Fourteen children with ASDs between the ages of 5 and 11 participated in this study (9 male and 5 female). In order to participate, a formal diagnosis of ASD from a medical doctor was required. The Autism Spectrum Quotient: Children’s Version (AQ-Child; Auyeung et al., 2008) was used to quantify autistic traits in children with ASD (see Table 1). The performance of children with ASD was compared to 14 age-, hand preference- and sex-matched typically-developing children who were randomly selected from a larger study that included 75 typically-developing children ages 5–11 (Scharoun and Bryden, 2014). Parents of typically-developing children were not asked to complete the AQ-Child. All children were recruited from summer camps at the institution, and elementary schools in the region. All data were collected outside of the camp or classroom setting. The institution Research Ethics Board approved all recruitment and testing procedures. Written informed consent was obtained from parents/guardians, and assent was obtained from child participants.

Figure 2 depicts the number of trials where children with ASD displayed end-state comfort when grasping the overturned glass to pour a cup of water. A Mann-Whitney test was used to compare end-state comfort in children with ASD and their typically-developing peers. Evidence differed significantly in children with ASD (Mdn = 1.00) and their typically-developing peers (Mdn = 3.00), U = 152.500, p = 0.011, r = −0.51. Next, a Kruskal-Wallis Test was used to assess end-state comfort when children with ASD and their typically-developing peers were separated into younger (i.e., ages 5–8) and older (i.e., ages 9–11) groups. End-state comfort was significantly affected by group, H₍₃₎ = 8.68, p = 0.034 (see Figure 3). Pairwise comparisons with adjusted p-values showed that there were no significant differences between 5- to 8- and 9- to 11-year-olds with ASD (p = 0.428, r = −0.15), 9- to 11-year-olds with ASD and 5- to 8-year-old typically-developing children (p = 0.267, r = −0.21), and 5- to 8- and 9- to 11-year-old typically-developing children (p = 0.408, r = −0.16). There was a significant difference between 5- to 8-year-olds with ASD and 9- to 11-year-old typically-developing children (p = 0.006, r = −0.52). Comparisons between 5- to 8-year-olds with ASD and 5- to 8-year-old typically-developing children (p = 0.057, r = −0.36), and 9- to 11-year-olds with ASD and 9- to 11-year-old typically-developing children (p = 0.053, r = −0.37) were approaching statistically significant differences.

The number of trials where children with ASD displayed beginning-state comfort is found in Figure 4. A Mann-Whitney test was used to compare beginning-state comfort in children with ASD and their typically-developing peers. The evidence did not differ significantly in children with ASD (Mdn = 1.00) and their typically-developing peers (Mdn = 3.00), U = 69.00, p = 0.063, r = −0.35. Next, a Kruskal-Wallis Test was used to assess end-state comfort when children with ASD and their typically-developing peers were separated into younger (i.e., ages 5–8) and older (i.e., ages 9–11) groups. Beginning-state comfort was not affected by group, H₍₃₎ = 3.92, p = 0.271 (see Figure 5).

Pearson correlations were used to assess the relationship between AQ-child scores and measures of anticipatory planning. No significant relationships emerged between AQ-child scores and end-state comfort or beginning-state comfort (see Table 2 and Figure 6) overall or when separated into younger (ages 5–8) and older groups (ages 9–11).

The first stage of analyses compared evidence of end-state comfort in children with ASD and their typically-developing peers. In line with our hypothesis, a significant difference emerged, such that children with ASD displayed less end-state comfort. To examine this further, children were separated into younger (ages 5–8) and older (ages 9–11) groups, based on previous reports of adult-like patterns in typically-developing children at approximately ages 9 or 10 (e.g., Wunsch et al., 2013). Five to eight-year-olds with ASD displayed less end-state comfort than 9- to 11-year-old typically-developing children. Furthermore, comparisons between 5- to 8-year-olds with ASD and 5- to 8-year-old typically-developing children (p = 0.057), and 9- to 11-year-olds with ASD and 9- to 11-year-old typically-developing children (p = 0.053) were close to the benchmark (p < 0.05) typically used to infer a significant difference. In line with our hypothesis, children with ASD displayed less end-state comfort. These results are contrary to Hamilton et al. (2007) and van Swieten et al. (2010), who observed identical performance in age-matched children with ASD and typically-developing children. Findings thus provide support for Hughes (1996) who suggested planning problems exists in children with ASD at the level of motor control.

Motor planning in object manipulation is challenging, because it requires consideration of the entire movement sequence, as opposed to simply the first step (Gowen and Hamilton, 2013); therefore, an action plan must be created before the action is completed (Manoel and Moreira, 2005; Adalbjornsson et al., 2008). Research has noted that tasks involving action chaining—movements linked in overlapping segments (Graybiel, 1998; Gobet et al., 2001)—are challenging for children with ASD; therefore, these children are observed performing each element of the task individually. As evidenced in the current study, rather than planning for secondary task requirements (i.e., object use), children with ASD displayed behaviors indicative of first-order planning (i.e., immediate reach-to-grasp; Rosenbaum et al., 2012); therefore displaying start-state comfort as opposed to end-state comfort.

Hughes (1996) referred to deficits in sequencing abilities, further explaining how impairment in sequential activity is not limited to, multiple movements, but can be observed in single purposeful movements as well. Hughes (1996) also emphasized the importance of prediction in movement (Poulton, 1957), such that children with ASD are restricted in their understanding of the consequences associated with one’s own actions as they pertain to the decision to plan that action. Finally, once an action plan is formulated, it is both difficult and costly for children with ASD to deviate from the original plan (Hill, 2004). As evidenced in the current study, difficulties planning, organizing and coordinating movements emerge when children with ASD are challenged to plan according to end-state comfort (Glazebrook et al., 2008; Gowen and Hamilton, 2013).

In comparison to Hughes (1996), and van Swieten et al. (2010) observed that the majority of 9- to 14-year-olds with and without ASD displayed end-state comfort. These results are in line with recent reports of typical development, where end-state comfort was comparable to adults by age 9 (Stöckel et al., 2012; Scharoun and Bryden, 2014). With respect to Hughes (1996) findings, van Swieten et al. (2010) hypothesized that differences in findings are likely attributable to experimental design. More specifically, van Swieten et al. (2010) had participants grasp and rotate a cylinder either clockwise or counterclockwise, whereas Hughes (1996) task required participants to grasp a rod and transport it to a disc so it stood upright. It was argued that Hughes (1996) task was more complex; therefore, planning deficits emerged in children with ASD (van Swieten et al., 2010; see also Sacrey et al., 2014).

In the present study, children were required to grasp the overturned glass and the pitcher in anticipation of pouring water. It is thus likely that the bimanual nature of the task increased complexity. Compared to a unimanual task, more extensive cognitive processing and motor planning are required in a bimanual task (Logan and Fischman, 2011). Bimanual actions require an interplay between intra- and inter-limb coordination (Bobbio et al., 2009), while simultaneously processing multisensory information from the environment (Brakke et al., 2007). Therefore, it is likely that the increased task complexity in the current study led to findings similar to Hughes (1996). Nevertheless, it is also important to note that Hamilton et al. (2007) implemented a bar-transport task similar to Hughes (1996), but failed to replicate the findings. Recent work with typically-developing children comparing object manipulation tasks note end-state comfort emerges earlier in the overturned glass task compared to the bar-transport task (Knudsen et al., 2012). Future work assessing the effect in children with ASD would benefit from the comparison of multiple tasks, to identify the nature of inconsistent results.

Unlike the differences that emerged between the groups in the assessment of end-state comfort, when asked to pick-up the overturned cup and pass it to the researcher to pour water, behavior of children with ASD did not differ from their typically-developing peers. This was surprising, considering it is suggested that impaired motor planning skills may impede a child’s ability to investigate their physical and social environment (Glazebrook et al., 2008). Furthermore, deficits in social-motor behaviors have been linked to malfunctioning of the mirror neuron system (Williams et al., 2004); where the autistic mirror neuron dysfunction hypothesis predicts behaviors such as goal inference, to be atypical in children with ASD (Hamilton et al., 2007).

Despite evident difficulties in social environments, behaviors in children with ASD—communication skills (Bernard-Opitz, 1982), physical and eye contact (Kasari et al., 1993) and social interaction (Knott et al., 1995)—have been shown to improve during interactions with familiar people (parent, sibling) in comparison to a stranger or peer. As such, it is suggested that dysfunctions in the mirror neuron system may be resultant of an underlying deficit in the ability to identify with unfamiliar things (Oberman et al., 2008). Furthermore, it is suggested that, in contrast to typically-developing children who identify all things (familiar and unfamiliar) as socially congruent, children with ASD may only denote familiar things to be socially compatible (Oberman et al., 2008).

Observing the familiarity of the current task, previous studies investigating the development of feeding behaviors suggest that children slowly gain experience with cups prior to their second birthday. Therefore, cup-use is a familiar action for children (Carruth and Skinner, 2002), where certain habitual objects, are perceived to afford specific functional movements (Herbort and Butz, 2011). As such, it is likely that children with ASD were displaying a habitual response, as opposed to an action indicative of motor planning. Future research with other objects is necessary to confirm or refute this hypothesis.

Interestingly, no relationship emerged between measures of anticipatory planning (i.e., end- and beginning-state comfort) and AQ-child scores. This it likely attributed to the heterogeneity of ASD that is commonly reported in the literature, genetically, neurologically, and behaviorally (e.g., Happé et al., 2006; Jeste and Geschwind, 2014). With respect to motor behavior, it has been argued that a better understanding of such processes will enable the heterogeneity of ASD to be disentangled (e.g., Gowen and Hamilton, 2013). As such, continued research in this area is of utmost importance.

Recent investigations have noted children with ASD perform identical to their typically-developing counterparts in motor planning tasks (Hamilton et al., 2007; van Swieten et al., 2010). However, planning problems at the level of motor control (e.g., Hughes, 1996) have led to motor impairments being identified as a cardinal feature of ASD (Fournier et al., 2010; Gowen and Hamilton, 2013). Although the literature is inconclusive, it was hypothesized that children with ASD would demonstrate significantly less end- and beginning-state comfort than their typically-developing counterparts. The results of this study are in agreement with our hypothesis with respect to end-state comfort; however, no differences emerged in the assessment of beginning-state comfort. The findings support the notion that children with ASD have difficulties in planning and controlling movement in independent task. However, when acting with a familiar object in joint action, behavior does not differ, likely indicative of a habitual response as opposed to a planned behavior.

It is well known that “neurological diseases exact an exorbitant health cost on our population” (Ajemian and Hogan, 2010, p. 337). Furthermore, other aspects of life—social (Gray, 1993), health (Allik et al., 2006) and overall life/lifestyle (Sen and Yurtsever, 2007)—of families and caregivers (Fletcher et al., 2012) of children with ASD are negatively affected. In order to mitigate these costs, we must develop a better understanding of the neurological foundations of these diseases (Ajemian and Hogan, 2010). It is suggested that motor symptoms may play a larger role in autism than originally believed. Therefore, transitioning towards a movement perspective may have implications for the development of new diagnostic criteria and intervention protocol (Leary and Hill, 1996; Provost et al., 2007), as recent investigations have suggested motor impairments are a cardinal feature of ASD (Fournier et al., 2010). Furthermore, in order to provide the best services for children with ASD, therapists must be aware of motor difficulties as they pertain to motor control (Mari et al., 2003; Provost et al., 2007); thus highlighting the importance of current and future research studies, which aim to delineate motor difficulties in children with ASD.