An examination of the relationship between motor coordination and executive functions in adolescents

Authors


Ms Daniela Rigoli at School of Psychology and Speech Pathology, Curtin University of Technology, GPO Box U1987, Perth 6845, Western Australia, Australia. E-mail: d.rigoli@curtin.edu.au

Abstract

Aim  Research suggests important links between motor coordination and executive functions. The current study examined whether motor coordination predicts working memory, inhibition, and switching performance, extending previous research by accounting for attention-deficit–hyperactivity disorder (ADHD) symptomatology and other confounding factors, in an adolescent normative sample.

Method  Ninety-three adolescents (38 females, 55 males) aged 12 to 16 years (mean age 4y 2mo, SD 1y 1mo) were assessed on the Movement Assessment Battery for Children-2 (MABC-2), Wechsler Intelligence Scale for Children-IV, N-back task, the inhibition subtest from the NEPSY-II: A Developmental Neuropsychological Assessment, second edition, and the parent-rated Strengths and Weaknesses of ADHD Symptoms and Normal Behaviour Questionnaire.

Results  The MABC-2 total score accounted for a significant proportion of the variance in visuospatial working memory (p=0.041) but not for verbal working memory. The MABC-2 aiming and catching component, however, was found to account for unique variance in both verbal (p=0.019) and visuospatial working memory (p=0.016). The MABC-2 total score was found to account for a significant proportion of the variance in inhibition total completion time (p=0.017). Finally, balance skills accounted for unique variance in a NEPSY-II inhibition total errors variable (p=0.020).

Interpretation  The results provide support for an overlap between motor coordination and executive functions, which has important practical implications. The study also suggests shared mechanisms underpinning the relationship between these areas, including possible cerebellar involvement.

Abbreviations:
DCD

Developmental coordination disorder

MABC-2

Movement Assessment Battery for Children-2

SES

Socio-economic status

SWAN

Strengths and Weaknesses of ADHD Symptoms and Normal Behaviour

VCI

Verbal Comprehension Index

WISC-IV

Wechsler Intelligence Scale for Children-IV

WMI

Working Memory Index

What this paper adds

  • The results show that motor coordination may be more closely linked to visuospatial working memory than to verbal working memory.
  • ‘Aiming and catching’ skills may be linked to both verbal and visuospatial working memory.
  • Motor coordination is related to performance speed on inhibition tasks.
  • Balancing skills are related to interference control.

It has been noted that motor control involves cognitive processes such as inhibiting frequently used movements, anticipating and updating aspects of the task to allow forward planning, resisting interference due to automatic postural control and fatigue, and the monitoring and correction of incorrect movements.1 However, although there is some suggestion that complex cognitive processes (i.e. executive functions) affect motor performance, causal evidence regarding the direction of the relationship is limited.

The notion that motor development may predict cognitive functioning is partly supported by research highlighting that it is the sensory and motor functioning regions of the brain that are typically the first to mature.2 Furthermore, longitudinal studies have found that early motor development predicts later performance on complex cognitive tasks, including working memory.3 Conversely, in a study of preschool children, Niederer et al.4 found that baseline memory was not associated with an improvement in motor skills 9 months later.

Diamond5 argued that the close association between motor and cognitive development is mediated by the coactivation of the cerebellum and the prefrontal cortex. It is also important to note the role of individual differences when understanding this relationship.6 For example, there are a number of studies suggesting that physical activity and high levels of aerobic fitness during childhood may enhance neurocognition.7 This provides further evidence that motor coordination may predict executive functions.

Evidence for the relationship between motor performance and executive functions also exists from behavioural studies. Normative studies1 as well as those examining developmental coordination disorder (DCD)8 have demonstrated a link between motor coordination and working memory. Baddeley’s9 model of working memory comprises separable components for the temporary storage of verbal (i.e. the phonological loop) and visuospatial (i.e. visuospatial sketchpad) information, and research in the area of DCD has found that motor coordination may be more closely linked to visuospatial working memory than to verbal working memory.8 This may be partly understood in terms of the visuospatial processing deficit found in individuals with DCD.10

Regarding other executive function domains, studies have found that children with coordination difficulties are slower in performing inhibition and attention shifting tasks but are not less accurate than typically developing children.6 It is possible that this reflects an automatization deficit in children with motor impairments, suggesting that cerebellar mechanisms may be implicated in the slower performances on these tasks.

The available literature on the relationship between motor functioning and executive functions leaves a number of issues needing to be addressed. First, it is important to control for attention and/or hyperactivity–impulsivity (attention-deficit–hyperactivity disorder [ADHD]) symptomatology as ADHD has been linked with motor problems11 as well as executive function areas such as working memory and inhibition.12 Very few studies have employed normative samples of children. Normative studies are important given methodological problems associated with clinical samples such as overestimating associations between domains.1 In addition, as there is evidence from normative samples1 and studies examining motor impairment6 that specific components of motor coordination have a relationship with certain aspects of executive function, whereas others do not, it is important to examine these components separately. Furthermore, research is needed in adolescent samples given that previous studies have involved younger children or a mixed sample of children and adolescents.13,14

The current study examined the relationship between motor coordination (namely overall motor performance, manual dexterity, aiming and catching, and balance) and executive functions (namely working memory, response inhibition, and switching) in an adolescent normative sample, whilst controlling for ADHD symptomatology, age, gender, socio-economic status (SES), and verbal ability. It is hypothesized that motor coordination will show a significant relationship with working memory, and this may be stronger for visuospatial working memory than for verbal working memory.8 It is also hypothesized that a significant relationship will be found between motor coordination and the timing measures from the response inhibition and switching tasks, but not with motor coordination and the accuracy variable.

Method

Participants

Recruitment occurred across five randomly selected secondary schools and through public advertisements (e.g. community newspapers). Adolescents aged 12 to 16 years were eligible for inclusion and had a minimum Verbal Comprehension Index (VCI) of 80 as measured by the Wechsler Intelligence Scale for Children-IV (WISC-IV), in order to exclude any adolescent whose difficulties might be attributed to general delayed development.15 Furthermore, a parent-rated developmental history questionnaire was used to ascertain the absence of physical disability, chronic illness, pervasive developmental disorder, and neurological disorder. Ninety-four participants responded and consented to the project; however, one participant with undiagnosed hand tremor was excluded. The final sample included 93 adolescents (38 females and 55 males) with a mean age of 14 years 2 months (SD 1y 1mo). The Australian Prestige Scale16 was used to provide SES scores. The scale assesses the prestige of occupations, with scores ranging from 1 (reflecting high prestige) to 6.9 (reflecting low prestige). For the current study, the occupation rated as most prestigious out of mother’s and father’s occupation was used as the SES score (mean 3.8, SD 1.0, range 1.8–6.6).

Measures

Movement Assessment Battery for Children-2 (MABC-2)

The MABC-2 is a standardized test used for the identification and description of children with movement difficulties.17 Age-based standard scores for manual dexterity, aiming and catching, and balance components and a total test score are provided (mean 10, SD 3), with higher scores demonstrating better performance. A total test score at or below the 5th centile indicates significant movement difficulty, whereas a score between the 5th and 15th centile indicates that a child is ‘at risk’.

Henderson et al.17 provide evidence suggesting favourable psychometric properties for the MABC-2. A reliability coefficient of 0.80 for the total test score and coefficients ranging from 0.73 to 0.84 for the individual component scores are reported.17

Wechsler Intelligence Scale for Children-IV – Australian

The WISC-IV is a measure of cognitive ability for children aged 6 years to 16 years 11 months.18 The 10 core subtests yield a full-scale IQ and four indices of verbal comprehension (i.e. VCI), perceptual reasoning, working memory (i.e. WMI), and processing speed. For the current study, the VCI was used as a potential control variable and to exclude any adolescent whose difficulties might be attributed to general delayed development. The WMI was employed as measure of verbal working memory. The WISC-IV is widely used and has excellent internal consistency, test–retest reliability, criterion validity, and construct validity.18

N-back task

The N-back task was used to assess visuospatial working memory, designed after Gevins and Cutillo19 and Jansma et al.20 The task has also been adapted to make it more attractive and appropriate for children.21 An apple with four holes from which a caterpillar appears is presented on the computer screen. Respondents are required to press one of the four buttons that corresponds spatially with the hole from which the caterpillar emerged. There are four conditions of graded difficulty requiring the respondent to indicate where the caterpillar was one move back, two moves back, three moves back, or four moves back. Each condition comprises a practice block (10 trials) and an experimental block in which performance is measured (32 trials). Respondents move to the next level of difficulty only if they score a minimum of eight correct responses (indicating performance above chance levels) in the experimental blocks. For this study, task performance is measured by the total number of correct responses across the conditions (maximum raw score of 128). The N-back task is a widely used measure of working memory, and in a study involving a sample of adolescents test–retest reliabilities of 0.70 and 0.66 were reported for 3- and 4-back, respectively.21

NEPSY-II: a developmental neuropsychological assessment

The NEPSY-II provides a comprehensive neuropsychological assessment for children and adolescents aged 3 to 16 years.22 The ‘naming’, ‘inhibition’, and ‘switching’ sections of the inhibition subtest were administered for the purposes of this study. These sections assess, respectively, simple naming skills, the ability to inhibit automatic responses in favour of novel responses, and the ability to switch between response types. The age-standardized total completion time scaled score for the inhibition and switching sections were utilized for this study, with higher scores representing faster completion times. A total errors scaled score was also used, which combines errors across all sections in the inhibition subtest (namely naming, inhibition and switching sections). A higher total errors scaled score corresponds to better performance (i.e. fewer errors made).

The inhibition subtest has shown adequate to high internal consistency, for example average reliability coefficients for the inhibition and switching combined scaled scores range from 0.73 to 0.87 for ages 12 to 16 years.22 The NEPSY-II also demonstrates adequate stability across time, as well as good content, construct, and criterion-related validity.22

Strengths and Weaknesses of ADHD Symptoms and Normal Behaviour (SWAN)

The parent-rated SWAN scale includes 18 items based on ADHD symptoms listed in DSM-IV.23 Parents are asked to rate the items based on observations from the last month and with reference to age-matched peers. Scores for each item range from +3 (i.e. ‘far below average’) to −3 (i.e. ‘far above average’). For the current study, attention and hyperactivity/impulsivity scores were calculated by averaging the total of the nine corresponding items, with positive scores indicating presence of symptoms and negative scores indicating absence of symptoms.

The SWAN scale has been found to yield a normal distribution of scores, making it useful for examining variability in (hyper)activity and (in)attention in the general population.24 Martin et al.25 found the prevalence rate of ADHD, as assessed using the SWAN scale, to be similar to what has been reported in previous studies.

Procedure

The Curtin University Human Research Ethics Committee granted approval for the project and National Health and Medical Research Council of Australia ethical guidelines were followed. Approval was also granted from the participating schools’ representative bodies and, subsequently, from interested school principals in Perth, Western Australia. Adolescents and their parents provided written consent and were then individually tested by a trained examiner at home or at the university, depending on family preference. Testing duration was 4.5 hours over two sessions. Measures were administered in a standard manner. Parents completed questionnaires including a developmental history questionnaire and the SWAN scale.

Statistical analysis

A series of hierarchical regressions were conducted to determine whether the MABC-2 total score or its component scores (manual dexterity, aiming and catching, and balance) accounted for incremental variance in working memory (N-back accuracy, WISC-IV WMI), inhibition (NEPSY-II inhibition total completion time scaled score), switching (NEPSY-II switching total completion time scaled score), and the NEPSY-II total errors scaled score, after controlling for covariates (WISC-IV VCI, SWAN attention and hyperactivity/impulsivity symptoms). In a hierarchical regression analysis, ΔR2 represents the increase in the proportion of variance in the criterion variable explained from step N−1 to step N. The term sr2 represents the unique amount of variance that a predictor brings to the model. In a hierarchical regression analysis where just one predictor is added at step N, then the ΔR2 from step N−1 to step N will be equivalent to the sr2 for the added predictor.

The most complex regression model included three control variables and three primary predictors. Our sample size of 93 was sufficient to detect moderate relationships (i.e. f2=0.12) between the criterion variables and the primary predictors.26

Results

Descriptives

Table I shows the means, standard deviations, and ranges for the study variables. Five adolescents scored at or below the 5th centile on the MABC-2 total score, indicating significant movement difficulty. The prevalence of significant movement difficulty was 5.4%, which is similar to previous estimates of 6%.27 Two adolescents scored between the 6th and 15th centiles, suggesting that they were ‘at risk’ of movement difficulty.

Table I.   Means, SDs, and range of scores for the study variables
 MeanSDRange
  1. aAge-standardized score. bRaw score. cTotal number of correct responses. dScores are calculated by averaging the total of the nine attention or hyperactivity/impulsivity items. eThe occupation rated as most prestigious out of mother’s and father’s occupations.

  2. MABC-2, Movement Assessment Battery for Children-2; WISC-IV, Wechsler Intelligence Scale for Children-IV; SWAN, Strengths and Weaknesses of ADHD Symptoms and Normal Behaviour; SES, socio-economic status.

MABC-2 total scorea10.632.563.0–16.0
MABC-2 manual dexteritya9.572.473.0–15.0
MABC-2 aiming and catchinga11.032.734.0–16.0
MABC-2 balancea11.422.984.0–14.0
WISC-IV Working Memory Indexa103.7512.4759.0–141.0
N-back accuracyb,c88.1719.696.0–124.0
NEPSY-II inhibition completion timea10.682.94.0–19.0
NEPSY-II switching completion timea10.642.423.0–16.0
NEPSY-II total errorsa8.883.191.0–16.0
SWAN attentionb,d−0.831.17−3.0 to 2.33
SWAN hyperactivity/impulsivityb,d−1.171.03−3.0 to 1.11
WISC-IV Verbal Comprehension Indexa106.6311.2581.0–132.0
SESb,e3.771.001.80–6.60

Bivariate correlations

The correlations between the criterion variables, predictors, and control variables are shown in Table II.

Table II.   Zero-order correlation matrix for the key and control variables
 MABC-2 total scoreMABC-2 manual dexterityMABC-2 aiming and catchingMABC-2 balanceSWAN attentionSWAN hyperactivity/impulsivityVCIAgeSexSES
  1. a p<0.01 (two-tailed). bp<0.05 (two-tailed).

  2. MABC-2, Movement Assessment Battery for Children-2; SWAN, Strengths and Weaknesses of ADHD Symptoms and Normal Behaviour; VCI, Verbal Comprehension Index; SES, socio-economic status; WISC-IV, Wechsler Intelligence Scale for Children-IV; WMI, Working Memory Index.

MABC-2 total score          
MABC-2 manual dexterity0.657a         
MABC-2 aiming and catching0.656a0.071        
MABC-2 balance0.780a0.264b0.423a       
SWAN attention−0.178−0.252b−0.052−0.106      
SWAN hyperactivity/impulsivity0.020−0.0700.0930.0070.724a     
VCI0.1520.0750.0480.155−0.382a−0.209b    
Age−0.114−0.069−0.066−0.0950.077−0.036−0.167   
Sex−0.0690.235b−0.397a−0.007−0.251a−0.184b−0.0180.021  
SES−0.074−0.1220.032−0.0990.1920.000−0.384a0.244b−0.083 
WISC-IV WMI0.2010.1130.251b0.112−0.298a−0.1350.442a−0.089−0.020−0.174
N-back accuracy0.271a0.1290.281a0.146−0.242b−0.1130.253b0.150−0.051−0.039
NEPSY-II inhibition completion time0.276a0.229b0.1760.163−0.0770.0270.1420.008−0.0920.061
NEPSY-II switching completion time0.237b0.1890.215b0.101−0.1820.0250.219b−0.158−0.102−0.017
NEPSY-II total errors0.1900.0960.0360.259b−0.325a−0.229b0.222b−0.0760.1630.007

Multiple linear regression analyses

As expected, there were strong correlations between the MABC-2 total score and each of its component scores (see Table II). Because the MABC-2 total score was a reliable predictor of the component scores, it was included as a proxy for the component scores in the primary analysis, thereby reducing the complexity of the regression model and optimizing statistical power. Because there was no correlation between the MABC-2 aiming and catching and manual dexterity components (r=0.071), individuals with the same total score can be fundamentally different at the component level. It was important to conduct secondary regression analyses that replaced MABC-2 total score with its component scores. Only those outcomes that were significantly associated with the MABC-2 score total, or at least one component score, were analysed. The VCI was entered first, followed by SWAN attention and hyperactivity/impulsivity (Table II indicates that these were the only covariates), and then the MABC-2 total score or its component scores.

Working memory

After controlling for the three covariates, the MABC-2 total score explained a significant 4.2% of the variance in N-back accuracy (ΔR2=0.042; p=0.041). When the MABC-2 total score was replaced by its component scores, however, the combined scores explained no additional variance over and above that already explained by the covariates (ΔR2=0.069; p=0.077), although aiming and catching uniquely explained a significant 5.8% of the variance in N-back accuracy (sr2=0.058; p=0.016).

After controlling for the three covariates, the MABC-2 total score explained no additional variance (ΔR2=0.011; p=0.272) in WISC-IV WMI performance. Similarly, when the MABC-2 total score was replaced by its component scores, the combined scores explained no additional variance over and above that already explained by the covariates (ΔR2=0.050; p=0.123), although aiming and catching uniquely explained a significant 4.8% of the variance in WMI performance (sr2=0.048; p=0.019), and VCI uniquely explained 12.2% of the variance (sr2=0.122; p<0.001). Table III summarizes the regression results for the N-back and WISC-IV WMI tasks.

Table III.   Step 3 statistics for hierarchical multiple regression analyses predicting working memory performance from MABC-2 total score or component scores (n=93)
 Working memory outcomes
WMIN-back
B95% CI sr 2 p-valueB95% CI sr 2 p-value
  1. a<0.01. b<0.05. c<0.001.

  2. WMI, Working Memory Index; B, unstandardized regression coefficient; CI, confidence interval; sr2, the part correlation squared, VCI, Verbal Comprehension Index; MABC-2, Movement Assessment Battery for Children-2.

Model 1 predictors
 VCI0.410.18, 0.630.1140.001a0.29−0.09, 0.660.0230.134
 Attention−2.13−5.27, 1.010.0160.181−2.86−8.13, 2.400.0110.282
 Hyperactivity/impulsivity1.03−2.31, 4.360.0030.5430.76−4.83, 6.360.0010.787
 MABC-2 total score0.52−0.42, 1.460.0110.2721.650.07, 3.220.0420.041b
 Total R2  0.197<0.001c  0.0970.011b
Model 2 predictors
 VCI0.430.20, 0.650.122<0.001c0.32−0.06, 0.700.0280.094
 Attention−1.82−4.99, 1.350.0110.257−2.67−8.04, 2.710.0090.327
 Hyperactivity/impulsivity0.61−2.71, 3.930.0010.7170.37−5.25, 5.990.0000.895
 Manual dexterity0.23−0.77, 1.230.0020.6490.52−1.17, 2.220.0040.540
 Aiming and catching1.120.19, 2.060.0490.019b1.950.37, 3.540.0580.016b
 Balance−0.30−1.18, 0.580.0040.502−0.20−1.70, 1.290.0010.788
Total R2  0.221<0.001c  0.1040.016b

Inhibition and switching

After controlling for two of the three covariates (VCI was not correlated with inhibition completion time), the MABC-2 total score explained a significant 6.1% of the variance in inhibition completion time (ΔR2=0.061; p=0.017). When the MABC-2 total score was replaced by its component scores, however, the combined scores explained no additional variance over and above that already explained by covariates (ΔR2=0.063; p=0.120).

After controlling for covariates, the MABC-2 total score explained no additional variance in switching completion time (ΔR2=0.023; p=0.128). Similarly, when the MABC-2 total score was replaced by its component scores, the combined scores explained no additional variance over and above that already explained by the covariates (ΔR2=0.043; p=0.236).

After controlling for covariates, the MABC-2 total score explained no additional variance in the total errors score (ΔR2=0.016; p=0.199). Similarly, when the MABC-2 total score was replaced by its component scores, the combined scores explained no additional variance over and above that already explained by the covariates (ΔR2=0.055; p=0.137), although the MABC-2 balance component uniquely explained a significant 5.4% of the variance in total errors (sr2=0.054; p=0.020). Table IV summarizes the regression results for the inhibition and switching tasks.

Table IV.   Step 3 statistics for hierarchical multiple regression analyses predicting inhibition completion time, switching completion time, and total errors from MABC-2 total score or component scores (n=93)
  Executive functioning
Inhibit timeSwitch timeTotal errors
B95% CI sr 2 p-valueB95% CI sr 2 p-valueB95% CI sr 2 p-value
  1. B, unstandardized regression coefficient; CI, confidence interval; sr2, the part correlation squared, VCI, Verbal Comprehension Index; MABC-2, Movement Assessment Battery for Children-2.

  2. a<0.05.

Model 3 predictors
VCI0.03−0.02, 0.080.0160.2050.03−0.03, 0.090.0090.340
Attention−0.24−0.99, 0.510.0040.521−0.60−1.25, 0.050.0330.070−0.62−1.48, 0.230.0210.148
Hyperactivity/impulsivity0.26−0.58, 1.100.0040.5370.61−0.08, 1.300.0300.082−0.13−1.04, 0.770.0010.769
MABC-2 total score0.290.05, 0.530.0610.017a0.15−0.04, 0.350.0230.128−0.17−0.09, 0.420.0160.199
Total R2  0.0500.056  0.0880.016a  0.0940.013a
Model 4 predictors
VCI0.04−0.01, 0.080.0220.1410.02−0.04, 0.090.0050.457
Attention−0.19−0.96, 0.590.0020.630−0.53−1.19, 0.140.0240.121−0.75−1.62, 0.120.0280.089
Hyperactivity/impulsivity0.23−0.62, 1.090.0030.5880.55−0.15, 1.250.0240.120−0.03−0.94, 0.870.0000.947
Manual dexterity0.23−0.03, 0.490.0320.0840.13−0.08, 0.340.0150.218−0.06−0.33, 0.220.0020.676
Aiming and catching0.14−0.10, 0.380.0140.2590.17−0.03, 0.370.0290.090−0.11−0.36, 0.150.0060.409
Balance0.05−0.18, 0.270.0020.683−0.06−0.24, 0.130.0040.5490.290.05, 0.530.0540.020a
Total R2  0.0310.173  0.0880.029a  0.1140.011a

Discussion

In the current study, the MABC-2 total score accounted for a significant proportion of the variance in a visuospatial working memory task but not in verbal working memory. These results suggest that motor coordination may be more closely related to visuospatial working memory than to verbal working memory, supporting previous findings of a specific deficit in visuospatial memory in children with DCD.8

However, aiming and catching (but not manual dexterity or balance) accounted for statistically significant unique variance in both visuospatial and verbal working memory. This supports previous research revealing specific relationships between working memory and certain aspects of motor coordination. For example, Piek and colleagues3 found a relationship between early gross motor (but not fine motor) development and later working memory ability in a normative sample of school-aged children. Although speculative in nature, it is possible that the specific association found in the current study may be partly explained by shared underlying cerebellar processes. The lateral zone of the cerebellum is important for the rapid, aimed movements required in aiming and catching tasks.28 Research has also implicated the cerebellum in working memory.29 Furthermore, Diamond5 highlighted the close coactivation of the cerebellum and prefrontal cortex when understanding the relationship between complex motor and cognitive domains. Consequently, it is also possible that the complex nature of ball skills assessed in the current study coactivates greater prefrontal cortex activity than the tasks solely assessing manual dexterity or balance skills.

Research has also demonstrated how individuals with motor difficulties tend to avoid participation in sporting domains.30 Given the reported cognitive benefits of physical activity through physiological (e.g. increased cerebral blood flow) and learning/developmental mechanisms,7 it is possible that the lack of opportunity to learn and practise the skills associated with aiming and catching games may play an important role in understanding the link found in the current study.

The MABC-2 total score was also found to account for a significant proportion of variance in inhibition completion time, supporting previous research in normative and motor impairment samples.6,14 The slower performance speed on inhibition tasks for children with DCD may be understood in terms of an automatization deficit, most likely linked to cerebellar dysfunction.6 Querne and colleagues31 also demonstrated slower responses for children with DCD and showed that children with DCD demonstrated abnormal hemispheric lateralization for attentional and inhibitory functions. It is also important to note that the current findings of a significant association between motor coordination and inhibition completion time may simply reflect the involvement of speed in both tasks, as, upon inspection of the MABC-2 components, manual dexterity (comprising two timed tasks) appeared to be the strongest predictor in explaining this relationship. However, Michel and colleagues,6 who reported slower performance in children with motor impairment, argued that their results were unlikely to be due to differences in information processing speed, as the children with motor impairment did not perform more slowly in a simple reaction time task. Rather, it was suggested that the slower performance of children with motor impairments was due to the complex demands of the task.6

In the current study, a non-significant relationship between motor coordination and switching completion time was found. This may suggest possible differences in the neural processes underlying inhibition and switching. Switching is a complex task requiring various cognitive processes in addition to the inhibitory demands inherent in the task.22 Thus, the switching task may require the recruitment of additional prefrontal regions and/or may be mediated by different cortical areas. In fact, functional magnetic resonance imaging studies have shown that localization within the frontal cortex is task dependent.32 This may explain the divergent findings between inhibition and switching in the current study.

An unexpected result in the current study was the specific link found between balancing ability and total errors (a composite score including inhibition and switching errors). This does not support previous research linking motor coordination to only the timing components of such tasks. However, it is important to note that these previous studies involved a composite score of movement ability14 or a group of children defined by fine motor difficulties.6

The link between balance and total errors supports accumulating evidence for the attentional requirements of young and older children during postural tasks.33 Woollacott and Shumway-Cook34 argued that postural control requires significant attentional resources depending on the complexity of the postural task and the individual’s age and balance abilities. The NEPSY-II inhibition subtest used in the current study is based on the Stroop paradigm and, thus, examines interference control (i.e. the ability to ignore irrelevant information). Therefore, this suggests that interference control may be important when understanding balancing ability.

The current study has some limitations. The study cannot provide information on the directional relationship between the motor and cognitive domains. There is some evidence to suggest that motor development may predict cognitive performance; however, further longitudinal research is needed.3,4 In addition, performance accuracy was measured by a variable combining simple naming, inhibition, and switching errors, which introduces the problem of process specificity. It is also important, in attempting to interpret the results of the present study, to note that other variables may have played a role (e.g. processing speed, motivation). In addition, the digit-span forward component of the WISC-IV WMI and the 1-back level of the N-back task may be considered measures of storage rather than of the processing component of working memory. However, Unsworth and Engle35 have suggested that short-term memory and working memory tasks largely measure the same basic processes and therefore argue against the notion that short-term memory and working memory are different constructs.

Finally, although the current sample size was sufficient to detect important relationships, upon closer inspection of the predictors for inhibition completion time, manual dexterity appeared to be the strongest predictor, although this did not reach statistical significance (p=0.084). This suggests that future research could benefit from examining these relationships with a larger sample.

Conclusion

The results of this adolescent normative study suggest specific relationships between aspects of motor coordination and executive functions. It is possible that the specific relationships found in the current study (e.g. between aiming and catching skills and working memory) may be understood through shared neural mechanisms, namely, cerebellar processes.

The current results have practical implications when considering interventions for motor and/or executive functioning difficulties. For example, the current study highlights the importance of assessing executive functions in individuals who present with motor difficulties and, subsequently, tailoring the intervention accordingly. Similarly, it may also be important to screen for motor difficulties in those who present with executive function problems.

Acknowledgements

We are very grateful to the parents and adolescents who were willing to participate in this study. We also thank Sean Piek, Linda Pannekoek, and Eva Kuhry for their assistance with data entry.

Ancillary