It has been hypothesized that the underlying mechanism of clumsy motor behaviour in children with Developmental Coordination Disorder (DCD) is caused by a deficit in the internal modelling for motor control. An internal modelling deficit can be shown on a behavioural level by a task that requires motor imagery. Motor imagery skills are suggested to be related to anticipatory action planning, but motor imagery and action planning have not been tested within the same child. In the present study, action planning and motor imagery skills were assessed in 82 children between 7 and 12 years of age. Twenty-one of these children met the criteria for DCD, which was assessed by the McCarron Assessment of Neuromuscular Development and 56 of these children were used in the control group. Motor imagery was tested by a mental rotation task of hands that were shown from a back and palm point of view. The results show that motor imagery is affected in children with DCD but only in conditions with complex task constraints (i.e., rotation of hand stimuli presented in palm view). These results provide partial support for the internal modelling deficit hypothesis. We were not able to elicit motor planning deficits in this group, however, and argue that more complex planning tasks may be needed to identify such deficits.
Motor clumsiness in children (or Developmental Coordination Disorder – DCD) affects around 5–6% of the children of primary school age (Zwicker, Missiuna, Harris, & Boyd, 2012). The disorder has no identifiable medical cause and is not explained by low intelligence (American Psychiatric Association, 2000). However, the underlying mechanism for DCD is still much debated (Wilson, Ruddock, Smits-Engelsman, Polatajko, & Blank, 2012). One hypothesis receiving converging support is the internal modelling deficit hypothesis (IMD) (Maruff et al., 1999, Wilson et al., 2004 and Wolpert, 1997). What remains unclear is whether this putative deficit is related to aspects of action planning in children. In the study presented here we addressed this question by looking at the relationship between motor imagery (used previously to examine internal modelling) and end-state comfort planning in DCD. The concept of internal modelling has become influential in neurocomputational models of motor control and learning. Broadly defined, it comprises two aspects: an inverse modelling process that maps the necessary motor parameters (like force, timing, and trajectory) to achieve a desired goal state, and forward modelling that uses a predictive estimate of the sensory consequences of an action as a means of error correction (Wolpert, 1997). It is in the latter sense that internal modelling has been investigated in a number of studies. More precisely, the latter involves use of a forward model of the efference copy to correct for errors in the motor command. The output of the forward model provides a template against which real-time feedback can be compared under tight temporal constraints, and motor output signals can be corrected if needed (Shadmer, Smith, & Krakauer, 2010). This occurs before slower sensory-based feedback can be processed, resulting in stability of the motor system, particularly when the movement is perturbed in some way (e.g., when the visual target changes during the course of an action or when the moving limb is subjected to an unexpected, external force). In the case of overt movement, feedback is generated that is based on both sensory information and the efference copy. In contrast, if a movement is imagined, no sensory-based feedback is available, only that based on putative internal feedback loops (Desmurget & Grafton, 2003). Therefore, motor imagery paradigms have been used to assess the contents of the internal model and how they might be used to implement and control overt behaviour (Crammond, 1997, Lewis et al., 2008, Williams et al., 2011, Wilson et al., 2004 and Wilson et al., 2001). Two paradigms regularly used to assess motor imagery capacity in children with DCD are the mental chronometry and the mental rotation paradigm.
3.1. Motor imagery
In the motor imagery task with the back view of the hands, four (19%) children from the DCD group and 12 (21%) from the control group were excluded from the repeated measures ANOVA due to errors. A significant effect of angle of rotation was found (Wilks’ λ = .364, F(3,57) = 33.265, p < .001). However, no difference was found between children in the DCD group and controls.
To examine the difference in response time among medial and lateral oriented rotations into more detail a 2 (laterality) by 2 (orientation) by 2 (group) repeated measures ANOVA was performed. One (5%) child from the DCD group and one child (2%) from the control group were excluded due to errors. The analyses yielded a significant main effect of orientation (Wilks’ λ = .858, F(1,73) = 12.114, p = .001), an interaction between orientation and hand laterality (Wilks’ λ = .915, F(1,73) = 6.767, p = .011), and an interaction between orientation, hand laterality and group (Wilks’ λ = .940, F(1,73) = 4.692, p = .034). No main group or hand laterality effects were found. The results are shown in Fig. 2.
