A new perspective on the development of motor variability during middle childhood Golenia, Laura
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.
Document Version
Publisher's PDF, also known as Version of record
Publication date:
2018
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):
Golenia, L. (2018). A new perspective on the development of motor variability during middle childhood.
Rijksuniversiteit Groningen.
Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.
Laura Golenia, Raoul M. Bongers, Jessika F. van Hoorn, Egbert Otten, Leonora J. Mouton & Marina M. Schoemaker
Human Movement Sciences (2018). 60: 202-213.
CHAPTER 4
Variability in coordination patterns in children with
Developmental Coordination Disorder (DCD)
Abstract
High motor variability is an often-found characteristic of Developmental Coordination Disorder (DCD). Still, the role of high motor variability in DCD needs further examination. This study focused on variability in coordination patterns, which is essential considering that DCD is a coor- dination disorder. We examined variability in coordination patterns of the arm over repetitions of trials in goal-directed reaching movements. This variability was partitioned into variability that does not affect the index fingertip position (Vucm) and variability that does affect the index fingertip position (Vort). This study aimed to increase the understanding of motor variability in DCD by comparing Vucm and Vort between children with DCD and typically developing (TD) children in a goal-directed reaching task. Twenty-two children (eleven with DCD) ages 6 – 11 performed 30 reaching movements. The Uncontrolled Manifold method was used to quantify Vucm and Vort. Results showed that children with DCD had more Vucm than TD children while Vort was similar between groups, showing that coordination patterns in children with DCD are more variable, but interestingly, this higher variability does not affect performance. This study indicates that high motor variability in DCD is not necessarily negative. Possible roles of motor variability in DCD are discussed.
CHAPTER 4
Introduction
Developmental Coordination Disorder (DCD) is an idiopathic condition that affects an estimated 6% of school-aged children and is characterized by impaired fine and gross motor coordination [1]. Children with DCD exhibit impairment while performing many activities of daily living. For example, children with DCD have problems balancing on one leg, directing their arms to the right location to catch a ball or grasping a glass of water without spilling. Findings from two decades of research on DCD consistently demonstrated that children with DCD show higher motor variability than typically developing (TD) children [2–7]. Most studies have focused on variability in perfor- mance measures and show, for example, that children with DCD are more variable than TD children in timing multi-limb coordination [8], generating manual force [2], and producing continuous and discontinuous drawing movements [4]. Unfortunately, even though motor variability is considered an important characteristic of DCD, its role is still poorly understood [9–11].
Until now, high performance variability in children with DCD has often been attributed to an high level of neuromotor noise [e.g., 2,9,12,13]. Noise in this case is seen as signal error, which means that high performance variability refers to inconsistency. High performance variability in a context that requires consistency can therefore be viewed as detrimental, hindering successful performance of the activities of daily living in children with DCD. However, an increasingly prominent view in the developmental literature conceptualizes motor variability as more than simply reflecting noise [e.g., 11,14–18]. In this view it is emphasized that motor variability can also be functional and even necessary for flexible and skilful development [c.f., 19,20,21,22]. This conception of variability has received limited attention in research on DCD, which is partly due to the fact that studies have mainly focused on performance variability and not on variability in coordination patterns (see for exceptions [23–26]). This is surprising considering that DCD is a coordination disorder. Inves- tigating variability in coordination patterns might reveal new insights into the origins of motor variability in DCD, which might affect the implementation of interventions [c.f., 11,27]. Increased attention to variability in coordination patterns has also been proposed by recent reviews of the possible causes of motor problems of children with DCD [10,11,28,29]. In the current paper, we aim to take a first step in understanding variability in coordination patterns in DCD by focusing on goal-directed reaching movements.
What do we know about differences in the variability of reaching movements between children with DCD and TD children? Until now, most studies on reaching in DCD focused on motor variability in performance measures (i.e., precision and movement time). Inconsistent results were found for target precision, as measured by constant error (mean deviation) and variable error (consistency of the deviation). Ameratunga et al. [30] found a higher constant error in children with DCD com- pared to TD children, whereas Van der Meulen et al. [31] found no differences between groups in variable error. Schoemaker et al. [32], on the other hand, found differences in variable error, but
not for constant error. More consistent results were found for movement time: Longer movement times [30,31,33–35] and more variability (standard deviation of movement time; [33]) were found in children with DCD compared to TD children. In sum, motor variability in performance measures appears to be increased for children with DCD. Here, we examine measures reflecting the coordi- nation patterns that produce these outcomes.
Focusing on the coordination patterns of reaching movements implies focusing on the degrees of freedom (DoF) that contribute to reaching. The contributing DoF are the joint angles of the arm, which are defined as the relative orientations of the different segments of the arm and hand (finger, wrist, elbow and shoulder joints). The wrist has, for example, two DoF (i.e. two joint angles): the wrist can flex or extend and abduct or adduct. We distinguished nine joint angles of the arm [c.f., 36,37]. If we take this number into account, there are far more DoF available in the arm than the minimum required to successfully perform the reaching task (i.e., reaching the 3D target position in space with the tip of the index finger; [c.f., 38]) This means that DoF are abundant, which, according to the principle of motor abundance [39–43]), is a luxury instead of a problem [38]. That is, motor abundance allows movements to be both flexible and stable [43]. For example, imagine sitting in front of a table and keeping your index finger tip stable at one position on the table.
While the finger position is stabilized, it is possible to vary the joints of the arm, implying that you are using multiple joint angle configurations. However, there are also joint angle configurations that result in a shift of the index finger position [43]. For instance, an extreme extension of your elbow joint cannot be compensated by another joint, which means that your index finger tip will move from its original position on the table. This demonstration shows that there are two kinds of variability in joint configurations; variability that does and variability that does not affect the position of the index finger.
To examine possible differences in these two types of variability between children with DCD and TD children, we need to distinguish them. This can be done with the Uncontrolled Manifold (UCM) method [42,44–46]. In our reaching task, the UCM is a manifold in joint space that represents the set of joint angle configurations in the arm with the tip of the index finger in one position.
Variance in joint angles over repeated trials projected onto the UCM corresponds to the joint angle configurations that do not affect the mean position of the index finger in space (Vucm). Variance orthogonal to the UCM subspace (the ORT subspace) corresponds to the joint angle configurations that lead to a deviation from the mean position of the index finger (Vort). Using the UCM method, we can test how variability in coordination patterns is structured by separating total variability into Vucm and Vort and examining whether this structure is different for children with DCD compared to TD children [47,48]. Studying structure in variability is generally considered an important way to reveal underlying coordination processes [e.g., 8,22].
CHAPTER 4 Until now, the UCM method has been primarily used to study the structure in joint angle variability
in adult behavior in various tasks [e.g., 44,48]. Only recently have the coordination patterns of 5-to 10-year-old TD children been studied with this method [Chapter 2,Chapter 3,19]. In all tasks, more Vucm than Vort was found for both adults and children [Chapter 2,Chapter 3,37,44,50–52].
Black et al. [53] found that children with Down Syndrome also had more Vucm than Vort in a walking task. Together, these studies show that, in all populations, task performance is stabilized as more joint angle combinations that do not affect task performance are used. Interestingly, children with Down Syndrome had a different structure in joint angle variability than their TD peers. Children with Down Syndrome showed higher values of Vucm and similar values of Vort [53], suggesting that their increased variability in coordination patterns represents a larger manifold of task-stabilizing joint angle configurations. Importantly, several studies with adult participants have shown that Vucm increases in challenging task conditions [e.g., 54,55]. Latash et al. [41] stated that this sug- gests that variability in coordination patterns is not simply a by-product of a noisy system that hinders performance.
In the current study, we examined differences between children with DCD and TD children in the structure of variability in coordinative patterns when performing goal-directed reaching move- ments. We hypothesized that Vucm is larger than Vort in both groups. Based on Black et al. [53], we hypothesized that Vucm is higher in children with DCD compared to TD children. We hypothesized that Vort is larger for children with DCD compared to TD children because reaching performance is often deteriorated in DCD [e.g., 30,33]. We also examined differences between DCD and TD children in performance measures (i.e., movement time and accuracy). As heterogeneity is an important feature of DCD [11], we explored individual differences in addition to group differences [21,56,57].
Method
Participants
The group of children with DCD consisted of 11 children aged between 6 and 11 years (M age = 9.1, SD = 1.7, range = 6.7-11.8, 10 boys). All children in the DCD group were referred to a rehabilitation clinic for outpatient diagnostics and/or treatment for their motor problems. All children met the operationalized DSM-V criteria for DCD [1], which were checked during the intake for outpatient diagnostics. The DSM-V criteria include the following: (A) a score below the 16th percentile on the Movement Assessment Battery for children-2nd Edition (MABC-2, [58]), (B) problems with the performance of motor tasks in daily life, (C) an onset of motor problems at an earlier age, and (D) exclusion of a medical condition that might explain these motor problems. During outpatient diagnostics, a physical therapist conducted the MABC-2 test, a rehabilitation physician quantified criterion B and C during an interview and excluded other medical conditions (criterion D) during
a medical examination. If there was any indication of cognitive developmental delay during the diagnostic session, children had a psychological examination (children diagnosed with DCD scored above 70 on an intelligence test). All children with DCD included in the current study attended mainstream primary schools. All included children performed the MABC-2 test again during the experimental session to reassess criterion A. The MABC-2 provides an indication of motor func- tioning across fine and gross motor tasks for children aged 3 to 16 years. The test consists of three age-related item-sets, measuring manual dexterity, ball skills, and balance. Children get a score on each item and total scores are transformed to centile scores considering the age of the concerned child.
Aged-matched controls were selected from a group of TD children who participated in a develop- mental study in which the same task was used. The results of that study are described in a different paper (Chapter 3). TD children with the age closest to a particular child with DCD and of the same sex were selected for the control group. This resulted in 11 age-matched controls (M age = 8.9, SD = 1.6, range = 6.9 - 11.0, 10 boys). Age-matched couples did not differ more than ±3 months, except for one couple. This couple had an 8-month difference, because the recruitment of the TD children was only done until an age of 11 years, whereas one child with DCD was 11 years and 8 months. Typical motor development was assessed with the MABC-2 test (score above the 16th percentile). TD children were recruited from local sport clubs and schools in the neighborhood of the university. TD children and children with DCD were right handed.
The study with children with DCD was approved by the Medical Ethical Committee of the University Medical Center Groningen, whereas the study with TD children was approved by the local ethics committee of the Center for Human Movement Sciences, University Medical Center Groningen.
Both studies were conducted according to the principles expressed in the Declaration of Helsinki.
Children’s parents or legal guardians provided written informed consent prior to the experiment.
Apparatus reaching task
To obtain joint angles of the shoulder, elbow, wrist, and index finger, 3D position data of all seg- ments of the right arm were collected with two Optotrak 3020 system sensors (Northern Digital, Waterloo, Canada). Three LED markers were attached to one rigid body to make attachment to the body easier. In total, six rigid bodies were attached to the participants’ right arm and trunk [59].
One rigid body was attached to the index finger so that it splinted the finger to prevent motion of the inter-phalangeal joints, i.e., the finger was considered as one segment in the analysis. The other five rigid bodies were triangular shaped and attached to the sternum, the acromion, the upper arm, the lower arm, and to the dorsal surface of the hand (Figure 1). Nineteen anatomical landmarks were digitized using a standard pointer device [59]. These landmarks were linked to the positions of the rigid bodies, which allowed extraction of the positions of the fingertip and the computation of the joint angles.
CHAPTER 4 The task was performed at a black table (height = 72cm). A large television screen (Panasonic,
62*111 cm) was incorporated in the tabletop to present the task display (Figure 1). The task display was developed using Presentation (Neurobehavioral systems, Berkely, CA). Children sat in a chair (Tripp Trapp, Stokke, Sweden) adjusted to their height, so that the relation between the table and participant was similar across participants. The back of the chair was extended with a board and children´s trunk was gently strapped against it. This was done to prevent movements of the torso, while allowing for free movements of the shoulder and elbow joints. To keep the starting posture of the upper extremity similar over trials, the elbow of the right arm was placed on a rest that was positioned at a comfortable height on the right side of the participant (Figure 1).
Figure 1. Experimental setup. Bird’s-eye view of a participant sitting at the experimental table. The partic- ipant was gently strapped to the chair (grey straps). The posture represents the posture at the start of each trial. The elbow of the participant was placed on the elbow rest and the index finger was positioned on the start position. An experimenter sat next to the child. Triangles and the rectangle on the finger represent rigid bodies (the rigid body on the sternum cannot be seen in the bird´s-eye view).
Procedure reaching task
Prior to the experiment, participants completed three practice trials to make sure that participants understood the instructions. One experimenter conducted all data collections and gave instruc- tions to all participants. An assistant experimenter sat next to the child to ensure that the hand and the arm were in the required position at the beginning of each trial, and to check that the child was attentive to the task.
At the beginning of each trial, participants were instructed to touch the start location with their right index finger while the elbow was positioned on the elbow rest. The start position (black, 2 cm diameter) was located 10 cm away from the body at the midline of the long side of the screen, which was aligned to the body midline (see Figure 1). A green target (2 cm diameter) appeared 25 cm away from the start location at the same midline. After hearing a beep that was emitted at a random interval of 1.0–1.5 seconds after the target location has appeared, participants could initiate the movement at their convenience. Participants were instructed to perform the reaching movement as fast and accurately as possible. The trial ended with holding the finger on the target location for a short period of time. This implies that children could see whether they reached the target or not. No additional knowledge of results was provided. Participants performed 30 reaching trials. If necessary, a short break was given after 15 trials.
Data analysis
For all analyses, customized data analysis programs were developed in Matlab (MathWorks; Natick, Massachusetts). To determine the initiation and the termination of the reaching movement, a backward (movement initiation) and forward (movement termination) search was performed from the maximum in the velocity profile of the forward direction (x-direction) of the index finger until a threshold of 5 cm/s, respectively. The first points below threshold were taken as the initiation and termination of the reaching movement, respectively. Correct determination of movement initiation and termination was checked visually for each trial.
Performance measures of the reach
To examine how DCD and TD children performed the reaching task, we calculated error variables and movement time. For each trial, the difference vector between the center of the target posi- tion and fingertip position at movement termination was calculated. Based on this, the constant error (CE; mean difference) and the variable error (VE; within-subject standard deviation of the difference) were determined. For each trial, movement time (MT) was determined as the time from movement initiation to movement termination. Based on this measure, the mean (referred to as MTmean) and the within-subject standard deviation (referred to as MTstdev) were calculated. For each individual participant, linear regression between MT and the difference vector between the target and the fingertip position for all trials was computed. The intercept and the slope of the regression line were determined for each individual. This was done to check whether the DCD group had a different relation between speed and accuracy than the TD group. Note that this calculation differs from the conventional speed-accuracy analysis. CE, VE, MTmean, MTstdev and the intercept and the slope of the regression lines were analyzed with independent t-tests to compare the TD and the DCD group, using SPSS version 20.0 (IBM, Armonk, New York). The level of significance was set at α < 0.05. Cohen´s d was used as effect size. For the TD group, Pearson’s correlations between CE, VE, MTmean, MTstdev and the MABC-2 percentile scores were examined.
CHAPTER 4 Variability in coordination patterns (UCM measures)
To examine the structure in joint angle variability, we computed Vucm and Vort with the UCM method as it has been described in detail previously [42–44,60]. Based on previous studies [36,37,44,52,61], the elemental variables selected were nine joint angles of the arm: shoulder plane of elevation, shoulder elevation, shoulder inward–outward rotation, elbow flexion–exten- sion, forearm pronation–supination, wrist flexion–extension, wrist abduction–adduction, index finger flexion–extension, and index finger abduction–adduction. These joint angles were calculated as proposed in the ISB standardization proposal for the upper extremity by Wu et al. [62]. The position of the index fingertip was selected as performance variable and had three dimensions in this study. The relation between changes in elemental variables and changes in the performance variable were computed using multiple regression [63–65] and united in a Jacobian (J) matrix [44,60]. Its null-space was used as a linear approximation of the UCM. Vucm and Vort were computed by projecting the total variance in joint space onto the null-space of J (Vucm) and the orthogonal complement (Vort), respectively. Equations 1 and 2 show the computation of Vucm and Vort, where tr denotes the trace of a matrix, J denotes the Jacobian matrix, C denotes the covariance matrix of all joint angles, n denotes the dimension of the joint space (n = 9) and d denotes the dimension of the task space (d = 3). Vucm and Vort were normalized by its number of DoF.
Eq (1)
Eq (2)
To correct for non-normal data distributions, Vucm and Vort were log transformed prior to the statistical analysis [66] and are displayed before log transformation in the figures in the results section. Vucm and Vort were analyzed at the instant of movement termination because reaching the target was the only constraint in the current study. This also enabled a comparison with the errors at the end-point of the index finger. Two paired t-tests were performed to compare Vucm and Vort within the DCD group and within the TD group. One independent t-test was performed to compare Vucm between groups and another one was performed to compare Vort between groups.
For the TD group, Pearson’s correlations between Vucm and Vort and the MABC-2 percentile scores were examined.
Results
61 out of 330 trials were removed from the DCD group and 74 out of 330 trials were removed from the TD group, leaving 525 trials for analysis. The minimum number of reaching trials per child included in the analysis was 18. Trials were excluded when at least one variable could not be 𝑉𝑉"#$=𝑡𝑡𝑡𝑡(𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛(𝐽𝐽).∗ 𝐶𝐶 ∗ 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛(𝐽𝐽))
𝑛𝑛 − 𝑑𝑑
𝑉𝑉45.=𝑡𝑡𝑡𝑡(((𝐽𝐽.)6).∗ 𝐶𝐶 ∗ (𝐽𝐽.)6) 𝑑𝑑
𝑉𝑉"#$=𝑡𝑡𝑡𝑡(𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛(𝐽𝐽).∗ 𝐶𝐶 ∗ 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛(𝐽𝐽))
𝑛𝑛 − 𝑑𝑑
𝑉𝑉45.=𝑡𝑡𝑡𝑡(((𝐽𝐽.)6).∗ 𝐶𝐶 ∗ (𝐽𝐽.)6) 𝑑𝑑
determined because of the occlusion of rigid bodies. The rigid body placed on the sternum was most often occluded, because children´s small sternum was easily occluded by the head when looking down to the target on the table. Another reason to exclude trials was when the initial posture was not correct or when the task was not performed according to the instructions, because reliable calculations of the UCM measures required stable experimental conditions.
Movement-ABC-2 test
Results of the Movement-ABC-2 test showed that all children with DCD scored below the 16th percentile (M percentile = 2.66, SD = 1.69, range = 0.1 - 5). Actually, all children in the DCD group scored below the 5th percentile which stresses the severity of the coordination problems of the included children. All children in the TD group scored above the 16th percentile on the Move- ment-ABC-2 test (M percentile = 53.45, SD = 13.53, range = 37 - 75), indicating typical motor development. For the TD group, no significant correlations were found between the M-ABC per- centile scores and the other examined dependent variables, all p´s > 0.37.
Performance measures of the reach
T-tests for both error measures, VE (p = 0.24) and CE (p = 0.26), showed no significant differences between the DCD group and TD group (Figure 2A and B, left columns). Results of the t-test of MTmean showed that the DCD group moved significantly slower than the TD group, t(20)= 2.48, p
= 0.011, d = 1.06 (Figure 2C, left column). Visual inspection of individual data for MTmean revealed that most (8 out of 11) of the children with DCD had longer MTmeans than their age-matched controls (Figure 2C, right column). MTstdev was significantly higher for the DCD group than for the TD group, t(20)= 2.61, p = 0.009, d = 1.11 (Figure 2D, left column). Visual inspection of the data revealed that 7 out of 11 children with DCD had higher MTstdevs than their age-matched controls (Figure 2D, right column). The t-test for the slope of the regression line between MT and accuracy revealed no differences between groups (p = .21). The intercept of the regression line between MT and accuracy revealed a significant difference between groups, t(20)= 2.95, p = 0.004, d = 1.25, demonstrating that the intercept was higher in the DCD group (M = 0.74, SEM = 0.06) than in the TD group (M = 0.52, SEM = 0.04).
Variability in coordination patterns within groups
In the DCD group, Vucm (M = 0.014, SEM = 0.003) was larger than Vort (M = 0.002, SEM = 0.001) at movement termination, t(10)= 11.41, p < 0.001 d = 1.31. Each child with DCD showed this structure. In the TD group, Vucm (M = 0.003, SEM = 0.001) was also larger than Vort (M = 0.001, SEM < 0.001) at at movement termination, t(10)= 10.79, p < 0.001, d = 1.77. Larger Vucm than Vort was also seen in each TD child.
CHAPTER 4
Figure 2. Performance measures of the reach for children with DCD and TD children. A. Constant error, B. Variable error, C. mean movement time, D. standard deviation of movement time. Black bars represent children with DCD, grey bars represent TD children. The left column shows group data, error bars represent standard error of the mean (SEM). The right column shows individual data for each matched pair, presented from youngest to oldest. The number on the x-axis below each pair represents the age of the child with DCD of the concerned pair.
Comparison of variability in coordination patterns between groups
Vucm was significantly higher in the DCD group compared to the TD group, t(20)= 2.02, p = 0.031, d = 0.86 (Figure 3A, left column). Visual inspection of the data revealed that almost all (9 out of 11) children with DCD had higher Vucm than their age-matched controls (Figure 3A, right column).
One child with DCD had lower Vucm than his age-matched control, and another pair had practically equal Vucm. Vort was not different for the two groups (p = 0.23).
Figure 3. Mean values and individual data for Vucm (A) and Vort (B). Black bars represent children with DCD, grey bars represent TD children. Data at movement termination are shown. The left column shows the mean for the DCD group and TD group, error bars represent standard error of the mean (SEM). The right column shows individual data for each matched pair, presented from youngest to oldest pair. The number on the x-axis below each pair represents the age of the child with DCD of the concerned pair. Note that the x-axis is scaled differently in A and B.
CHAPTER 4
Discussion
In this study, we approached motor variability in children with DCD in a novel way. Using the UCM method, we studied the structure of variability in coordination patterns, because examining structure in variability is generally considered an important way to reveal underlying coordination processes [e.g., 14,49]. Variability in coordination patterns in a simple reaching task was divided into variability not affecting the index finger position over repetitions of trials (Vucm), and vari- ability affecting this position (Vort). Results showed that children with DCD and TD children had similar reaching errors, but MT of children with DCD was longer and more variable. Concerning variability in coordination patterns, we found that children with DCD and TD children showed more Vucm than Vort. This shows that task performance was stabilized, as variability in joint angle combinations was mostly variability not affecting the position of the fingertip. As hypothesized, children with DCD had more Vucm than TD children. However, contrary to the initial hypothesis Vort was similar between groups. These findings shed an interesting light on the concept of variability in DCD: until now most studies have revealed higher variability in performance measures reflecting inconsistency, giving a negative connotation to variability in actions of children with DCD. The current findings show that children with DCD also demonstrate more variability in coordination patterns than TD children, but interestingly, this higher joint angle variability was variability not affecting performance. It is therefore not a sign of inconsistency. Consequently, variability in coor- dination patterns in DCD does not necessarily have to be interpreted as negative. The next step is to reveal the exact role of this higher variability. In the following section we therefore provide some suggestions for what the role of this variability might be and possible ways to assess this role in future studies.
One possibility is that the high Vucm reflects an high level of neuromotor noise, like it has been suggested by previous studies in DCD [e.g., 12]. Arguably, if the higher Vucm in children with DCD compared to age-match controls is a result of a higher level of noise, then also Vort and VE should be higher, because noise would affect the whole system. However, results showed that Vort and VE were not increased. We therefore argue that our results indicate that high variability in coordina- tion patterns in DCD reflects not just noise. Such an interpretation of variability is in line with other developmental and adult studies [e.g., 11,14,67]. Importantly, recent studies in DCD also emphasize that variability in DCD can be more than noise [11,29]. But how then, can the increased variability in coordination patterns in DCD be understood?
In motor learning and developmental literature it has been proposed that (parts of) variable behavior might reflect exploration processes [c.f., 16,19,20,21,68]. In the joint space, exploration would occur within the space spanned by the joint angles (the joint space) to find the manifold representing the solutions of the task. This exploration would manifest itself partly in Vucm and partly in Vort. Vucm would reflect the search within the solution manifold, whereas Vort represents the search outside the solu-
tion manifold. Although the understanding of these UCM measures in terms of exploration seems appealing, the current study is not up to the task to provide a final conclusion on this matter. First of all, the UCM method cannot quantify systematic trial to trial changes in coordination patterns that define exploration [19], because the UCM is estimated for a block of trials. When interested in examining changes over time, one would need to conduct more of these blocks and assess whether Vucm changes over these blocks [51]. Second, the reaching task we used was rather simple; it is there- fore questionable how much exploration could be expected to occur when performing this task.
However, taking into account that children with DCD still performed worse (longer MTmean) than TD children, it seems that the task is still somewhat difficult for children with DCD.
Assuming that the increased Vucm in children with DCD is exploration, what could it tell us about the nature of DCD? It is generally known that movements of children with DCD are less coordinated, that is, they often fail in activities of daily life [69]. Exploration for new movement solutions might indicate searches to overcome such clumsiness. Higher exploration might generate information that can be exploited for increasing task success or for tailoring actions to the constraints at hand.
Children with DCD had only higher level of Vucm, which might indicate that they searched within the solution space. As Vucm does not affect the position of the index finger, exploration within Vucm does not lead to increased accuracy. Instead, exploration within Vucm could concern exploration of task constraints that are not directly coupled to accuracy, such as requirements related to move- ment effort or energy. In the current study such processes might be reflected in slower MT found for children with DCC. Children with DCD and TD children performed the reaching task at the same level of accuracy, but for children with DCD this level of accuracy went at cost of MT. It could, therefore, be that the higher Vucm in children with DCD reflects a search for faster movement speed.
The increased standard deviation in MT in children with DCD might emphasizes this suggestion as exploration of different reaching speeds implies varying movement speed across repetitions.
Future studies could conduct a learning experiment and use different time analysis techniques [19,20,70] to quantify whether children with DCD keep searching (indicating problems) or whether a stable pattern gradually emerges with learning (indicating exploration to overcome clumsiness).
A different option would be to constrain the duration of the reaching movement. If the higher Vucm in children with DCD comes from exploring different movement speeds, then invoking a time constraint would restraint this exploration, hence, Vucm should not be higher in such conditions.
The constraints of the task and their effect on variability in coordination patterns seems to be an important aspect to consider in future studies. For example, we found similar levels of accuracy for the TD and the DCD group and the accuracy level was rather high (CE errors did not exceed 8mm), which shows that the target was reached in most trials. Our target was rather large; it could be that a smaller target would result in different accuracy levels and thereby in a different variability structure. The results of the regression coefficient analysis revealed interesting features about the relation between MT and accuracy. The relation between MT and accuracy was similar between
CHAPTER 4 groups, but children with DCD performed at a general lower speed. Different speed constraints
might have different effects on the relation between MT and accuracy. Considering the study by Utley et al. [71], different speed constraints might also result in different coordination patterns.
They examined an interception task and results showed that children with DCD coupled their DoF in a rigid way when catching a ball.
Another step that should be taken in future studies is to build and use theories that expand our knowledge of the role and consequences of motor variability. We propose that the Dynamic Sys- tems (DS) approach can fulfill this mission [72], and has recently increased in popularity in research about DCD [10,11,28]. The DS approach states that development is the result of self-organiza- tion processes emerging from interactions between organism, task and environment [22,72–74].
Within this view, variability and structure therein characterizes the system. For instance, in contrast to other approaches, DS approach also considers low variability as an indicator of devel- opmental problems [15].
Clinical implications
Our results may also have clinical implications. Assuming that motor variability reflects noise implies rehabilitation practices to reduce motor variability and to increase performance consis- tency. Our results showed that high variability in DCD is not necessarily negative as the high variability in joint angles did not affect reaching performance. Irrespective of what the role of variability might be, these results imply that not all motor variability should be reduced during rehabilitation. How variability should be incorporated in rehabilitation has to be investigated in future studies. Studies in adults have started to investigate exploration of redundancy during interventions and how this effects performance [36,75]. Future studies should also identify how training perspectives that advocate variable practice [c.f. 76,77] link with DCD.
Conclusion
Our results showed that the structure in variability in coordination patterns was different in chil- dren with DCD compared to TD children. Variability not affecting performance was increased in children with DCD whereas the variability affecting performance was similar in children with DCD and TD children. Even though the results obtained do not tell us what the high variability reflects, our results pave the way for a novel view on motor variability in children with DCD. That is, high motor variability in DCD is not necessarily negative. Our data thereby support the current DCD literature, which recently proposed that variability might be more than noise [11,29]. In addition, our findings demonstrate how the UCM method can help to unravel the structure of variability in populations at risk. Altogether, broadening the concept of motor variability in DCD may offer a new understanding into possible underlying coordination mechanisms of DCD.
Acknowledgments
We thankfully acknowledge the children and parents who participated. We thank Inge Tuitert, Iris Erpelinck and the technical support of Human Movement Sciences Groningen for assistance in data collections. We also thank Justine Hoch and Joanne Smith for their feedback on the man- uscript.
CHAPTER 4
References
1. American Psychiatric Association. DSM 5 diagnostic and statistical manual of mental disorders. Washington, DC: APA; 2013.
2. Smits-Engelsman BCM, Westenberg Y, Duysens J. Children with developmental coordination disorder are equally able to generate force but show more variability than typically develop- ing children. Hum Mov Sci. 2008;27: 296–309.
doi:10.1016/j.humov.2008.02.005
3. Roche R, Viswanathan P, Clark JE, Whitall J. Children with developmental coordina- tion disorder (DCD) can adapt to perceptible and subliminal rhythm changes but are more variable. Hum Mov Sci. 2016;50: 19–29.
doi:10.1016/j.humov.2016.09.003
4. Bo J, Bastian AJ, Kagerer FA, Contreras-Vidal JL, Clark JE. Temporal variability in con- tinuous versus discontinuous drawing for children with Developmental Coordination Disorder. Neurosci Lett. 2008;431: 215–220.
doi:10.1016/j.neulet.2007.11.040
5. Volman MJM, Geuze RH. Relative phase sta- bility of bimanual and visuomanual rhythmic coordination patterns in children with a Devel- opmental Coordination Disorder. Hum Mov Sci. 1998;17: 541–572. doi:10.1016/S0167- 9457(98)00013-X
6. Geuze RH, Kalverboer AF. Inconsistency and adaptation in timing of clumsy children. Hum Mov Sci. 1987;13: 421–432.
7. Williams HG, Woollacott MH, Ivry R. Timing and motor control in clumsy children. J Mot Behav. 1992;24: 165–172.
8. Mackenzie SJ, Getchell N, Deutsch K, Wilms- Floet A, Clark JE, Whitall J. Multi-limb coordination and rhythmic variability under varying sensory availability conditions in children with DCD. Hum Mov Sci. 2008;27:
256–269. doi:10.1016/j.humov.2008.02.010 9. Wilson PH, Caeyenberghs K, Dewey D,
Smits-Engelsman B, Steenbergen B. Hybrid is not a dirty word: Commentary on Wade and Kazeck (2017). Hum Mov Sci. 2017;57: 510–
515. doi:10.1016/j.humov.2017.02.004 10. Wilson PH, Smits-Engelsman BC, Caeyenber-
ghs K, Steenbergen B, Sugden DA, Clark JE, et al. Cognitive and neuroimaging findings in Developmental Coordination Disorder : New insights from a systematic review of recent research. Dev Med Child Neurol. 2017;59:
1117–1129. doi:10.1111/dmcn.13530 11. Wade MG, Kazeck M. Developmental coordi-
nation disorder and its cause: The road less travelled. Hum Mov Sci. 2017;57: 489–500.
doi:10.1016/j.humov.2016.08.004
12. Smits-Engelsman BCM, Wilson PH. Noise, variability, and motor performance in devel- opmental coordination disorder. Dev Med Child Neurol. 2013;55: 69–72. doi:10.1111/
dmcn.12311
13. Licari MK, Reynolds JE. Understanding Performance Variability in Developmental Coordination Disorder: What Does It All Mean?
Curr Dev Disord Reports. 2017;4: 53–59.
doi:10.1007/s40474-017-0112-3
14. Adolph KE, Cole WG, Vereijken B. Intraindi- vidual Variability in the Development of Motor Skills in Childhood. In: Diehl M, Hooker K, Sli- winsk M, editors. Handbook of Intra-individual Variability Across the Lifespan. New York:
Routledge/Taylor & Francis Group; 2015. pp.
59–83.
15. Hadders-Algra M. Variation and variability:
key words in human motor development.
Phys Ther. 2010;90: 1823–1837. doi:10.2522/
ptj.20100006
16. Thelen E, Corbetta D, Spencer JP. Develop- ment of reaching during the first year: role of movement speed. J Exp Psychol Hum Percept Perform. 1996;22: 1059–1076.
doi:10.1037/0096-1523.22.5.1059
17. Slifkin AB, Newell KM. Noise, information transmission, and force variability. J Exp Psy- chol Hum Percept Perform. 1999;25: 837–51.
18. Vereijken B. The complexity of childhood development: variability in perspective. Phys
Ther. 2010;90: 1850–1859. doi:10.2522/
ptj.20100019
19. Lee M, Farshchiansadegh A, Ranganathan R.
Children show limited movement repertoire when learning a novel motor skill. Dev Sci.
2017; e12614. doi:10.1111/desc.12614 20. Pacheco M, Newell K. Transfer as a function of
exploration and stabilization in original prac- tice. Hum Mov Sci. 2015;44: 258–269. doi:doi:
10.1016/j.humov.2015.09.009
21. Adolph K, Robinson SR. Motor development.
In: Lerner RM, Liben L, Muller U, editors.
Handbook of child psychology and develop- mental science: Vol 2: Cognitive processes. 7th ed. New York: Wiley; 2015. pp. 114–157.
22. Thelen E, Smith LB. A dynamic systems approach to the development of cognition and action. Cambridge, MA: MIT Press; 1994.
23. Wilmut K, Du W, Barnett AL. Gait patterns in children with Developmental Coordination Disorder. Exp Brain Res. 2016;234: 1747–
1755. doi:10.1007/s00221-016-4592-x 24. Rosengren KS, Deconinck FJA, DiBerardino
LA, Polk JD, Spencer-Smith J, De Clercq D, et al. Differences in gait complexity and vari- ability between children with and without Developmental Coordination Disorder. Gait Posture. 2009;29: 225–229. doi:10.1016/j.
gaitpost.2008.08.005
25. Wilmut K, Gentle J, Barnett AL. Gait sym- metry in individuals with and without Developmental Coordination Disorder. Res Dev Disabil. 2017;60: 107–114. doi:10.1016/j.
ridd.2016.11.016
26. Castra Ferracioli M, Yukiko Hiraga C, Pelegrini AM. Emergence and stability of interlimb coordination patterns in children with develop- mental coordination disorder. Res Dev Disabil.
2014;35: 348–356.
27. Henderson SE, Geuze RH. Ten Conferences on Developmental Coordination Disorder (DCD):
A Brief Commentary on 20 Years of Research.
Phys Occup Ther Pediatr. 2015;35: 97–102. doi :10.3109/01942638.2015.1008896
28. Whitall J, Clark JE. Developmental Coordi- nation Disorder from a Dynamic Systems
Perspective: What is on offer? Curr Dev Disord Reports. 2016;3: 94–96.
29. Sugden DA. Developmental coordination disor- der and its cause: The road less travelled. Hum Mov Sci. 2017;57: 501–504. doi:10.1016/j.
humov.2017.04.006
30. Ameratunga D, Johnston L, Burns Y. Goal-di- rected upper limb movements by children with and without DCD: a window into percep- tuo-motor dysfunction? Physiother Res Int.
2004;9: 1–12. doi:10.1002/pri.295
31. van der Meulen JH, Denier van der Gon JJ, Gielen CC, Gooskens RH, Willemse J. Visuo- motor performance of normal and clumsy children. I: Fast goal-directed arm-movements with and without visual feedback. Dev Med Child Neurol. 1991;33: 40–54.
32. Schoemaker MM, van der Wees M, Flapper B, Verheij-Jansen N, Scholten-Jaegers S, Geuze RH. Perceptual skills of children with devel- opmental coordination disorder. Hum Mov Sci. 2001;20: 111–133. doi:10.1016/S0167- 9457(01)00031-8
33. Biancotto M, Skabar A, Bulgheroni M, Carrozzi M, Zoia S. Neuromotor deficits in developmen- tal coordination disorder: Evidence from a reach-to-grasp task. Res Dev Disabil. 2011;32:
1293–1300. doi:10.1016/j.ridd.2011.02.007 34. Wilmut K, Byrne M, Barnett AL. Reaching to
throw compared to reaching to place: A com- parison across individuals with and without Developmental Coordination Disorder. Res Dev Disabil. 2013;34: 174–182. doi:10.1016/j.
ridd.2012.07.020
35. Smyth MM, Anderson HI, Churchill a C.
Visual information and the control of reaching in children: a comparison between children with and without developmental coordination disorder. J Mot Behav. 2001;33: 306–320.
doi:10.1080/00222890109601916
36. Tuitert I, Bootsma RJ, Schoemaker MM, Otten E, Mouton LJ, Bongers RM. Does practicing a wide range of joint angle configurations lead to higher flexibility in a manual obstacle-avoid- ance target-pointing task? PLoS One. 2017;12:
e0181041.
CHAPTER 4 37. Van Der Steen MC, Bongers RM. Joint angle
variability and co-variation in a reaching with a rod task. Exp Brain Res. 2011;208: 411–422.
doi:10.1007/s00221-010-2493-y
38. Bernstein N. The Coordination and Regulation of Movements. Oxford: Pergamon Press; 1967.
39. Latash M. There is no motor redundancy in human movements. There is motor abundance.
Motor Control. 2000;4: 259–260. doi:10.1123/
mcj.4.3.259
40. Gelfand IM, Latash ML. On the problem of adequate language in motor control. Motor Control. 1998;2: 306–313. doi:10.1123/
mcj.2.4.306
41. Latash ML. The bliss (not the problem) of motor abundance (not redundancy). Exp Brain Res. 2012;217: 1–5. doi:10.1007/s00221-012- 3000-4
42. Scholz JP, Schöner G. The uncontrolled mani- fold concept: Identifying control variables for a functional task. Exp Brain Res. 1999;126:
289–306. doi:10.1007/s002210050738 43. Latash ML, Scholz JP, Schöner G. Toward a
new theory of motor synergies. Motor Control.
2007;11: 276–308. doi:17715460
44. Domkin D, Laczko J, Djupsjöbacka M, Jaric S, Latash ML. Joint angle variability in 3D biman- ual pointing: Uncontrolled manifold analysis.
Exp Brain Res. 2005;163: 44–57. doi:10.1007/
s00221-004-2137-1
45. Scholz JP, Schöner G, Latash ML. Identi- fying the control structure of multijoint coordination during pistol shooting. Exp Brain Res. 2000;135: 382–404. doi:10.1007/
s002210000540
46. Tseng Y, Scholz JP, Schöner G. Goal-equivalent joint coodination in pointing:affect of vision and arm dominance. Motor Control. 2002;6:
183–207.
47. Latash ML, Scholz JP, Schöner G. Motor Con- trol Strategies Revealed in the Structure of Motor Variability. Exerc Sport Sci Rev. 2002;30:
26–31. doi:10.1097/00003677-200201000- 00006
48. Latash ML, Scholz JP, Danion F, Schöner G.
Structure of motor variability in margin-
ally redundant multifinger force production tasks. Exp Brain Res. 2001;141: 153–165.
doi:10.1007/s002210100861
49. Deutsch KM, Newell KM. Age Differences in Noise and Variability of Isometric Force Production. J Exp Child Psychol. 2001;80:
392–408. doi:10.1006/jecp.2001.2642 50. Wu J, McKay S, Angulo-Barroso R. Center of
mass control and multi-segment coordination in children during quiet stance. Exp Brain Res.
2009;196: 329–339. doi:10.1007/s00221-009- 1852-z
51. Yang JF, Scholz JP, Latash ML. The role of kine- matic redundancy in adaptation of reaching.
Exp Brain Res. 2007;176: 54–69. doi:10.1007/
s00221-006-0602-8
52. Greve C, Hortobàgyi T, Bongers RM. Physical Demand but Not Dexterity Is Associated with Motor Flexibility during Rapid Reaching in Healthy Young Adults. PLoS One. 2015;10:
e0127017. doi:10.1371/journal.pone.0127017 53. Black DP, Smith BA, Wu J, Ulrich BD. Uncon- trolled manifold analysis of segmental angle variability during walking: preadolescents with and without Down syndrome. Exp brain Res.
2007;183: 511–21. doi:10.1007/s00221-007- 1066-1
54. de Freitas SMSF, Scholz JP, Stehman AJ. Effect of motor planning on use of motor abundance.
Neurosci Lett. 2007;417: 66–71. doi:10.1016/j.
neulet.2007.02.037
55. Dutta GG, Freitas SMSF, Scholz JP. Diminished joint coordination with aging leads to more variable hand paths. Hum Mov Sci. 2013;32:
768–784. doi:10.1016/j.humov.2013.04.002 56. Golenia L, Schoemaker MM, Mouton LJ,
Bongers RM. Individual Differences in Learn- ing a Novel Discrete Motor Task. PLoS One.
2014;9: e112806. doi:10.1371/journal.
pone.0112806
57. King AC, Ranganathan R, Newell KM. Indi- vidual differences in the exploration of a redundant space-time motor task. Neuro- sci Lett. 2012;529: 144–9. doi:10.1016/j.
neulet.2012.08.014
58. Henderson S, Sugden D, Barnett A. Movement Assessment Battery for Children. 2nd ed.
Oxford: Pearson; 2007.
59. van Andel CJ, Wolterbeek N, Doorenbosch CAM, Veeger DHEJ, Harlaar J. Complete 3D kinematics of upper extremity functional tasks.
Gait Posture. 2008;27: 120–7. doi:10.1016/j.
gaitpost.2007.03.002
60. Verrel J. A formal and data-based comparison of measures of motor-equivalent covariation.
J Neurosci Methods. 2011;200: 199–206.
doi:10.1016/j.jneumeth.2011.04.006 61. Domkin D, Laczko J, Jaric S, Johansson H,
Latash ML. Structure of joint variability in bimanual pointing tasks. Exp Brain Res.
2002;143: 11–23. doi:10.1007/s00221-001- 0944-1
62. Wu G, Van Der Helm FCT, Veeger HEJ, Makhsous M, Van Roy P, Anglin C, et al. ISB recommendation on definitions of joint coordi- nate systems of various joints for the reporting of human joint motion - Part II: Shoulder, elbow, wrist and hand. J Biomech. 2005;38:
981–992. doi:10.1016/j.jbiomech.2004.05.042 63. de Freitas SMSF, Scholz JP. A comparison
of methods for identifying the Jacobian for uncontrolled manifold variance analysis. J Biomech. 2010;43: 775–777. doi:10.1016/j.
jbiomech.2009.10.033
64. Krishnamoorthy V, Scholz JP, Latash ML. The use of flexible arm muscle synergies to perform an isometric stabilzation task. Clin Neuro- physiol. 2007;118: 525–537. doi:10.1016/j.
pestbp.2011.02.012.
65. Tuitert I, Valk TA, Otten E, Golenia L, Bongers RM. Comparing different methods to create a linear model for the uncontrolled manifold method analysis. Motor Control. 2018;
66. Verrel J. Distributional properties and vari- ance-stabilizing transformations for measures of uncontrolled manifold effects. J Neurosci Methods. Elsevier B.V.; 2010;191: 166–170.
doi:10.1016/j.jneumeth.2010.06.016 67. Riley MA, Turvey MT. Variability and Determin-
ism iln Motor Behavior. J Mot Behav. 2002;34:
37–41. doi:10.1080/00222890209601934
68. Ranganathan R. Reorganization of finger coor- dination patterns through motor exploration in individuals after stroke. J Neuroeng Rehabil.
2017;14: 90. doi:10.1186/s12984-017-0300-8 69. van der Linde BW, van Netten JJ, Otten B,
Postema K, Geuze RH, Schoemaker MM.
Activities of Daily Living in Children With Developmental Coordination Disorder: Perfor- mance, Learning, and Participation. Phys Ther.
2015;95: 1496–1506.
70. Abe MO, Sternad D. Directionality in distribu- tion and temporal structure of variability in skill acquisition. Front Hum Neurosci. 2013;
doi:10.3389/fnhum.2013.00225
71. Utley A, Steenbergen B, Astill SL. Ball catching in children with developmental coordination disorder: control of degrees of freedom. Dev Med Child Neurol. 2007;49: 34–38.
doi:10.1111/j.1469-8749.2007.0096a.x 72. Turvey MT, Fitzpatrick P. Commentary: Devel-
opment of perception–action systems and general principles of pattern formation. Child Dev. 1993;64: 1175–1190.
73. Newell KM. Constraints on the development of coordination. In: Wade MG, Whitin HTA, edi- tors. Motor development in children: Aspects of coordination and control. Dordrecht:
Nijhoff; 1986. pp. 341–360.
74. Molenaar PCM, Lerner RM, Newell KM. Hand- book of developmental systems theory and methodology. New York: Guilford Press; 2014.
75. Ranganathan R, Newell KM. Emergent flexibil- ity in motor learning. Exp Brain Res. 2010;202:
755–764. doi:10.1007/s00221-010-2177-7 76. Frank TD, Michelbrink M, Beckmann H, Schöll-
horn WI. A quantitative dynamical systems approach to differential learning: Self-organiza- tion principle and order parameter equations.
Biol Cybern. 2008;98: 19–31. doi:10.1007/
s00422-007-0193-x
77. Shea JB, Morgan RL. Contextual interference effects on the acquisition, retention, and trans- fer of a motor skill. J Exp Psychol Hum Learn Mem. 1979;5: 179–187. doi:10.1037/0278- 7393.5.2.179
