Association between motor skills and a healthy diet: a cross-sectional study of rst-grade schoolchildren in Japan

Background: Childhood motor skills are important not only for the physical and mental health of children, but also for the prevention of future lifestyle diseases. This study aimed to investigate how motor skills among rst-grade children in Japan are associated with dietary and lifestyle habits, after adjustment for various confounding factors. Methods: First-grade children (aged 6-7 years) attending three public elementary schools in Tokyo, Japan (n=884), participated in this cross-sectional study. Homeroom teachers distributed self-administered questionnaires to parents and children. Questionnaires focused on lifestyle habits and required completion of a 1-day dietary record of meals that children ate at home. School lunch consumption was also assessed. Motor skills were measured by the New Physical Fitness Test (NPFT). Physique was calculated using Rohrer’s index formula: weight (kg) / height (cm) 3 × 10 7 . Multiple regression analysis was used to investigate the association between NPFT score and determinant factors. We also examined the association between NPFT score and the amount of energy derived from a healthy diet versus snacks. Results: NPFT scores were signicantly and positively correlated with involvement in exercise lessons (boys, β = 0.131, P = 0.006; girls, β = 0.121P = 0.012), total energy intake (boys, β = 0.096, P = 0.041; girls, β = 0.145, P = 0.003), and outside playtime in boys (β = 0.135), and negatively correlated with Rohrer’s index in girls (β = -0.097, P = 0.047). Moreover, the amount of energy derived from a healthy diet showed positive correlations with NPFT score (boys, β = 0.120, P = 0.011; girls, β = 0.137, P = 0.005). Conclusions: Children’s motor skills were associated with the Rohrer’s index, involvement in sports lessons, outside playtime, and total energy intake, particularly that derived from a healthy diet. These results suggest that a well-balanced diet including grains, vegetables, sh and meat, fruits, and milk, is important for improving children’s motor skills. signicant. Participant characteristics were compared in relation to NPFT score quintile groups by gender with weighted one-way analyses of variance for continuous variables, and the Mantel–Haenszel test for trend for categorical variables. A multiple regression analysis was conducted using the forced entry method for the investigation of items related to NPFT score quintile. In the multiple regression model, NPFT score was entered as an independent variable and the following were entered as dependent variables: Rohrer’s index, sleep duration, frequency of outside playtime, time spent time spent watching time spent video gaming or participation in lessons (exercise, classroom learning, or frequency of eating per and total daily When the covariates, Rohrer’s index was classied into eight quintiles and sleep duration classied into eight categories with 30-minute increments. The frequency of outside playtime as ‘1’ for one of timeframes of after all responses time eight with 15-minute increments. Total time spent watching television or participating in video gaming or reading comics was aggregated and classied into four categories with 60-minute increments. A value of ‘1’ was allocated for children engaged in exercise, classroom learning, and/or music, whereas children not engaged in lessons were allocated dummy variables with a value of ‘0.’ A value of ‘1’ was allocated for frequency of eating breakfast per week responses of “7 times a week,” whereas responses of “0– “3–4,” or times per week” were allocated dummy variables with a value of ‘0.’

We hypothesized that children who receive more energy from healthy food groups have a higher physical tness. To test this, the present study aimed to investigate how motor skills among rst-grade children in Japan are associated with a healthy diet, after adjustment for lifestyle habits and various confounding factors.

Methods
A cross-sectional study of children and their guardians during the rst year of compulsory education was conducted.

Participants
Self-administered questionnaires and dietary records of meals that children ate at home were distributed to 1,982 pairs of rst-grade students aged 6-7 years and their guardians. Participating children were those who started school at three public primary schools in Tokyo between April 2013 and April 2018. Responses were obtained from 1,018 pairs (51.4%). Of these, 884 pairs (51.2 % boys and 48.8 % girls) were analysed after questionnaires with missing answers were excluded (Fig 1).

Data collection
The researcher responsible for the survey explained the aim and content of the survey to the principal, deputy principal, and class teachers in each school. The class teachers distributed envelopes containing consent forms, lifestyle habit questionnaires (to be lled out by the participants), and dietary records of meals that children ate at home to all rst-grade children, except those in the special class for handicapped children. The survey forms were distributed annually between late May and early June, from 2013 to 2018, and were collected two weeks later. Consent forms and questionnaires were lled out at home, placed in an unsigned envelope, and sent to the researcher by each school.

Measures
Motor skills (New Physical Fitness Test score) Each school measures eight items during the rst term of every school year, based on the MEXT New Physical Fitness Test Implementation Guideline [16]. Included items are grip strength, sit-up exercises, sit and reach, side-step, 20-metre shuttle run, 50-metre run, standing long jump, and softball throw. Each measurement is allocated a score according to sex, in accordance with the Implementation Guideline. The total score for these eight items, or "NPFT score," was used for analysis of motor skills in the present study.
(3) Sit and reach Sit and reach measurements were performed using a stand made of two boxes approximately 25 cm in height that were joined across by a piece of cardboard across the top of both boxes. Children sat with both legs straight in front of them, between the two boxes, with their back and buttocks ush against a wall. They were instructed to place both hands at on the nearest side of the cardboard, and to expand the chest with elbows straight to lengthen the spine. Children then bent forward from the starting position, without removing their hands, in order to slide the box as far forward as they could reach. The distance that the box was moved from the starting position was then measured with a ruler. This exercise was repeated twice, and the best result was recorded. Boys were allocated the following point values: ≤ 14 cm = 1 point,   15-18 cm = 2 points, 19-22 cm = 3 points, 23-26 cm = 4 points, 27-29 cm = 5 points, 30-33 cm = 6 points, 34-37 cm = 7 points, 38-42 cm = 8 points, 43-48 cm = 9 points, and ≥ 49 cm = 10 points (16) . Girls were allocated the following point values: ≤ 17 cm = 1 point, 18-20 cm = 2 points, 21-24 cm = 3 points, 25-28 cm = 4 points, 29-32 cm = 5 points, 33-36 cm = 6 points, 37-40 cm = 7 points, 41-45 cm = 8 points, 46-51 cm = 9 points, and ≥ 52 cm = 10 points [16].
(4) Side-step A central line was drawn on the oor with two parallel lines drawn 100 cm to either side of the central line. Children were instructed to stand with their feet on either side of the central line and, when given the signal to start, to step sideways to straddle the right line, return to straddle the central line, and similarly step sideways to straddle the left line. The exercise was continued until children stepped either outside of or on the lines. Children repeated this exercise as many times as possible for 20 seconds, and the number of times any line was crossed was recorded. This test was repeated twice, and the best result was recorded. Boys were allocated the following point values: ≤ 17 times = 1 point, 18-21 times = 2 points, 22-25 times = 3 points, 26-29 times = 4 points, 30-33 times = 5 points, 34-37 times = 6 points, 38-41 times = 7 points, 42-45 times = 8 points, 46-49 times = 9 points, and ≥ 50 times = 10 points (16) . Girls were allocated the following point values: ≤ 16 times = 1 point, 17-20 times = 2 points, 21-24 times = 3 points, 25-27 times = 4 points, 28-31 times = 5 points, 32-35 times = 6 points, 36-39 times = 7 points, 40-42 times = 8 points, 43-46 times = 9 points, and ≥ 47 times = 10 points [16].
(5) 20-metre shuttle run Children were instructed to run back and forth between two parallel lines placed 20 m apart in time with auditory prompts from a CD player. When prompts were played at xed intervals, the children ran 20 m to the opposite line. When their feet touched or went past the line, they turned around to await the next prompt, and repeated the exercise. The intervals between auditory prompts were shortened approximately every minute, and children attempted to reach the opposite line before hearing the next prompt. The number of times that each child was able to run 20 m in time with the prompts was recorded as the total number of complete times. This exercise was performed once. Boys were allocated the following point values: ≤ 7 times = 1 point, 8-9 times = 2 points, 10-14 times = 3 points, 15-22 times = 4 points, 23-32 times = 5 points, 33-44 times = 6 points, 45-56 times = 7 points, 57-68 times = 8 points, 69-79 times = 9 points, and ≥ 80 times = 10 points (16) .
Characteristics and lifestyle habits of children Participating children were asked how frequently they played outside (every day, sometimes, or never). Guardians were asked to provide the following information about the children: height (cm); weight (kg); waking time; bedtime; frequency of eating breakfast per week (0-2, 3-4, 5-6, or 7 days); whether or not breakfast and dinner are eaten with an adult (parent, grandparent, and/or any other adult); participation in weekly lessons; time spent watching television per day; participation in video gaming or reading comics; time spent studying (minutes per day); and guardian's employment status (stay-at-home parent, part-time/casual work, or works ve or more days per week). Weekly lessons were categorised into the following three groups, depending on lesson content: exercise (e.g., baseball, soccer, swimming, gymnastics, basketball, tennis, or ballet), music (e.g., violin, piano, orchestra, drums, or chorus), and classroom learning [e.g., painting, science, language, juku (cram school), calligraphy, abacus, or religious studies]. Lesson time per week was calculated based on the number of times each session was completed per week and the time spent at each session, as indicated by the guardian. A child's physique was calculated using the following Rohrer's index formula: weight (kg) / height (cm) 3 × 10 7 .

Children's dietary habits
(1) Breakfast, dinner, and between-meal eating Guardians of participating children provided information regarding the menu, ingredients, weight, and estimated dietary intake (including breakfast, lunch, dinner, and between-meal eating) of each child. This information was entered into the dietary record in as much detail as possible for an average school day.
(2) Lunch (school lunch) Guardians of participating children entered information about meals eaten at home into the dietary record, but were unable to detail what the children ate during school lunches. In order to gather this information, 2-3 investigators were assigned to each class at each primary school to observe the children and survey their school lunch consumption.
Japanese school lunches are provided every day based on school lunch implementation standards [17] in accordance with Dietary Reference Intakes for Japanese [11] set forth by the Ministry of Health, Labour and Welfare, and these standards are followed with regard to the amount of energy, protein, lipid, sodium, calcium, iron, vitamins A, B1, B2, and C, and dietary bre present in the lunches. Each school was surveyed for one day, and days with popular menus and thus higher eating rates than usual were avoided to ensure similar conditions across schools. Surveys were performed on days that white rice was served as the children's grain dish. Investigators were registered dieticians or students in the process of becoming registered dieticians. The amount of leftover food was recorded based on the amount of food in containers before serving, the amount of food left in containers after serving, and the amount of leftover food when the entire class nished eating their lunch. The intake of each individual child was ascertained in four stages for each dish, together with the number of second helpings for each child. Investigators observed children from the time they were served each dish up until the point they returned the trays. Consumption rate for each dish was evaluated using ve categories (0% (food untouched), 25%, 50%, 75%, and 100% (completed)). Lunch intake per child was calculated using the following formulas and was considered to be the lunch intake for each individual child. The energy content of surveyed meals was calculated using nutritional value calculation software (Excel Eiyo-kun, Version 8.0), with values based on the 2015 Standard Tables of Food Composition in Japan (Seventh Revised Version) [18]. Dietary intake was categorized by food group based on the Japanese Food Guide Spinning Top, which was formulated by the Ministry of Agriculture, Forestry and Fisheries and the Ministry of Health, Labour and Welfare [18]. Food consumed by participating children was divided into six food groups, according to the criteria in the Japanese Food Guide Spinning Top, and numbers of servings were calculated. One serving (SV) was de ned as the following: 40 g of carbohydrates for grain dishes; 70 g of the main ingredient in vegetable dishes; 6 g of protein in sh or meat dishes; 100 mg of calcium in milk or dairy products; 100 g of the main ingredient in fruit dishes; or 80 kcal of energy in snacks, confectionaries, or sweetened beverages. Children's intakes of breakfast, dinner, and between-meal eating were obtained from dietary records and combined with lunch intake to calculate daily intake.
(4) Breakdown of energy intake Information on energy intake was examined in detail to assess the food content of children's diets. Energy intake from grain dishes, vegetable dishes, sh and meat dishes, milk and dairy products, and fruit was de ned as the amount of energy derived from a healthy diet, based on food groups designated by the Japanese Food Guide Spinning Top [19]. Energy intake from snacks was de ned as the amount of energy derived from snacks, confectionaries, and sweetened beverages. The combination of these two categories was de ned as total energy intake. Confectionaries were classi ed according to the 2015 Standard Tables of Food Composition in Japan (Seventh Revised Version) [18]. Although the Standard Tables of Food Composition [18] classi ed ice cream as a dairy product, this study classi ed it as a confectionary. Similarly, this study categorized juice beverages that were not 100% fruit juice or that were from concentrate as sweetened beverages, in accordance with the Standard Tables of Food Composition [18] de nition of sweetened beverages.

Statistical analysis
Statistical analysis was performed using IBM SPSS Statistics for Windows (Version 23.0). A P value less than 0.05 was considered statistically signi cant. Participant characteristics were compared in relation to NPFT score quintile groups by gender with weighted one-way analyses of variance for continuous variables, and the Mantel-Haenszel test for trend for categorical variables.
A multiple regression analysis was conducted using the forced entry method for the investigation of items related to NPFT score quintile. In the multiple regression model, NPFT score was entered as an independent variable and the following were entered as dependent variables: Rohrer's index, sleep duration, frequency of outside playtime, time spent studying, time spent watching television, time spent video gaming or reading comics, participation in lessons (exercise, classroom learning, or music), frequency of eating breakfast per week, and total daily energy intake. When inputting the covariates, Rohrer's index was classi ed into eight quintiles and sleep duration was classi ed into eight categories with 30-minute increments. The frequency of outside playtime was set as '1' for responses of "almost every day (4-5 times a week)" for even one of the timeframes of morning, early break, lunch break, or after school; all other responses were assigned '0.' Study time was classi ed into eight categories with 15-minute increments. Total time spent watching television or participating in video gaming or reading comics was aggregated and classi ed into four categories with 60-minute increments. A value of '1' was allocated for children engaged in exercise, classroom learning, and/or music, whereas children not engaged in lessons were allocated dummy variables with a value of '0.' A value of '1' was allocated for frequency of eating breakfast per week responses of "7 times a week," whereas responses of "0-2," "3-4," or "5-6 times per week" were allocated dummy variables with a value of '0.' Moderator variables for investigation of the relationship between NPFT score and daily energy intake included the following: Rohrer's index, sleep duration, frequency of outside playtime, time spent studying, time spent watching television, time spent video gaming or reading comics, participation in lessons (exercise, classroom learning, or music), and frequency of eating breakfast per week. Model 1 de ned the explanatory variable as the amount of energy derived from a healthy diet. Model 2 de ned the explanatory variable as the amount of energy derived from snacks. Model 3 de ned the explanatory variable as the amount of energy derived from both a healthy diet and snacks.
This study complied with the Strengthening the Reporting of Observational Studies in Epidemiology -nutrition epidemiology (STROBE-nut) guidelines (see STROBE checklist, Additional le).
[Insert Table 1 here] Comparison of children's characteristics and lifestyle habits in the ve NPFT score quintiles indicated that boys with higher NPFT scores had signi cantly higher heights (P < 0.001) and weights (P < 0.001). Boys also had signi cantly longer exercise lesson times (P < 0.001) and signi cantly higher frequencies of outside playtime (P < 0.001; Table 1). [Insert Table 2 here] Girls had signi cantly longer exercise lesson times (P = 0.010) and signi cantly lower Rohrer's index values (P = 0.039; Table 2). [Insert Table 3 here] Comparison of children's dietary intake in the ve NPFT score quintiles indicated that boys with higher NPFT scores had signi cantly higher total energy intake at lunch (P < 0.003), total daily energy intake (P = 0.013), energy intake derived from a healthy diet (P = 0.003), vegetable dish intake (P = 0.001), and sh and meat dish intake (P = 0.028; Table 3). [Insert Table 4 here] Girls with higher NPFT scores had signi cantly higher values for total energy intake at breakfast (P = 0.008), energy intake derived from a healthy diet at breakfast (P = 0.030), total energy intake at lunch (P < 0.001), total daily energy intake (P = 0.001), energy intake derived from a healthy diet (P = 0.005), and sh and meat dish intake (P = 0.037; Table 4).
[Insert Table 5 here] The results of multiple regression analysis for investigation of factors related to NPFT scores are shown in Table 5. In boys, signi cant positive correlations were observed between frequency of outside playtime [standardised partial regression coe cient (β) = 0.135, P = 0.004], exercise lessons (β = 0.131, P = 0.006), and total daily energy intake (β = 0.096, P = 0.041). In girls, signi cant positive correlations were observed with exercise lessons (β = 0.121, P = 0.012), and total daily energy intake (β = 0.145, P = 0.003), while a signi cant negative correlation was observed with Rohrer's index (β = -0.097, P = 0.047). [Insert Table 6 here] Multiple regression analysis was also conducted to assess the relationships between NPFT scores and energy derived from a healthy diet and snacks. In boys, signi cant positive correlations were observed with energy intake derived from a healthy diet (β = 0.122, P = 0.009; Model 1). This relation was signi cant even after adjusting for snack intake (β = 0.120, P = 0.011; Model 3). Similarly, in girls, signi cant positive correlations were observed with energy intake derived from a healthy diet (β = 0.125, P = 0.010; Model 1). This relation was also signi cant even after adjusting for snack intake (β = 0.137, P = 0.005; Model 3). Snacks were not signi cantly correlated with NPFT score in boys or girls.

Discussion
Investigation of the factors related to motor skills in rst-grade boys revealed signi cant positive correlations with the frequency of outside playtime, exercise lessons, and total daily energy intake. In girls, exercise lessons and total daily energy intake had positive correlations with NPFT score, while Rohrer's index had a signi cant negative correlation. Investigation of total energy intake revealed a signi cant positive correlation between dietary energy and motor skills in both boys and girls. Previous studies have reported that 13-year-old children who exercise three or more times per week have improved grip and 50-metre run results [7], and that moderate-to-vigorous physical activity (associated with ≥ 3 metabolic equivalents) by school-age and adolescent children is positively correlated with physical tness [10,20]. In this study, exercise lessons (one of the indices for daily physical activity) were positively correlated with NPFT scores, in keeping with previous research ndings. Exercise lesson programmes such as baseball, soccer, and swimming often include basic skill practice in addition to competition. Thus, children attending exercise lessons regularly participate in activities such as side-stepping and sit-ups, as well as endurance enhancing exercises such as jogging. The strength developed during these activities is then displayed in NPFT exercises and may be positively correlated with NPFT scores. Reports have shown that children with three or more days per week of outside playtime (another index of daily physical activity) participate in signi cantly longer periods of light and total physical activity, as compared to children who do not play outside as frequently [21]. This study also found outside playtime to have a signi cant positive correlation with NPFT score in boys and, although the correlation was not signi cant among girls, a positive tendency was observed. Higher frequencies of outside playtime, corresponding to larger amounts of physical activity, affected NPFT scores.
A previous study reported a positive correlation between BMI and tests of physical tness [10]. In contrast, this study found a signi cant negative correlation between NPFT score and Rohrer's index in girls and a tendency toward a negative correlation in boys. Comparison of Rohrer's index between the groups with the lowest and highest NPFT scores, however, showed that the mean values in all groups of both boys and girls were within the normal range for physique (Rohrer's index; 115 to < 145). Comparison of physical tness measurements among children classi ed as thin, standard, or obese in the same previous study [10] indicated that children with standard body types had the best results. Therefore, it cannot be suggested that thinner children have higher NPFT scores.
This study did not observe signi cant correlations between NPFT scores and sleep duration or frequency of eating breakfast per week. Sandercock et al. [9] reported higher cardiorespiratory endurance values in adolescent children who ate breakfast every day, as compared to values among children who skipped breakfast. However, only 67.9% of the study subjects in the same study [9] ate breakfast every day, as compared to approximately 98% of subjects in the present study. The percentage of children who missed breakfast one or more days per week in the current study was approximately 2%, which is a substantially smaller proportion than that observed previously. This difference may explain why the current study did not nd breakfast consumption to be signi cantly correlated with NPFT scores. In the MEXT survey, children with eight or more hours of sleep per night had the highest NPFT scores [22]. However, no signi cant correlation between these elements was observed in the current study. Previous studies that did not indicate correlations between sleep duration and physical tness found mean sleep durations to be 9.6-9.7 hours per night [10,20]. The present study found the mean sleep duration to be 9.6 hours per night for both boys and girls, which supports the results of previous studies. Therefore, the lack of an observed correlation in this study may have been due to the children receiving adequate sleep. These ndings suggest that not only physique and exercise habits, but also total daily energy intake, are correlated with motor skills in children.
Based on the results suggesting a correlation between NPFT score and total daily energy intake, further investigation of the total energy breakdown was performed. Energy intake derived from a healthy diet indicated a signi cant positive correlation, demonstrating that simply having a high daily energy intake was not correlated with NPFT score. Asghari et al. reported that higher consumption of confectionaries was associated with higher fat-energy ratios and lower protein-energy ratios [23]. The present study also found that sh and meat intake was positively correlated with NPFT scores in both boys and girls. The intake of confectionaries and sweetened beverages elevates blood sugar levels, with previous studies reporting that blood sugar levels peak within 30 minutes after consuming high-sugar foods before returning to fasting levels 60 minutes after eating [24]. In contrast to this rapid uctuation, blood sugar gradually increases for up to 45 minutes after consuming non-sugar and medium-sugar foods, before steadily declining for up to 120 minutes after eating [24]. Previous studies have also reported higher grip and lower limb muscle strength in groups with higher protein intake [25] and positive correlations between physical tness and frequency of fruit and vegetable consumption [10]. The consumption of energy from grains, vegetables, and sh and meat rather than from confectionaries and sweetened beverages promotes healthier children's bodies and provides the required energy and nutrients.
No signi cant correlations were observed with energy derived from snacks and NPFT scores. Comparison of the ve NPFT quintiles in the present study showed a mean intake of approximately 200 kcal from confectionaries and sweetened beverages among both boys and girls in all groups (q1 to q5). Previous studies on the consumption of sweets by children have shown correlations with mental health problems [14] and lack of sleep [26]. Consumption of sweets and sweetened beverages by adults has also been correlated with psychological disorders [27]. Previous reports consisted of surveys on the frequency of consumption of sweets and sweetened beverages and, to our knowledge, there are no reports on these correlations with inclusion of speci c quantities ingested. The Japanese Food Guide Spinning Top de nes an adequate intake of sweets and sweetened beverages as ≤ 200 kcal [18]. Although the quantitative effect of sweets and sweetened beverages on humans is unknown, it is considered important to maintain an appropriate level of intake for these food groups.
The results of this study suggest that it is important to consume sweets and sweetened beverages in moderation, and not to use these foods as a primary energy source. Energy should instead be consumed in the form of meals.

Strengths and limitations
The strength of this study was that it investigated the correlations between motor skills and nutritional intake among approximately 900 primary school students. The use of dietary records allowed for consideration of the mutual effects of meal quantity and quality, in addition to lifestyle habits. An additional strength was that multiple measurements of motor skills were performed in primary schools according to the MEXT guidelines.
This study had several limitations. First, the cross-sectional design prevented assessment of the potentially causative relationships between NPFT scores and lifestyle habits. In addition, the present study used a self-administered dietary record, which allowed for input of children's heights and weights based on the judgement of the guardian. Because food intake was only surveyed for one day, the meals consumed on that day may not be indicative of children's habitual meals. The dates on which dietary records were completed and the date of the school lunch survey were different; thus, the total intakes of breakfast, dinner, between-meal eating, and lunch were taken from different days. The combined results did not record all meals from a complete day, so caution is needed when interpreting the results. However, participating children and guardians were asked to enter data into the dietary record of meals that children ate at home that most closely represented meals on an average day, and school lunches were provided in accordance with Japanese school lunch implementation standards [17] and thus represent an average daily meal.
The present study was unable to obtain actigraphy or pedometer measurements as indices of physical activity, due to the signi cant burden this would place on participating teachers and children. These indices were instead set as the frequencies of outside playtime and exercise lessons. In addition, data on the guardians' educational backgrounds and household nancial status could not be obtained. Future studies are needed to investigate the relationships between motor skills and actual measurements of the amount of physical activity among children, including various factors such as guardian educational background and household nancial status. Finally, the response rate in this study was low at 44.6%, and many of the guardians were stay-at-home parents. The survey may thus have had a selection bias, with a signi cant number of respondents who had a high degree of involvement in the survey and a high level of interest in participating children's meals and health.

Conclusions
The present study investigated the relationship between children's motor skills and their dietary and lifestyle habits. Study ndings indicated that children's NPFT scores were signi cantly and positively correlated with involvement in sports lessons, total energy intake, and outside playtime in boys, and negatively correlated with Rohrer's index in girls. Further investigation of daily energy intake breakdown revealed a correlation between NPFT scores and daily energy intake derived from a healthy diet, independent of children's physique or exercise habits.
These results demonstrated that children's physique and exercise habits, in addition to consumption of a healthy diet that includes grains, vegetables, fruits, milk, sh and meat, are important for children's motor skills. It is essential that dietary education be conducted to encourage the intake of a healthy diet, and to investigate whether this leads to improvements in motor skills. Availability of data and materials: The datasets generated and analysed during the current study are not publicly available, but some data are available from the corresponding author on reasonable request.

List Of Abbreviations
Competing interests: The authors declare that they have no competing interests.
Funding: This study was supported by Health and Labor Sciences Research Grants (H29Sukoyaka-Ippan 003). The funders (Ministry of Health, Labor and Welfare of Japan) had no role in the study design, data collection/analysis, decision to publish, or preparation of the manuscript.
Authors' contributions: YT, TF, AH, and YK designed the research; YT, NH, MK, RK, and TH conducted the research; YT and NH analyzed the data; YT and NH wrote the manuscript; YT assumes primary responsibility for the nal content. All authors read and approved the nal manuscript.  Energy intake derived from grain dishes, vegetable dishes, fish and meat dishes, milk and dairy products, and fruit (based on the food groups defined by the Japanese Food Guide Spinning Top), as opposed to snacks, confectionaries, and sweetened beverages Mean ± SD Rohrer's index was classified into eight quintiles. Sleep duration was classified into eight categories with 30-minute increments. The frequency of outside playtime was set as '1' for responses of "almost every day (4-5 times a week)" for even one of the timeframes of morning, early break, lunch break or after school; all other responses were set as dummy variables with a value of '0.' Study time was classified into eight categories with 15-minute increments. The total time spent watching television or participating in video gaming or reading comics was aggregated and classified into four categories with 60-minute increments. A value of '1' was allocated for children engaged in exercise, classroom learning, and/or music, whereas children not engaged in lessons were allocated dummy variables with a value of '0.' A value of '1' was allocated for frequency of eating breakfast per week responses of "7 times a week," whereas responses of "0-2," "3-4," or "5-6 times per week" were allocated dummy variables with a value of '0.' Each model was adjusted for Rohrer's index, sleep duration, the frequency of outside playtime, study time, the total time spent watching television or participating in video gaming or reading comics, participation in lessons (exercise, classroom learning, and/or music), and the frequency of eating breakfast per week. † Standardized regression coefficient Energy intake derived from grain dishes, vegetable dishes, fish and meat dishes, milk and dairy products, and fruit (based on the food groups defined by the Japanese Food Guide Spinning Top), as opposed to snacks, confectionaries, and sweetened beverages