Secondary Logo

Journal Logo


The Effects of Aerobic Exercise on Academic Engagement in Young Children With Autism Spectrum Disorder

Oriel, Kathryn N. PT, EdD; George, Cheryl L. PhD; Peckus, Rebecca DPT; Semon, Amanda DPT

Author Information
doi: 10.1097/PEP.0b013e318218f149
  • Free


Autism is defined as “a neurobehavioral syndrome marked by qualitative impairments of social interaction and communication, and by restricted, repetitive, and stereotyped patterns of behavior.”1(p257) Stereotypic behavior may involve toe walking, hand-and-arm flapping, rocking or spinning the body, or finger flicking.2 According to the National Institute of Child Health and Human Development, health care providers envision autism on a spectrum, as a group of disorders with similar features.3 This group includes autistic disorder, Asperger syndrome, pervasive developmental disorder not otherwise specified, Rett syndrome, and childhood disintegrative disorder. With the prevalence of autism spectrum disorder (ASD) on the rise, the need for effective educational interventions for this population is increasing.4

While the importance of physical activity in children with ASD has been described in the literature and is currently recommended by the National Center for Physical Activity and Disability, children with ASD typically present with decreased physical activity levels and tolerance and developmental delay.58 This may be a result of the limited physical activity opportunities available for children with ASD.5 The limited opportunities may be attributed to the social and behavioral deficits that are often present in children with ASD that make participation with peers difficult.9 Although few studies have explored the physical activity patterns of children with ASD,5 social constraints, more than existing impairments, could be the limiting factor affecting physical activity.10

The effect of exercise on stereotypic behaviors in children with ASD has been explored in the literature. Kern et al11 examined the effect of exercise on 3 children with ASD between 7 and 11 years of age. Results indicated that stereotypic behaviors decreased following vigorous jogging (mildly strenuous, with the child showing increased breathing rate and/or flushed face) for 15 minutes, although little or no reduction was observed following mild exercise (ball playing, where an adult gently threw a ball to a child from a distance of 2–3 m, at a rate of 10–12 times per minute).11 Watters and Watters12 found similar results in 5 boys with ASD between 9 and 11 years of age. They found that the greatest decrease in stereotypic behavior was found following physical activity (jogging in the schoolyard alongside teachers for 8–10 minutes) as compared with academic work or television viewing. Similarly, a study examining three 11-year-old children (2 males and 1 female) with ASD found that stereotypic behaviors decreased after 15 minutes of continuous jogging. This study also examined carryover of the decreased stereotypic behaviors and found that the behavior carryover was limited to 90 minutes postintervention.13

Petrus et al4 conducted a current, systematic review of studies measuring the effects of exercise on stereotypic behaviors of children with ASD. The 7 studies included a total of 25 subjects. Twenty-two were older children (6–15 years old) with ASD, and 3 were younger children (3–5 years old). In the 3 studies including younger children, the exercise programs involved walking versus jogging, jogging, and jogging versus ball playing.11,14,15 Only jogging resulted in a decrease in stereotypic behaviors.15 Three of the 7 studies also measured academic performance following exercise.4 Watters and Watters12 reported no effect on academic responding following exercise, but 2 studies did report improved academic responding following jogging.15,16 In addition, Rosenthal-Malek and Mitchell16 documented an increase in on-task behavior after aerobic exercise.

The effect of exercise on academic performance was examined by Kern et al.15 They found that jogging, at a speed that increased the subjects' breathing rate and elicited flushed faces, decreased stereotypic behavior and increased appropriate play (ie, on-task time) and academic performance.15 Similar results were found by Elliot et al17 in their study examining 6 adults with ASD. These researchers found that aerobic activity increased appropriate behavior and decreased off-task time in this population. Powers et al18 found similar results in an 8-year-old child with ASD. They reported a decrease in stereotypic behavior and an increase in on-task performance following a roller skating intervention for 10 minutes.

Aerobic exercise has been shown to improve correct responding, improve on-task time, and reduce stereotypic behaviors in older children (6–15 years old) with ASD, although few studies have explored such an effect in young children (3–5 years old).4,1118 The purpose of this study was to determine whether participation in aerobic exercise before classroom activities improves academic engagement and reduces stereotypic behaviors in young children with ASD.


Study Design

This study used a within-subjects crossover design, with a treatment condition and a control condition, across 4 early intervention classrooms. Two of the 4 classes were randomly assigned to the treatment condition and 2 to the control condition for the first 3 weeks of the study. During the subsequent 3 weeks of the study, each class received the opposite condition. The treatment condition included 15 minutes of running/jogging followed by a classroom task. The control condition included participation in a classroom task that was not preceded by aerobic exercise. The number of stereotypic behaviors demonstrated, the percentage of on-task behavior, and the numbers of correct/incorrect responses given during academic tasks for each child were recorded throughout the duration of the study, regardless of condition, to allow for comparison.

According to Spencer et al,19 the crossover design minimizes the effects of preexisting differences between classrooms, such as academic levels of students, age/grade levels, as well as differences in curricula. In this design, each student receives both treatments and serves as his or her own control, which also minimizes the possibility of differential attrition within treatment conditions. Finally, teacher effects are accounted for since teachers provide the instruction in both conditions.19 In special education classroom research studies, the crossover design has been used successfully “to address problems inherent in between-classrooms matching.”1923


This study was approved by the Institutional Review Board at Lebanon Valley College. A purposive sample of convenience was used. The participants in this study were recruited from 4 early intervention autistic support classrooms. The 24 children in these early intervention classrooms had a diagnosis of ASD and were between the ages of 3 and 6 years. The families of all children attending these classes were sent a letter explaining the study and a consent form. All children who returned a signed parental consent form were included in the study. Ultimately, 7 males and 2 females between the ages of 3 and 6 years (mean age, 5.2 years) participated. Seven of the children had a formal diagnosis of autism, while 1 had a diagnosis of intellectual disability, and 1 had a diagnosis of developmental delay. Although 2 of the children had a primary diagnosis of intellectual disability and developmental delay, all children in the early intervention autistic support classrooms met the educational criteria for ASD as defined by the Individuals with Disabilities Education Improvement Act of 2004.24 Otherwise, they would not be enrolled in these classrooms.

To qualify for enrollment in these early intervention autistic support classrooms, children must present with more severe symptoms along the continuum of ASD. Characteristics of the children participating in this study included resistance to change in environment or change in routine, lack of social interaction with peers, lack of eye contact, and absence or delay in communication. Some of the children in this study used speech to communicate, while others were nonverbal. In addition, some children used sign language as a means to communicate.

Treatment Condition

The treatment condition consisted of 15 minutes of running/jogging as a group. A consistent period of running/jogging was maintained through the use of prompts that were established by the child's classroom teacher (ie, edible reinforcements, verbal cueing). If the researchers were unable to prompt a child to run, jumping on a minitrampoline was used as an alternate mode of sustained aerobic activity. The researchers returned the child to running/jogging as quickly as possible. An increased breathing rate and observation of a flushed face were used to substantiate sufficiently strenuous activity,16 since measurement of heart rate was not tolerated by the participants in the study. The researchers also completed an exercise log after each session, which included the amount of time exercised, motivational strategies used to maintain running/jogging, and any additional comments regarding the child's performance.

When the 15 minutes of running or jogging was complete, the children participated in gentle seated stretching and were given a cup of water. The children were then assisted back to their classroom by either a classroom aide or their teacher. The children began their classroom activity immediately upon returning to the class.

Control Condition

The control condition consisted of participation in a classroom task that was not preceded by exercise. Classroom teachers were asked to discourage aerobic activity prior to the classroom task being assessed while children participated in the control condition.

The early intervention classrooms were highly structured, beginning with either 1:1 academic probing of students or “table time” with up to 4 children at each table. Academic probes addressed a variety of skill sets (ie, learning readiness, communication, motor skills, and academic).25 Examples of academic probes included solving a 4- to 6-piece puzzle, putting shapes in a container, cutting a piece of paper, communicating a word given a teacher prompt (eg, “say (or sign) ‘cereal’,” or asking student to hand the teacher the correct picture from 3 to 5 picture cards). During academic probing, the student was also asked to respond to teacher directives such as “stand up” and “go to table.” Table time primarily consisted of writing tasks. Children would write or trace their names, or Velcro letters to spell their names.

Classroom tasks remained the same in both the treatment and control conditions. No attempt was made to change classroom tasks in either condition.

Classroom Data Collection

Before beginning this study, the second researcher received permission to videotape children with ASD during academic tasks. This videotape was used solely to train undergraduate student observers. Training began with review of definitions for the behaviors to be recorded (ie, correct and incorrect academic responding, on-task behavior, stereotypic behavior). Then, student observers practiced recording data while watching the videotape. Practice continued until students were reliably recording classroom data.

Undergraduate student observers and the second researcher were randomly assigned to observe 1 or 2 children per session. Whenever possible, 2 observers would be assigned to observe and independently record data for the same child, to ensure interrater reliability. Observers were assigned to different children each session. In the event that a child was absent, that observer would then independently collect data on another child, which provided additional reliability data.

The second researcher and student observers were all blind to who exercised prior to classroom instruction. All observers collected 2 sets of classroom observation data per session; each set was 15 minutes in length.

Dependent Measures

Four dependent variables were measured: (1) correct academic responses, (2) incorrect academic responses, (3) stereotypic behaviors, and (4) on-task behavior. A correct academic response was scored, using a frequency count, when a child responded correctly to a directive given by the teacher. An incorrect academic response involved an incorrect response or no response to a directive given by the teacher. A frequency count was also taken of stereotypic behaviors observed; in these classrooms, the stereotypic behaviors included hand-and-arm flapping, body rocking, and toe walking.

Children were considered to be on task when seated and producing academic responses. Children were considered to be off task when they did not produce an academic response and were engaged in disruptive behavior (eg, crying, out of seat, playing with objects). When off-task behavior would begin, the observer would note the time, followed by noting the time again when the child resumed on-task behavior. Then, the percentage of time on task was calculated by subtracting the time off task and dividing by 15 minutes. Table 1 displays the operational definitions.

Operational Definitions

Data Analysis

Interrater reliability. Intraclass correlation coefficients were calculated to examine reliability between raters in the measurement of correct/incorrect responses, on-task time, and number of stereotypic behaviors observed.

Treatment versus control condition. The Wilcoxon signed rank test was used to compare differences between correct/incorrect responses, on-task time, and number of stereotypic behaviors in the traditional and exercise condition. This nonparametric test was used because the parametric assumptions of homogeneity and normality were not met. The SPSS statistical program was used and the level of significance was set at α = .05.


Interrater Reliability

Participants were observed by more than 1 rater to establish interrater reliability on 83% of observations (10 of the 12 days). Intraclass correlation coefficients (2,k) were calculated for correct responses, incorrect responses, time on task, and number of stereotypic behaviors observed, and were found to be 0.97, 0.84, 0.96, and 1.0, respectively. These rates were consistent with other studies that reported interrater reliability. Researchers reported collecting interrater reliability data in 50% to 100% of the sessions and reported reliability rates ranging from 70% to 100%.11,1315

Treatment Condition Versus Control Condition

Correct/incorrect responses. Seven of the 9 participants improved in correct responding following the exercise condition. The mean percentage of correct responding was found to be 71.49 on control days versus 82.57 on treatment days. These improvements were found to be statistically significant (P < .05). Figure 1 displays the mean differences for correct responding between the control and treatment phases of the study for each of the 9 participants.

Fig. 1
Fig. 1:
Percentage of correct responding during the control and treatment phases for each participant.

On-task time. No significant differences were found between the 2 conditions for on-task behavior. The mean percentage of time on task during control days was 94.48 versus 95.92 on treatment days. While statistical significance was not observed, visual analysis of the data revealed that 5 of the 9 participants improved in on-task time following the exercise condition. Figure 2 displays the mean differences for time on task between the control and treatment phases of the study for each of the 9 participants.

Fig. 2
Fig. 2:
Percentage of time spent on task during the control and treatment phases for each participant.

Stereotypic behaviors. No significant differences were observed between conditions for stereotypic behaviors, although only 5 of the 9 participants demonstrated stereotypic behaviors on at least 1 observation day. Table 2 provides a summary of the total number of stereotypic behaviors observed during the treatment and control conditions. While no statistical significance was observed, 4 of the 5 participants demonstrated fewer stereotypic behaviors during the intervention phase of the study. Table 3 summarizes the results of this study.

Total Number of Stereotypic Behaviors Observeda
Summary of Results


In this study, 7 of the 9 participants improved in correct responding following the treatment condition. It is important to note that the 2 children who did not improve in correct responding may not have expended as much energy when exercising. Exercise logs indicate that these 2 children were more difficult to motivate to exercise and, thus, spent less time exercising than their peers.

While we found statistically significant improvements in correct responding following 15 minutes of running, similar to other studies described in the literature in which jogging continued for 20 minutes, shorter durations of exercise should be explored.15,16 It would be beneficial to educators and school-based physical therapists to know whether shorter durations of exercise are capable of producing the same gains in academic responding, as shorter durations of exercise would be even more feasible to implement during the school day. Although academic responding was not measured by Celiberti et al,14 they found decreased stereotypic behaviors following jogging for 6 minutes in a 5-year-old boy with ASD. Further research is warranted to determine the duration of exercise needed as well as the intensity to produce academic and behavioral changes in young children with ASD.16

Whereas statistically significant improvements were not observed in the amount of time on task or in stereotypic behaviors, the trends observed were encouraging. Five of the 9 participants (55.6%) improved in on-task time. This trend is consistent with the few other studies that have examined the effect of exercise on on-task time in children with ASD.17,18 Similar findings of improved on-task time have also been described in children developing typically after exercise.26

Given that the children participating in this study were all enrolled in early intervention autistic support classrooms, it was unexpected that only 5 of the 9 children demonstrated stereotypic behaviors. Even among the 5 children who demonstrated stereotypic behaviors, the stereotypic behaviors observed were few (Table 2). Those stereotypic behaviors observed included hand-and-arm flapping, rocking, and toe walking. Whereas stereotypic behaviors were observed, there were less stereotypic behaviors observed following exercise. The total number of stereotypic behaviors across participants decreased from 11 during the control days to 2 during the treatment days, which may be a clinically significant finding.

The few stereotypic behaviors observed may have been a direct result of the structure provided within these autistic support classrooms. Teachers used immediate and consistent verbal prompts (eg, “quiet hands”) when stereotypic behavior began. Use of these prompts may be reducing or minimizing the demonstration of stereotypic behaviors, such as hand flapping, over time.

Further research should also examine the duration of treatment effects. In this study, improvements in academic responding continued for 30 minutes following aerobic exercise. Kern et al15 found that self-stimulatory behavior always decreased following jogging and did not return to baseline until 90 minutes later. While these researchers also reported that on-task behavior and academic responding always increased following jogging, the duration of these behaviors was not reported. Similarly, Rosenthal-Malek and Mitchell16 found statistically significant increases in correct responding and number of tasks completed following aerobic exercise but did not examine the duration of these effects.

Although the results of this study are encouraging, limitations must be noted. This study was of a short duration, so it is unclear whether improvements in academic responding would increase, maintain, or decrease over time. The lack of measurement of exercise intensity was also a limitation of this study. Whereas the measurement of heart rate using a pulse oximeter was the original intent of the authors, the participants in this study were unable to tolerate the use of the device. The device was given to the classroom teachers 4 weeks before the initiation of the study in an attempt to improve participant tolerance of its use, although when the study began, the teachers continued to report an increase in inappropriate behaviors with the use. While we used observation to monitor increased breathing and flushed faces; consistent with Rosenthal-Malek and Mitchel,16 using a more objective means to quantify exercise intensity should be considered in future studies.

Finally, since few studies have examined the effect of aerobic exercise on the academic performance of young children with ASD, further research is recommended to replicate these results. The National Autism Center's National Standards Project lists exercise as an emerging treatment for individuals with ASD and recommends further research in this area.25 Findings from this report also recommend that treatments be applied in the “real world” versus laboratory settings to determine whether the results found in laboratory settings will generalize.25 Therefore, it is recommended that future studies of the effects of aerobic exercise on the academic performance of young children with ASD be conducted by physical therapists and teachers who educate and support these children daily.


Results of this study indicate that aerobic exercise prior to classroom activities may improve academic responding in young children with ASD. These preliminary results are consistent with similar findings in older children. Results of this study help substantiate the importance of aerobic exercise in young children with ASD. The 15 minutes of aerobic exercise was feasible and easy to implement, offering a strategy for improving academic responding.


1. Heward W. Exceptional Children: An Introduction to Special Education. 9th ed. Upper Saddle River, NJ: Merrill; 2009.
2. Hooper SR, Umansky W. Young Children With Special Needs. 5th ed. Upper Saddle River, NJ: Merrill; 2009.
3. National Institute of Child Health and Human Development. National Institutes of Health. Published 2008. Accessed April 21, 2009.
4. Petrus C, Adamson SR, Block L, Einarson SJ, Sharifnejad M, Harris SR. Effects of exercise interventions on stereotypic behaviours in children with autism spectrum disorder. Physiother Can. 2008;60(2):134–145.
5. Pan C, Frey GC. Physical activity patterns in youth with autism spectrum disorders. J Autism Dev Disord. 2006;36:597–606.
6. Todd R, Reid G. Increasing physical activity in individuals with autism. Focus Autism Other Dev Disord. 2006;21(3):167–176.
    7. Berkley S, Zittel L, Pitney L, Nichols S. Locomotor and object control skills of children diagnosed with autism. Adapt Phys Act Q. 2001;18(4):405–416.
    8. Manjiviona J, Prior M. Comparison of Asperger's syndrome and high-functioning autistic children on a test of motor impairment. J Autism Dev Disord. 1995;25(1):23–39.
    9. Fox KR, Riddoch C. Charting the physical activity patterns of contemporary children and adolescents. Proc Nutr Soc. 2000;59:1–8.
    10. Llewellyn A, Hogan K. The use and abuse of models of disability. Disabil Soc. 2000;15(1):157–165.
    11. Kern L, Koegel RL, Dunlap G. The influence of vigorous versus mild exercise on autistic stereotyped behaviors. J Autism Dev Disord. 1984;14(1):57–67.
    12. Watters RG, Watters WE. Decreasing self-stimulatory behavior with physical exercise in a group of autistic boys. J Autism Dev Disord. 1980;10(4):379–387.
    13. Levinson LJ, Reid G. The effects of exercise intensity on the stereotypic behaviors of individuals with autism. Adapt Phys Act Q. 1993;10:255–268.
    14. Celiberti DA, Bobo HE, Kelly KS, Harris SI, Handleman JS. The differential and temporal effects of antecedent exercise on self-stimulatory behaviour of a child with autism. Res Dev Disabil. 1997;18(5):139–168.
    15. Kern L, Koegel RL, Dyer K, Blew PA, Fenton LR. The effects of physical exercise on self-stimulation and appropriate responding in autistic children. J Autism Dev Disord. 1982;12(4):399–412.
    16. Rosenthal-Malek A, Mitchell S. Brief report: the effects of exercise on self stimulatory behaviours and positive responding of adolescents with autism. J Autism Dev Disord. 1997;27(2):193–201.
    17. Elliot R, Dobbin A, Rose G, Soper H. Vigorous, aerobic exercise versus general motor training activities: effect on maladaptive and stereotypic behaviors of adults with both autism and mental retardation. J Autism Dev Disord. 1994;24(5):565–576.
    18. Powers S, Thibadeau S, Roase K. Antecedent exercise and its effect on self stimulation. Behav Residential Treat. 1992;7(1):15–22.
    19. Spencer VG, Scruggs TE, Mastropieri MA. Content area learning in middle school social studies classrooms and students with emotional or behavioral disorders: a comparison of strategies. Behav Disord. 2003;28(2):77–93.
    20. Brigham FJ, Scruggs TE, Mastropieri MA. The effect of teacher enthusiasm on the learning and behavior of learning disabled students. Learn Dis Res Pract. 1992;7:68–73.
      21. George CL. Effects of response cards on performance and participation in social studies for middle school students with emotional and behavioral disorders. Behav Disord. 2010;35:200–213.
      22. Mastropieri MA, Scruggs TE. Mnemonic social studies instruction: classroom applications. Rem Spec Educ. 1989;10:40–46.
        23. Scruggs TE, Mastropieri MA, Bakken JP, Brigham FJ. Reading versus doing: the relative effects of textbook-based and inquiry-oriented approaches to science learning in special educational classrooms. J Spec Educ. 1993;27:1–15.
        24. Individuals with Disabilities Education Improvement Act of 2004, 20 USC §300.8 (2004).
        25. The National Autism Center's National Standards Project: Findings and Conclusions. 2009.
        26. Mahar MT, Murphy SK, Rowe DA, Golden J, Shields AT, Raedeke TD. Effects of a classroom-based program on physical activity and on-task behavior. Med Sci Sports Exerc. 2006;38(12):2086–2094.

        aerobic exercise; attention; autism spectrum disorder; child; educational activities; physical therapy/methods; special education; stereotypic movement disorder/therapy

        © 2011 Lippincott Williams & Wilkins, Inc.