Online control in children with developmental coordination disorder

Developmental coordination disorder (DCD) is characterized by impaired motor performance in the absence of any identifiable neurological or medical condition. Despite affecting ≈6% of children, the aetiology of the disorder remains unclear. I argue here that one aspect of motor control that shows great promise in clarifying the nature of DCD is rapid online control (ROC) of reaching. However, data on DCD have been conflicting: where some recent work using double-step reaching tasks suggests no difficulty in online control, others suggest deficits. The aim of the present thesis was to clarify the nature of ROC in DCD using the well-validated double-step reaching task interpreted using a neuro-computational framework.

In Study 1, chronometric analysis of double-step reaching was conducted to determine the integrity of ROC in DCD. Compared to controls, children with DCD showed increased movement time and error following unexpected target perturbation at movement onset. Interpreted from a computational perspective, I argue that this decreased capacity for accounting for target perturbation may reflect a reduced capacity for correcting predictive models of movement. While chronometric analysis provides important information about the integrity of the global properties of ROC, it does not inform the subtle transitions in control that support the un-folding movement. To better identify the motor control properties responsible for the inefficient double-step reaching shown by children with DCD, Study 2 adopted both kinematic and chronometric analysis of double-step reaching.

In Study 2, the chronometric pattern of performance shown in Study 1 was replicated. Early kinematic markers failed to discriminate between groups. Importantly, children with DCD took significantly longer to correct the trajectory of their reaching towards a new target following perturbation. Time to correction (ToC) is thought to reflect the stage in reaching when error information which arises following a mismatch between the expected (according to the predictive model) and actual sensory consequences of reaching has been successfully integrated with the on-going motor command. This pattern of results was consistent with the view that children with DCD experience difficulty correcting predictive models of movement online.

What remained unclear was whether the demonstrated atypical pattern of double-step reaching (and apparent deficit in predictive modeling) in DCD reflects a developmental immaturity or deviance from the typical developmental path. To address this issue, Study 3 compared the double-step reaching of older children with DCD (8-12 years) with both a group of age-matched controls and younger control children (5-7 years). Analysis revealed that the movement time of children with DCD and younger controls was equally affected by target perturbation, while kinematic analysis showed overlap in their early kinematic profiles; specifically ToC. This supports the view that difficulties in ROC (and predictive modeling) in DCD may reflect a developmental immaturity.

I argue that this group of studies provides compelling evidence that ROC is impaired in DCD. Interpreted from a computational perspective, this impairment appears to reflect disruption to predictive control (ala the IMD account of DCD). These difficulties likely reflect developmental immaturity possibly at the level of the posterior parietal cortex or cerebellum.