The Effect of Physical Activity Classes on Motor Skill in 12 - 24-Month-Old Children

Background: Children with enhanced fundamental movement skills may bene-fit from improved physical, social and psychological development, resulting in an increased likelihood of an active lifestyle in later years. Aim: We inves-tigated the effects of a nine-week, child-centred, physical activity programme on cognitive and motor skills in typically developing 12 - 24-month-old toddlers. Methods: In a randomised control trial, 90 toddlers (age 17.0 ± 2.6 months; 52.2% male) were split into two groups stratified by age and sex. The intervention completed was either nine weeks of one-hour per week physical activity classes (n = 45; EXP) or normal physical activity (n = 45; control). Prior to and following the intervention period, safety skills (nine-skill test battery), anthropometric measures (mass and height), motor and cognitive development (Bayley Scales of Infant Development) were assessed. Results: EXP improved overall safety skills score (P = 0.04), toddlers’ abilities to climb over a small-runged A-frame while using a cylinder grip and safe face-the-slope dismount (P = 0.001), and the execution of a safety roll down a foam wedge (P = 0.02). Improvements in development as measured by the Bayley’s Scales were attributed to typical development rather than the intervention. Conclu-sions: A 9-week, physical activity programme improved toddlers’ safety skills but not overall cognitive or motor development.


Introduction
Promoting young children's physical activity levels and motor skill proficiency positively contributes to their physical, social and psychological development, resulting in an increased likelihood of an active lifestyle (Lubans et al., 2010). Physical activity is associated with improving children's motor skill proficiency (Barnett et al., 2008b;Sääkslahti et al., 2004) and, when they grow to adolescence, young children with better motor skill proficiency have 10% -20% higher chance of participating in vigorous physical activity (Barnett et al., 2009), enhanced cardiovascular fitness (Barnett et al., 2008b) and greater perceived sports competence (Barnett et al., 2008a). Children with better motor skills also show enhanced academic (Bittmann et al., 2005) and cognitive skills (Piek et al., 2006). Additionally, associations between motor skill development and injury rate have been reported (Agran et al., 2003). Infants 15 -17 months old have the highest injury rates of all children before 15 years of age, possibly because the development of the upper body occurs earlier than that of the lower body, thereby enabling children to access more hazards, without having the motor skills to avoid injury (Agran et al., 2003). Fall-related injury accounts for 56% of the hospitalisations of children aged birth to four years in New Zealand (Safekids New Zealand, 2006). Thus, enhancing balance and coordination in infants may be an effective strategy to reduce the risk of fall-related injury.
The effects of physical activity on motor skills and physical activity levels of typically developing young children have been examined in a few intervention studies (Matvienko & Ahrabi-Fard, 2010;Reilly et al., 2006). Reilly et al. (2006) found that a 24-week physical activity intervention of three 30-min sessions per week promoted a significant improvement in 4-year-olds' (n = 545 at baseline) gross motor skills 6 and 12 months after the intervention. Matvienko & Ahrabi-Fard (2010) reported significant improvements in 5 -7-year-old children's (n = 70) fundamental motor skills following a 4-week intervention that involved a 90-min physical activity session and a 15-min walk per day. A literature review by Altunsöz (2015) suggests that at least 540 min spent on developing children's fundamental movement skills are likely to produce a medium to large effect size. Altunsöz (2015) suggests that developmentally appropriate intervention programmes can improve young children's motor skills and that early intervention is desirable, but more research is needed regarding retention effects and interventions for typically developing children. Furthermore, little is known regarding the effects of physical activity interventions on younger children (birth-5 years).
Research suggests that between 4% -10% of preschool children are not meeting physical activity guidelines that recommend 180 min of physical activity a day (Goldfield et al., 2012;Hnatiuk et al., 2012). Although there is limited research regarding New Zealand preschool children's physical activity, previous research suggests that teachers may lack knowledge and confidence of how to promote it in childcare (Kolt et al., 2005) and statistics on childhood obesity in New Zealand indicate that sedentary behaviour in homes and early childhood settings is a mitigating factor (Ministry of Health, 2014 Therefore, we aimed to examine the effects of a physical activity intervention on cognitive, motor, and safety skill development and on balance in typically developing toddlers in New Zealand. We hypothesised that the intervention would positively affect cognitive and motor skill development and improve safety skills and balance in these toddlers. A further aim was to examine the effect of age on typical development of cognitive and motor skill in these New Zealand children.

Participants
Sample size was estimated using data from Eickmann et al. (2003) and an ap-

Inclusion Criteria
Participants were required to be available for the 9-week duration of the study and to attend one 60-min class per week with their toddler, if assigned to the intervention group. Participants received their randomised group allocation, achieved with randomisation software (https://www.randomizer.org/form.htm) where separate randomisations based on gender were completed. In addition, participants needed to be available for both baseline and post-intervention assessments and be proficient in English. Toddlers were required to meet the criteria for typical development of balance function adapted from Hsu et al. (2009), needed to be able to stand unaided for approximately 20 s, and had not attended Jumping Beans classes or similar programmes prior to the study.

Demographic and Anthropometric Information
Demographic information about the toddlers and parents and anthropometric  (Table 2). Body mass of each toddler was measured using scales accurate to 0.1 kg and standing height was measured using a stadiometer.  (Bayley, 2006). Administration of the test took between 15 -40 min depending on the age and ability of the child, with older, more developed children requiring a longer testing period due to their further progression through the scales. The total sample was split into three age groups that corresponded to the ages set out by the BSID-Screening Test Manual (Bayley, 2006) that relate to the cut scores i.e. three categories of development: "competent", "emerging" and "at-risk" corresponding to certain raw scores for each subscale.

Safety Skills
The toddlers were assessed in nine safety skills (Table 1), and attempted each skill a maximum of three times, and without assistance from the researcher or parent. The assessment criteria for each skill were based on pilot testing and input from various physical activity instructors. The level of competency for each skill was assessed using a 6-point Likert scale: 0. not attempted; 1. fully assisted; 2. partially assisted; 3. supported; 4. independent; 5. mastery.

Physical Activity Intervention
The intervention involved nine weeks of Jumping Beans "Toddler Beans" classes. Child-centred "Toddler Beans" classes cater to children aged approximately 12 -24 months and are 60 min in duration. These classes incorporate movement to music, ball skills, movement through the custom-designed gym equipment and games using fine manipulative equipment, and have a special focus on learning safety skills (Foster & Hartigan, 2006; Figure 1). Activity classes led by qualified instructors begin with free play on the custom-designed equipment. Movement to music, that includes cross-lateral patterning and socialisation aspects, follows

Statistical Analysis
Statistical analysis was completed using the Statistical Package for the Social Sciences (Chicago, IL). All continuous data were inspected visually and statistically for normality and variance. Normally distributed data with equal variance is described using mean ± SD and non-normally distributed data using median [25, 75 percentiles]. Categorical data is described using frequency percentages. Independent t-tests (for parametric data) and Mann-Whitney tests (non-parametric data) were used to determine any differences between descriptive characteristics at baseline. Independent t-tests were used to determine the differences between treatment groups in the change variable of each major dependent variable.
Analysis of variance (ANOVA) was used to determine the difference between conditions across various age groups. Post-hoc analysis with Bonferroni adjustments determined where the differences lay.
Correlation between outcome measures and anthropometric and demographic variables was assessed using either Pearson's or Spearman's correlation coefficients, dependent on whether data were parametric or not. An r-value of ±0.1 represents a weak correlation, ±0.3 a moderate correlation, and ±0.5 a strong correlation (Field, 2009

Descriptive Characteristics
The intervention and control groups were closely matched at baseline, with the only notable difference being birth order (Table 2).

Safety Skills
Total Safety Skill Score On average, EXP demonstrated a significant improvement in the post-intervention safety skills total score (P = 0.04). There was a medium effect of the physical activity programme on the post-intervention safety skills total score in EXP (P = 0.02, d = 0.53) which was also reflected in the change in total score.
Safety skill scores at post-intervention were greater than the total safety skills scores at baseline in both the intervention group (P < 0.01) and the control group (P < 0.01; Figure 2). Age group did not account for the changes in total safety skills (P = 0.12), however, the change in total safety skills score was significantly correlated with being the youngest child in the family (r pb = 0.22, P = 0.04).

Individual Safety Skills Scores
All individual safety skills improved over the duration of the study within both treatment groups (P < 0.01). The ability to climb over the small-runged A-frame while using a cylinder grip and safe face-the-slope, leg-over dismount (Skill F; P = 0.001) and the execution of a safety roll (Skill G; P = 0.02) were significantly improved for those toddlers in the intervention group (Table 3).
While age group did not account for the change in Skill F (P = 0.19) nor on the change in Skill G (P = 0.32), the changes in Skill F and G were both positively correlated with being the youngest in the family (r pb = 0.25, P = 0.02 and r pb = 0.29, P = 0.01 respectively).

Regression analysis
A hierarchical multiple linear regression showed that birth order significantly predicted the change in total safety skills score (P = 0.04, adjusted R 2 = 0.038), with only 3.8% of the variance in the change in total safety skills score predicted by knowing whether the child was the youngest born or not. Knowing the sex of the toddler and whether they were the youngest in the family (P = 0.02, adjusted R 2 = 0.064) predicted 6.4% of the variance in the change in total safety skills score. A forward step-wise linear regression demonstrated that birth order significantly predicted the change in Skill F score (P = 0.02, adjusted R 2 = 0.051), however, as indicated by the adjusted R 2 , only 5.1% of the variance in the change in Skill F score could be predicted by knowing whether the child was the youngest born or not. Birth order also significantly predicted the change in Skill G score (P = 0.007, adjusted R 2 = 0.072). However, only 7.2% of the variance in the change in Skill G score could be predicted by knowing whether the child was the youngest born or not.

BSID-Screening Test
Cognitive ability subscale One significant post-intervention difference was found between treatment  groups, in terms of cognitive ability: in Age Group 2 only, toddlers from the control group demonstrated greater improvements in cognitive ability than those from the intervention group (EXP vs. CON, 1.6 ± 4.5 vs. 5.2 ± 2.4, P = 0.01, d = −0.25). No other significant cognitive differences were found between treatment groups, however, within both the intervention and control groups, Age Group 1 toddlers demonstrated a significantly greater improvement in cognitive ability (TOTAL, 4.9 ± 4.7, P < 0.001; EXP, 4.8 ± 5.0, P = 0.001; CON, 5.0 ± 4.6, P = 0.001) than those in Age Group 3 (TOTAL, −1.2 ± 3.7; EXP −1.7 ± 2.6; CON, −0.8 ± 4.5). Within the control group and within the total sample, Age Group 2 cognitive ability improved more than Age Group 3 (TOTAL, 3.3 ± 4.0, communication were also greater for children in Age Group 2 (TOTAL, 2.9 ± 4.3; CON, 4.0 ± 3.4) than those in Age Group 3, within the control group and total sample.

Correlations
There were high positive correlations between the change scores of all subscales of the BSID (P < 0.01). All subscales were negatively correlated with children's height and age at baseline (r = −0.35 to r = −0.59; P < 0.01). Changes in cognitive ability (r = 0.24), receptive communication (r = 0.21) and expressive communication (r = 0.31) were positively correlated with the child being in-home care (P < 0.05). Due to the high correlations between the subscales of the BSID Test, further analysis was completed using the change in total development score. Similarly, due to the high correlation between height at baseline and age group (r = 0.57, P < 0.001) further analysis was completed using age group only. Age Group significantly predicted the change in total development score (P < 0.01, adjusted R 2 = 0.24). Being cared for at home improved prediction of the change in total development score (P < 0.01, adjusted R 2 = 0.28). Therefore, being in the younger age groups contributes most to predicting the improvement in development, but the day-to-day environment of the child also contributes to this prediction.

Discussion
This was the first randomised control trial that examined the effects of a child-centred physical activity programme on overall development, safety skills and balance in typically developing 12 -24-month-old toddlers, in New Zealand.
The main finding was that the intervention improved toddlers' score within a battery of safety skills, but did not affect overall cognitive or motor development.
Moreover, younger children (12 -18 months) tended to be more likely to show greater improvements in all developmental subscales as compared to older Safety skills improved in the toddlers as a result of the physical activity classes.
In particular, the ability to climb over a small-runged A-frame while using a cylinder grip, safe face-the-slope dismount, and the execution of a safety roll down a foam wedge was significantly improved. Although this study did not examine the effect of learning safety skills on the risk of injury from falls, anecdotal evidence suggests that attending these types of physical activity classes has some success in both reducing falls and injury from falls (Foster & Hartigan, 2006).
Notably, toddlers who were youngest in their family, regardless of treatment group, were more likely to have a greater improvement in the total safety skills score and in particular Safety Skill F and G. The reasons underpinning the association between birth order and safety skills development in this study are unknown, however, it is possible that factors such as time spent with parent or siblings and older peers or enhanced parental learning may contribute to this association. More research is warranted to examine these issues further.
Non-significant improvements in cognitive and motor development occurred between baseline and post-intervention for the total sample. Possibly, these improvements can be attributed to typical development. This is supported by the age-group analysis that showed younger toddlers (12 -18 months) were more likely to have greater improvements in all subscales of development than older toddlers (18 -24 months). As the early work of Bayley (Bayley, 1936) suggests, development, particularly motor control, occurs very rapidly until approximately 21 months and then slows down. Therefore, it may be more appropriate to design interventions after children reach 21 months to help factor out changes due to typical development. However, as development of neural pathways occurs most rapidly in the first two years, there is an argument for the earlier the better (with regards to physical activity intervention).
A longer or more intense intervention may have had more impact on the toddlers' cognitive and motor skills (Altunsöz, 2015;Matvienko & Ahrabi-Fard, 2010;Reilly et al., 2006). While many intervention studies have concentrated on children with neurological or motor deficits or as prevention programmes for children at risk of obesity (Angulo-Barroso et al., 2008;Bluford et al., 2007;Valvano & Rapport, 2006), more research is needed regarding the design and effect of interventions for typically developing children (Altunsöz, 2015), particularly young children and toddlers. Moreover, the BSID screening test, which was designed predominantly to detect if children are progressing according to normal expectations (Bayley, 2006), may not have been sensitive enough to detect the impact of a nine-week intervention in this particular typically developing sample.
Another key result of this study was that almost one third (27.8%) of the change in overall development can be predicted by knowing the age of the child and whether they are primarily in a home-care environment. Children in childcare are less likely to achieve daily requirements for physical activity than child-  (Campbell & Hesketh 2007). There is an established link between parental education, parental physical activity and eating habits and child outcomes (Østbye et al., 2013) however literature regarding the associations between attending childcare and physical activity is equivocal. Knowing whether a child attends childcare or not, may not be sufficient to determine the contributing effect on physical activity or development. A number of studies have reported that children attending early childhood centres with supportive environments, including opportunities to be active, appropriate play equipment, and a sufficient amount of space both indoors and outdoors, are more likely to achieve higher levels of moderate to vigorous physical activity (MVPA) than children attending less supportive early childhood centres (Bower et al., 2008;Gubbels et al., 2012). Gubbels et al. (2012) reported that attending an early childhood centre between the ages of one and two years was positively associated with a greater increase in body mass index. Sugiyama et al. (2012) reported that children who attended early childhood centres were mostly sedentary and spent between 12 -36 min per day in MVPA, results supported by Pate et al. (2008) who reported that children were engaged in MVPA for less than 3% of observation intervals. Further research is required to examine the effectiveness of teachers' practices in supporting infants' and toddlers' physical activity and development in the childcare setting. There were a number of limitations with this study that should be addressed for future research. The short-duration intervention may have limited the potential improvements in all of the safety skills and overall development, or, it may be that toddlers require longer interventions than older children to enhance fundamental movement skills. Matvienko & Ahrabi-Fard (2010) showed improvements in fundamental movement skills in 5 -7-year-olds following a 4-week physical activity intervention; we see a need for further research regarding the effectiveness of interventions, particularly for typically developing toddlers. Due to the time constraints of this study, it was not possible to achieve reliability and validity testing prior to data collection. Therefore, unestablished methodology was used to assess safety skills, ergo, data presented is not necessarily reliable and valid. However, the testing was conducted by an investigator who was blinded to whether children were in the experimental or control groups, and not involved in delivery of the classes. Finally, due to the differences between the sample in this study and the demographics of New Zealand, any results from this study cannot be generalised to wider populations. However, it is tempting to speculate that the physical activity intervention may have been more beneficial in lower socioeconomic groups as our cohort came from a predominantly affluent, New Zealand European background. Research into the impact of physical activity classes within different socioeconomic and ethnic groups is certainly warranted; in New Zealand children living in socioeconomically deprived areas (65% of Māori and 78% of Pacific Island people live in socioeconomically deprived neighbourhoods) are more likely to be obese and tend to

Conclusion
The aim of this study was to examine the effects of a 9-week, child-centred physical activity programme on overall development, safety skills and balance in 12 -24-month-old toddlers. The physical activity programme significantly improved total safety skills score, particularly the ability to climb over the small-runged A-frame while using a cylinder grip and safe face-the-slope dismount and the execution of a safety roll down a foam wedge. However, there were no significant improvements in overall development that can be attributed to the intervention. Younger children (12 -18 months) tended to be more likely to show greater improvements in all developmental subscales as compared to older children (18 -24 months) and it is possible that any development improvements in overall development are a result of typical development. Almost one third (27.8%) of the change in overall development can be predicted by knowing the age of the child and whether their day-to-day environment is mostly home care with their parent or other adults, or not. This suggests that future studies should examine children's day-to-day environments, which have the potential to affect a variety of outcome variables.