Developmental Coordination Disorder pathophysiology

Contents lists available at ScienceDirect

Brain and Cognition

journal homepage:

Impaired motor inhibition in developmental coordination disorder

J.L. Hea,⁎, I. Fuelschera, J. Coxonb, P. Barhouna, D. Parmara, P.G. Enticotta, C. Hydea

a Cognitive Neuroscience Unit, School of Psychology, Deakin University, Geelong, Australia bMonash Institute of Cognitive and Clinical Neurosciences (MICCN), School of Psychological Sciences, Monash University, Clayton, Australia


Keywords: Developmental coordination disorder Developmental disorders Movement disorders Motor control Motor inhibition Response inhibition Go/no-go task Stop-signal task


This study aimed to evaluate the ‘inhibitory deficit’ hypothesis of developmental coordination disorder (DCD). We adopted a multifaceted approach, investigating two distinct, yet complimentary facets of motor inhibition: action restraint and action cancellation. This was achieved using carefully constructed versions of the ‘Go/No-go’ and ‘Stop-signal’ tasks, respectively. The sample comprised 11 young adults with DCD aged between 18 and 30 years of age and 11 typically developing, age-matched controls. Participants completed both the ‘Go/No-go’ and ‘Stop-signal’ tasks to assess action restraint and action cancellation respectively. Individuals with DCD were less efficient than their typically developing peers at performing both action restraint and action cancellation, indicated by significantly reduced action restraint efficiency index scores on the ‘Go/No-go’ task and a trend towards longer stop-signal reaction times on the ‘Stop-signal’ task. This work clarifies disparate evidence speaking to the integrity of action restraint in DCD and provides the first account of action cancellation in DCD using a purpose-built measure. In support of the inhibitory deficit hypothesis of DCD, our results suggest that young adults with DCD experience broad difficulties with engaging inhibitory mechanisms during motor be- haviour.

1. Introduction

Developmental coordination disorder (DCD) is a neurodevelop- mental condition primarily characterized by poor motor skill in the absence of any identifiable medical or neurological causes (Association, 2013). Difficulties with motor coordination affect the ability of diag- nosed individuals to complete day-to-day tasks, such as writing, getting dressed and self-care (Summers, Larkin, & Dewey, 2008). Recent meta- analyses have highlighted a growing body of evidence reporting com- promised inhibitory control in DCD (Biotteau, Chaix, & Albaret, 2016; Gomez & Sirigu, 2015; Wilson, Ruddock, Smits-Engelsman, Polatajko, & Blank, 2013), leading some to suggest that reduced inhibition may underpin or contribute to compromised motor skill in this group (Mandich, Buckolz, & Polatajko, 2003; Tsai, Yu, Chen, & Wu, 2009; Wilson et al., 2013).

While individuals with DCD do indeed show widespread deficits in a number of neuropsychological measures of inhibition (Piek, Dyck, Francis, & Conwell, 2007), the shared motor and cognitive demands of

these tasks are such that it is often difficult to discern the degree to which poor performance can be attributed to motor or cognitive pro- cessing (Hyde, Rigoli, & Piek, 2016; Leonard, Bernardi, Hill, & Henry, 2015). In contrast to many of the neurocognitive mechanisms that have been implicated in DCD, inhibitory control has been shown amenable to improvement through intervention (e.g. Tsai et al., 2009). Accordingly, there is a great need to clarify the integrity of inhibitory control in DCD, particularly as it relates to motor control.

1.1. Motor inhibition

Motor inhibition refers to a form of response inhibition specifically pertaining to the volitional cancellation or suppression of unwanted movement (Coxon, Stinear, & Byblow, 2006). The motor control lit- erature often distinguishes between two broad forms of motor inhibi- tion: the inhibition of prepotent, yet unwanted motor responses (re- ferred to as action restraint) and the cancellation of prepared or ongoing movements (referred to as action cancellation; (Barkley, 1997; Eagle, Received 1 May 2018; Received in revised form 9 August 2018; Accepted 11 September 2018

Abbreviations: DCD, Developmental Coordination Disorder; ISI, inter-stimulus interval; RTs, response times; SSD, stop-signal delay; TMS, Transcranial Magnetic Stimulation; M1, Primary Motor Cortex; TD, Typically Developing; BOT-2, Bruininks-Oseretsky Test of Motor Proficiency 2nd edition; Go-RT, infrequent-go response times; Nogo-PC, percentage accuracy on No-go trials; AREI, action restraint efficiency index; MATLAB, matrix laboratory; LT, lift time; ANOVA, analysis of variance; IPL, inferior parietal lobe; SFG, superior frontal gyrus; CI95%, 95% confidence interval ⁎ Corresponding author at: Cognitive Neuroscience Unit, School of Psychology, Deakin University, Burwood Campus, 221 Burwood Hwy, Burwood, Victoria 3125,

Australia. E-mail address: (J.L. He).

Brain and Cognition 127 (2018) 23–33

Available online 20 September 2018 0278-2626/ © 2018 Elsevier Inc. All rights reserved.


Bari, & Robbins, 2008; Rubia et al., 2001; Schachar et al., 2007). While each engage circuitry critical for motor inhibition (e.g., pre-motor and insula cortices), evidence from neuroimaging studies suggest that action restraint relies more heavily on the dorsolateral and medial prefrontal cortices (Rubia, Smith, Brammer, Toone, & Taylor, 2005), while action cancellation tends to depend more on the right inferior frontal gyrus and basal ganglia (Aron, Fletcher, Bullmore, Sahakian, & Robbins, 2003; Chambers et al., 2006; Chikazoe et al., 2008). Given this, action restraint and action cancellation are best conceptualised as overlapping yet dis- tinct motor inhibitory processes (Dambacher et al., 2014a, 2014b; Eagle et al., 2008; Schachar et al., 2007). Experimentally, each can be ex- amined using classic behavioural tasks: the ‘Go/No-go’ task and ‘Stop- signal’ task respectively (Chikazoe et al., 2008; Coxon et al., 2006; Coxon, Stinear, & Byblow, 2007, 2009; Eagle et al., 2008; Picton et al., 2006).

1.1.1. Action restraint While a variety of ‘Go/No-go’ tasks exist, action restraint is tradi-

tionally indexed when participants are pseudo-randomly presented with stimuli that either indicate that a pre-determined motor response (i.e., a button press) should be performed (‘Go’ trials) or withheld (‘No- go’ trials; (Johnstone, Pleffer, Barry, Clarke, & Smith, 2005; Simmonds, Pekar, & Mostofsky, 2008; Smith, Johnstone, & Barry, 2008)). In- hibitory load is manipulated by increasing the frequency of ‘Go’ relative to ‘No-go’ trials and/or by decreasing the time between each stimulus presentation (viz inter-stimulus interval [ISI]; (Benikos, Johnstone, & Roodenrys, 2013; Lindqvist & Thorell, 2008; Simmonds et al., 2008)). As the ratio of ‘Go’ relative to ‘No-go’ trials increase, participants de- velop a greater pre-potency towards executing the ‘Go’ motor response (Bruin & Wijers, 2002). Where ‘Go’ trial frequency is sufficiently high (i.e., > 50% (Barry & De Blasio, 2015; Barry & Rushby, 2006)), the associated ‘Go’ response becomes the default response upon stimulus presentation (Simpson & Riggs, 2006). In such cases, pseudo-random presentation of infrequently occurring ‘No-go’ trials requires the in- hibition of the prepotent motor response, with failure to do so referred to as a ‘commission error’ (Bezdjian, Baker, Lozano, & Raine, 2009; Yechiam et al., 2006). Indeed, as the percentage of ‘Go’ trials increases, commission errors become more frequent (Bedi, 2015).

Still, while a sufficiently high proportion of ‘Go’ relative to ‘No-go’ trials is required to elicit an inhibitory response, such designs make it difficult to discern the degree to which ‘No-go’ performance reflects one’s ability to enact action restraint, relative to their ability to process infrequently occurring stimuli (Chikazoe et al., 2008). Accordingly, it has become increasingly common to include both frequent (75%) and infrequent (e.g., 12.5%) ‘Go’ trials, with the latter occurring in equal frequency to ‘No-go’ trials (e.g., 12.5%; Hirose et al. (2012)). An effi- ciency index can then be calculated based on response times (RTs) from ‘Infrequent-go’ trials and accuracy from ‘No-go’ trials. This index pro- vides a measure of action restraint efficiency which controls for in- dividual differences in ability to process infrequently occurring stimuli (see Section 2 for a detailed explanation: Fig. 2). Finally, reducing ISIs can also increase the automaticity of the ‘Go’ response, further in- creasing the chances of ‘commission errors’ during ‘No-go’ trials (Zamorano et al., 2014). Taken together, while the ‘Go/No-go’ task is a valid and reliable measure of action restraint, task demands must be sufficiently high to engage inhibitory systems. This appears to be re- flected at a neural level, with recent evidence demonstrating that neural activity in fronto-motor networks critical for action restraint during ‘Go/No-go’ tasks is 75% lower when equiprobable ‘Go’ and ‘No-go’ trials and long ISIs are used, compared to rare ‘No-go’ trials with short ISI’s (i.e., ∼0–1500ms; (Wessel, 2017)).

1.1.2. Action cancellation The cancellation of prepared or ongoing movements (aka action

cancellation) can be measured using a Slater-Hammel (1960) inspired version of the ‘Stop-signal’ task (Cowie, MacDonald, Cirillo, & Byblow, 2016; Coxon et al., 2006, 2007; MacDonald, McMorland, Stinear, Coxon, & Byblow, 2017; Rubia et al., 2001). Here, participants begin a trial by depressing a computer key and, after a short delay, an indicator bar begins rising at a constant velocity from a starting line towards a fixed target line (see Section 2 for an illustration: Fig. 2). On the fre- quently occurring ‘Go’ trials, participants are instructed to try to stop the rising bar at the target line by lifting their finger from the computer key. In a small percentage of trials (∼30%), the rising bar stops un- expectedly before reaching the target line, indicating a ‘Stop’ trial. In ‘Stop’ trials, participants are instructed to refrain from releasing their finger from the computer key. As with the ‘Go/No-go’ task, when ‘Stop’ trials are rare, the associated ‘Go’ trial response (i.e., the finger lift) becomes the prepotent response. Thus, successful ‘Stop’ trial perfor- mance requires the cancellation of a prepared motor response. A major advantage of the Slater-Hammel (Slater-Hammel, 1960) inspired Stop- signal task is that the ‘Go’ response does not involve a choice compo- nent, allowing for more direct assessment of motor inhibition without the confound of decision making processes which are otherwise re- quired in more classical choice-reaction Stop-signal tasks and Go/No-go tasks.

Inhibitory load on the ‘Stop-signal’ task can be manipulated by delaying the time at which the rising bar stops within a ‘Stop’ trial (also known as the stop-signal). By increasing the stop-signal delay (SSD; i.e., time between onset of ‘Go’ and ‘Stop’ signals) on a ‘Stop’ trial, greater inhibitory errors are observed (Verbruggen & Logan, 2008). Neuro- physiologically, evidence from transcranial magnetic stimulation (TMS) studies indicates that primary motor cortex (M1) activity increases immediately prior to the finger lift on ‘Go’ trials, signifying the pre- paration of the finger lift command (Coxon et al., 2006; Macdonald, Coxon, Stinear, & Byblow, 2014). Conversely, M1 excitability decreases approximately ∼140ms after the presentation of the stop-signal on a ‘Stop’ trial. As this occurs, a concurrent increase in cortical inhibition of M1 activity can be observed upon successful cancellation of action (Coxon et al., 2006; Macdonald et al., 2014), suggesting that cortical inhibition within the M1 may subserve successful suppression of the prepared finger lift (Coxon et al., 2006). At the neural level, these in- hibitory processes within the motor cortices are thought to reflect projections from upstream sub-cortical structures including the basal ganglia and subthalamic nucleus (Aron, Robbins, & Poldrack, 2014).

1.2. Motor inhibition in DCD

Though the integrity of action cancellation in DCD is yet to be probed using the previously described ‘Stop-signal’ task, a small group of stu- dies have investigated action restraint in DCD using the ‘Go/No-go’ task, with some suggesting atypical performance (Cousins & Smyth, 2003; Piek et al., 2004; Querne et al., 2008) and others not (Rahimi- Golkhandan, Steenbergen, Piek, & Wilson, 2015; Rahimi-Golkhandan, Steenbergen, Piek, Caeyenberghs, & Wilson, 2016). We argue that this disparity may be due to the variant of the ‘Go/No-go’ tasks adopted. Most saliently, the ‘Go/No-go’ tasks used in these previous studies contained blocks with equal ‘Go’ and ‘No-go’ trials (e.g., (Piek et al., 2004; Querne et al., 2008; Thornton, Bray, Langevin, & Dewey, 2018)), or long ISIs (up to 3000ms) in-between trials (e.g., (Cousins & Smyth, 2003; Thornton et al., 2018)). As noted, while ‘Go/No-go’ tasks can provide a reliable and valid measure of action restraint, task demands must be sufficiently high to engage inhibitory systems. Accordingly, we must be circumspect about drawing inferences about action restraint in

J.L. He et al. Brain and Cognition 127 (2018) 23–33


DCD from these earlier studies. More recently, Rahimi-Golkhandan and colleagues used a modified

‘Go/No-go’ task to explore both ‘hot’ and ‘cold’ executive functioning in DCD. Here, a greater ‘Go’ (70%) to ‘No-go’ (30%) trial frequency was adopted (Rahimi-Golkhandan et al., 2015, 2016). The initial study counterbalanced ‘happy’ and ‘sad’ faces to measure ‘hot’ executive functioning. This version of the ‘Go/No-go’ task saw an emotional processing load superimposed onto mechanisms of action restraint that would typically be indexed by the more classic version of the ‘Go/No- go’ task. The latter was investigated via a ‘cold’ executive functioning condition whereby neutral ‘male’ and ‘female’ facial expressions were counterbalanced as ‘Go’ and ‘No-go’ trials (Rahimi-Golkhandan et al., 2015). Contrary to earlier findings, Rahimi-Golkhandan and colleagues (Rahimi-Golkhandan, Steenbergen, Piek, & Wilson, 2014) did not find differences in performance between DCD and TD groups on the ‘cold’ condition of the ‘Go/No-go’ task, suggesting preserved action restraint. However, where ‘No-go’ trials occur rarely (as was the case here), it can be difficult to discern the degree to which performance on such tasks reflects inhibitory control or an individual’s ability to process infre- quently occurring stimuli (Chikazoe et al., 2008). Taken as a whole, the status of action restraint in DCD remains unclear, paving the way for a much needed and highly controlled study to elucidate this important issue.

1.3. The present study

The aim of the present study was to examine the inhibitory deficit model of DCD by employing two carefully constructed measures of motor inhibition. To best index action restraint, we employed a modified version of the ‘Go/No-go’ task with sufficient task difficulty to engage inhibitory systems. To index action cancellation, we used a well vali- dated iteration of the ‘Stop-signal’ task. In keeping with the inhibitory deficit hypothesis of DCD, we predicted that motor inhibition profi- ciency would be reduced in DCD relative to typically developing con- trols, shown by decreased performance efficiency on both the ‘Go/No- go’ and ‘Stop-signal’ tasks respectively.

2. Methods

2.1. Participants

The sample consisted of 22 participants, 11 adults with DCD, aged 18–30 (3 males and 8 females; Mage=23.45, SD=2.21), and 11 ty- pically developing (TD) controls (4 males and 7 females; Mage=24.55, SD=4.20). Written informed consent was obtained from all partici- pants and all participants were financially compensated for their time. Ethical clearance was received from the Deakin University Human Research Ethics Committee.

All participants were screened using methods that have been suc- cessfully utilized by several research groups in identifying adults with DCD (e.g., (Du, Wilmut, & Barnett, 2015; Hyde et al., 2014; Hyde et al., 2018; Williams, Kashuk, Wilson, Thorpe, & Egan, 2017; Wilmut, Du, & Barnett, 2015). Participants were first recruited through advertisements placed on the website of an Australian University and social media outlets (i.e., Facebook). The posters contained a link to the Adult Dyspraxia/Developmental Coordination Checklist (ADC: (Kirby, Edwards, Sugden, & Rosenblum, 2010)), a self-report measure of pro- ficiency performing everyday tasks that involve movement. Participants that were deemed eligible were then contacted, and a time was orga- nized for participants to complete the two behavioral measures (i.e., the ‘Go/No-go’ and ‘Stop-signal’ tasks [see below]) and the Bruininks- Oseretsky Test of Motor Proficiency (BOT-2; (Bruininks, 2005)). The

BOT-2 is a well-validated standardized measure of motor ability which contains four domains: Fine Manual Control, Manual Coordination, Body Coordination and Strength and Agility. A ‘Total Motor Composite score’ (M=50, SD=10) can then be calculated for each participant by combining their performance on each of the four domains. The Total Motor Composite provides an index of each participant’s age-normed motor ability. The BOT-2 was adopted rather than other standardized measures of motor ability (e.g., the McCarron Assessment of Neuro- muscular Development (McCarron, 1982) or the Movement Assessment Battery for Children-2 (Henderson, Sugden, & Barnett, 2007)) as it has been recently found to be the most valid and reliable battery available for identifying motor difficulties in young adults in Australia (Hands, Licari, & Piek, 2015).

Participants with DCD had motor proficiency significantly below that expected given their age (Criterion A), indicated by BOT-2 Total Motor Composite scores< 15th percentile. These motor difficulties significantly impacted their ability to undertake daily activities invol- ving movement (Criterion B) and arose in childhood (Criterion C) as determined by their ADC checklist responses. Since the ADC contains separate ‘total’ and ‘child’ scale scores, it is commonly used to identify current (Criterion B) and childhood (Criterion C) difficulties completing daily living tasks related to motor function (Hyde et al., 2014; Kashuk, Williams, Thorpe, Wilson, & Egan, 2017; Wilmut et al., 2015). While the ADC is well-validated and reliable, standard cut-off scores for ad- dressing Criterion B and C are yet to be established. As per recent work (Hyde et al., 2018; Kashuk et al., 2017), we used cut-off scores based on our earlier study (Christian Hyde et al., 2014), which had developed 95% confidence intervals for total (CI95%: 21.26 Mean± 3.27) and child (CI95%: 4.26 Mean± 0.86) ADC scores using a sample of 47 healthy young Australian adults. Participants were considered to have met Criterion B if they scored above the 95% confidence interval cut off for the total ADC score of healthy young Australians (i.e., 25 or above), and Criterion C if they scored above the 95% cut off for the child subscale (i.e., 6 and above). One participant in our DCD group reported having had a prior diagnosis of DCD/Dyspraxia. This participant still com- pleted the ADC and BOT-2 to ensure they met criterion for the DCD group. Otherwise, participants with DCD did not report having any previous diagnosis of a neurological or medical condition affecting their movement (e.g., cerebral palsy), and were deemed to have had in- telligence at least in the normal range since they were recruited through the University setting and/or had completed an undergraduate degree (Criterion D).

All participants in the control group had Total Motor Composite scores above the 20th percentile (indicating age-appropriate motor ability) and did not report any medical or neurological impairments.

2.2. Measures

2.2.1. Action restraint The ability to restrain a prepotent action was assessed using a ‘Go/

No-go’ task originally developed by Chikazoe and colleagues (Chikazoe et al., 2008; Hirose et al., 2012). The task was programmed with E- prime software (Version 2.0, Psychology Software Tools, Pittsburgh, PA, USA), and was presented from a distance of approximately 50 cm at eye level on a 15-inch Acer (New Taipei, Taiwan) computer monitor. Participants were seated comfortably in a height adjustable chair in front of the computer monitor, with a keyboard centered approximately 6 in. from their midline.

Each trial began with the presentation of a fixation cross for 400ms (the ISI), followed by a colored circle (10 cm×10 cm in size) presented for another 400ms (see Fig. 1). The color of the circle denoted the trial- type, with white circles indicating ‘Frequent-go’ trials (occurring 75%

J.L. He et al. Brain and Cognition 127 (2018) 23–33


of trials), yellow circles indicating ‘Infrequent-go’ trials (occurring 12.5% of trials) and blue circles indicating ‘No-go’ trials (occurring 12.5% of trials). The color indicating ‘Infrequent-go’ and ‘No-go’ trials were counterbalanced across subjects to avoid any sequence effects (Chikazoe et al., 2008; Hirose et al., 2012). For both ‘Frequent’ and ‘Infrequent-go’ trials, participants were instructed to respond as quickly and accurately as possible by depressing the spacebar of the keyboard using the distal aspect of their right index finger. For ‘No-go’ trials, participants were instructed to withhold responding.

All participants completed a practice block prior to completing two test blocks. All blocks contained fifteen ‘Frequent-go’ trials which were presented at the beginning of each block to allow task acclimation before performance was recorded (Chikazoe et al., 2008; Hirose et al., 2012). These trials were not included in subsequent analyses and were not part of the total trial counts. Each test block contained 192 trials, consisting of 144 ‘Frequent-go’ (75%), 24 ‘Infrequent-go’ (12.5%), and 24 ‘No-go’ (12.5%) trials. RTs to ‘Frequent-go’ and ‘Infrequent-go’ trials were recorded when a response was made within the stimulus duration (400ms). Thus, if a response was not recorded in this 400ms window, it was assumed that the participant had responded during the proceeding fixation cross of the subsequent trial (Bezdjian et al., 2009). In these instances (∼8.5% and ∼19% of ‘Frequent-go’ trials and ‘Infrequent-go’ trials respectively), RTs for these trials were replaced with the max- imum possible RT for ‘Go’ trials (i.e., 400ms). ‘Go’ trials with RTs of< 150ms were excluded from analyses to exclude premature re- sponses (Chikazoe et al., 2008; Hirose et al., 2012). Mean RTs were then calculated for both ‘Frequent-go’ and ‘Infrequent-go’ trials. For ‘No-go’ trials, a participant was considered to have accurately completed the trial if they did not depress the spacebar during the duration of the trial. Thus, for each participant, the percentage of accurately inhibited ‘No- go’ trials was calculated.

The ability to inhibit a prepotent motor response (i.e., action re- straint) was evaluated using an efficiency index as per (Chikazoe et al., 2008; Hirose et al., 2012), which accommodated both accuracy on ‘No- go trials’ with consideration of RTs on ‘Infrequent-go’ trials. To control for potential individual differences in ability to process infrequently occurring stimuli, only individual mean RTs from ‘Infrequent-go’ trials (which occurred with equal frequency to ‘No-go’ trials) were used in the calculation of the efficiency index. As per Hirose et al. (2012) a linear regression was conducted between individual mean RTs on ‘Infrequent- go’ trials (Go-RT) and individual percentage accuracy on ‘No-go’ trials (Nogo-PC). The line of best fit (see Fig. 2A) was used to represent the performance of the average participant. The vertical distance between each participant and the regression line (i.e., their standardized re- sidual) was used to indicate their respective action restraint efficiency index (AREI). Thus, participants who scored above the regression line had positive AREIs, and were considered to be more efficient perfor- mers of action restraint than the average participant. Likewise, partici- pants who scored below the regression line had negative AREIs, and were considered to be less efficient performers (see Fig. 2B).

2.2.2. Stop-signal task The ability to cancel a prepared movement (i.e., action cancellation)

was assessed using a ‘Stop-signal’ task based on the work of Coxon and colleagues (Coxon et al., 2006). See also (Cowie et al., 2016, 2006; Macdonald et al., 2014, 2017). The task was programmed using

Fig. 1. Visual representation of the Go/No-go task used in the present study.

Fig. 2. (A) The distribution of response times in ‘Infrequent-go’ trials and percentage accuracy in ‘No-go’ trials. Black dots represent TD controls, while white dots represent DCD participants. (B) Performers who higher above the regression line were considered to be better performers of action restraint.

J.L. He et al. Brain and Cognition 127 (2018) 23–33


MATLAB, and was presented on the same computer monitor used for the ‘Go/No-go’ task. Participants’ forearms were positioned midway between pronation and supination, with the keyboard raised to elim- inate any radial or ulnar deviation of the wrist (Coxon et al., 2006).

The stimulus display consisted of a fixed, vertically orientated in- dicator bar (15 cm high × 1 cm wide) centered in the middle of the screen (see Fig. 3). Each trial commenced approximately 1 s after the participant fully depressed the 0 key with the medial aspect of their right index finger (Coxon et al., 2006, 2007; MacDonald et al., 2017). Upon trial commencement, the indicator bar filled at a constant velo- city from the bottom of the indicator bar (starting at 0ms) finishing at the top of the indicator bar (at 1000ms). A fixed target line was pre- sented 200ms from the top of the indicator (800ms from the bottom of the bar). Participants were informed that lifting their finger from the key would stop the indicator from rising, and that their primary task was to stop the indicator as close to the target line as possible (‘Go’ trials). Participants were also informed that in some of the trials (‘Stop’ trials) the indicator would stop unexpectedly prior to reaching the target line. Here, participants were instructed to keep their finger de- pressed on the computer key, rather than lifting their finger as they would for a ‘Go’ trial.

‘Stop’ trial difficulty was modulated by altering the time at which the rising bar stopped within a ‘Stop’ trial (i.e., the SSD). As per the horse race model of inhibition (Logan & Cowan, 1984), the later the rising indicator bar stopped relative to the start line (i.e., the later the SSD), the more difficult it would be for participants to inhibit their prepared finger lift. SSDs were set at seven different stop-times: 500ms, 525ms, 550ms, 575ms, 600ms, 625ms, and 650ms from the bottom of the indicator (see Fig. 4). SSDs resulting in a 50% accuracy rate on ‘stop’ trials are typically thought to index moderate task difficulty (Verbruggen & Logan, 2008, 2009). Our preliminary piloting demon- strated that this point consistently fell between 550 and 575ms, which we used as a mid-point for our range of SSDs. We expanded our SSDs in two 25ms increments, above 550ms and below 575ms, to ensure sufficient variation in task difficulty across stop trials. These values were strikingly similar to those reported in earlier accounts using this task in healthy adults (Coxon et al., 2006).

Participants first observed the performance of approximately 15 trials by the experimenter, before completing a minimum of 20 practice trials. Once participants had stated that they had understood the pur- pose of the task, they completed 450 self-paced, pseudorandomly pre- sented trials over 9 blocks (50 trials per block). Of the 450 trials, 310 were ‘Go’ trials (∼70%) and 140 were ‘Stop’ trials (∼30%). Of the 140 ‘Stop’ trials, there were 20 trials for each of the seven SSD times. For ‘Go’ trials, the time at which a participant lifted their finger (referred to as lift time [LT]) from the computer key relative to when the trial began (0ms) was recorded in ms.

As with the Go/No-go task, ‘Go’ trials with LTs of< 150ms were excluded from analysis to exclude premature responses. Mean LT for ‘Go’ trials, mean error, absolute error, and variable error were then calculated for each participant (Coxon et al., 2006). Mean LT referred to the average LT. Mean error refers to the mean of the deviation of LTs from the target line. Absolute error refers to the mean of the absolute deviations from the target line and variable error referred to 1 SD of the distribution of all lift times. For ‘Stop’ trials, participants were con- sidered to have successfully completed the trial if they did not lift their finger from the computer key after the rising indicator bar had stopped.

The ability to inhibit a prepared motor response (i.e., action can- cellation) was evaluated using stop-signal reaction times (SSRT). SSRTs are commonly used as an index of action cancellation efficiency (Verbruggen & Logan, 2009). Linear interpolation was first used to determine the SSD where the probability of responding was approxi- mately 50% for each participant. This value was then subtracted from the participant’s mean ‘Go’ trial LT to determine their individual SSRTs

Fig. 3. Visual representation of the Stop-signal task used in the present study.

Fig. 4. Visual representation of the range of SSD times presented during ‘stop’ trials.

J.L. He et al. Brain and Cognition 127 (2018) 23–33


(Coxon et al., 2006). Here, shorter SSRTs indicated greater action can- cellation ability, and longer SSRTs indicated worse action cancellation ability.

2.3. Design and analysis

Prior to analysis, all outcome variables were assessed for violations of normality. No violations of normality that could have reasonably influenced interpretation of the results were observed. Prior to all group comparisons, outliers (considered to be any data point that fell above or below 2.5 SD from the group mean (Miller, 1991)) were removed be- fore assumptions of homogeneity of variance were tested for. Where assumptions of homogeneity of variance were met, independent sam- ples t-test were used. Where assumptions of homogeneity of variance were violated, Welch’s t-test were used instead, as it has been shown to be more reliable in the presence of unequal variance (Ruxton, 2006).

2.3.1. Go/No-go task To confirm that participants were indeed slower to respond in

‘Infrequent-go’ trials compared to ‘Frequent-go’ trials (as per (Chikazoe et al., 2008; Hirose et al., 2012)), and to confirm this pattern of per- formance was comparable between DCD and control groups, a 2 (Trial type: ‘Frequent-go’, ‘Infrequent-go’) × 2 (Group: DCD, Control) re- peated measures ANOVA was conducted for mean RT. While RTs on ‘Frequent-go’ and ‘Infrequent-go’ trials and percentage accuracy on ‘No- go’ trials independently provide limited insight into action restraint, mean values for each dependent variable were presented for each group for the purpose of comparison with earlier and future accounts. Indeed, our RTs for both ‘Frequent-go’ and ‘Infrequent-go’ trials were very si- milar to that of previous work using the same ‘Go/No-go’ task (Chikazoe et al., 2008; Hirose et al., 2012). To compare efficiency of action restraint between groups, an independent samples t-test was conducted where the dependent variable was AREI, and the in- dependent variable was group (Group: DCD or TD).

2.3.2. Stop-signal task Preliminary analyses were conducted to confirm that accuracy on

‘Stop’ trials reduced as SSDs increased (as per our piloting data), and to determine whether this pattern of performance was comparable be- tween DCD and control groups. To this end, a 2 (Group: DCD, Control) × 7 (500ms, 525ms, 550ms, 575ms, 600ms, 625ms, 650ms) re- peated measures ANOVA was conducted on mean accuracy on ‘Stop’ trials.

To probe for group differences in regards to lift performance, in- dependent samples t-tests were conducted to compare the groups (DCD,

TD) on lift performance measures (i.e., mean lift time, mean error, absolute error and variable error). Finally, in order to compare DCD and TD groups on their ability to perform action cancellation, SSRTs were compared between groups using a Welch’s t-test.

3. Results

3.1. Action restraint: Go/ No-Go task analysis

Descriptive statistics for performance on ‘Frequent’ and ‘Infrequent- go’ trials, and ‘No-go’ trials are presented in Table 1. Results of the 2- way mixed design ANOVA on RTs during ‘Go/No-go’ performance showed that participants were significantly slower when responding to ‘Infrequent-go’ trials (M=298.69, SD=30.77) compared to ‘Frequent- go’ trials (M=274.99, SD=29.90), as indicated by a main effect of trial-type F(1,40) = 6.57, p= .014, η2 = 0.706. No significant main effect for group, F(1, 40) = 0.91, p= .345, partial η2 = 0.022 nor in- teraction was observed, F(1, 40) = 0.18, p= .677, partial η2 = 0.069.

The independent samples t-test comparing DCD and TD groups on mean AREI found that the DCD group (M=−5.54, SD=12.43) were significantly less efficient than their TD peers (M=5.54, SD=9.88), t (20) = 2.32, p= .031, partial η2 = 0.211 (see Fig. 5).

3.2. Action cancellation: Stop-signal task analysis

2 (Group: DCD, Control) × 7 (SSD: 500, 525, 550, 575, 600, 625, 650) repeated measures ANOVA on total accuracy demonstrated a significant linear trend for stop-time, suggesting that accuracy de- creased with increased SSD (see Fig. 6), F(6, 120) = 209.84, p= .000, partial η2 = 0.913. No significant interaction effect F(6, 120) = 0.40, p= .875, partial η2 = 0.020, or group effect F(1, 20) = 0.04, p= .838, partial η2 = 0.002, was observed.

Table 1 RTs on ‘Frequent’ and ‘Infrequent-go’ trials, and accuracy on ‘No-go’ trials of DCD and TD groups.

TD (N=11)

DCD (N=11)


Frequent-go 272.52 (33.40) – 277.47 (27.35) – Infrequent-go 292.33 (36.60) – 305.06 (23.66) – Nogo-PC – 50.57 (20.16) – 45.83 (19.87)

TD, typically developing, DCD; developmental coordination disorder; RT, re- sponse time; ACC, accuracy.

Fig. 5. Boxplots comparing AREI between TD and DCD groups.

J.L. He et al. Brain and Cognition 127 (2018) 23–33


Results of the independent samples t-tests comparing lift perfor- mance measures (mean lift time, mean error, absolute error and vari- able error) between DCD and control groups are presented in Table 2. No group differences were observed for any of the lift performance

measures. The Welch’s t-test comparing DCD and TD groups on SSRT found a

non-significant trend which suggested that the DCD group (M=223.56, SD=9.17) were slower at inhibiting their prepared re- sponses and thus less efficient at restraining prepotent actions com- pared to their TD peers (M=210.97, SD=17.72), t(15.29) = −2.07, p= .056, partial η2 = 0.176. See Fig. 7.

4. Discussion

The aim of the present study was to investigate the integrity of motor inhibition in individuals with DCD using two carefully con- structed measures of motor inhibition: The ‘Go/No-go’ task and ‘Stop- signal’ task. As hypothesised, individuals with DCD were less efficient at performing both action restraint and action cancellation, indicated by significantly reduced AREI scores on the ‘Go/No-go’ task and a non- significant trend of longer SSRTs on the ‘Stop-signal’ task compared to TD peers. Critically, the present study clarifies disparate evidence speaking to the integrity of action restraint in DCD and provides the first account of action cancellation using a purpose-built measure. In support of the inhibitory deficit hypothesis of DCD, these findings suggest that young adults with DCD experience broad difficulties with engaging inhibitory mechanisms for the purposes of action.

4.1. Less efficient action restraint in individuals with DCD

As predicted, individuals with DCD were less efficient at restraining prepotent actions compared to controls, evidenced by significantly lower AREI scores on the ‘Go/No-go’ task. This finding is consistent with a number of previous studies which had used the ‘Go/No-go’ task to assess action restraint in DCD (Cousins & Smyth, 2003; Querne et al., 2008), although contradictory to some others (Rahimi-Golkhandan et al., 2014, 2016). We argue that our finding of reduced AREIs in the DCD group is well placed to clarify these previously inconsistent find- ings. Specifically, our iteration of the ‘Go/No-go’ task incorporated rare ‘no-go’ trials, as well as short ISIs, both of which are considered ne- cessary to engage inhibitory systems at behavioural and neurological levels (Wessel, 2017). Further, the purposeful inclusion of the ‘in- frequent-go’ trials allowed us to control for individual differences in participant ability to process infrequently occurring stimuli, a

Table 2 Lift performance on the Stop-signal task.

TD DCD p-value Effect size (partial η2)

Mean Lift Time 819.09 (18.20)

823.39 (17.33)

0.577 0.016

Mean Error 19.09 (18.20) 23.39 (17.33) 0.577 0.016 Absolute Error 37.66 (11.41) 42.18 (14.94) 0.434 0.031 Variable Error 35.92 (13.31) 35.77 (15.05) 0.981 0.000

TD, typically developing, DCD; developmental coordination disorder.

Fig. 7. Boxplots comparing SSRT between TD and DCD groups.

Fig. 6. Linear trend between Accuracy and SSD for both DCD and TD groups.

J.L. He et al. Brain and Cognition 127 (2018) 23–33


mechanism considered to be unrelated to motor inhibition that can otherwise confound the interpretation of traditional ‘Go/No-go’ task performance (Chikazoe et al., 2008; Hirose et al., 2012). While these design characteristics are paramount for the valid measurement of motor inhibition using the ‘Go/No-go’ task, ours is the first study to incorporate all in an investigation of motor inhibition in DCD. Indeed, we can be confident that our task had elicited sufficient inhibitory demand to require an inhibitory response since similarly to earlier studies with comparable task parameters (Chikazoe et al., 2008; Hirose et al., 2012), both individuals with DCD and healthy controls failed to inhibit their responses in approximately half of the ‘No-go’ trials.

Importantly, we also argue that the profile of ‘Go/No-go’ task per- formance shown here by those with DCD is consistent with specific difficulties engaging motor inhibition, unlikely to be unduly influenced by general cognitive and/or psychomotor ability (c.f. traditional neu- ropsychological measures of inhibition). Indeed, successful completion of ‘Go’ trials required participants to correctly identify the colour of a basic geometric shape prior to executing or withholding a predefined basic movement (i.e., simple depression of a computer key). Hence, the perceptual, cognitive, and motor demands of the task were low and unlikely to have unduly affected task performance. This argument is supported by the finding that the groups did not differ on RTs in either ‘Frequent-go’ or ‘Infrequent-go’ trials. Further, the AREI measure on which groups differed provides an indication of the ability of partici- pants to withhold an inappropriate movement (as per accuracy on ‘No- Go’ trials), whilst accounting for RTs on ‘Go’ trials. In doing so, it controls for those participants who may have slowed their responses in ‘Go’ trials in order to be more accurate in ‘No-go’ trials (i.e., a speed- accuracy trade off), a strategy that can potentially mask group differ- ences in motor inhibition ability (Seli, Jonker, Cheyne, & Smilek, 2013). Thus, the pattern of data presented here support the idea that in- dividuals with DCD are less efficient at restraining prepotent actions compared to their TD peers. Taken together, our work provides a much- needed clarification on the nature of action restraint in DCD.

4.1.1. Action cancellation is inefficient in DCD Similarly, we found that individuals with DCD were also less effi-

cient at engaging action cancellation compared to TD controls, indicated by longer SSRTs on the well-validated ‘Stop-signal’ task. Although this comparison fell just short of statistical significance (p= .056), the ob- served effect size was large (partial η2 = 0.176) according to Cohen (1988) taxonomy. While there has been indirect evidence suggesting that individuals with DCD have difficulties cancelling and re-directing ongoing reaching actions in double-step reaching tasks (Fuelscher, Williams, Enticott, & Hyde, 2015; Hyde & Wilson, 2011a, 2011b; Wilmut, Wann, & Brown, 2006), there has been no direct investigation of action cancellation in DCD. Hence, here our study is the first to di- rectly assess this particular facet of motor inhibition using a purpose- built measure.

Importantly, like our findings from the ‘Go/No-go’ task, we argue that the group differences in task performance here are unlikely to have been unduly influenced by general cognitive and psychomotor pro- cesses. Like the ‘Go/No-go’ task, the motor demands of the ‘Stop-signal’ task are relatively simple. To successfully perform ‘Go’ trials, partici- pants were only required to depress and release a computer key in order to stop a rising indicator at a fixed target line. It is clear that both groups were able to perform the basic perceptual-motor processes ne- cessary for the task, as the results show that individuals in the DCD group were just as consistent and accurate as their TD peers on ‘Go’ trials, with comparable mean errors, absolute errors, and variable er- rors. Instead, group differences were only observed on SSRTs, a mea- sure that takes into account both RTs on ‘Go-trials’ and accuracy on ‘Stop-trials’. Like the AREI of the ‘Go/No-go’ task, SSRTs are used to control for those participants who slow their responses in ‘Go-trials’ in order to be more accurate in ‘Stop-trials’ when performing the ‘Stop- signal’ task, providing a more valid measure of action cancellation ability

than accuracy on ‘Stop-trials’ alone. Taken as a whole, like our finding of reduced action restraint in DCD, we argue that the longer SSRTs of the DCD group here specifically reflect a reduced ability to efficiently perform action cancellation, rather than difficulties with the general cognitive and/or psychomotor processes required by these tasks. Taken together, our finding of reduced action cancellation in DCD provides further support for the inhibitory deficit hypothesis of DCD.

4.1.2. Adults with DCD show broad deficits in motor inhibition By adopting two gold-standard measures of motor inhibition, we

were able to assess the integrity of two related, yet distinct, facets of motor inhibition in adults with DCD: action restraint and action cancel- lation. As a result of this multifaceted approach, our results provide compelling evidence to support the idea that motor inhibition is com- promised in DCD. Our work extends earlier accounts by demonstrating that those difficulties with action restraint previously reported using the ‘Go/No-go’ task (Cousins & Smyth, 2003; Querne et al., 2008) also extend to the cancellation of prepared actions. Together, our results suggest that individuals with DCD have broad difficulties with the ef- ficient inhibition of actions, in-keeping with the inhibitory deficit hy- pothesis of DCD.

Interestingly, our suggestion that individuals with DCD may have broad difficulties with engaging inhibitory systems for the purposes of action is supported by a pair of recent investigations into executive functioning (EF) in children with DCD, which demonstrated that the EF deficits commonly observed in this group are mostly pronounced when motor demands are high (Bernardi, Leonard, Hill, Botting, & Henry, 2017; Leonard et al., 2015). For example, using a wide range of verbal and non-verbal measures to assess EF in children with DCD, Bernardi and colleagues (Bernardi et al., 2017) found that the children with DCD mostly had difficulties completing nonverbal measures of EF where the motor and visuospatial demands of the tasks were high. Conversely, on verbal measures of EF where the motor and visuospatial demands were low, children with DCD were able to perform comparably to their TD peers on the majority of the measures. Based on these pattern of results, the authors speculated that the EF difficulties experienced by children with DCD may be more related to their primary difficulties with motor control rather than a more general deficit with cognitive processing. Thus, while the neuropsychological assessments adopted in this study were not explicitly designed to measure motor inhibition as con- ceptualised here, the results of their study nonetheless provide support for the notion that individuals with DCD have broad difficulties with engaging inhibitory processes specifically for the purpose of action.

Our behavioural data here are consistent with neuroimaging evi- dence suggesting that individuals with DCD may activate executive systems differently to healthy controls during movement. Indeed, re- sults of a recent activation likelihood estimation meta-analysis showed that children with DCD demonstrate atypical activation of executive neural systems during the performance of tasks of manual dexterity, including the inferior parietal lobe (IPL) and superior frontal gyrus (SFG; (Fuelscher et al., 2018)). Hence, it may be that the atypical motor inhibition observed in our sample of young adults with DCD is sub- served by differential activation of such executive or ‘inhibitory’ cir- cuity. In support, those regions that were shown to be activated dif- ferentially in DCD during manual control (e.g., the IPL and SFG) are also known to be central to ‘Go/No-go’ task performance in healthy controls (Simmonds et al., 2008), and hence, motor inhibition. Further, given the important role of the basal ganglia in motor inhibition (Aron et al., 2014) and the direct connections between the basal ganglia and the cerebellum via the subthalamic nucleus (Bostan & Strick, 2018), our findings also show support for the body of studies which have pre- viously implicated both the basal ganglia (Gheysen, Van Waelvelde, & Fias, 2011; Pitcher, Piek, & Barrett, 2002; Smits-Engelsman, Westenberg, & Duysens, 2008) and the cerebellum (Mariën, Wackenier, De Surgeloose, De Deyn, & Verhoeven, 2010; Zwicker, Missiuna, Harris, & Boyd, 2011) in the pathophysiology of DCD.

J.L. He et al. Brain and Cognition 127 (2018) 23–33


It is worth highlighting that our finding of a broad deficit in motor inhibition in DCD here has specifically come from a sample of young adults. While it might be tempting to conclude that the motor inhibition difficulties identified here are a continuation of childhood inhibitory deficits, given previous reports of motor inhibitory difficulties in chil- dren with DCD (Cousins & Smyth, 2003; Querne et al., 2008), further longitudinal work assessing motor inhibition in DCD across the devel- opmental spectrum needs to be conducted before we can make this assumption. Moreover, a longitudinal approach investigating the re- lationship between reduced motor inhibition and motor skill acquisi- tion may provide insight into if and/or how compromised motor in- hibition contributes to atypical development of motor skill in DCD. As canvassed earlier, because inhibition has been shown to be amenable to treatment (Tsai, 2009), determining whether reduced inhibition plays a causal role in the development of poor motor skill in DCD is critical, as it could help establish whether further development of interventions aimed at improving inhibition in this disorder group would be justified.

Finally, while the results of the present study are promising, they must be considered in the light of the study’s limitations. First, while no participants in our DCD group reported having been diagnosed with ADHD, it is possible that some participants may have had co-occurring yet undiagnosed ADHD, especially since the comorbidity rates between DCD and ADHD are suggested to be as high as 50% (Pitcher, Piek, & Hay, 2003). Indeed, in an investigation of inhibitory functioning in DCD, it is important to consider such a possibility. Further, given our modest sample size, we must be circumspect about the degree to which our findings can be generalized towards all individuals diagnosed with DCD. That is, inspection of the AREIs and SSRTs do suggest that some participants in the DCD group were able perform motor inhibition as efficiently as their TD counterparts. For example, only seven of the eleven (or 74%) participants with DCD had AREIs outside the 95% confidence intervals of the AREIs of the control group (CI95%: 5.54Mean± 6.64). Similarly, only six of the eleven (or 55%) of the participants in the DCD group had SSRTs outside the 95% confidence intervals for SSRTs of the control group (CI95%: 210.97Mean± 11.9). Thus, despite the significant group difference in AREIs, or non-sig- nificant trend in the case of SSRTs, our results show that a subset of participants in the DCD group were still able to perform motor inhibi- tion as efficiently as their TD counterparts, suggesting that impaired motor inhibition may only be present in a subset of diagnosed in- dividuals. Importantly, this finding is also commonly reported in other studies (He et al., 2018; Hyde & Wilson, 2011a, 2011b) which show that even where neuro-cognitive mechanisms are found to be atypical in DCD, a subset of the participants in the DCD group appear to be able to perform similarly to controls, perhaps highlighting the heterogeneity of symptom presentation in this disorder group.

4.1.3. Conclusion In summary, the present study clarifies previously inconsistent

findings speaking to the nature of action restraint in DCD and provides the first account of action cancellation in this disorder group. In parti- cular, we found that individuals with DCD were less efficient at per- forming action restraint compared to their TD peers, as evidenced by reduced AREIs on the ‘Go/No-go’ task. Further, novel to this study was the finding that individuals with DCD were also less efficient at per- forming action cancellation, evidenced by longer SSRTs in the ‘Stop- signal’ task. Together, these results suggest that individuals with DCD have broad deficits engaging inhibitory mechanisms for the purpose of action. Since our sample consisted of young adults with DCD, our work is consistent with the view that deficits in motor inhibition may persist beyond childhood in those with atypical motor skill. However, future longitudinal evidence is necessary to confirm.


The authors would like to thank all the participants who

volunteered their time and participated in this study.

Appendix A. Supplementary material

Supplementary data to this article can be found online at https://


Aron, A. R., Fletcher, P. C., Bullmore, E. T., Sahakian, B. J., & Robbins, T. W. (2003). Stop- signal inhibition disrupted by damage to right inferior frontal gyrus in humans. Nature Neuroscience, 6(2), 115.

Aron, A. R., Robbins, T. W., & Poldrack, R. A. (2014). Inhibition and the right inferior frontal cortex: One decade on. Trends in Cognitive Sciences, 18(4), 177–185.

Association, A. P. (2013). Diagnostic and statistical manual of mental disorders (DSM-5®). American Psychiatric Pub.

Barkley, R. A. (1997). ADHD and the nature of self-control. Guilford Press. Barry, R. J., & De Blasio, F. M. (2015). Performance and ERP components in the equi-

probable go/no-go task: Inhibition in children. Psychophysiology, 52(9), 1228–1237. Barry, R. J., & Rushby, J. A. (2006). An orienting reflex perspective on anteriorisation of

the P3 of the event-related potential. Experimental Brain Research, 173(3), 539–545. Bedi, A. (2015). The effects of response probability on commission errors in high go low

no-go dual response versions of the sustained attention to response task (SART). Benikos, N., Johnstone, S. J., & Roodenrys, S. J. (2013). Varying task difficulty in the Go/

Nogo task: The effects of inhibitory control, arousal, and perceived effort on ERP components. International Journal of Psychophysiology, 87(3), 262–272. https://doi. org/10.1016/j.ijpsycho.2012.08.005.

Bernardi, M., Leonard, H., Hill, E. L., Botting, N., & Henry, L. A. (2017). A two-year follow-up study of executive functions in children with Developmental Coordination Disorder. Developmental Medicine and Child Neurology.

Bezdjian, S., Baker, L. A., Lozano, D. I., & Raine, A. (2009). Assessing inattention and impulsivity in children during the Go/NoGo task. British Journal of Developmental Psychology, 27(2), 365–383.

Biotteau, M., Chaix, Y., & Albaret, J.-M. (2016). What do we really know about motor learning in children with Developmental Coordination Disorder? Current Developmental Disorders Reports, 3(2), 152–160.

Bostan, A. C., & Strick, P. L. (2018). The basal ganglia and the cerebellum: Nodes in an integrated network. Nature Reviews Neuroscience, 1.

Bruin, K., & Wijers, A. (2002). Inhibition, response mode, and stimulus probability: A comparative event-related potential study. Clinical Neurophysiology, 113(7), 1172–1182.

Bruininks, R. H. (2005). Bruininks-Oseretsky Test of Motor Proficiency (BOT-2). Chambers, C. D., Bellgrove, M. A., Stokes, M. G., Henderson, T. R., Garavan, H.,

Robertson, I. H., … Mattingley, J. B. (2006). Executive “brake failure” following deactivation of human frontal lobe. Journal of Cognitive Neuroscience, 18(3), 444–455.

Chikazoe, J., Jimura, K., Asari, T., Yamashita, K.-I., Morimoto, H., Hirose, S., … Konishi, S. (2008). Functional dissociation in right inferior frontal cortex during performance of go/no-go task. Cerebral Cortex, 19(1), 146–152.

Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale: Erlbaum Associates.

Cousins, M., & Smyth, M. M. (2003). Developmental coordination impairments in adulthood. Human Movement Science, 22(4), 433–459.

Cowie, M. J., MacDonald, H. J., Cirillo, J., & Byblow, W. D. (2016). Proactive modulation of long-interval intracortical inhibition during response inhibition. Journal of Neurophysiology, 116(2), 859–867.

Coxon, J. P., Stinear, C. M., & Byblow, W. D. (2006). Intracortical inhibition during vo- litional inhibition of prepared action. Journal of Neurophysiology, 95(6), 3371–3383.

Coxon, J. P., Stinear, C. M., & Byblow, W. D. (2007). Selective inhibition of movement. Journal of Neurophysiology, 97(3), 2480–2489.

Coxon, J. P., Stinear, C. M., & Byblow, W. D. (2009). Stop and go: The neural basis of selective movement prevention. Journal of Cognitive Neuroscience, 21(6), 1193–1203.

Dambacher, F., Sack, A. T., Lobbestael, J., Arntz, A., Brugman, S., & Schuhmann, T. (2014b). The role of right prefrontal and medial cortex in response inhibition: Interfering with action restraint and action cancellation using transcranial magnetic brain stimulation. Journal of cognitive neuroscience, 26(8), 1775–1784.

Dambacher, F., Sack, A. T., Lobbestael, J., Arntz, A., Brugman, S., & Schuhmann, T. (2014a). A network approach to response inhibition: Dissociating functional con- nectivity of neural components involved in action restraint and action cancellation. European Journal of Neuroscience, 39(5), 821–831.

Du, W., Wilmut, K., & Barnett, A. L. (2015). Level walking in adults with and without Developmental Coordination Disorder: An analysis of movement variability. Human Movement Science, 43, 9–14.

Eagle, D. M., Bari, A., & Robbins, T. W. (2008). The neuropsychopharmacology of action inhibition: Cross-species translation of the stop-signal and go/no-go tasks. Psychopharmacology, 199(3), 439–456.

Fuelscher, I., Caeyenberghs, K., Enticott, P., Williams, J., Lum, J., & Hyde, C. (2018). Differential activation of brain areas in children with developmental coordination disorder during tasks of manual dexterity: An ALE meta-analysis. Neuroscience & Biobehavioral Reviews.

Fuelscher, I., Williams, J., Enticott, P. G., & Hyde, C. (2015). Reduced motor imagery efficiency is associated with online control difficulties in children with probable developmental coordination disorder. Research in Developmental Disabilities,

J.L. He et al. Brain and Cognition 127 (2018) 23–33