Regular physical activity (PA) improves the health of many bodily systems, including the cardiovascular and musculoskeletal systems (Di Stasi et al., 2010; Leung et al., 2008), whereas factors such as physical inactivity lead to a sedentary lifestyle that in turn leads to increased weight (Canoy and Bundred, 2011; Park et al., 2004; Thomas and Albert, 2002) and adverse health effects such as musculoskeletal disorders, cardiac diseases, hypertension, stroke, type II diabetes, obesity, some types of cancers, and metabolic syndrome (Eckel et al., 2005; World Health Organization, 2011). Therefore, increasing PA is considered as an important approach for preventing adult diseases. Many studies have reported that increasing PA is beneficial for improving health outcomes (Baker et al., 2011; Daniels et al., 2005; Keteyian, 2011).

Recently, several studies have reported that increased PA is associated with improved cognitive and memory functions of the brain (Flöel et al., 2010; Ploughman, 2008). The precise mechanisms underlying this association are unclear and inadequately understood. However, there are 3 hypotheses that can explain how PA affects brain function. The first hypothesis is that PA increases oxygen saturation and angiogenesis in the brain areas essential for performing tasks (Kleim et al., 2002; Kramer et al., 1999). The second hypothesis is that PA also increases the activity of brain neurotransmitters such as serotonin and norepinephrine, thereby facilitating information processing (Kubesch et al., 2003; McMorris et al., 2008; Winter et al., 2007). The third hypothesis is that PA upregulates neurotrophins such as brain-derived neurotrophic factor, insulin-like growth factor-I, and basic fibroblast growth factor. These neurotrophins support neuronal survival and differentiation in the developing brain (Schinder and Poo, 2000).

These studies indicated that because PA is related to brain function, increasing PA may improve school performance and academic achievements in adolescent students. However, although PA has been shown to increase brain function and some practical evidence exists on the association between PA and academic performance (Sibley and Etnier, 2003), there is no epidemiological evidence supporting this finding in the literature on adolescent students. Furthermore, although a physical education (PE) class represents a type of PA in school, there is no epidemiological evidence showing that such classes are beneficial for adolescent students. Therefore, the purpose of this study is to examine whether the number of PE classes attended per week is related to school performance in Korean adolescent students. The results will be used to further assess the differences in the school performance of Korean adolescent students according to the frequency of PE classes per week and will then be used to recommend increased PA as an approach for enhancing school performance and achievements.

Participants
The 5^th Korea Youth Risk Behavior Web-based Survey (KYRBWS-V) was conducted in 2009 as a cross-sectional epidemiological study with a complex sample design that included multistage sampling, clustering, and stratification. The KYRBWS-V consisted of questions about PA based on the International Physical Activity Questionnaire (IPAQ) of the World Health Organization (Craig et al., 2003). It was a national school-based survey conducted by the Korea Centers for Disease Control and Prevention (KCDCP) to determine the prevalence of health risk behavior in Korean adolescent students (Korea Centers for Disease Control and Prevention, 2010). A representative sample of students from middle school first to high school third (age, 13-18 years) was selected; this sample included 76,937 students from 24,000 classrooms (secondary sampling units) of 800 middle and high schools (primary sampling units) and from 192 strata that were identified using the stratified multistage cluster sampling method. After sample determination, classroom teachers assigned each participating student a unique identification number that was used by the students to access the survey web page. On the web page, the students were first asked about their willingness to participate in the survey. Students who were willing to participate in the survey were asked to answer a self-administered questionnaire, which they completed anonymously in school. Students who did not wish to participate were not asked to answer the questionnaire. Furthermore, because private information was not collected by the KYRBWS-V, ethical approval was not required for this study.

All the details pertaining to data collection procedures have been reported by the KCDCP (Korea Centers for Disease Control and Prevention, 2010), and the reliability and validity of the data obtained using the questionnaire have been evaluated in other studies (Bae et al., 2010a; 2010b).

The present study obtained data from the KYRBWS-V to evaluate the association between the number of PE classes attended per week and school performance, taking into consideration potential covariate variables such as gender, age, body mass index (BMI), parents' education level, family's economic status, vigorous and moderate PA, and muscle strengthening exercises. Thus, 75,066 adolescent students (39,612 boys and 35,454 girls) participated in the study, and the response rate was 97.6%. The characteristics of the subjects are shown in Table 1.

Independent variables
To determine the number of PE classes attended per week by all the subjects, the following question was asked in the KYRBWS-V survey: "(Q1) How many PE classes do you attend per week in school?," with the following response options: (1) no PE class, (2) once per week, (3) twice per week, and (4) >3 times per week. Next, these responses were classified into the following 2 groups for multivariate logistic regression analyses: (1) >3 PE classes per week and (2) <3 PE classes per week.

Dependent variables
Self-reported school performances were evaluated for each adolescent student by asking 1 question: "(Q2) In the past 12 months, how has your average school performance been?" The response options were (1) very good, (2) good, (3) average, (4) poor, and (5) very poor.

Covariate variables
The covariate variables were as follows:

1. Gender: The 2 response options were (1) male and (2) female.

2. Age: Data pertaining to the ages of the students, as defined by the KYRBWS-V, were used without any modifications.

3. BMI: The students were asked to self-record their height and weight. BMI (kg/m²) was calculated from the data recorded by each student.

4. Parents' education level: The 4 response options were (1) unknown, (2) middle school or lower, (3) high school and (4) college or higher.

5. Family's economic status: The 5 responses ranged from (1) very rich, (2) rich, (3) average, (4) poor, and (5) very poor.

6. Frequency of vigorous PAs such as digging, aerobics, heavy lifting, or fast cycling during the week: The 6 responses ranged from (1) no, (2) once per week, (3) twice per week, (4) thrice per week, (5) 4 times per week, and (6) over 5 times per week.

7. Frequency of moderate PAs such as bicycling at a regular pace, carrying light loads, or playing doubles tennis during the week: The 6 responses ranged from (1) no, (2) once per week, (3) twice per week, (4) thrice per week, (5) 4 times per week, and (6) over 5 times per week.

8. Frequency of muscle strengthening exercises such as sit-ups, push-ups, and weight lifting or weight training during the week: The 6 responses ranged from (1) no, (2) once per week, (3) twice per week, (4) thrice per week, (5) 4 times per week, and (6) over 5 times per week.

Statistical analysis
All the results are presented in terms of mean ± standard deviation values. Multivariate logistic regression analyses were performed to determine whether the number of PE classes attended per week was related to school performance in the participants of this study, after adjusting for the covariate variables. Statistical significance was set at p < 0.05, and all the analyses were performed using SPSS ver. 12.0 (SPSS, Chicago, IL, USA).

The results of the multivariate logistic regression analyses of school performance for the groups of Korean adolescents participating in >3 and <3 PE classes per week are shown in Table 2. This table compares the information about the Korean adolescents who attended <3 PE classes per week to that of Korean adolescents who attended >3 PE classes per week, according to school performances that ranged from (1) very good to (5) very poor after adjusting for covariate variables such as gender, age, BMI, parents' education level, family's economic status, vigorous and moderate PA, and muscle strengthening exercises.

Results show that for students attending <3 PE classes per week, good school performance increased by 12.5%, average school performance increased by 14.7%, poor school performance increased by 14.6%, and very poor school performance increased by 19.1% as compared to very good school performance.

The aim of this study is to investigate the association between school performance and attending <3 PE classes per week in Korean adolescents. The results of our study show that attending <3 PE classes per week was associated with poor school performance in Korean adolescent students, despite adjusting for school performance-related covariate variables.

Several cross-sectional studies have shown that less PA is associated with poor cognitive function and memory functions of the brain (Flöel et al., 2010; Ploughman, 2008). Thus, the results of our epidemiological study indicate that PE classes represent PA, which may increase the cognitive and memory functions of the brain. Furthermore, they indicate that increasing the frequency of PE classes positively affects school performance in Korean adolescent students.

School performance may be highly associated with the reading, speaking, writing, and understanding levels of individuals, which in turn are related to brain activities and functioning. Here, our cross-sectional study showed that school performance was associated with the frequency of attending PE classes.

Limitations and future research
This study has several limitations. First, the KYRBWS-V was conducted using online methods; consequently, the parents' education level and the family's economic status were recorded by the students themselves, and these data may therefore be inaccurate. Second, this study was a cross-sectional epidemiology study. Therefore, we could not elucidate the cause-and-effect relationship but could only assess the interrelationship between school performance and the frequency of attending PE classes. However, because our study sample comprised 75,066 students from all parts of Korea and was thus highly representative of the Korean adolescent population, the relationship between school performance and the number of PE classes attended per week could be generalized to all Korean adolescents.

However, in the future, better-designed studies should be performed to determine the extent to which these variables, including the covariate variables, affect the school performance of Korean adolescents.

• Korean adolescents, attending >3 PE classes per week was positively correlated with school performance.
• Korean adolescents, attending <3 PE classes per week was negatively correlated with school performance.