Investigating the Matching Relationship Between Physical Exercise and Stereotypic Behavior in Children With Autism

Andy C. Y. Tse; Venus H. L. Liu; Paul H. Lee


Med Sci Sports Exerc. 2021;53(4):770-775. 

In This Article

Abstract and Introduction


Purpose: Physical exercise has been shown to be effective in reducing stereotypic behaviors in children with autism spectrum disorder. One possible mechanism concerns the matching hypothesis between exercise and behavior. The present study sought to examine this matching exercise–behavior relationship.

Methods: Participants (N = 21, 17 males and 4 females, M age = 11.07 ± 1.44 yr, M height = 1.46 ± 0.99 m, and M weight = 40.60 ± 8.25 kg), with observable forms of hand-flapping and body-rocking stereotypic behaviors, underwent three separate days of conditions, one for the control condition, one for the 10-min ball-tapping exercise condition, and one for the 10-min jogging condition, in a randomized order. The frequency of each type of stereotypic behavior was videotaped from 15 min before to 60 min after the exercise.

Results: Results revealed that only hand-flapping stereotypic behaviors were significantly reduced in the ball-tapping exercise condition (P < 0.017), whereas only body-rocking stereotypic behaviors were significantly reduced in the jogging exercise condition (P < 0.017). However, the behavioral benefit diminished at 45 min after the respective exercise.

Conclusion: Physical exercise should be topographically matched with stereotypic behavior to produce desirable behavioral benefits in children with autism spectrum disorder.


Stereotypic behaviors (SB), also known as repetitive and restrictive behaviors, are one of the iconic symptoms in children with autism spectrum disorder (ASD). According to DSM-5 diagnostic criteria,[1] individuals with ASD must exhibit at least one symptom from a heterogeneous set of SB of interest. These behaviors include repetitive speech and motor movements, such as hand flapping, body rocking, and body spinning in circles.[2,3] The detrimental effects of these behaviors are well evidenced in the population. For example, many studies showed that SB could interfere with children's social engagement with their peers[4,5] and disrupt their learning ability and the learning environment.[6–8] Therefore, it is important to develop an effective intervention that can ameliorate the problems of SB.

Currently, certain interventions, such as operant conditioning,[9] differential reinforcement,[10] and sensory integration,[11] are commonly used to reduce SB in children with ASD. Although the effectiveness of the interventions is clearly demonstrated in research, the interventions are generally costly because of professional requirements. Physical exercise, a type of intervention that is of low cost and easy to implement, has received growing research attention.[12,13] A recent review conducted by Tarr et al.[14] confirmed the beneficial effects of physical exercise for reducing SB with moderate effect sizes (ES), which supported physical exercise as a potential treatment for SB in children with ASD.[14] Given the behavioral benefits of physical exercise, a logical subsequent research question is to determine the optimal exercise type.

Previously, we conducted a study to investigate the matching relationship between physical exercise and SB.[15] We used a ball-tapping exercise as an intervention for a group of children with ASD who exhibited both hand-flapping and body-rocking SB. Results revealed that only hand-flapping behaviors were significantly reduced after the intervention. We contended that this may be due to the matching stimulation effect.[16] According to the theory, the "matched stimuli' (i.e., the matched exercise) may have aroused the participants with desired sensory stimulation and therefore decreased the need to engage in SB.[16] This notion was supported by other studies.[17–19] For example, Piazza et al.[18] showed that the hand-mouthing behavior of the participants was effectively reduced with the presence of the stimuli (e.g., green ball or rocking dinosaur) that provided similar kinesthetic sensory consequences to the behavior as shown by functional behavior analysis.[18] In our previous study,[15] however, there are some serious methodological limitations existed in our study. For example, only one exercise intervention (i.e., ball tapping) was used to examine its effect on two SB (i.e., hand-flapping and body-rocking behaviors). Without an additional exercise intervention, we were unable to confirm whether the reduction of hand-flapping behaviors was due to the matching characteristic of the exercise intervention or was simply attributable to the physical arm fatigue of the participants. Other limitations included the non-counterbalanced sequence of the conditions (i.e., the control and the intervention), the lack of baseline motor proficiency test, and the limited assessment time points (only pre- and posttest), all of which may also have diminished the reliability of the findings.

The present study addressed these limitations by adding one more exercise intervention (jogging), counterbalancing the sequence of conditions, and conducting baseline motor skill assessments. We also used HR monitors and RPE[20] to continuously measure the physical activity level of each participant. In addition, more assessment time points were inserted to examine the sustainability of behavioral effects of each exercise. The purpose of the study was to determine the behavioral effects of exercise that topographically matched the SB. Considering the findings of our previous study,[15] we hypothesized that physical exercise that topographically matched the SB would produce a larger effect in decreasing that particular SB and that the effect would be reduced across time.