Identifying the signature of prospective motor control in children with autism

Abstract

Failure to develop prospective motor control has been proposed to be a core phenotypic marker of autism spectrum disorders (ASD). However, whether genuine differences in prospective motor control permit discriminating between ASD and non-ASD profiles over and above individual differences in motor output remains unclear. Here, we combined high precision measures of hand movement kinematics and rigorous machine learning analyses to determine the true power of prospective movement data to differentiate children with autism and typically developing children. Our results show that while movement is unique to each individual, variations in the kinematic patterning of sequential grasping movements genuinely differentiate children with autism from typically developing children. These findings provide quantitative evidence for a prospective motor control impairment in autism and indicate the potential to draw inferences about autism on the basis of movement kinematics.

Figure 1

(A) Image sequence of a reach-to-grasp movement with hand model overlay. (B–F) Kinematic profiles of reach-to-grasp movements performed with three prospective intentions (to place, to pour, to pass). Wrist velocity (B), wrist acceleration (C), wrist jerk (D), grip aperture (E) and wrist height (F) were extracted. Thick lines represent group average (red = ASD group, green = TD group). Thin lines represent trial-average of individual participants.

Figure 2

(A) Classification accuracy of group (ASD, TD) using subject-wise, record-wise 10 folds cross-validation and record-wise 40 folds cross-validation. Histograms represent mean ± SEM. (B,C) Permutation null distribution generated by splitting data in a record-wise fashion (B) or subject-wise fashion (C). The permutation null distribution is represented by the grey histograms. The purple line in (B) represents the observed record-wise accuracy. The dark blue line in (C) represents the observed subject-wise accuracy. The dashed grey line indicates the 0.5 chance level. (D) Group classification accuracy as a function of sample size computed using record-wise and subject-wise cross-validation. (E) Group classification accuracy of unseen subjects obtained using subject-wise hyper-parameter tuning and record-wise hyper-parameter tuning. (F,G) Movement data colored according to group classification of one exemplar unseen subject with ASD using subject-wise hyper-parameter tuning (F) and record-wise hyper-parameter tuning (G). Data are visualized as a scatterplot of velocity at 80% of movement duration and grip aperture at 90% of movement durations.