Motor planning and control in autism. A kinematic analysis of preschool children

https://doi.org/10.1016/j.rasd.2010.09.013Get rights and content

Abstract

Kinematic recordings in a reach and drop task were compared between 12 preschool children with autism without mental retardation and 12 gender and age-matched normally developing children. Our aim was to investigate whether motor anomalies in autism may depend more on a planning ability dysfunction or on a motor control deficit. Planning and control processes were separately investigated by examining kinematic recordings divided into primary movement- (planning-based) and corrective submovement- (control-based) phases.

Despite longer movement durations, participants with autism were as accurate in their movements as normally developing children were and showed a preserved movement structure. No differences were observed for the initial movement phases for hand velocity, accuracy and inter-trial variability.

Our main finding was that of a group difference in proximity of the target, at transition from planning-based to control-based movement guidance. At primary movement conclusion, the normally developing children had already reduced velocity and begun orienting their hands for ball drop. Also, they tended to terminate movements within the same movement unit that had transported the hand into the target box. Compared to this group, participants with autism reached this stage with less preparation: their speed was significantly higher, wrist inclination reduced and they showed further movement units after entering the box over the vast majority of trials. These additional movement units were presumed to represent late control-based spatial adjustments. Hence, our data support the hypothesis that children with autism have a greater need for corrective submovements.

We provide evidence that motor anomalies in autism might be determined either by a disruption in planning-control integration, or by a limited planning process capacity, as participants with autism might have been able to plan only the very beginning of the movement, leaving its final phases to further planning on the fly, with important consequences on movement time optimization.

Introduction

Autism is classified as a pervasive developmental condition defined by impairment in communication, social reciprocity, and repetitive-stereotyped behavioural patterns (American Psychiatric Association, 1994). Yet, these symptoms are also frequently associated with anomalies in movement skills, which may be evident as early as the age of 6 months (Zwaigenbaum et al., 2005) and which may cause delays in the attainment of developmental motor milestones (Teitelbaum, Teitelbaum, Nye, Fryman, & Maurer, 1998). Impaired motor skills in autistic spectrum disorders also emerge on standardized tests, such as the Bruininks–Oseretsky Test of Motor Proficiency (BOMTP, Bruininks and Bruininks, 2005, Provost et al., 2007) and the Movement Assessment Battery for Children (M-ABC, Green et al., 2002, Henderson et al., 2007).

In terms of understanding movement behaviour, fundamental movement skills consist of goal-directed movements such as locomotion and object-control skills (e.g., throwing a ball), whereas motor abilities refer to underlying capacities to movement skill performance (Magill, 1998). The latter are therefore not directly observable and must be inferred via movement kinematics analysis.

The present study was aimed at examining early motor abilities underlying object control skills in preschool children with autism in comparison with a group of normally developing children.2 The specific purpose was to investigate the children's ability to transport an object from one location to another and drop it into a hole; this action resembles a reaching movement requiring the hand to move from a set location to a target destination. In fact, some studies have shown reaching movement anomalies in autism. For example, Mari, Castiello, Marks, Marraffa, and Prior (2003) described two different motor patterns in function of intellectual ability after observing that low functioning children (IQ < 80) displayed a bradykinesic motor pattern, with longer movement durations and decelerations, lower peak velocities, and delayed maximum grip apertures for grasping. By contrast, average and high functioning children (IQ > 80) reached for objects more rapidly than typically developing children of the same age did. The authors hypothesized that the high functioning children's behavioural pattern was linked to movement modulation and control incapacity, due to problems in using online feedback for regulation. In a subsequent study conducted with a young adult sample (which was also more heterogeneous for IQ: 65–119), Glazebrook, Elliott, and Lyons (2006) recorded reduced acceleration velocities and doubled inter-trial variability in the amount of time required to reach peak velocity. These researchers conversely hypothesized a dysfunction in planning processes associated with the specification and timing of muscular force involved in the initial parts of reaching.

This latter planning-control dichotomy was firstly suggested by Woodworth as early as 1899 and has represented the foundation of some of the most influential models of limb control (Glover, 2004) and speed-accuracy trade-off in movements (Meyer, Abrams, Kornblum, Wright, & Smith, 1988). Glover's theory proposes that planning processes are responsible for selecting the appropriate motor program for an intended action, which is based on both the action goal and target characteristics, and that control processes conversely support movement execution, by monitoring discrepancies among the motor plan, the actual movement, and the target, as well as by quickly generating corrections for spatial errors. This need for corrections moreover may derive from planning errors or from noise in the neuromuscular system during execution. Along the same lines, Meyer et al. (1988) parsed movements into their component submovements, and identified an initial impulse, called “primary movement”, followed by an error correction phase. They also assumed that the initial impulse might be planned to end at the location of the target (or, just short of the target: see Elliott, Helsen, & Chua, 2001). The subsequent error correction phase was presumed to be based on feedback information and to be observable in discrete submovements with new accelerations (Chua & Elloitt, 1993). The combination of primary movement and corrective submovements represents the integration of planning and control and serves to optimize movement time. According to Meyer et al., this process is the result of a compromise between initial movement speed, in which faster movements are more likely to induce spatial errors, and corrective sub-movements, which may increase accuracy by decreasing limb velocity.

The extent to which movement is determined by these planning/initial impulse and control/corrective phases can be determined by examining kinematic variables at various stages of an action. In Glover's model, all the initial movement parameters are determined by planning mechanisms, whereas kinematic measures observed later in the movement are linked to control processes that smoothly and progressively take charge of the action.

With the aim of analyzing planning and control processes in pre-school children with autism, we identified an arbitrary end location for a primary movement where corrective submovements should begin. We therefore combined a spatial criterion in proximity of the target hole with a kinematic criterion, yielded by significant variations in the acceleration profile, as signalled by new “movement units”. The procedure required that participants transport an object (a rubber ball) from a given location to a target hole (located inside a see-through, Plexiglas box) in which the ball was to be dropped. The box edges represented an obstacle in proximity of the target and were considered a spatial marker to divide the initial phase from the final movement phases. Hence, all kinematic parameters in the initial parts of the movement were considered to depend on initial planning (primary movement) processes. Conversely, new movement units recorded within the box area were considered to represent corrective submovements and to depend on control processes.

Furthermore, the target box was located higher than the starting position and required a parabolic trajectory thereby: an upward movement to raise the hand above the box edges and a downward movement to bring the hand inside the box for the ball drop. This variable was presumed to increase the probability of corrections, which are more frequent in downward aiming movements (Lyons, Hansen, Hurding, & Elliott, 2006).

Lastly, the accuracy indicator was not determined by a precise hand position, because the ball could be thrown into the hole from any location within the box, but by wrist inclination, as this was a crucial variable for a successful drop.

Section snippets

Participants

We examined 12 pre-school children with autism (AS) (9 boys and 3 girls; M = 3.5 years; SD = 9.0 months with no associated mental retardation. These participants’ performance was compared to that of a control group (CONT) of 12 normally developing children and matched by gender and age; matching validity was confirmed by independent sample t tests (both t(22) < 1.0).

The AS participants were recruited from the Child Psychiatry Unit of our institute over a 12-month period. Inclusion criteria were: age

Calculation

Kinematic parameters were collected from the instant of ball contact to ball release. Movement onset was established as the beginning of an uninterrupted movement from ball support in the direction of the target hole; movement offset was established as the instant of ball drop, indicated by a g acceleration in ball velocity. Raw data were filtered with a 12 Hz Butterworth filter-zero phase 5th order. All variables were computed in Matlab®.

We calculated the following as general movement skill

Results

Movements showed comparable between-group trajectories, given that no differences in hand-distance-travelled were observed (F(1,23) < 1.0). Conversely, a consistent group effect for movement duration emerged, as the AS group's actions lasted nearly twice as long as those executed by the CONT group participants (M = 1495 ms vs. M = 813 ms; F(2,21) = 18.3; p < 0.001). Movement duration showed an inverse relationship with IQ (r = −0.54; p < 0.01), but group differences were significant irrespectively of the

Discussion

Kinematic recordings in a reach and drop task were compared between 12 preschool children with autism without mental retardation (AS) and twelve gender and age-matched normally developing children (CONT). Our aim was to investigate whether motor anomalies in autism may depend more on a planning ability dysfunction, as suggested by Glazebrook et al. (2006), or on a motor control deficit, as suggested by Mari et al. (2003) based on the results they obtained with their average-high functioning

Acknowledgements

We kindly thank all our medical doctors who collaborated in this research with the clinical assessment of children with autism: Dr. Elisa Mani, Dr. Laura Villa, Dr. Catia Rigoletto and Dr. Silvia Borini. We also thank Dr. Roberto Ripa for his important help in recruiting control participants. We are also grateful to Prof. Flavio Keller, Prof. Claes Von Hofsten and Prof. Umberto Castiello for their special suggestions.

This research has been funded by the FP6-NEST Adventure activities Specific

References (34)

  • R. Chua et al.

    Visual regulation of manual aiming

    Human Movement Science

    (1993)
  • L. Zwaigenbaum et al.

    Behavioral manifestations of autism in the first year of life

    International Journal of Developmental Neuroscience

    (2005)
  • T.M. Achenbach et al.

    Manual for the ASEBA preschool forms & profiles

    (2000)
  • American Psychiatric Association

    Diagnostic and statistical manual of mental disorders, 4th edition (DSM IV)

    (1994)
  • M.K. Belmonte et al.

    Autism and abnormal development of brain connectivity

    Journal of Neuroscience

    (2004)
  • R.H. Bruininks et al.

    Bruininks–Oseretsky test of motor proficiency

    (2005)
  • W.M. Bump

    The homogeneity of regression assumption in the analysis of covariance

  • O. Dinca et al.

    Systematic review of randomized controlled trials of atypical antipsychotics and selective serotonin reuptake inhibitors for behavioural problems associated with pervasive developmental disorders

    Journal of Psychopharmacology

    (2005)
  • D. Elliott et al.

    A century later: Woodworth's (1899) two component model of goal directed aiming

    Psychological Bulletin

    (2001)
  • M. Fabbri-Destro et al.

    Planning actions in autism

    Experimental Brain Research

    (2009)
  • C.M. Glazebrook et al.

    How do individual with autism plan their movements?

    Journal of Autism and Developmental Disorders

    (2008)
  • C.M. Glazebrook et al.

    A kinematic analysis of how young adults with and without autism plan and control goal-directed movements

    Motor Control

    (2006)
  • C. Glazebrook et al.

    The role of the vision for online control of manual aiming movements in person with autism spectrum disorders

    Autism

    (2009)
  • S. Glover

    Separate visual representations in the planning and control of action

    Behavioral and Brain Sciences

    (2004)
  • D. Green et al.

    The severity and nature of motor impairment in Asperger's syndrome: A comparison with specific developmental disorder of motor function

    Journal of Child Psychology and Psychiatry

    (2002)
  • R. Griffiths

    The ability of young children. A study in mental measurement

    (1970)
  • S.E. Henderson et al.

    Movement assessment battery for children

    (2007)
  • Cited by (0)

    1

    Tel.: +39 031 877111; fax: +39 031 877499.

    View full text