Neuroscience 286 (2015) 13–26

CHILDREN WITH DEVELOPMENTAL COORDINATION DISORDER AREDEFICIENT IN A VISUO-MANUAL TRACKING TASK REQUIRINGPREDICTIVE CONTROL

G. D. FERGUSON, a,b* J. DUYSENS b ANDB. C. M. SMITS-ENGELSMAN a,b

aUniversity of Cape Town, Faculty of Health Sciences, Department

of Health and Rehabilitation Sciences, Suite F45: Old Main

Building, Groote Schuur Hospital, Main Road, Observatory

7925, Cape Town, 8000, South Africa

bKatholieke Universiteit Leuven, Faculty of Kinesiology and

Rehabilitation Sciences, Department of Kinesiology,

Movement Control and Neuroplasticity Research Group,

Tervuursevest 101, Postbox 1501, B-3001 Heverlee, Belgium

Abstract—The aim of this study was to examine how feed-

back, or its absence, affects children with Developmental

Coordination Disorder (DCD) during a visuo-manual track-

ing task. This cross-sectional study included 40 children

with DCD and 40 typically developing (TD) children between

6 and 10 years old. Participants were required to track a tar-

get moving along a circular path presented on a monitor by

moving an electronic pen on a digitizing tablet. The task was

performed under two visibility conditions (target visible

throughout the trajectory and target intermittently occluded)

and at two different target velocities (30� and 60� per sec-

ond). Variables reflecting tracking success and tracking

behavior within the target were compared between groups.

Results showed that children with DCD were less proficient

in tracking a moving target than TD children. Their perfor-

mance deteriorated even more when the target was

occluded and when the target speed increased. The mean

tracking speed of the DCD group exceeded the speed at

which the target rotated which was attributed to accelera-

tions and decelerations made during tracking. This sug-

gests that children with DCD have significant difficulties in

visuo-manual tracking especially when visual feedback is

reduced. It appears that their impaired ability to predict

together with impairments in fine-tuning arm movements

may be responsible for poor performance in the intermit-

tently occluded visuo-manual tracking task.

� 2014 IBRO. Published by Elsevier Ltd. All rights reserved.

http://dx.doi.org/10.1016/j.neuroscience.2014.11.0320306-4522/� 2014 IBRO. Published by Elsevier Ltd. All rights reserved.

*Correspondence to: G. D. Ferguson, University of Cape Town,Faculty of Health Sciences, Department of Health and RehabilitationSciences, Suite F45: Old Main Building, Groote Schuur Hospital,Main Road, Observatory 7925, Cape Town 8000, South Africa. Tel:+27-214066045, +27-82974394 (mob); fax: +27-214066323.

E-mail address: gillian.ferguson@uct.ac.za (G. D. Ferguson).Abbreviations: ADD, Attention Deficit Disorder; ADHD, Attention Deficitwith Hyperactivity Disorder; APAs, anticipatory postural adjustments;DSST, Double Step Saccade Task; DCD, Developmental CoordinationDisorder; fMRI, functional Magnetic Resonance Imaging; IMD, internalmodeling deficit; LD, Learning Disorder; MABC-2, MovementAssessment Battery for Children-second edition; PPC, posteriorparietal cortex; TD, typically developing; TV, tracking variability.

13

Key words: developmental coordination disorder, visuo-man-

ual tracking, manual pursuit, feed forward control, predictive

control.

INTRODUCTION

Children with Developmental Coordination Disorder

(DCD) are reported to have difficulties executing

everyday manual tasks that require visual information

(visuo-manual coordination) such as writing, tying

shoelaces, aiming and reaching for objects or catching a

ball (Wang et al., 2009; Magalhaes et al., 2011;

Ferguson et al., 2014). In comparison to their peers, the

motor performance of children with DCD during these

tasks lacks fluency and precision. Their inaccurate and

error-prone movements have been attributed to variability

in controlling the temporal and spatial aspects of move-

ment and to inconsistencies in force production and regu-

lation, all pointing toward an underlying deficit in motor

control (Piek and Skinner, 1999; Smits-Engelsman

et al., 2003a; Van Waelvelde et al., 2006).

Motor control includes the use of a combination of

both feed-forward and feedback control strategies

(Wolpert and Miall, 1996). The feedback system is reliant

on adequate sensory information, error detection and

integration (Scott, 2012). Impairments in one or more of

these processes may be regarded as a source of poor

motor performance. Feed-forward motor control is

defined as the ability to estimate the temporal and spatial

requirements of a motor task and predict the sensory con-

sequences of the impending action (Wolpert and Miall,

1996). It is hypothesized that feed-forward control is

sub-served by internal forward models (Blakemore and

Sirigu, 2003; Kawato et al., 2003). Limb state-estimation

(i.e., estimating the location and position of a limb at

any given moment), target-state estimation (i.e., estimat-

ing the coordinates of external objects or targets) and

information related to the relationship between limb and

object/target are all inputs into an internal model (Scott,

2012). This information is then used to organize and mon-

itor movements. Information via internal models is avail-

able much more rapidly than afferent feedback signals,

resulting in greater efficiency of the motor system

(Wolpert et al., 1998).

To understand the extent to which deficits in motor

control processes contribute to the poor motor

14 G. D. Ferguson et al. / Neuroscience 286 (2015) 13–26

performance that is seen in children with DCD, various

studies have focused on the interaction between feed-

forward and feedback control processes as the basis for

efficient motor control (Wilson et al., 2004, 2013;

Gabbard, 2009). In terms of the underlying motor control

deficits associated with DCD, evidence from an informa-

tion processing paradigm has implicated impairments in

the use of feedback systems such as motion detection,

tactile perception and visuo-spatial processing (Wilson

et al., 2013). However, recent research, using computa-

tional neuroscience paradigms, proposes that impaired

motor control in children with DCD may also be due to

deficits in generating and using internal forward models,

which is referred to as an internal modeling deficit (IMD)

(Wilson et al., 2004; Williams et al., 2006; Hyde and

Wilson, 2011).

Evidence suggests that children with DCD are more

reliant on online visual feedback information and that

they take longer to utilize visual information for the

generation and control of actions than their typically

developing (TD) peers (Wilmut et al., 2006). Successful

use of predictive control strategies may also be con-

founded by deficits in kinesthetic processing ability and

by problems with error detection and correction which

has been reported in children with DCD (Wilson and

McKenzie, 1998; Kagerer et al., 2006). Hence, there is

a need to re-examine these questions preferably by using

a paradigm that allows various degrees of feedback and

feed-forward control. Since DCD involves deficits in man-

ual dexterity (Smits-Engelsman et al., 2001, 2003b,

2008), the choice was made to study specific visuo-man-

ual tracking or manual pursuit tasks.

When tracking a slow-moving target that follows a

predictable trajectory, feedback and feed-forward

strategies may be used to prevent and correct errors.

However, when target motion is faster, the use of

feedback is more difficult and one has to rely more on

feed-forward control (Wolpert et al., 1998). Making appro-

priate adjustments in a tracking task entails the detection

of visual information about the current position of the tar-

get and making estimates of that position in the near

future (referred to as prospective control). Prospective

control is an essential component of effective motor con-

trol and has been reported to be impaired in children with

DCD (Debrabant et al., 2013). In cases where the visual

stimulus is removed or transiently occluded, then the

most important source of feedback is lacking, thereby

forcing the subjects further in the direction of enlisting

feed-forward control.

Pursuit tasks with temporary occlusion of the target

have been used primarily in oculomotor studies

(Newsome et al., 1988; Mrotek and Soechting, 2007;

Orban de Xivry et al., 2008). For example, Newsome

et al. (1988) introduced an ocular pursuit task with visual

occlusion in monkeys to demonstrate that certain cells in

the middle temporal area continue firing after occlusion.

They thus demonstrated that these cells used extraretinal

input, most likely derived from an efference copy, pro-

duced by an internal model. Two recent studies have

investigated ocular pursuit in children with DCD and found

a reduced gain with regard to horizontal pursuit in a young

DCD group (n= 8, 5–7 years) (Langaas et al., 1998). In a

larger study with older children (n= 27, 8–12 years), it

was shown that the vertical pursuit, not the horizontal

one, is significantly impaired in children with DCD

(Robert et al., 2014).

While the tasks used thus far in studies of pursuit in

DCD have focused on ocular pursuit, the addition of

manual pursuit tracking makes this paradigm directly

related to functional visuo-manual skills used in daily

life. However, studies using manual pursuit tasks in

children are rare and they have focused either on TD

children (van Roon et al., 2008) or children with learning

disabilities (van Roon et al., 2010) and not on children

with DCD. Recently, a functional Magnetic Resonance

Imaging (fMRI) study, which adopted a visually guided

tracking task, reported that children with DCD were signif-

icantly less accurate than control children when guiding a

cursor toward an easy to track, continuously moving, hor-

izontal target (Kashiwagi et al., 2009). More importantly,

differences in brain activation were shown, in the left pos-

terior parietal cortex (PPC) and left post-central gyrus.

These areas are known to cause visuo-motor deficits

affecting hand-eye coordination. Moreover, the PPC plays

an important role in generating mental representations of

movement (Sirigu et al., 1996).

In contrast to ocular pursuit, no study using a manual

pursuit task with temporary occlusion in children with DCD

was available at the time of our study. As it was expected

that DCD subjects would be poor in tracking, it was

hypothesized that their pursuit trace would be less

optimal than that of the TD controls. To allow these

differences to be revealed as clearly as possible, the

pursuit task was executed at two speeds and with

intermittent occlusion of the target. It was hypothesized

that tracking would show even more deficits in DCD

children under these more difficult conditions. The aim

of this study was thus to examine how feedback, or its

absence, affects visuo-manual tracking in children with

DCD. The specific objectives were to examine the effect

of visual information and target speed on (1) manual

tracking performance, (2) parameters of the trajectory

trace while in the target and (3) the differences between

DCD and TD on these variables.

EXPERIMENTAL PROCEDURES

Participants

In this cross-sectional study, children between the ages of

6 and 10 years old were recruited from two mainstream

primary schools situated in Cape Town, South Africa

using convenience sampling.

Children were identified as having DCD by using the

four criteria based on the Diagnostic and Statistical

Manual-4th Edition (American Psychiatric Association,

2000). Children were included in the DCD group if they

met all four criteria.

The Movement Assessment Battery for Children-2

(MABC-2) (Henderson et al., 2007) was used to evaluate

the motor performance of the participants (criterion A).

The MABC-2 is considered to be a reliable and valid

measure for assessing motor performance in DCD

G. D. Ferguson et al. / Neuroscience 286 (2015) 13–26 15

(Wuang et al., 2012). Children whose motor performance

was at or below the 5th percentile on the MABC-2 (indic-

ative of poor motor coordination ability) were included.

A questionnaire was developed by the researchers

and administered to teachers and parents to determine

whether the child presented with any difficulty performing

motor tasks at school or at home (criterion B). The

questionnaire contained a single closed-ended question

in which teachers were asked to consider the motor

performance of each child in their class. Teachers could

select one of three options which stated that the child

under consideration (i) had a motor coordination

problem, (ii) possibly had a motor coordination problem

or (iii) did not have a motor coordination problem.

Teachers were advised to base their decision on their

observation of the performance of each child in the

classroom, on the playground and during physical

education lessons. These options were marked on a

class list and teachers were given the option of providing

a reason for their decision as an additional comment.

Parents were given an information leaflet regarding

DCD and were asked to indicate on a questionnaire

whether they believed their child had a motor

coordination problem or not. Parents were also given

the opportunity to explain why they believed their child

had a motor coordination problem. A positive report

from either the teacher or parent was used to establish

criterion B.

A reported diagnosis of cerebral palsy or other

significant neurological disorder (e.g. severe epilepsy,

acquired brain injury or spinal cord lesions) as reported

by a parent (criterion C) and failing a grade level at

school more than once (criterion D) were used as

exclusion criteria.

An age and gender-matched sample of TD children

was selected to form the comparison group. A ratio of

1:1 was used to select one TD child randomly for every

child identified with DCD. Children were identified as TD

if: (i) their motor function was believed to be within

normal range according to their teacher and/or the

parent(s), (ii) they scored above the 16th percentile on

the MABC-2, (iii) they had not failed a grade level at

school and (iv) they displayed no significant neurological

pathology as reported by a parent.

The effective sample thus consisted of 40 children with

DCD (23 boys, 17 girls) with a mean age of

8.03 ± 1.25 years and 40 TD children (24 boys, 16

girls), with a mean age of 8.20 ± 1.36 years. All children

had normal, or corrected to normal, vision. Handedness

was established by identifying the child’s preferred

Table 1. Comparison of the motor performance scores of DCD (n= 40) and

MABC-2 TD group

Total standard score 12.05 ± 2.46

Manual dexterity 10.88 ± 2.65

Aiming and catching 11.10 ± 2.41

Balance 11.75 ± 2.35

The mean motor performance scores with standard deviation (± SD) of the DCD (n= 40

(MABC-2) with the results of the statistical tests for differences (t-test). Note: The mean Tot

and Balance) were all significantly lower in the DCD group.* Degrees of freedom adjusted for unequal variance.

writing hand during a brief handwriting task (e.g. writing

name or drawing a picture) as recommended in the

MABC-2 instructors’ manual (Henderson et al., 2007).

Most children were right-handed (n= 75), with two left-

handed children in the DCD group and three in the TD

group. Comparisons between DCD and TD groups were

tested and no significant differences were found for age

(t= 0.60, df= 78, p= 0.55), gender (v2 = 0.05,

p= 0.82) and handedness (v2 = 0.21, p= 0.64). The

mean motor performance scores on the MABC-2 are

reflected in Table 1.

Written informed consent and assent was obtained

from parents and children, respectively. Authorization for

conducting research in the two schools was granted by

the Western Cape Education Department. Ethical

approval was granted by the University of Cape Town,

Health Sciences Faculty, Human Research Ethics

Committee (HREC: 218/2012). The study was performed

in accordance with the ethical standards outlined in the

Declaration of Helsinki (World Medical, 2013).

Apparatus

Movements were measured by means of a digitizing

tablet (Intuos UD-1218-R, Kiko Software, Doetinchem;

The Netherlands) and an electronic pen of normal

appearance and weight. The pen had a white plastic tip

that left no trace. The 2D position of the pen on the

surface of the digitizer was sampled at a frequency of

206 Hz and with an accuracy of 0.1 mm.

Task

The task used in this study was similar to that used by van

Roon et al. (2008, 2010) except that in the current study

tracking of the target was conducted under two visibility

conditions and at two different speeds. Participants were

required to manually track a red circular target (diame-

ter = 2.70 cm) presented on a 48.3-cm flat screen, verti-

cally mounted computer monitor positioned at

approximately 40 cm from and level with the participant’s

eye. The target rotated clockwise along a circular path

(diameter = 10 cm) at constant speeds. Using the pen

as a cursor, tracking was executed on a sheet of paper

(A3) placed on a horizontally mounted digitizing tablet

positioned in front of the participant (Fig. 1). During the

task, the participants were required, using their preferred

hand, to keep the cursor, represented by a small yellow

dot (diameter = 0.53 cm), inside the target. If the child

TD (n = 40) groups

DCD group Statistics

3.25 ± 1.45 t= 19.51, df= 1.78, p< 0.001

4.23 ± 2.11 t= 12.42, df= 1.78, p< 0.001

6.48 ± 3.40 t= 7.03, df= 1.70*, p< 0.001

4.58 ± 2.37 t= 13.58, df= 1.78, p< 0.001

) and TD (n= 40) groups using the Movement Assessment Battery for Children

al Standard Score and component scores (Manual Dexterity, Aiming and Catching

Fig. 1. Representation of the manual tracking task. The image shows the target (red circle) with the cursor (yellow dot) presented on the monitor.

The digitizer tablet, with the sheet of paper on the surface, is positioned in front of the screen. (For interpretation of the references to color in this

figure legend, the reader is referred to the web version of this article.)

16 G. D. Ferguson et al. / Neuroscience 286 (2015) 13–26

became inattentive or distracted during a trial, the task

was repeated at the end of the experiment.

Procedure

The tasks were conducted under two target visibility

conditions, viz. visible and occluded. Each of these was

conducted at two different speeds (slow and fast),

resulting in a mixed 2 by 2 design. Speed was

randomized but all children started with the visible task

(easiest) followed by the occluded task. This approach

was adopted to keep the participants motivated.

In the visible condition, the target was evident

throughout the duration of the task (60 s). The target

rotated at a constant speed (either fast or slow) and

participants were instructed by the examiner to stay

within the center of the target at all times. In the

occluded condition, the target disappeared intermittently

at random intervals (i.e., every 2–4 s) and then

continued invisibly along the path (duration 1–1.5 s).

While the target was occluded, participants were told to

continue pursuing where they thought the target would

be and to keep the yellow dot in the center of the

(invisible) target so that when it reappeared the cursor

would still be inside the target.

The tasks were conducted at two speeds. In the slow

condition, the red target moved along the circular path at

30�/s (corresponding to 2.62 cm/s) and in the fast

condition, the target moved at 60�/s (corresponding to

5.24 cm/s). These two fixed speeds were based on

findings determined in an earlier study (van Roon et al.,

2008) and our own pilot study to ensure that both groups

of children would be able to perform the task.

All participants were allowed one practice trial

(duration 30 s) for each combination of speed (2) and

visibility (2) condition to get acquainted with the task

procedure. This was followed by two trials lasting one

minute in each of the four combinations of the speed

and visibility conditions leading to 8 min of data

recording per child. We averaged the dependent

measures for each participant, over the two trials

performed in a given condition. The whole experiment

took approximately 20 min. Performance feedback was

given after each trial (i.e., the average distance from the

center of the target to the center of the cursor).

Participants were allowed to rest between trials.

Dependent variables

Data obtained from the digitizer were analyzed using a

custom-made script in OASIS software and variables

were defined in line with previous studies using the

same arrangement (Smits-Engelsman et al., 2002; van

Roon et al., 2008, 2010; Caeyenberghs et al., 2009a).

The 2D positional data of the pen were low-pass filtered

using a zero phase lag, 2nd-order Butterworth filter with

a cut-off frequency of 10 Hz and then differentiated to cal-

culate movement velocity (cm/s).

The trajectory was segmented into parts where the

cursor was within and outside the boundaries of the

target. Movement velocities were calculated as the first

derivative of the pen position. Using these data and the

known coordinates of the target trajectory; we computed

and then analyzed both temporal and spatial measures

of performance of the two groups.

Tracking performance. Tracking performance

variables reflect the extent to which a participant

succeeds in keeping the cursor within the target. First,

we calculated the time on target (s) because this

Group

D)

Statistics

DCD

(M±

SD)

TD

(M±

SD)

Statistics

12.16

F(1,78)=

222.13,

p<

0.001,g2

=0.74

39.10±

12.17

49.85±

8.33

F(1,78)=

37.15,

p<

0.001,g2

=0.32

12.32

F(1,78)=

299.10,

p<

0.001,g2

=0.79

27.19±

11.99

16.47±

11.96

F(1,78)=

41.86,

p<

0.001,g2

=0.35

0.38

F(1,78)=

2.97,

p=

0.09,g2

=0.04

0.79±

0.51

0.69±

0.40

F(1,78)=

3.59,

p=

0.06,g2

=0.04

upsandtheresultsofthestatisticaltests

(ANOVA)reflectingthedifferences.s=

seconds,cm

=centimeter,

G. D. Ferguson et al. / Neuroscience 286 (2015) 13–26 17

provides a single unambiguous measure of performance

that relates directly to the explicitly stated goal of the

task. We also examined two spatial kinematic

performance measures, which reflect positional error:

the frequency of being out-of-target (Number of Times

out) and the time required to re-position the cursor in

the target (Time to Reacquire).

Within-target kinematics. For the second part of the

analysis, we analyzed the segments in which the cursor

was in the target, to separate the pursuit part of the task

from the fast ‘‘saccade-like’’ movements used to re-enter

the target. As a measure of tracking behavior, the

tracking variability (TV) of the pen position within the

target per task was calculated in terms of the variability in

distance to the middle of the target (cm). Hence, if a child

stayed on the same spot relative to the moving target, the

TV would be zero. If the cursor stayed within the target

but moved around from the front to the back and left to

right (or vice versa) inside the target, the TVwould be large.

Next, the mean velocity (cm/s) of the cursor while

inside the target was calculated. The gain (ratio of

cursor velocity to target velocity) was calculated for the

periods during which the cursor was inside the target.

The out-of-target segments were excluded, because

catch up movements outside the target would inflate the

gain. A gain value equal to one suggests perfect unity

between the cursor velocity and target velocity and is

considered a measure of efficient tracking.

edto

overalltrackingperform

ance

tvisibility

Targetspeed

e SD)

Occluded

(M±

SD)

Statistics

Slow

(M±

SD)

Fast

(M±

S

±11.37

42.39±

11.72

F(1,78)=

69.53,

p<

0.001,g2

=0.47

48.94±

9.36

40.02±

±15.27

21.25±

10.51

F(1,78)=

2.47,

p=

0.12,g2

=0.03

15.11±

10.11

28.55±

±0.55

0.80±

0.34

F(1,78)=

7.95,

p=

0.006,g2

=0.09

0.77±

0.52

0.71±

ancevariableswithstandard

deviation(±

SD)oftheDCD

(n=

40)andTD

(n=

40)gro

Data analysis

Statistical analysis was conducted using SPSS 20 (IBM,

2011). Levene’s test was used to assess the equality of

variances between groups. Independent t-test, and Pear-

son’s Chi squared tests were used to examine the differ-

ences between groups on descriptive variables (i.e.,

MABC -2 performance scores, age, gender, and handed-

ness). If the Levene’s test indicated unequal variance,

degrees of freedom were adjusted.

Of the total number of experimental trials planned

(n= 640), 28 were missing. Since the children

performed two trials per condition and because we

averaged performance over the two trials in a given

condition per participant, no data were missing for the

statistical analysis.

Repeated measures ANOVAs were used to compare

two groups (i.e., TD or DCD) using a factorial design with

two visibility conditions (visible and occluded) and two

speeds (30�/s and 60�/s). Group allocation was also

used as the between subject factor. All statistical tests

were completed with alpha set at 0.05. Only significant

interactions are reported.

Table

2.Main

effects

relat

Variable

Targe

Visibl

(M±

Tim

ein

target

(s)

46.56

Tim

esoutof

target(n)

22.40

Tim

eto

reacquire

target(s)

0.68

Themean(M

)trackingperform

n=

number.

RESULTS

Tracking performance

The means (SD) and statistics of main effects related to

tracking performance are summarized in Table 2.

Means (SD) of the two and three-way interactions are

reported per variable, if significant.

18 G. D. Ferguson et al. / Neuroscience 286 (2015) 13–26

Time in target. A 2 � 2 � 2 ANOVA (group �visibility � speed) showed a significant interaction of

time in target between group and visibility, F(1, 78) =

6.99, p= 0.01, g2 = 0.08, (Fig. 2A) and an interaction

of group and speed (Fig. 2B), F(1, 78) = 6.15, p=

0.02, g2 = 0.07.

Overall, children with DCD spent a shorter time in

the target (39.10 s ± 12.17) than the TD children

(49.85 s ± 8.33). Moreover, both groups performed

better when the target was visible (46.56 s ± 11.37)

compared to when it was occluded (42.39 s ± 11.72).

Similarly, groups performed better overall when the

target rotated at a slow speed (48.94 s ± 9.36)

compared to a fast speed (40.02 s ± 12.16).

The difference between groups increased when the

target was occluded, confirmed by a significant

interaction (group � visibility). A larger difference was

detected between TD (48.43 s ± 7.83) and DCD

(36.36 s ± 11.87) groups when the target was occluded

compared to a smaller difference when it was visible

(DCD= 41.85 s ± 11.87, TD = 51.27 s ± 8.60).

The difference between groups also increased when

the target was moving faster (group � speed) F(1,78) = 6.15, p= 0.02, g2 = 0.07 (Fig. 2B). A difference

of 12.23 s was detected between groups when the target

was moving faster (DCD= 33.90 s ± 11.93,

TD = 46.13 s ± 8.89) compared to a difference of 9.27 s

when the target moved slowly (DCD= 44.30 s ± 10.01,

TD = 53.57 s ± 5.70).

Number of times out of target. The DCD group made

significantly more excursions out of the target

(27.19 ± 11.99) compared to the TD group

(16.47 ± 11.96). In the fast tracking speed condition,

both groups drifted out of the target more often

(28.55 ± 12.32) than in the slow speed condition

(15.11 ± 10.11). Performance (number of excursions

out of the target) was not affected by visibility condition

(visible: 22.40 ± 15.27, occluded: 21.25 ± 10.51).

Fig. 2. Time in target: performance of groups in visibility (A) and speed (B)

between the two groups in the different visibility (A) and speed (B) conditions is

than the DCD group. Error bars represent the standard deviation.

Time to reacquire target. Overall, no differences

between groups were found regarding the time taken for

the cursor to return to the target but there were

interaction effects. Participants generally performed

better when the target was visible (0.68 s ± 0.55)

compared to when it was occluded (0.80 s ± 0.34). In

contrast, there was no difference overall, in performance

related to the speed at which the target rotated (slow:

0.77 s ± 0.52, fast: 0.71 s ± 0.38).

The interaction of group � visibility condition, F(1,76) = 5.73, p= 0.02, g2 = 0.07 and group � speed

condition were significant, F(1, 76) = 4.64, p= 0.03,

g2 = 0.06. As shown in Fig. 3A, the DCD group was

slower to reacquire the target in the occluded condition

(0.90 s ± 0.38) than the TD group (0.69 s ± 0.25). The

TD group was also faster (0.61 s ± 0.25) at re-entering

the target in the fast speed condition compared to the

DCD group (0.80 s ± 0.46) as shown in Fig. 3B.

In short, children with DCD showed impaired tracking

performance on all variables tested, they became even

less successful when the target moved faster and when

it was occluded.

Within-target kinematics

Means (SD) and statistics of the main effects related to

within-target kinematics are summarized in Table 3.

Means (SD) of the two and three-way interactions are

reported per variable, if significant.

Velocity. The groups showed significantly different

tracking velocities while the cursor was inside the target.

The DCD group demonstrated a higher velocity

(4.87 cm/s ± 1.91) than the TD group (4.24 cm/

s ± 1.45) as they tracked the target.

Overall, participants moved significantly faster when

tracking a target in the occluded condition (4.68 cm/

s ± 1.90) than in the visible condition (4.43 cm/

s ± 1.52). Moreover, participants showed increased

tracking velocity in the fast condition (5.84 cm/s ± 0.88)

conditions. The comparison in tracking performance (time in target)

shown. In both conditions, the TD group spent more time in the target

Fig. 3. Time to reacquire target: performance in visibility (A) and speed (B) conditions. The difference between groups in the visibility condition (A)

and speed condition (B) for the time to return to target. In the more difficult conditions (occluded and fast), the DCD group took a longer time to return

to target. Error bars represent the standard deviation.

G. D. Ferguson et al. / Neuroscience 286 (2015) 13–26 19

compared to the slower speed condition (3.28 cm/

s ± 1.38) as they attempted to remain inside the faster

moving target.

A two-way interaction was found between groups and

the different visibility conditions F(1, 78) = 5.16,

p= 0.03, g2 = 0.06. Moreover a significant interaction

for group � speed � visibility condition was found F(1,78) = 4.92, p= 0.03, g2 = 0.06.

Fig. 4 shows that the difference was greatest when

comparing the groups during the occluded condition

when the target was moving slowly (2.62 cm/s). The

DCD group was clearly faster (3.93 cm/s ± 2.27) in the

occluded condition when the target was moving slowly

in comparison to the TD group (2.85 cm/s ± 0.36).

The maximum velocity in the target showed the same

main effects (for group, visibility and speed) but no

interactions with group (see Table 3). The maximum

speed for the DCD group (13.54 cm/s ± 13.21) was

almost twice as high as for the TD group (7.07 cm/

s ± 5.21); F(1, 78) = 20.06, p< 0.001; g2 = 0.21). It is

evident from the difference in means, the large standard

deviations in the DCD group and the statistics reported

that the DCD group was more variable than the TD group.

Tracking variability. The TV was significantly different

between groups. The DCD group showed more

variability in their cursor trajectory within the target

(0.39 cm± 0.62) compared to the TD group

(0.25 cm± 0.09). No significant effect of visibility

condition was found, but groups responded differently to

increased target speed (group � speed: F(1,

78) = 5.63, p= 0.02, g2 = 0.07). The TD children

showed little difference in their regularity between slow

(0.23 ± 0.06) and fast (0.26 ± 0.10) moving targets.

Children with DCD showed a far more erratic trajectory

while in the target, as shown by almost double the

variability in distance to the middle of the target (TV) in

the slow condition (DCD= 0.43 cm± 0.74 and

TD= 0.23 cm± 0.06). Interestingly, the children with

DCD tended to move more regularly when the target

moved faster (0.36 cm± 048) in comparison to their

trajectory in the slow condition (0.43 cm± 0.74).

In Figs. 5 and 6, comparisons in spatial and temporal

variability are made evident on examination of the pursuit

traces in the visible and occluded conditions.

Gain. Overall, the mean speed of the DCD group while

in target was higher than the target as demonstrated by

their gain value (1.28 ± 0.52), while the gain of the TD

group was very close to unity with a lower spread

(1.09 ± 0.17).

An effect of visibility and speed condition was also

found, with groups showing a gain closer to one in the

visible (1.15 ± 0.28) compared to the occluded

condition (1.22 ± 0.49) as well as in the fast target

speed condition (1.12 ± 0.17) compared to the slow

condition (1.25 ± 0.53).

The interaction between group � visibility condition

was significant, F(1, 78) = 5.93, p= 0.017, g2 = 0.07

as was group � speed F(1, 78) = 8.68, p= 0.004,

g2 = 0.10). The difference in gain between groups was

greater in the occluded condition than in the visible

condition. Moreover, the difference between groups was

also greater in the slow condition than in the fast

condition.

In addition, a significant interaction between visibility,

group and speed was found, F(1, 78) = 6.05,

p= 0.016, g2 = 0.07). The highest gain was seen in

the DCD group in the occluded condition when the

target moved slowly.

In summary, children with DCD showed less efficient

behavior characterized by many directional changes at

velocities that were higher than the target.

DISCUSSION

The main findings of this study are that children with DCD

are less proficient in tracking a moving target and that

they show more variability with regard to position and

speed than TD children during a smooth manual pursuit

task. Poor manual pursuit among children with DCD is

Table

3.Main

effects

relatedto

overallkinematicperform

ancewhile

inthetarget

Variable

Targetvisibility

Targetspeed

Group

Visible

(M±

SD)

Occluded

(M±

SD)

Statistics

Slow

(M±

SD)

Fast

(M±

SD)

Statistics

DCD

(M±

SD)

TD

(M±

SD)

Statistics

Velocityin

Target

(cm/s)

4.43±

1.52

4.68±

1.90

F(1,78)=

9.85;

p<

0.002;g2

=0.11

3.28±

1.38

5.84±

0.88

F(1,78)=

937.24,

p<

0.001,g2

=0.92

4.87±

1.91

4.25±

1.45

F(1,78)=

14.86,

p<

0.0001;g2

=0.16

Maxim

um

Velocity

inTarget(cm/s)

9.70±

10.53

10.90±

10.54

F(1,78)=

6.40;

p=

0.013;g2

=0.08

7.93±

11.94

12.68±

8.30

F(1,78)=

32.52,

p<

0.001,g2

=0.29

13.54±

13.21

7.07±

5.21

F(1,78)=

20.06,

p<

0.001;g2

=0.21

Gain

inTarget

1.15±

0.28

1.22±

0.49

F(1,78)=

6.80;

p=

0.01;g2

=0.08

1.25±

0.53

1.12±

0.17

F(1,78)=

16.28,

p<

0.001,g2

=0.17

1.28±

0.52

1.09±

0.17

F(1,78)=

14.44,

p<

0.001;g2

=0.16

TrackingVariability

inTarget

2.36±

2.40

1.62±

1.67

F(1,78)=

2.65;

p=

0.11;g2

=0.03

0.33±

0.53

0.31±

0.35

F(1,78)=

1.04,

p=

0.31,g2

=0.01

0.39±

0.62

0.25±

0.09

F(1,78)=

5.78,

p=

0.02;g2

=0.07

Themean(M

)kinematicperform

ancevariableswithstandard

deviation(±

SD)andtheresultsofthestatisticaltests

(ANOVA)reflectingthedifferencesbetweenconditionsandthegroupsDCD

(n=

40)andTD

(n=

40).

20 G. D. Ferguson et al. / Neuroscience 286 (2015) 13–26

evidenced by four key findings in this study. Children with

DCD (i) spend a shorter time within the boundaries of the

target, (ii) drift out of the target more often, (iii) take a

longer time to reacquire the target when they are out,

and (iv) within-target behavior is characterized by a

highly variable trace. Some of these characteristics

were exacerbated when the target was occluded or

when the same task was performed at a higher speed.

Performance in visible condition

In the visible condition, both TD children and children with

DCD made fewer errors in tracking performance in their

pursuit of the target than in the occluded condition. This

was expected, as visual feedback plays a key role in

manual tracking and online control. Our hypothesis that

children with DCD would make more tracking errors

than controls was confirmed by examining their

performance in the visible condition. While tracking the

target, the DCD group made more changes in their

trajectory trace (Fig. 5) and more catch-up movements

leading to higher mean and maximum velocity

compared to the controls (Table 3).

The greater TV and greater number of target misses

noted in the DCD group can be partially explained by

two related phenomena: the impaired ability to grade

muscle forces accurately in the upper limb leading to

increased variability of the corrective movements, which

is commonly reported in DCD and the presence of noise

in the sensorimotor system (Smits-Engelsman et al.,

2001; Biancotto et al., 2011; Smits-Engelsman and

Wilson, 2013).

In the current study, we propose that greater TV in the

visible condition among the DCD group may be indicative

of an inability to fine-tune the force output of the smaller

muscles in the hand resulting in a higher incidence of

over- and undershooting of the target. This impaired

ability to calibrate force output appropriately could be

due to deficits in precise control between agonists and

antagonist muscle activation in children with DCD. The

high tracking speeds noted in the children with DCD

therefore could be a sign of a compensatory movement

strategy to gain more control and reduce errors.

Previous studies have described how children with DCD

use a dynamic stiffness strategy to dampen their

increased noise levels (Smits-Engelsman et al., 2001,

2008). This corroborates the finding in this study that chil-

dren with DCD showed less variance in their trajectory

when they had to move faster.

Variability in DCD

Unwanted variability in movement execution may be the

result of impaired fine motor control. Specifically, fine-

tuning of muscle contraction forces or a coordination

deficit causing impaired co-articulation results in

increased errors and movement trajectories that are

less well formed (van Galen et al., 1993). Movement var-

iability thus emerges, in the literature, as a source of error

that should be eliminated or reduced (Fitts, 1954; Harris

and Wolpert, 1998; van Beers, 2009).

Fig. 4. Mean tracking velocity of groups in different visibility and speed conditions. Children with DCD move faster than TD children in all conditions,

however the greatest difference between TD and DCD occurs in the occluded condition when the target is moving slowly. Error bars represent the

standard deviation.

Fig. 5. Sample trace of a TD child (A) and child with DCD (B) in the visible condition. (A) is a sample trace of a TD child in the visible condition. It

shows very small velocity changes and hardly any spatial variability. (B) is a sample trace of a child with DCD that shows greater spatial and

temporal variance. Although most of the movements were within the 2.5-cm target, increased within-target variability is clear.

G. D. Ferguson et al. / Neuroscience 286 (2015) 13–26 21

Evidence of the impact of unwanted variability on

movement execution in DCD was demonstrated in

another smooth pursuit experiment by Lee and

colleagues (2013) who assessed the performance of chil-

dren with DCD and co-morbid Attention Deficit with

Hyperactivity Disorder (ADHD) (Lee et al., 2013). Partici-

pants were required to draw circles within a constrained

outlined using a stylus to assess fine motor fluency. To

evaluate smooth pursuit, participants were required to

move a pen along a track of asterisks that moved at a

constant speed across a monitor. In both these tasks,

where visual information was provided throughout the

task, children with DCD+ ADHD performed poorer than

children with ADHD only. Specifically, the DCD+ ADHD

group had difficulty maintaining consistent velocity during

pursuit and used extra force when executing the move-

ment compared to the controls and their movements were

not smooth. These findings indicate that fine motor control

deficits affect performance of children with DCD in visuo-

manual tasks.

Herzfeld and Shadmehr (2014), however, view vari-

ability as an intrinsic attribute of motor learning. In this

sense, variability is considered a positive and important

feature. Motor variability is seen as a form of exploration

in the early stages of learning (Herzfeld and Shadmehr,

2014). When a skill is mastered it is possible to deliver

stable performance even under changing conditions

(Bernstein, 1967). Variance in movement execution can

also be a drive for learning. In the current study, it appears

that the DCD group, given the same amount of practice,

was unable to learn from the discrepancy between pre-

dicted and observed outcome of the movement despite

being confronted with repeated distortions in their perfor-

mance. This suggests that for DCD the variability

observed is not a positive feature in their case.

Several studies suggest that children with DCD show

greater variability across movement trials in many tasks,

including postural and gait activities (Wilson et al.,

2013). While this variability is regarded by some as a

means of exploring movement strategies and optimizing

Fig. 6. Sample trace of a TD child (A) and a child with DCD (B) in the occluded condition. (A) shows traces of a TD child when the target is

temporally occluded. Although the spatial accuracy diminishes, the trace is still relatively fluent. To make these kinds of smooth movement

trajectories, a movement has to be planned in advance (i.e., using feed-forward, predictive control) with only small in-flight corrections during the

movements. (B) shows a trace of a child with DCD in the occluded condition. The trace shows that the child made many errors (overshooting) and

how the child tries to recapture the target when he sees it again which leads to large velocity changes (peaks in this example are five times larger

compared to the TD child).

22 G. D. Ferguson et al. / Neuroscience 286 (2015) 13–26

control strategies (Hadders-Algra, 2000; Herzfeld and

Shadmehr, 2014), this may not be true for children with

DCD, who demonstrate an impaired motor learning strat-

egy and fail to establish adequate internal movement rep-

resentations for future use (Smits-Engelsman and Wilson,

2013).

Performance in the occlusion condition

As visual feedback is interrupted intermittently in the

occluded condition children may have to rely more on

predictive control to track the target successfully.

Hence, one would expect children with DCD to have

greater difficulty with this task than TD children (Wilson

et al., 2013). In contrast, if only force grading were the

issue, one would not expect this additional deficit.

Our data points toward increased deficits in the DCD

group. In other words, children with DCD ventured much

further out of the target in the occluded condition and

demonstrated high variability in their movement

trajectory and their speed. In comparison, the TD

children had distances out of the target that were

relatively stable in both conditions (occluded and

visible). Velocity profiles also differed. Children with

DCD were faster and more variable in the occluded

condition. The high tracking velocity could be attributed

to the high catch-up velocities they had to achieve in

order to reacquire the target when it reappeared,

leading to gains well above unity.

Impact of speed

In general, both the TD and DCD groups performed more

poorly when the target moved faster. At higher speeds,

both groups spent less time inside the target, went out

of the target more often and went out further than they

had at lower speed. Nevertheless, in the DCD group,

the pursuit of the target was affected to a greater extent

by speed than in the TD group. The TD group was also

quicker to re-enter the target in the fast speed condition.

Evidence of deficits in predictive control

Our findings in the visible, occluded and speed conditions

suggest that the performance of children in the DCD

group is representative of an underlying deficit in

predictive control. Four intrinsic processing deficits

support our hypothesis, including (1) inability to build

and store an internal model of dynamic target position

and motion, (2) inability to estimate hand state, (3)

deficiency in smooth oculomotor pursuit and (4) a

combination of these processes.

Evidence for a deficit in internal modeling of targetmotion. The ability to predict target motion is to an extent,

reliant on information from prior storage of the object’s

motion. The first step therefore, is the construction of an

accurate internal representation of dynamic target

position and motion that can be used to successfully

pursue the moving target. The next step concerns the

ability to update this model (i.e., knowing the movement

trajectory) which enables the smooth pursuit of a

moving target even in the presence of occlusion.

Wilson and colleagues (2013) support the theory that

children with DCD have a reduced ability to build and

update internal models and, as such, require more time

and practice in order to build adequate representations

(Wilson et al., 2013). It is hypothesized that the IMDs

seen in DCD may be related to immature neurodevelop-

ment (Hyde and Wilson, 2013), more specifically, in areas

G. D. Ferguson et al. / Neuroscience 286 (2015) 13–26 23

of the brain considered to process and store action repre-

sentations such as the PPC and cerebellum (Desmurget

et al., 1999; Desmurget and Sirigu, 2009; Zwicker et al.,

2009, 2012).

The present study shows that children with DCD were

unable to successfully predict the trajectory of the moving

target despite having multiple trials or opportunities to

build an accurate representation. The fact that children

with DCD made more tracking errors when the target

speed changed or was temporarily occluded is taken as

evidence to support an inability to update their internal

representation of dynamic target motion.

Evidence for how short-term velocity memory stores

are used as a basis for smooth pursuit eye tracking is

provided in a review by Barnes (Barnes, 2008). It is pre-

sumed that an internal positive feedback loop might serve

as the basis for the target motion memory, which is then

released as a predictive estimate of the required velocity.

In the current pursuit task, subjects were given one prac-

tice trial (30 s) and two test trials (60 s each), which

allowed them to establish an internal representation of

the target motion. Despite the length of the trial, children

with DCD were unable to either generate an accurate rep-

resentation or utilize the model effectively compared to

the TD group as evidenced by their inaccurate tracking

performance.

Evidence for deficiency in smooth oculomotor pur-

suit. Wilmut et al. (2006) found that 7- and 8-year olds

with DCD foveated longer on new targets compared to

age-matched controls, and both groups foveated longer

than adults. They interpreted these findings as evidence

of decreased reliance on a feed-forward control strategy

in children with DCD. Further evidence of deficits in

smooth pursuit is seen in another study looking at a group

of children with DCD (Langaas et al., 1998). Interestingly,

this study found that in most cases the eyes were ahead

of the target indicating that these children used primarily a

feed-forward strategy, but that they did so inappropriately.

Their findings are reinforced by the current data, which

shows the failing predictive control strategy used by chil-

dren with DCD.

Other studies have largely confirmed these findings.

Eye movement control seemed relatively spared in

children with DCD but deficiencies were present when

predictive control was required (Tsai et al., 2009, 2010).

An example is the study of Katschmarsky et al. (2001),

using the Double Step Saccade Task (DSST). This task

requires participants to make rapid eye movements to

two targets presented sequentially. These authors con-

firm that the ability to move the eye toward a target is pre-

served in DCD (Katschmarsky et al., 2001). Participants

were thus able to make accurate first saccades. However,

differences were apparent during the second saccade

and suggest that the DCD group performed worse as a

result of a diminished ability to program movements on

the basis of predictive control.

Evidence for inability to estimate and predict hand

state. While the present data are compatible with a deficit

in oculomotor prediction, it can be assumed that there

was also a more generalized deficit in hand-state

prediction. Temporary occlusion of the target forces the

child, not only to predict visually the location of the

target but also to predict the location of the hand (i.e.,

hand state-estimation) during the occlusion period. In

the present study, it is evident that children with DCD

were unable to use information regarding limb and

target state-estimation to build and refine their internal

model. Whether children with DCD have difficulty with

target and/or hand-state information is unclear as the

present paradigm does not allow us to distinguish

between these two abilities. It would be of interest

therefore to combine recordings of eye and hand

movements in future studies to disentangle these

various proposed mechanisms.

The present study is the first to look in detail at visuo-

manual target tracking. Nevertheless, there are some

data available related to aiming arm movements.

Contreras-Vidal (2006) found that TD 6-year-olds did

not accurately update movement trajectories in an

aiming task compared to 8- and 10-year-olds.

Internal representation of the consequences of ones

own actions slowly develops during childhood.

Researchers have used various paradigms to examine

the maturation of sensorimotor representations in

childhood. The acquisition of an optimal, coordinated

reaching and grasping strategy, for example, occurs

only around 12 years of age and this pattern shows a

high variability at age 6 (Kuhtz-Buschbeck et al., 1998;

Olivier et al., 2007). Similarly, in oculomotor studies it

was found that maturation requires about 12 years. For

example, Alahyane et al. (2014) used a rapid event-

related fMRI design in three age groups (8–12, 13–17

and 18–25 years). Participants were asked to make

either a prosaccade (look toward peripheral target) or

an antisaccade (look away from target) during the

task. The younger children (8–12 years) showed reduced

fronto-parietal activity during prosaccade and antisaccade

preparation compared to the other groups suggesting a

developmental trend (Alahyane et al., 2014).

Another paradigm used to examine the maturation of

sensorimotor representations is through the examination

of anticipatory postural adjustments (APAs) in bimanual

load-lifting tasks. This paradigm requires one to make

accurate predictions regarding the timing and magnitude

of postural stabilization, to enable a loaded limb to

remain relatively stable during unloading. In the study by

Schmitz et al. (2002) it was shown that the development

of APAs is characterized by early emergence, and slow

maturation through childhood, which the authors specu-

late, is linked to the slow maturation of internal models

of action (Schmitz et al., 2002).

Wilmut et al. using a Double Step Pointing Task,

described that problems occurred during the preparation

of sequential aiming movements, indicative of

impairment in children with DCD when updating an

existing forward model (Wilmut et al., 2006). Similar con-

clusions were reached in another study on aiming move-

ment of the upper limb, again suggestive of the reduced

functionality of predictive networks among children with

DCD (Kagerer et al., 2006).

24 G. D. Ferguson et al. / Neuroscience 286 (2015) 13–26

Underlying brain areas

The question arises whether the presently described

deficits of the DCD group reflect immaturity of the motor

system or pathology. Ballistic movement strategies, as

observed here in the DCD group, have been reported to

be a control mode typically seen in younger children

under eight years of age (Caeyenberghs et al., 2009b).

Predictive motor control is reported to develop over time

and is largely influenced by experience and maturation

(van Roon et al., 2008; Caeyenberghs et al., 2009b).

There is also a difference between eye and hand move-

ments. For example, there is evidence that predictive

eye movement to future locations develops early as

reflected by the ability of young babies to track and follow

moving objects visually (Tsai et al., 2009), and to predict

visually where an object will reappear after temporary

occlusion (von Hofsten et al., 2007). In contrast, manual

tracking appears to develop later depending on the task

(van Roon et al., 2008; Alahyane et al., 2014). The partic-

ipants in the present study had a mean age of eight years,

which means that the TD group presumably had matured

sufficiently to perform the task well. In contrast, the DCD

group might not yet have developed the neuronal net-

works required for the current task.

Both the parietal lobe and cerebellum have been

identified as working together for predicting the sensory

consequences of movement and in monitoring and

correcting movement (Wolpert et al., 1998; Kawato

et al., 2003). Studies using fMRI showed under-activation

of the cerebellum and parietal regions in children with

DCD relative to TD children (Kashiwagi et al., 2009;

Zwicker et al., 2009). The cerebellum is essential for

online corrections of smooth pursuit since it constructs

internal models that control movement output in relation

to input from both internal forward predictions and senso-

rimotor feedback (Tseng et al., 2007).

Limitations and recommendations

One of the limitations of the current study is the absence

of knowledge on the co-morbidity profile of the current

sample (Alloway, 2012). It is well-known that there are fre-

quent associations between DCD and other disorders,

such as Attention Deficit Disorder (ADD), Learning Disor-

ders (LD) and ADHD (Jongmans et al., 2003). The pres-

ence of these co-morbidities may have influenced the

performance of children with DCD. Although none of the

parents in our study reported a confirmed diagnosis of

either ADD or ADHD, it was not possible to confirm this

through formal testing. We attempted to limit the influence

of LD in the sample by conducting the study at a main-

stream school and excluding children who had failed a

grade level more than once. Hence, additional studies

are recommended to determine whether children with

DCD would be able to learn how to perform the task suc-

cessfully if allowed more time.

CONCLUSIONS

The ability to anticipate action is particularly important in

daily life especially in dynamic environments. The

perception, interpretation and use of visual cues are

essential elements that guide our behaviors, and any

maladaptive decision or choice made in a specific

situation might have serious consequences.

The poor tracking performance of the DCD group both

in the visual and in the occluded condition suggests

impairments in predictive control and poor dynamic

internal representation of target or hand motion.

Even in the absence of visual information, we

observed a good match between target and pursuit

directions during the periods of occlusion in the TD

group. Therefore, we suggest that the hand motor

system, like the ocular motor system (Whittaker and

Eaholtz, 1982) has access to a dynamic internal repre-

sentation of target motion, i.e., a representation that

evolves over time and that closely represents current tar-

get motion. In children with DCD, this ability is probably

deficient, thereby providing hints as to possible interven-

tions that could improve their predictive skills.

Acknowledgments—The authors acknowledge Dr. Eugene Ram-

eckers, Mrs Annelies Vos and Mrs Wies van Arkel for their assis-

tance in data collection. We also acknowledge the University of

Cape Town Research Committee for funding the first author.

The funders were neither involved in the study design, data col-

lection, analysis or interpretation of data, writing of the report

nor the decision to submit for publication.

REFERENCES

Alahyane N, Brien DC, Coe BC, Stroman PW, Munoz DP (2014)

Developmental improvements in voluntary control of behavior:

effect of preparation in the fronto-parietal network? NeuroImage

98:103–117.

Alloway TP (2012) The neurophysiological performance of

visuospatial working memory in children with developmental

coordination disorder. Dev Med Child Neurol 54:1075–1076.

American Psychiatric Association (2000) Diagnostic and statistical

manual of mental disorders, fourth edition, text revision (DSM-IV-

TR). Arlington, Washington: American Psychiatric Association.

Barnes GR (2008) Cognitive processes involved in smooth pursuit

eye movements. Brain Cogn 68:309–326.

Bernstein NA (1967) The co-ordination and regulation of

movements. Oxford Pergamon Press.

Biancotto M, Skabar A, Bulgheroni M, Carrozzi M, Zoia S (2011)

Neuromotor deficits in developmental coordination disorder:

evidence from a reach-to-grasp task. Res Dev Disabil

32:1293–1300.

Blakemore SJ, Sirigu A (2003) Action prediction in the cerebellum

and in the parietal lobe. Exp Brain Res 153:239–245.

Caeyenberghs K, van Roon D, van Aken K, De Cock P, Linden CV,

Swinnen SP, Smits-Engelsman BC (2009a) Static and dynamic

visuomotor task performance in children with acquired brain

injury: predictive control deficits under increased temporal

pressure. J Head Trauma Rehab 24:363–373.

Caeyenberghs K, Wilson PH, van Roon D, Swinnen SP, Smits-

Engelsman BC (2009b) Increasing convergence between

imagined and executed movement across development:

evidence for the emergence of movement representations. Dev

Sci 12:474–483.

Debrabant J, Gheysen F, Caeyenberghs K, Van Waelvelde H,

Vingerhoets G (2013) Neural underpinnings of impaired predictive

motor timing in children with Developmental Coordination

Disorder. Res Dev Disabil 34:1478–1487.

G. D. Ferguson et al. / Neuroscience 286 (2015) 13–26 25

Desmurget M, Epstein CM, Turner RS, Prablanc C, Alexander GE,

Grafton ST (1999) Role of the posterior parietal cortex in updating

reaching movements to a visual target. Nat Neurosci 2:563–567.

Desmurget M, Sirigu A (2009) A parietal-premotor network for

movement intention and motor awareness. Trends Cogn Sci

13:411–419.

Ferguson G, Jelsma J, Versfeld P, Smits-Engelsman B (2014) Using

the ICF framework to explore the multiple interacting factors

associated with developmental coordination disorder. Curr Dev

Disord Rep:1–16.

Fitts PM (1954) The information capacity of the human motor system

in controlling the amplitude of movement. J Exp Psychol

47:381–391.

Gabbard C (2009) Studying action representation in children via

motor imagery. Brain Cogn 71:234–239.

Hadders-Algra M (2000) The neuronal group selection theory: a

framework to explain variation in normal motor development. Dev

Med Child Neurol 42:566–572.

Harris CM, Wolpert DM (1998) Signal-dependent noise determines

motor planning. Nature 394:780–784.

Henderson SE, Sugden DA, Barnett AL (2007) Movement

assessment battery for children (2nd ed.) (examiner’s manual).

London: Harcourt Assessment.

Herzfeld DJ, Shadmehr R (2014) Motor variability is not noise, but

grist for the learning mill. Nat Neurosci 17:149–150.

Hyde C, Wilson PH (2011) Dissecting online control in Developmental

Coordination Disorder: a kinematic analysis of double-step

reaching. Brain Cogn 75:232–241.

Hyde CE, Wilson PH (2013) Impaired online control in children with

developmental coordination disorder reflects developmental

immaturity. Dev Neuropsychol 38:81–97.

IBM (2011) SPSS Statistics for Windows Version 20.0. vol. 20.0

Armonk, New York: IBM Corp.

Jongmans MJ, Smits-Engelsman BC, Schoemaker MM (2003)

Consequences of comorbidity of developmental coordination

disorders and learning disabilities for severity and pattern of

perceptual-motor dysfunction. J Learn Disabil 36:528–537.

Kagerer FA, Contreras-Vidal JL, Bo J, Clark JE (2006) Abrupt, but not

gradual visuomotor distortion facilitates adaptation in children with

developmental coordination disorder. Hum Mov Sci 25:622–633.

Kashiwagi M, Iwaki S, Narumi Y, Tamai H, Suzuki S (2009) Parietal

dysfunction in developmental coordination disorder: a functional

MRI study. NeuroReport 20:1319–1324.

Katschmarsky S, Cairney S, Maruff P, Wilson PH, Currie J (2001)

The ability to execute saccades on the basis of efference copy:

impairments in double-step saccade performance in children with

developmental co-ordination disorder. Exp Brain Res 136:73–78.

Kawato M, Kuroda T, Imamizu H, Nakano E, Miyauchi S, Yoshioka T

(2003) Internal forward models in the cerebellum: fMRI study on

grip force and load force coupling. Prog Brain Res 142:171–188.

Kuhtz-Buschbeck JP, Stolze H, Johnk K, Boczek-Funcke A, Illert M

(1998) Development of prehension movements in children: a

kinematic study. Exp Brain Res 122:424–432.

Langaas T, Mon-Williams M, Wann JP, Pascal E, Thompson C

(1998) Eye movements, prematurity and developmental co-

ordination disorder. Vision Res 38:1817–1826.

Lee IC, Chen YJ, Tsai CL (2013) Kinematic performance of fine motor

control in attention-deficit/hyperactivity disorder: the effects of

comorbid developmental coordination disorder and core

symptoms. Pediatr Int 55:24–29.

Magalhaes LC, Cardoso AA, Missiuna C (2011) Activities and

participation in children with developmental coordination

disorder: a systematic review. Res Dev Disabil 32:1309–1316.

Mrotek LA, Soechting JF (2007) Predicting curvilinear target motion

through an occlusion. Exp Brain Res 178:99–114.

Newsome WT, Wurtz RH, Komatsu H (1988) Relation of cortical

areas MT and MST to pursuit eye movements. II. Differentiation of

retinal from extraretinal inputs. J Neurophysiol 60:604–620.

Olivier I, Hay L, Bard C, Fleury M (2007) Age-related differences in

the reaching and grasping coordination in children: unimanual and

bimanual tasks. Exp Brain Res 179:17–27.

Orban de Xivry JJ, Missal M, Lefevre P (2008) A dynamic

representation of target motion drives predictive smooth pursuit

during target blanking. J Vis 8(6):1–13.

Piek JP, Skinner RA (1999) Timing and force control during a

sequential tapping task in children with and without motor

coordination problems. J Int Neuropsychol Soc 5:320–329.

Robert MP, Ingster-Moati I, Albuisson E, Cabrol D, Golse B, Vaivre-

Douret L (2014) Vertical and horizontal smooth pursuit eye

movements in children with developmental coordination disorder.

Dev Med Child Neurol.

Schmitz C, Martin N, Assaiante C (2002) Building anticipatory

postural adjustment during childhood: a kinematic and

electromyographic analysis of unloading in children from 4 to

8 years of age. Exp Brain Res 142:354–364.

Scott SH (2012) The computational and neural basis of voluntary

motor control and planning. Trends Cogn Sci 16:541–549.

Sirigu A, Duhamel JR, Cohen L, Pillon B, Dubois B, Agid Y (1996)

The mental representation of hand movements after parietal

cortex damage. Science 273:1564–1568.

Smits-Engelsman BC, Niemeijer AS, van Galen GP (2001) Fine

motor deficiencies in children diagnosed as DCD based on poor

grapho-motor ability. Hum Mov Sci 20:161–182.

Smits-Engelsman BC, Van Galen GP, Duysens J (2002) The

breakdown of Fitts’ law in rapid, reciprocal aiming movements.

Exp Brain Res 145:222–230.

Smits-Engelsman BC, Westenberg Y, Duysens J (2003a)

Development of isometric force and force control in children.

Brain Res Cogn Brain Res 17:68–74.

Smits-Engelsman BC, Westenberg Y, Duysens J (2008) Children

with developmental coordination disorder are equally able to

generate force but show more variability than typically developing

children. Hum Mov Sci 27:296–309.

Smits-Engelsman BC, Wilson PH (2013) Noise, variability, and motor

performance in developmental coordination disorder. Dev Med

Child Neurol 55(Suppl. 4):69–72.

Smits-Engelsman BC, Wilson PH, Westenberg Y, Duysens J (2003b)

Fine motor deficiencies in children with developmental

coordination disorder and learning disabilities: an underlying

open-loop control deficit. Hum Mov Sci 22:495–513.

Tsai CL, Pan CY, Chang YK, Wang CH, Tseng KD (2010) Deficits of

visuospatial attention with reflexive orienting induced by eye-

gazed cues in children with developmental coordination disorder

in the lower extremities: an event-related potential study. Res Dev

Disabil 31:642–655.

Tsai CL, Pan CY, Cherng RJ, Hsu YW, Chiu HH (2009) Mechanisms

of deficit of visuospatial attention shift in children with

developmental coordination disorder: a neurophysiological

measure of the endogenous Posner paradigm. Brain Cogn

71:246–258.

Tseng YW, Diedrichsen J, Krakauer JW, Shadmehr R, Bastian AJ

(2007) Sensory prediction errors drive cerebellum-dependent

adaptation of reaching. J Neurophysiol 98:54–62.

van Beers RJ (2009) Motor learning is optimally tuned to the

properties of motor noise. Neuron 63:406–417.

van Galen GP, Portier SJ, Smits-Engelsman BC, Schomaker LR

(1993) Neuromotor noise and poor handwriting in children. Acta

Psychol (Amst) 82:161–178.

van Roon D, Caeyenberghs K, Swinnen SP, Smits-Engelsman BC

(2008) Development of feedforward control in a dynamic manual

tracking task. Child Dev 79:852–865.

van Roon D, Caeyenberghs K, Swinnen SP, Smits-Engelsman BC

(2010) Children with a learning disorder show prospective control

impairments during visuomanual tracking. Res Dev Disabil

31:195–202.

Van Waelvelde H, De Weerdt W, De Cock P, Janssens L, Feys H,

Smits Engelsman BC (2006) Parameterization of movement

execution in children with developmental coordination disorder.

Brain Cogn 60:20–31.

von Hofsten C, Kochukhova O, Rosander K (2007) Predictive

tracking over occlusions by 4-month-old infants. Dev Sci

10:625–640.

26 G. D. Ferguson et al. / Neuroscience 286 (2015) 13–26

Wang T-N, Tseng M-H, Wilson BN, Hu F-C (2009) Functional

performance of children with developmental coordination disorder

at home and at school. Dev Med Child Neurol 51:817–825.

Whittaker SG, Eaholtz G (1982) Learning patterns of eye motion for

foveal pursuit. Invest Ophthalmol Vis Sci 23:393–397.

Williams J, Thomas PR, Maruff P, Butson M, Wilson PH (2006)

Motor, visual and egocentric transformations in children with

Developmental Coordination Disorder. Child Care Health Dev

32:633–647.

Wilmut K, Wann JP, Brown JH (2006) Problems in the coupling of eye

and hand in the sequential movements of children with

Developmental Coordination Disorder. Child Care Health Dev

32:665–678.

Wilson PH, Maruff P, Butson M, Williams J, Lum J, Thomas PR

(2004) Internal representation of movement in children with

developmental coordination disorder: a mental rotation task.

Dev Med Child Neurol 46:754–759.

Wilson PH, McKenzie BE (1998) Information processing deficits

associated with developmental coordination disorder: a meta-

analysis of research findings. J Child Psychol Psychiatry

39:829–840.

Wilson PH, Ruddock S, Smits-Engelsman B, Polatajko H, Blank R

(2013) Understanding performance deficits in developmental

coordination disorder: a meta-analysis of recent research. Dev

Med Child Neurol 55:217–228.

Wolpert DM, Miall RC (1996) Forward models for physiological motor

control. Neural Netw 9:1265–1279.

Wolpert DM, Miall RC, Kawato M (1998) Internal models in the

cerebellum. Trends Cogn Sci 2:338–347.

World Medical A (2013) World Medical Association Declaration of

Helsinki: ethical principles for medical research involving human

subjects. JAMA 310:2191–2194.

Wuang YP, Su JH, Su CY (2012) Reliability and responsiveness of

the Movement Assessment Battery for Children-Second Edition

Test in children with developmental coordination disorder. Dev

Med Child Neurol 54:160–165.

Zwicker JG, Missiuna C, Boyd LA (2009) Neural correlates of

developmental coordination disorder: a review of hypotheses. J

Child Neurol 24:1273–1281.

Zwicker JG, Missiuna C, Harris SR, Boyd LA (2012) Developmental

coordination disorder: a pilot diffusion tensor imaging study.

Pediatr Neurol 46:162–167.

(Accepted 13 November 2014)(Available online 29 November 2014)