Cycling Interventions on Cognition in Children With Autism (Cycling&Autism)

Why Does Learning to Ride a Bike Improve Executive Functions in Children With ASD?

The goal of this clinical trial was to determine if a specific type of cycling exercise improved executive functions in school-aged children with autism spectrum disorder (both boys and girls, aged 8-10 years, no healthy volunteers).

The main questions it addressed were:

1. . Did learning to ride a real bicycle improve planning, flexibility, working memory, and inhibition more than stationary cycling?
2. .Was the benefit driven by (A) dynamic balance, (B) spatial updating, or both?

Researchers compared four arms to identify if dynamic balance and/or spatial updating were the active components using three interventions groups, namely learning to bicycle (LTB), bicycle treadmill (BT), cycling with training wheels (TW)-and one active control group (stationary cycling, SC).

Participants:

1. . Provided a small urine sample three times for BDNF brain-marker testing.
2. . Played four short tablet games (10 minutes each) at 1st intervention session , 4th intervention session and 8th intervention session to assess their Executive Functions.
3. . Attended four 45-minute cycling sessions per week for 2 weeks at their respective school.
4. .Continued all usual therapies; no medicine was administered

The trial was completed on 31 December 2024 (RGC Ref 18616522, HREC 2021-2022-0397).

Study Overview

• Status: Completed
• Conditions: Autism Spectrum Disorder (ASD)
• Intervention / Treatment: Behavioral: Learning to Bicycle; Behavioral: Bicycle Treadmill; Behavioral: Cycling with Training Wheels; Behavioral: Stationary Cycling

Detailed Description

BACKGROUND Executive functions (EF)-planning, cognitive flexibility, working memory, and inhibition-were markedly impaired in autism spectrum disorder (ASD). A 2021 pilot RCT by the same team (Med Sci Sports Exerc 2021;53:1417-24) demonstrated that only children who learned to ride a real bicycle (n=24) improved EF; stationary cycling did not. The cognitive gain therefore stemmed from skill acquisition, not aerobic fitness alone.

HYPOTHESIS

Improvements were driven by two mechanisms recruited during real-bike learning:

(A) dynamic balance (vestibular/proprioceptive load) (B) spatial updating during self-produced translation through space

TRIAL DESIGN This study was a four-arm randomized controlled trial (RCT) with equal allocation to three intervention groups-learning to bicycle (LTB), bicycle treadmill (BT), cycling with training wheels (TW)-and one active control group (stationary cycling, SC). Assessments were conducted at baseline (T1), mid-intervention (T2, one week), and post-intervention (T3, 2 weeks) to evaluate executive function (EF), BDNF and HRV.

Data Collection Each participant was asked to perform a set of neuropsychological assessments, where different EF components were recorded. Each assessment was lasted for 15 minutes followed by a 5-minute break to prevent cognitive fatigue in the participants. The neuropsychological assessments were conducted one hour before the commencement of the first exercise intervention (T1: baseline), the 4th intervention (T2: mid-intervention) and the last exercise intervention (T3: post-intervention). The mid-assessment provides valuable information about how much variance in EF is accounted for by variance in HRV records. The HRV assessment was conducted 10 minutes after the end of the first (T1), 4th (T2), and last (T3) interventions using Garmin Venus 2S watch, a wrist-based device validated for short-term HRV recordings. During a 5-minute seated rest period in a quiet environment, participants were asked to perform an on-demand Health Snapshot test (2-minute) to capture parasympathetic recovery post-exercise. To ensure comfortability and data quality, children were instructed to wear the Garmin Venus 2 watch only during the rest period and were asked to remain still and breathe naturally. The Garmin Venu 2's Elevate V4 optical heart rate sensor recorded inter-beat interval data, which were processed by the Garmin Connect app to compute RMSSD, These procedures were standardized across groups to minimize variability due to environmental or procedural factors. Meanwhile, urine samples from the participants were collected in 120-ml sterile cups at T1, T2 and T3. Upon completion, the samples were promptly transported from participants' homes to the laboratory of the Department of Chemical Pathology located at Princes Wales Hospital .

PARTICIPANTS Sixty-two participants with mild to moderate ASD, aged 8-12 years, were recruited from a special school for intellectual disabilities, Chinese YMCA of Hong Kong and a cycling company in Hong Kong with existing research links to the study team. Inclusion criteria were: (1) ASD diagnosis per DSM-5 and ADOS-2; (2) non-verbal IQ >40 (Wechsler Intelligence Scale for Children revised in China); (3) ability to follow instructions and perform interventions/assessments with assistance; (4) no additional regular physical activity (beyond school physical education) for 2 months prior; and (5) novice at two-wheel bicycle riding (<10 seconds independent riding). Exclusion criteria included: (1) color blindness; (2) medical conditions limiting exercise (e.g., asthma, seizures); (3) history of brain injury or complex neurologic disorders and (4) obesity [>95th percentile BMI]. Screening was conducted by a psychologist and a psychiatric physician, with parent-reported autistic traits assessed via the Social Responsiveness Scale, Second Edition (SRS-2). After screening, fifty-six eligible participants were randomized to one of four groups using block randomization (block size = 4) to ensure equal allocation. Upon the completion of the study, participant numbers per group were: LTB (n=13, 1 dropout at T2), BT (n=12, 2 dropouts at T2), TW (n=14), and SC (n=12, 2 dropouts at T2).

INTERVENTIONS All interventions were lasted for 2 weeks, comprising 8 sessions (4 sessions per week, 45 minutes each) in school halls/gymnasiums, led by a qualified cycling coach (Hong Kong Cycling Association) assisted by student helpers (1:1 staff-to-participant ratio). Each intervention session consisted of a 10-minute preparation and warm-up, 30-minute cycling training, and 5-minute cool-down. Exercise intensity was monitored by asking participants every 10 minutes during exercise to indicate their ratings of perceived exertion (target range:3-5). Cycling distance was measured by speedometers for non-stationary groups (LTB, TW) to ensure equivalent spatial exposure. Participants were positively reinforced verbally with compliments for their efforts during the interventions and their improvements were visualized through graphs kept in the child's bedroom at home (e.g., list of achievements: distance; able to use another set of wheels and able to self-launch). For the stationary cycling group, participants were provided a graph of the cumulative total time spend cycling or the equivalent distance. Apart from the LTB group, all participants in the other groups were taught how to ride a bicycle after T3 to recognize their contributions in the study.

Learning to Bicycle (LTB): Participants were asked to ride on a normal bicycle with training wheels. Participants were then progressed to a two-wheel bicycle. To keep participants on the learning curve, in case the participants have successfully mastered the cycling skill before the end of the intervention period, they were asked to ride through an obstacle course that WAS progressively more difficult to navigate. The obstacle course was designed by a focus group, which consisted of four physical education teachers from participating schools and one experienced cycling coach with more than 5 years of coaching experience.

Bicycle Treadmill (BT): Participants in this group first learned how to ride a normal bicycle with its back wheel and training wheels on a roller and the front wheel on the holder stand. Then they were progressed to ride the two-wheel bicycle with its back wheel on a roller (no training wheels), and finally to the conventional two-wheel bicycle on the rollers.

Cycling with Training Wheels (TW): Participants in this group were asked to ride a bicycle with training wheels throughout the whole study. They also rode through an obstacle course as they became more proficient at riding.

Stationary Cycling (SC): Participants in the group were asked to cycle on a stationary bicycle, with progress tracked via cumulative time/distance graphs.

TRIAL STATUS Recruitment: 62/62 (completed 30 Jun 2023) Last participant last visit: 30 Jun 2023 Data lock: 31 Aug 2024 Primary manuscript will be submitted to Autism Journal for review.

SAFETY No serious adverse events occurred; minor grazes (n=1) were treated on-site.

FUNDING & ETHICS Funded by RGC General Research Fund 18616522 (HK$814,430). Approved by HREC-EdUHK 2021-2022-0397 (15 Jun 2022).

• Study Type: Interventional
• Enrollment (Actual): 62
• Phase: Not Applicable

Contacts and Locations

Study Locations

• Hong Kong
  • Hong Kong, Hong Kong
    • Chinese University of Hong Kong

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: Child
• Accepts Healthy Volunteers: No

Description

Potential participants were recruited from the local special schools for intellectual disabilities. They werescreened with the following inclusion and exclusion criteria.

The inclusion criteria are:

1. age 8 - 10 years;
2. mild to moderate ASD diagnosis from physicians or psychologists based on the Diagnostic and Statistical Manual of Mental Disorders, 5th edition, (DSM-5)and Autism Diagnostic Observation Schedule, 2nd Edition (ADOS-2);
3. non-verbal IQ over 70 using a brief version of Wechsler intelligence scale for children (Chinese revised);
4. able to follow instructions with the assistance of research staff;
5. able to perform the requested physical activity intervention and executive function measures with the assistance of research staff;
6. no additional regular participation in physical activity other than school physical education classes for at least 3 months prior to the study;

(6) novice at riding a two-wheel bicycle (i.e., cannot ride the bicycle alone for more than 10 consecutive seconds).

Exclusion criteria are:

1. color blindness reported by parents;
2. history of other medical conditions that limit physical exercise capacities (e.g., asthma, seizure, cardiac disease);
3. history of brain injury including concussions, other complex neurologic disorders (e.g., epilepsy, phenylketonuria, fragile X syndrome);
4. on current or recent use of antidepressants and other psychotropic medications known to alter peripheral BDNF levels;
5. medical conditions of the bladder;
6. obesity (i.e. >95th percentile of age-gender specific BMI cutoff).

Study Plan

How is the study designed?

Design Details

• Primary Purpose: Basic Science
• Allocation: Randomized
• Interventional Model: Parallel Assignment
• Masking: Single

Number of Arms

4

Arms and Interventions

Participant Group / Arm	Intervention / Treatment
Experimental: Learning to Bicycle; Participants were asked to ride on a normal bicycle with training wheels. Participants were then progressed to a two-wheel bicycle. To keep participants on the learning curve, they were asked to ride through an obstacle course that was progressively more difficult to navigate.	Behavioral: Learning to Bicycle; 2 weeks, 4×45 min/week. Progressive removal of stabilisers on 16-inch bikes. Recruited both dynamic balance and spatial updating during real locomotion.; Other Names: LTB
Experimental: Bicycle Treadmill; Participants in this group first learned how to ride a normal bicycle with its back wheel and training wheels on a roller and the front wheel on the holder stand. Then they were progressed to ride the two-wheel bicycle with its back wheel on a roller (no training wheels), and finally to the conventional two-wheel bicycle on the rollers.	Behavioral: Bicycle Treadmill; 2 weeks, 4×45 min/week. Bike fixed on smart rollers. Recruited dynamic balance only.; Other Names: BT
Experimental: Cycling with Training Wheels; Participants in this group were asked to ride a bicycle with training wheels throughout the whole study. They also rode through an obstacle course as they became more proficient at riding.	Behavioral: Cycling with Training Wheels; 2 weeks, 4×45 min/week. Fixed stabilisers. Recruits spatial updating only.; Other Names: TW
Active Comparator: Stationary Cycling; Participants in the group were asked to cycle on a stationary bicycle, with progress tracked via cumulative time/distance graphs.	Behavioral: Stationary Cycling; 2 weeks, 4 ×45 min/week. Monark ergometer. Active control; recruited NEITHER dynamic balance nor spatial updating mechanism.; Other Names: SC

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Cognitive Planning	Planning: Participant's ability to plan was assessed using computerized PEBL Tower of London (PEBL TOL) task. Participants were presented a sequence of three different colored balls on three pegs of different lengths on a computer screen. During each of the 12 trials, participants were instructed to move the balls using a computer mouse to match the target peg arrangement in a minimum amount of moves according to the pre-specified rules. The level of the task difficulty and the restriction on number of moves were progressively higher over time. If there were three consecutive failed trial attempts, the task was halted. Scores for children's general planning ability were calculated by summing the correct tasks performed out of the 12 trials given, and the total score ranges from 0 to 36. Higher scores indicated better planning.	Baseline (T1: day 1), 4th Intervention Session (T2: Day 4), 8th Intervention Session (T3: Day 8)
Working Memory	Working memory: Participants' visual-spatial working memory was evaluated using the Corsi block tapping task. In the CBTT, participants observed and replicated a sequence of tapped blocks, starting with three blocks. Sequence length increased by one after every two trials. The task ended after two incorrect sequences of the same length, with the longest correctly repeated sequence recorded as the score. Higher scores indicated better working memory.	Baseline (T1: day 1), 4th Intervention Session (T2: day 4), 8th Intervention Session (T3: day 8)
Inhibition	The participants' ability to inhibit unwanted responses to changing stimuli was measured by a computerized Go/No-go (GNG) task. In the task, participants pressed a left or right key for corresponding arrows (Go response) but withheld responses for up arrows (No-go response). After 20 practice trials, they completed 300 trials: 220 Go trials (110 left, 110 right) and 80 No-go trials (26.7%), which were presented randomly via E-Prime 3.0 software. Each stimulus appeared for 500 ms, followed by a 1000 ms blank interval. Breaks of 2 minutes were offered after every 60 trials. No feedback was provided, and response times were recorded but not analyzed due to unreliable recording . False alarms (FA), defined as Go responses on No-go trials, measured inhibition, with lower FA rates indicating better inhibitory control.	Baseline (T1: day 1), 4th Intervention Session (T2: day 4), 8th Intervention Session (T3: day 8)
Flexibility	Children's Color Trail Making Test (CCTT) was employed to assess cognitive flexibility of the participants. Participants completed two parts of the TMT as quickly and accurately as possible. In Part 1 (CCTT-1), they connected numbered circles in ascending order using a single color (e.g., black 1-2-3). In Part 2 (CCTT-2), they alternated between connecting numbered circles of two colors (e.g., black 1, yellow 2, pink 3), requiring task-switching between color sets. Cognitive flexibility was quantified by the interference score, which was calculated by the difference of time between Part 2 and Part 1 (i.e. CCTT-2 minus CCTT-1), where lower scores indicate better cognitive flexibility, reflecting enhanced ability to shift between mental sets.	Baseline (T1), 4th Intervention Session (T2), 8th Intervention Session (T3)
Urinary BDNF level	Urine samples from the participants were collected in 120-ml sterile cups . Upon completion, the samples were promptly transported from participants' homes to the laboratory of the Department of Chemical Pathology located at Princes Wales Hospital.	Baseline (T1: day 1), 4th Intervention Session (T2: day 4), 8th Intervention Session (T3: day 8)

Secondary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Heart Rate Variability	The HRV assessment was conducted using Garmin Venus 2S watch, a wrist-based device validated for short-term HRV recordings. During a 5-minute seated rest period in a quiet environment, participants were asked to perform an on-demand Health Snapshot test (2-minute) to capture parasympathetic recovery post-exercise. To ensure comfortability and data quality, children were instructed to wear the Garmin Venus 2 watch only during the rest period and were asked to remain still and breathe naturally. The Elevate V4 optical heart rate sensor of the Garmin watch recorded inter-beat interval data, which were processed by the Garmin Connect app to compute RMSSD (Garmin, 2025),	Baseline (T1: day 1), 4th Intervention Session (T2: day 4), 8th Intervention Session (T3: day 8)

Collaborators and Investigators

• Sponsor: Chinese University of Hong Kong
• Collaborators: University of Leicester; San Francisco State University; Education University of Hong Kong; United Christian Hospital
• Investigators: Principal Investigator: Andy Choi-Yeung Tse, PhD, Chinese University of Hong Kong

Publications and helpful links

General Publications

• Ng JSK, Chau JPC, Chan AWK, Lui JKC, Cheng JWCH. A Nurse-Led Web-Based Home Asthma Education Program for Children and their Families: A Randomized Controlled Trial. J Pediatr Nurs. 2021 Jul-Aug;59:158-163. doi: 10.1016/j.pedn.2021.04.014. Epub 2021 Apr 21.

Study record dates

Study Major Dates

• Study Start (Actual): January 1, 2023
• Primary Completion (Actual): June 30, 2024
• Study Completion (Actual): December 31, 2024

Study Registration Dates

• First Submitted: December 4, 2025
• First Submitted That Met QC Criteria: December 19, 2025
• First Posted (Actual): December 22, 2025

Study Record Updates

• Last Update Posted (Actual): December 22, 2025
• Last Update Submitted That Met QC Criteria: December 19, 2025
• Last Verified: December 1, 2025

More Information

Terms related to this study

• Keywords: randomized controlled trial; children; autism spectrum disorder; BDNF; executive function; dynamic balance; cycling; actigraph; spatial updating
• Additional Relevant MeSH Terms: Mental Disorders; Neurodevelopmental Disorders; Child Development Disorders, Pervasive; Autism Spectrum Disorder
• Other Study ID Numbers: 2021-2022-0397; 18616522 (Other Grant/Funding Number: Research Grants Council (RGC), Hong Kong)

Plan for Individual participant data (IPD)

• Plan to Share Individual Participant Data (IPD)?: NO
• IPD Plan Description: According to the ethical approval form, we must not share any individual participant data to anyone other than the research team members.

Drug and device information, study documents

• Studies a U.S. FDA-regulated drug product: No
• Studies a U.S. FDA-regulated device product: No