did the exercise over a duration of 14 weeks. The intervention period was interrupted by one 2-week and one 1-week school holidays break. Version 1 was the original version reported by Seligman et al. (2005). Pupils were asked to write down 3 good things that had occurred each day and how or why the good things happened. Version 2 was a neutral version that was not restricted to positive events but encouraged respondents to simply write down three things that had occurred each day and how or why the things happened. Version 3 incorporated self-determination theory (Ryan & Deci, 2000): as in the original version, pupils were asked to write down 3 good things that had occurred each day and how or why the good things happened. In addition, pupils were asked whether these events were associated with the basic needs postulated by self-determination theory. These needs are competence, autonomy, and relatedness (Ryan & Deci, 2000). The question regarding competence was “Did you achieve something? If yes, which event?”. The question regarding autonomy was “Did you have a choice? If yes, which event?”. And the question regarding relatedness was “Did you spend time with your friends? If yes, which event?”. Version 4 was geared towards future planning: Pupils were asked to write down 3 events they were looking forward to the next day; they were also prompted to indicate who had planned the events. Diaries were administered confidentially, i.e., without teachers or researchers reading the diary entries. This took place during normal school hours, usually at the end of a day’s lessons. Diaries remained at school. Researchers received a log kept by pupils indicating the days on which “Three Good Things” was performed.

ProceduresStudy 1

Study 1 followed a correlational design. Participants were tested individually during school hours. Quiet testing spaces were provided on school premises. Measures were collected in the following sequence for all participants: PANAS-C, CAMM, SWLS-C, EEG-setup (for children participating in the EEG recordings), computerized task with EEG recording. EEG was recorded using a portable system consisting of acquisition and stimulus presentation laptops, a Brain Products actiCHamp amplifier, and actiCAP electrodes.

EEG signal was recorded with 30 Ag/AgCl electrodes; TP10 was the reference site and Fpz served as the ground. Data was obtained at a sampling rate of 1 kHz with a Brain Products actiCHamp system. Two electrodes, situated above and below the right eye, recorded ocular movements. The impedance of all electrodes were kept below 25 kΩ. Online, the EEG signal was filtered with a bandpass filter range of 0.01–200 Hz, 48 dB/Oct slope. ERP data was processed in BrainVision Analyzer. It was first cleaned automatically so that the maximal allowed difference within 200 ms intervals was 1500 μV and the lowest activity in 100 ms intervals was 0.5 μV. Independent Component Analysis (ICA) followed by an inverse ICA was employed to regress out eye-blinks. Offline bandpass filter with a range of 0.1–30 Hz, 48 dB/Oct slope, was applied. Residual artefacts were cleaned manually and then the data was re-referenced to the average of T7 and T8. For the 1st stimulus (cue or distractor) of a trial the data was epoched into 2100 ms segments starting at − 200 ms, and baseline corrected using the signal 200 ms before stimulus onset. For the 2nd stimulus of a trial (target or distractor) the data was epoched into 1000 ms segments starting at − 100 ms, and baseline corrected using the signal 100 ms before stimulus onset. Finally, averages for each condition and participant, and then grand averages across participants for each condition and group were computed.

Study 2

In Study 2, measures for all participants at T1, T2, and T3 were taken in the same sequence as in Study 1: PANAS-C, CAMM, SWLS-C, EEG-setup (for children participating in the EEG recordings), computerized task with EEG recording. The protocol for EEG recordings was the same as in Study 1. Participants in MG answered the questions about mindfulness practice at the beginning of each testing session, however only the data from T2 and T3 were used for analyses, since the T1 data on mindfulness practice were already analyzed in Wimmer and Dorjee (2020). Logs recording frequency of the positive psychology interventions were filled in at school during the intervention period, and not during testing sessions.

Mindfulness training was practiced continuously. Pupils reported that they had on average started mindfulness practice 2.11 years (SD = 1.01) before T1 (see Wimmer & Dorjee, 2020). T2 tests were conducted 6 months after pre-tests, T3 assessments were carried out 6 months after T2.

The positive psychology interventions started immediately after T1; T2 tests in this group happened 17 weeks after time 1, T3 assessments were taken 5 months after time 2.

Data AnalysisStudy 1

Questionnaire measures of well-being and emotion regulation, as well as behavioral indicators of EF, were analyzed using bivariate correlations with DMW.

ERP analysis was carried out using ANCOVAs assessing mean amplitude and peak latency data (except for the CNV since its latency is not commonly reported due to the broad peak nature of this component) for electrodes of interest. The following clusters of electrodes were selected for analyses for each of the components based on previous literature and observed maximal signal: Cue-P3—Pz, P3, and P4 across the time window 240–360 ms (comparable to Hämmerer et al., 2010); CNV—Fz, Cz, and Pz in the time windows (early) 800–1000 ms and (late) 1400–1800 ms (similar to Jonkman et al., 2003); NoGo-N2—P3, P4, P7, and P8, in the time window 160–260 ms (similar time windows were used by, e.g., Hämmerer et al., 2010, however at predominantly frontal sites); NoGo-P3—Pz, P3, P4, CP1, and CP2, in the time window 280–360 ms (comparable to Hämmerer et al., 2010).Footnote 2 For the two cue-locked components, i.e., Cue-P3 and CNV, ANCOVAs were run with condition (Cue vs. NonCue) as a factor and CAMM score as a covariate for peak latency; for mean amplitude the set of factors were complemented by an n-ary electrode factor.Footnote 3 An additional factor of time window (early vs. late) was included for CNV mean amplitude. For the two components locked to the 2nd stimulus of a trial, i.e., NoGo-N2 and NoGo-P3, ANCOVAs were run with condition (Go vs. NoGo) as a factor and CAMM score as a covariate for peak latency; for mean amplitude the set of factors were complemented by n-ary electrode factor. All artefact free correct trials were included in the ERP analyses for all participants with at least 15 trials per condition. Twelve out of the 55Footnote 4 participants did not meet this criterion and therefore were excluded from analyses resulting in final N = 43 for the ERP analyses.

Study 2

Questionnaire-based measures and behavioral performance in AX-CPT were analyzed using mixed factorial ANOVA, involving one between subjects factor, group (MG vs. CG), and one within subjects factor, time (T1, T2, T3). Due to a low retention rate of the ERC and ERP components for the control group, analysis of these indicators was restricted to MG. Hence, the factor of group was dropped for ERP analyses.

Electrode clusters and time windows used in ERP analyses were the same as in Study 1. For the two cue-locked ERP components, i.e., Cue-P3 and CNV, initial ANOVAs were run with condition (Cue vs NonCue) and time (T1 vs. T2 vs. T3) as factors for peak latency; for mean amplitude an electrode factor was added (as in Study 1). An additional factor of time window (early vs. late) was included for CNV mean amplitude. For the two components locked to the target of a trial, i.e., NoGo-N2 and NoGo-P3, initial ANOVAs were run with condition (Go vs. NoGo) and time (T1, T2, T3) as factors for peak latency; for mean amplitude an electrode factor was added. Just as in Study 1, all artefact free correct trials were included for all participants with at least 15 trials per condition. The final sample was n = 13 for the ERP analyses.

All dependent measures, practice enjoyment, and practice frequency were corrected for age by regressing each outcome on age. The resulting standardized residual served as an age-corrected variable in all analyses reported below.

ResultsStudy 1

Descriptive statistics for questionnaire-based and behavioral dependent measures are summarized in Table 1. An outlier analysis detected one participant with an extremely high score on PANAS-C negative affect that was more than 3SD above the sample mean. Therefore this data was removed from the dataset. In addition, less than 3.43% of data per measure was excluded from analyses because they were more than 3SD from the sample mean or because they had RTs in the AX-CPT below 200 ms.

Correlational Analyses

CAMM was significantly negatively correlated with PANAS-C negative affect, r = − 0.323, p = 0.011, but also significantly negatively correlated with subjective well-being measured by SWLS-C, r = − 0.266, p = 0.037 (remaining indicators of emotion regulation and well-being: ps > 0.43). Furthermore, CAMM was significantly positively correlated with percentage of impulses to the 1st stimulus of a trial, r = 0.31, p = 0.023, meaning that those with higher DMW showed more impulses (remaining behavioral indicators of EF: ps > 0.21).

ERP Components

For Cue-P3, there was a mean amplitude main effect of DMW, F(1, 244) = 25.30, p < 0.0001, η2p = 0.094, with children high in DMW displaying less positive amplitudes than individuals low in DMW (other ps > 0.20; cf. Fig. 2A). After exclusion of three outlying values for Cue-P3, significances remained the same. For CNV mean amplitude, the ANCOVA revealed only an interaction of condition with CAMM, F(1,484) = 4.58, p = 0.033, η2p = 0.009, indicating that children with high and low DMW did not differ in the NonCue condition, however in the Cue condition children high in DMW demonstrated more negative CNV amplitude than children low in DMW (cf. Fig, 2B; other ps > 0.22). After exclusion of two outliers for CNV, the interaction of condition with CAMM became marginal, p = 0.072 (other ps > 0.08). The ANCOVAs for NoGo-N2 mean amplitude and peak latency did not reveal any main effects or interactions, ps > 0.26 (cf. Fig. 2C). As for NoGo-P3, the ANCOVAs yielded a significant effect of DMW suggesting a negative relationship between