Engaging in appropriate volumes of physical activity is important to our current and future health status. Being able to develop habits for physical activity during early years of life may have long lasting benefits throughout individuals' life course. For preschool-aged children, the World Health Organization recommends engagement in a range of activity of varying intensities (from light to vigorous intensities) for 180 minutes every day. However, one local study conducted by the Principal Investigator found that only 12.8% of children from Hong Kong met these recommendations. One key contributing factor to physical activity is one's fundamental movement skills (FMS), such as running, jumping, throwing, and kicking. FMS are considered the prerequisites or "building blocks" of many physical activities and sports. In addition to supporting an active lifestyle, researchers have shown that proficiencies in these skills are associated with better cardiovascular fitness and healthier weight statuses. Building children's fundamental movement skills proficiency from a young age is thus important. "Physical Development" is a development objective under the Hong Kong Kindergarten Education Curriculum. Hence all local preschools provide some form of education to develop children's FMS. However, most preschool teachers have not received formal training in physical education, and standardized curricula in the physical domain is currently unavailable. As such, preschool teachers' ability to instruct and assess children's development in these skills may be limited and suboptimal. Parents are significant influencers in children's development, therefore appropriate support from parents also contribute to the potential successes of the programme. In this study, we aim to address the aforementioned challenges by (1) providing preschool teachers professional development training in pedagogical knowledge and methods in the physical domain, (2) creating a set of teaching materials (in printed and digital formats) and methods that could be incorporated into the overall and preschool-based curricula, (3) creating a parental toolkit with contents mirroring the school materials, to enhance parents' knowledge and promote family-school collaboration, and (4) developing a smartphone-based artificial intelligence rating system to help preschool teachers conveniently and accurately assess children's motor skills, and provide appropriate instructional feedback.

Physical activity (PA) is related to better health and well-being of children. Research has also shown that PA is associated with young children's physical, cognitive, and social-emotional developments, which encompass three out of five objectives of pre-primary education in Hong Kong. By contrast, physical inactivity accounts for paediatric overweight/ obesity, which is on a rise globally. Such phenomena emerge as soon as in the first 6 years of lifespan. In 2019, the World Health Organization released their first global PA guidelines for children below 5 years old. World Health Organization (WHO) recommends that preschool-aged children should spend at least 180 minutes in a variety of physical activities at any intensity (i.e., light, moderate, or vigorous) daily, of which 60 minutes should be at a moderate intensity or above. However, few preschool-aged children in Hong Kong meet these standards. To address the aforementioned difficulties and challenges, we propose to develop and evaluate an intervention to improve the quantity and quality of PA in preschool-aged children.

To this end, four core intervention components will be applied in the proposed study: 1. Teacher professional development and support. Training and on-going support will be provided to in-service preschool teachers to equip them with knowledge, pedagogical skills, and assessment literacy pertaining physical literacy and FMS instruction. Specific intervention activities and contents will be co-designed with each participating school, and will be integrated to the curriculum and monthly themes of each school.

2. Development of a smartphone-based FMS-rating machines learning system. To allow efficient and objective assessment of children's FMS, an application will be developed to allow teachers and parents to rate children's FMS using a smartphone application. The application will be based on a machine learning system that was developed, and filed for patent by project team members. Additional information regarding the system will be provided in the sections below.

3. Teaching materials for teachers. A series of print and digital materials will be developed to support teachers' FMS instruction. The materials will cover key knowledge, pedagogical skills, games and activity ideas pertaining to PA and FMS, where these materials could be adopted into existing school-based curricula and allow for crossdomain instruction (e.g., infusing PA and FMS instruction into storytelling activities).

4. Parent materials. A set of materials on PA and FMS instruction will be developed for parents. The key concepts and contents will match the developed teaching materials; teachers' and parents' understanding for each development stage will be aligned. In turn, children will be able to receive non-conflicting feedback and instruction from teachers and parents. This form of engagement can foster strong family-school collaboration.

The required sample size was calculated based on a prudent expectation to achieve a small effect (Cohen's d = 0.2) in terms of primary outcome of accelerometer-measured physical activity in preschool-aged children. Using G*Power 3.1.9.4 with a 1:1 experimental versus control ratio, an alpha level at .05 and a power of .8, the required total sample size will be 156. To avoid potential contamination of intervention effects within schools, randomization will be conducted by clusters, at the school level. Based on previous work conducted in Hong Kong, the compliance rate of deployed accelerometers is approximately at 70%. Based on this rate, and from a cluster size of 15 (i.e., 5 children per grade), the effective average cluster size is 10.5. With an intraclass coefficient of .07 (He et al., 2021), the design effect is 1+(10.5-1)*0.07 = 1.67. Therefore, the sample size required is 260, and therefore 26 schools will be recruited to the main trial. From these schools, 13 of them will be randomised to the experimental group, and the remaining 13 will be allocated to the wait-list control group. Participation consent will first be sought at the school level.

Eligible preschools should be 1) government funded or subsidised, 2) coeducational so that the potential findings could be more generalizable in Hong Kong. Parents of all K1-3 children attending the preschools who have agreed to take part will be invited to take part in the research study. To enhance sample representativeness, we will aim at recruiting at even numbers of kindergarten from each region (i.e., Kowloon, Hong Kong Island, the New Territories). We will also purposefully invite preschool-aged children that are attending half-day, full-day, and long full-day preschools. For each participant, data will be collected on three measurement occasions, i.e., at baseline, the end of Year 1 (primary endpoint of the study), and two years after baseline. Randomisation will take place at the school (i.e., cluster) level after baseline measures have been taken. A random number generation sequence conducted by the research team will be employed for the randomisation of preschools to the experimental or control groups. Preschools randomized to the experimental group will be exposed to the intervention between the first two time points. The control group will receive no intervention during the same period, but will be exposed to the same components the following year. Parents who agreed to take part will be invited to complete a questionnaire at each time point. The FMS competence of their children will also be measured at all three times. A subsample of participants (i.e., 5 per grade per school) who agreed to take part will be invited to provide accelerometer measures for their physical activity (MVPA). This study will be based on Social Cognitive theory (SCT; Bandura, 2004). Within SCT, personal factors, the social environment, and behaviours are related intricately and reciprocally. Behaviours of preschool-aged children, by and large, are shaped by key socialising agents (i.e., parents and teachers) and the environments created by these key members. Therefore, SCT is an appropriate framework for the design of PA interventions. In the proposed study, intervention components were designed based on the theoretical underpinnings of SCT. Specifically, intervention components were designed to target six key SCT constructs, namely self-efficacy, behavioural capacity, outcome expectations, self-regulation, social environment, and physical environment.

Quantitative outcomes will be measured at all time points of the proposed study to evaluate the effectiveness of the intervention, using a clustered randomised controlled trial. Analyses will be conducted by assessors blinded to group allocation. The primary outcomes of the current trial are MVPA and FMS of preschool-aged children. Secondary outcomes include parent-child co-PA, parents' PA, children and parents' sleep time (accelerometry), children and parents' screen time, emotional well-being, socioemotional outcomes, physical health, family PA routines (parent questionnaires), and executive functioning (Head-ToesShoulders task). Teachers' teaching self-efficacy and competence will also be measured as a secondary outcome of the trial.

Inclusion Criteria:

• preschool children aged 3-6 with no history of neurological, psychiatric or physical illness

Exclusion Criteria:

• - preschool children aged 3-6 with with physical illness