Prosopagnosia is highly comorbid in individuals with probable developmental coordination disorder

Abstract

Developmental co-ordination disorder (DCD) is characterised by difficulties in motor control and coordination from early childhood. While problems processing facial identity are often associated with neurodevelopmental conditions, such issues have never been directly tested in adults with DCD. We tested this possibility through a range of tasks and assessed the prevalence of developmental prosopagnosia (i.e. lifelong difficulties with faces), in a group comprising individuals who self-reported a diagnosis of, or suspected that they had, DCD. Strikingly, we found 56% of this probable DCD group met recently recommended criteria for a diagnosis of prosopagnosia, with 22% acquiring a diagnosis using traditional cognitive task-based methods. Moreover, their problems with faces were apparent on both unfamiliar and familiar face memory tests, as well as on a facial perception task (i.e. could they tell faces apart). Positive correlations were found between self-report measures assessing movement and coordination problems, and objective difficulties on experimental face identity processing tasks, suggesting widespread neurocognitive disruption in DCD. Importantly, some issues in identity processing in our probable DCD group remained even after excluding participants with comorbid conditions traditionally associated with difficulties in face recognition, that is, autism and dyslexia. We recommend that any diagnostic test for DCD should include an assessment for prosopagnosia. Given the high prevalence of prosopagnosia in our probable DCD group, and the positive correlations between DCD and prosopagnosia symptoms, there may be a stronger link between movement and facial identity abilities than previously thought.

Keywords: DCD; developmental coordination disorder; dyspraxia; face perception; face recognition; motor; prosopagnosia.

Figure 1.

Prosopagnosia Index results. Higher scores indicate more severe symptoms of prosopagnosia. PDCD participants reported significantly higher levels of symptoms than controls (p < .001). This was replicated in both the subgroups of pDCD participants without ASD and dyslexia (p = .003), and those with pDCD and comorbid dyslexia and autism (p < .001). Error bars indicate 95% confidence intervals.

Figure 2.

Probable DCD participants made significantly more errors on the upright Cambridge Face Perception Test (CFPT; p < .001), only after we added control data from Burns et al. (2023). This was replicated in a subgroup of those with pDCD and comorbid dyslexia and autism [p < .001], but not those without ASD and dyslexia (p = .09) although adding Burns et al. (2022) data did reveal problems in this subgroup (see text). Error bars indicate 95% confidence intervals.

Figure 3.

Probable DCD participants scores were significantly different than controls on the CFPT Holistic (p < .001) only after adding control data from Burns et al. (2023). This was replicated in the subgroup of pDCD with comorbid dyslexia or autism (p < .001) but not in pDCD participants without ASD and dyslexia (p = .07). Error bars indicate 95% confidence intervals.

Figure 4.

Probable DCD participants scored significantly fewer correct trials on the CFMT than controls (p = .002). This was replicated in the subgroup of pDCD participants with comorbid dyslexia and autism (p < .001), but not those without ASD and dyslexia (p = .14). Error bars indicate 95% confidence intervals.

Figure 5.

Probable DCD participants scored significantly lower on the FFT than controls (p < .001). Values are in percentages. This was replicated in the subgroup of those with dyslexia and autism (p < .001), but not pDCD participants without ASD and dyslexia (p = .12). Error bars indicate 95% confidence intervals.

Figure 6.

Top left panel indicates that increasing symptoms of DCD (AACQ) positively predicted prosopagnosia symptoms (PI20, r = .71, 95% CI = [0.57, 0.81], p < .001]. When analysed separately, the same correlation was not significant in the controls (white circles, r = −0.10, 95% CI = [−0.41, 0.23], p = .56), but was in DCD (blue circles, r = .51, 95% CI [0.19, 0.73], p = .003). Top right panel illustrates that increasing DCD symptomology (AACQ) negatively predicts ability to identify previously unknown faces (CFMT; r = −.33, 95% CI [−0.52, −0.1], p = .006). The lower left panel indicates that greater movement and coordination difficulty on the AACQ negatively predicted ability to recognise familiar faces on the FFT (r = −.34, 95% CI [−0.54, −0.12], p = .004).

Figure 7.

The effect sizes of impairments in pDCD across all face and movement measures. The perception related problems with faces (e.g. CFPT Upright and Holistic) were comparable to the unfamiliar face memory problems (i.e. the CFMT). Then the pDCD group’s familiar face memory (FFT) problems were midway between these and their self-reported complaints with faces in daily life (i.e. PI20). Importantly, there were minimal differences between the pDCD and control groups, as shown by the very small CFPT inverted effect size. The two movement measures are plotted for comparison purposes.