Research has shown deficits in motor planning and the predictive control of movements (i.e. internal modeling) «Adams IL, Lust JM, Wilson PH, ym. Compromised moto...»1, «Wilson PH, Smits-Engelsman B, Caeyenberghs K, ym. ...»2 in children with DCD. This has led to the intervention approaches that support the internal modeling process and thus may improve motor learning. In recent years, motor imagery (MI) training and action observation (AO) training have been used for DCD «Scott MW, Wood G, Holmes PS, ym. Combined action o...»3. Pilot studies have shown the use of motor imagery and action observation training to enhance motor performance in children with DCD «Adams ILJ, Smits-Engelsman B, Lust JM, ym. Feasibi...»4, «Bhoyroo R, Hands B, Wilmut K, ym. Motor planning w...»5, «Mielikuvaharjoittelu saattaa vahvistaa motorista suoriutumista harjoittelemattomuuteen verrattuna lapsilla, joilla on keskivaikea tai vaikea kehityksellinen koordinaatiohäiriö (DCD). Vaikutuksen suuruus lienee samankaltainen kuin havaintomotorisen harjoittelun.»C. Moreover, preliminary evidence based on experimental studies (see below) «Scott MW, Emerson JR, Dixon J, ym. Motor imagery d...»6, «Marshall B, Wright DJ, Holmes PS, ym. Combined act...»7, «Scott MW, Wood G, Holmes PS, ym. Combined action o...»8 suggests that combined use of action observation and motor imagery (AO+MI) may be more efficient in enhancing motor performance in children with DCD compared to either simulation performed in isolation.
1. Motor imagery during action observation enhances imitation of everyday rhythmical actions in children with and without developmental coordination disorder
Scott et al «Scott MW, Emerson JR, Dixon J, ym. Motor imagery d...»6 investigated the effects of combined action observation and motor imagery (AO+MI) on intentional imitation in children with DCD through the kinematics of familiar rhythmical actions.
Researchers selected 25 voluntary children aged 7–11 years of which thirteen met the DSM-V diagnostic criteria for DCD, while twelve met the inclusion criteria for TD. Children who scored ≥20th percentile in MABC-2 met the criteria for the TD group and those with ≤16th percentile were allocated to the DCD group (p<.001). DCDQ '07 was used to confirm disruption of daily activities. Exclusion criteria were neurological or visual impairment and ADHD. An IQ > 70 was assumed given that all participants attended mainstream education with no diagnosis of learning disorders.
The experiment consisted of 48 trials. For each trial participants viewed a picture or video of a target movements consisting of everyday rhythmical pantomime actions (window washing, tooth brushing, face washing and paint brushing) performed at either a fast or slow pace in the vertical and horizontal plane. Following observation and/or imagery participants pantomimed the target action as accurately as possible. A magnetic motion sensor was fitted to the distal end of the second metacarpal bone of the dominant hand. Participants' kinematic data were sampled at 103 Hz in 3-D space for 4 s periods using a Minibird Magnetic Tracking System. Three different trial conditions were included: (1) intentional imitation (‘copy the action as closely as possible'), (=observe then imitate the target action); (2) intentional imitation with action observation and motor imagery (AO+MI) (‘from your own point of view please imagine performing the action in time with the observed action'), (=observe while imagining the same action before imitating); (3) Intentional imitation with motor imagery (MI); After observing the movie (‘from your own point of view please imagine performing the action while staring at the purple cross'), (=observe then imagine the action before imitating).
Measures: The mean response cycle times (ms) were calculated between peak kinematic positions across all peak positions available within a 4 s time for both horizontal and vertical actions and the mean cycle time ratio (%) (M) between slow and fast trials within each habitual speed.
Results: DCD vs. TD children. Kinematic analyses revealed TD children (M = 132%, SE = 2.59) imitated the observed cycle times significantly better than children with DCD (M = 121%, SE = 2.42, CI = 3.43–17.88%) (F(1,23) = 9.30, p = .006, ηp2 = 0.28). Imitation improved for AO + MI compared to the other two instructions (F(2,46) = 6.10, p = .004, ηp2 = 0.21). The mean cycle time ratios were for intentional imitation with AO + MI (M = 132%, SE = 2.59), for the intentional imitation with MI (M = 122%, SE = 2.58, p = .004, CI = 3.36–16.1%) and for the intentional imitation condition (M = 126%, SE = 1.92, p = .028, CI = 0.74–11.5%).
Within-group analyses found a significant advantage in DCD for AO + MI (M = 127%, SE = 4.58) compared to observe then imagine (M = 116%, SE = 2.77, p = .021, CI = 1.99–20.21%), (F(2,24) = 3.90, p = .034, ηp2 = 0.24). In TD children, imitation was significantly enhanced for AO + MI (M = 137%, SE = 2.08) compared to observe then imitate (M = 130%, SE = 2.48, p = .032, CI = 0.72–13.11%), F(2,22) = 2.52, p = .103, ηp2 = 0.187.
Authors' conclusion: Combined AO + MI instructions enhanced imitation ability in children with and without DCD compared to their natural imitation strategies and compared to instructions AO followed by MI. The AO + MI instruction was therefore an effective tool for improving instantaneous imitation in children both with and without DCD.
2. Combined action observation and motor imagery facilitates visuomotor adaptation in children with developmental coordination disorder.
Marshall et al «Marshall B, Wright DJ, Holmes PS, ym. Combined act...»7 studied the effects of a combined action observation and motor imagery (AO+MI) training for facilitating visuomotor adaptation and eye-hand coordination in children with DCD.
Twenty children aged 7 to 11 years (13 male, 7 female) were recruited through local DCD support groups. Potential participants (DCDQ'07 scores within the range of 15–55) completed the MABC-2 test. Children who scored ≤ 5th percentile and who, based on parent reports, did not suffer from any other general medical condition known to affect sensorimotor function and had no diagnosis of learning difficulties or ADHD, were asked to take part in the study.
Participants performed a virtual radial Fitts task. For this task, a 90° counter-clockwise visual feedback rotation was used that resulted in stylus movements along the x-axis producing equivalent cursor movements along the y-axis and vice versa. The goal of the task was to use a stylus to guide a cursor from a central home square to a highlighted target square and then back to the home square. Six targets were presented sequentially from left to right with the next target becoming highlighted each time the cursor returned to the central square. Testing was performed on touchscreen monitor. Eye-movements were monitored using eye tracking glasses. Participants were instructed to perform the task as quickly and accurately as possible. The experiment included pre-test trials, AO+MI trials and post-test trials.
Participants in the AO+MI group (6 male, 4 female) performed motor imagery of executing the task whilst they simultaneously observed videos of an adult performer completing the same visuomotor rotation task (hand moving the stylus). The video series consisted of three videos recorded at different stages of the learning experience as they performed 50 trials of the task; Early T1(trials 1 to 10), Mid T2 (trials 11 to 30), and Late T3 (trials 31 to 50). At the start of each video, a motor imagery script was presented in written form on the screen along with an audio-recorded narration. (Early: "I am watching the video on the screen. The hand in the video is mine and I am making the movements that I see. I can feel myself holding the pen and I can feel my arm and hand moving the cursor to the yellow squares"). After each AO+MI trial, participants immediately performed a physical practice trial. Participants in the control group (7 male, 3 female) watched clips of a nature documentary that contained no human motor content followed by an immediate physical practice trial. After intervention trials participants completed post-test trials, identical to pre-test trials.
Measures included completion time, a target-locking score (TLS), total path length and normalized jerk. The time taken (in seconds) to finish the entire trial (12 target hits), from leaving the home square at the start to returning to the home square after hitting the sixth target, was used as a measure of completion time. A target-locking score was calculated by subtracting the percentage of cursor fixation time from the percentage of target fixation time to create a ratio measure of the allocation of visual attention. A positive score reflects more time fixating on targets whereas a negative score reflects more time spent fixating the cursor. Normalised jerk was calculated as a measure of movement smoothness.
Results: AO+MI group produced significantly faster completion times than the control group at T2 (p = 0.002), T3 (p = 0.007) and post-test (p = 0.009) and a significantly greater TLS at T1 (p < .001), T2 (p < .001), T3 (p = .002) and post-test (p = 0.012) compared to the control group. The movements of the control group, however, were significantly more jerky compared to the movements of the AO+MI group, F(1, 17) = 31.98, p < .001, ηp2 = .65. Whereas both groups exhibited a predominantly ‘cursor-focused' visual strategy at pre-test (target-locking score of approximately -60%), the AO+MI group became almost totally ‘target-focused' at post-test (target-locking score of approximately 40%). In contrast, the control group were unable to progress (target-locking score just above 0%).
Authors' conclusion: After training, the AO+MI group showed faster task completion times, smoother movement kinematics and more effective eye-hand coordination compared to the control group. These results suggest that AO+MI interventions may help to alleviate such deficits and improve motor performance in children with DCD.
3. Combined action observation and motor imagery improves learning of activities of daily living in children with Developmental Coordination Disorder.
Scott et al «Scott MW, Wood G, Holmes PS, ym. Combined action o...»8 investigated the effect of a home-based, parent-led, AOMI intervention on the performance and learning of four ADLs (shoelace tying, cutlery use, shirt buttoning, a cup stacking task) in children with DCD.
Twenty-eight children aged 7 to 12 years (21 male, 7 female) with confirmed (n = 23) or suspected (n = 5) DCD completed the MABC-2 test. Children who had an official diagnosis (according to DSM-5 criteria) and also scored below the 5th percentile on the MABC-2 were classified as confirmed DCD and those without a confirmed DCD diagnosis but who scored below the 5th percentile were classified as suspected DCD. Exclusion criteria were any co-occurring medical condition known to impair motor function, or learning difficulties, or ADHD.
This study used a mixed measures design, in which two groups (experimental and control) completed pre-test, post-test and retention test measurements to determine the efficacy of a homebased AOMI intervention for children with DCD. Interventions were delivered at the participants' home by parents after receiving training in their respective intervention. After the pre-test, children were randomly allocated to either an AOMI group or a control group (both n = 14) via minimisation. (for more details, see ref.).
Pre-test: After completing MABC-2 and Movement Imagery Questionnaire for Children (MIQ-C), children attempted to complete 5 trials for each of the four ADLs. Before this, they were shown two videos of the tasks being performed with a conventional strategy. Children were then instructed to attempt to copy the technique shown as quickly and accurately as possible.
Training phase: AOMI group. Both parents and their children were introduced to the concept of AOMI, and parents were trained on how to deliver an AOMI intervention at home using a tablet computer with the AOMI intervention uploaded. The intervention involved children first reading an imagery script tailored to the task they were training. (e.g."The hands I see are mine…I am making accurate movements…I can feel holding the laces between my fingers and thumbs and I can feel myself tying…"). Children were instructed to imagine, simultaneously, the feelings and sensations associated with performing the movement during the video where they come to see a proficient child model performing the same ADLs attempted at pre-test. Following an AOMI trial, on-screen instructions then prompted the child to physically execute one repetition of the ADL. Children were instructed to complete four 40-minute training sessions per week for four weeks. In each training session they were asked to spend 10 minutes practicing each ADL, adhering to the repetitive structure of one AOMI trial followed by one physical practice trial, before moving on to the next ADL.
Control group: Tablets of control children and their parents were provided with a video game called Cut the Rope downloaded to use for the duration of the study ("as ‘gamifying' training to alleviate boredom associated with repetitive task practice"). This game was selected as playing it required an element of fine motor planning and execution but was unrelated to the ADLs being practiced. The children of control group followed a similar training structure to the AOMI group, however, adhering to repetitive structure of one trial of Cut the Rope followed by one physical practice trial, before moving on to the next ADL. To ensure adherence throughout the intervention, participants in both groups maintained contact with the research team via weekly video calls and email.
Post-test: After a 4-week home-based intervention, participants were invited for post-test measurements. They completed the MIQ-C and the manual dexterity component of the MABC-2 to assess whether any improvements in ADLs transferred to these tasks. Participants then attempted each ADL five times with the same performance measures recorded as at pre-test.
A retention test was accomplished two weeks after the post-test. For the duration of this 2-week retention period participants were asked to cease specific training of the ADLs but to continue with their typical weekly routines. The retention phase involved replicating the post-test protocol.
Measures: Performance times. On each visit children completed 5 trials of each ADL while being timed (secs). If children were unable to complete one trial of an ADL, incompletion times were imputed to allow analysis of these (incompletion time = all participants' pre-test Mean + 2.5*SD). Technique rating scales were developed ranging from 0–5 (strategy quantification, movement quality). For each task 0 represented an incompletion and 5 represented use of the same technique demonstrated in the videos.
Results: The AOMI group (M = 35.51, SD = 23.42) performed shoelace tying task significantly faster than the control group (M = 61.4, SD = 33.41, p = 0.045) at post-term, but not at the retention phase (p = 0.162). However, within a sub-sample of 18 children (9 per group) unable to complete this task successfully, 89% (eight out of nine) of children following the AOMI intervention were able to complete this task successfully by the post-test, while only 44% (four out of nine) of those following the control intervention (F(2,32) = 4.5, p = 0.018, ηp2 = 0.22). No group differences in performance times were found in three other ADLs.
The shoelace tying technique used by the AOMI group (M = 4.1, SD = 1) at post-test found to be significantly better than the technique used by the control group (M = 3.09, SD = 1.02) (p = 0.002). This effect was also evident at retention (p = 0.011); the AOMI group was significantly better than the control group. The AOMI group (M = 3.4, SD = 1.29) had a better technique also in cup stacking task than the control group at post-test (M = 2.53, SD = 0.91, p = 0.008). This group difference was also present at retention (p = 0.01). No group differences in technique ratings were found in Shirt buttoning or Cutlery use.
Authors' conclusion: While both groups improved their performances, AOMI was more effective in supporting the more complex ADLs (shoelace tying and cup stacking tasks), and the greatest benefits were found for children with the lowest proficiency in these tasks. The findings indicate that home-based, parent-led, AOMI interventions can aid the learning of complex ADLs in children with DCD, and may be particularly effective for facilitating the learning of motor skills that do not currently exist within these children's motor repertoire.