Evaluating of the Axivity accelerometers algorithm in measurement of physical activity intensity in boys and girls

Background: Accelerometers are new but popular tools in measurement of physical activity level but most of them use the algorithms which are more suitable for adults rather than adolescents. Also accelerometers accuracy in assessing physical activity intensity in boys and girls is widely unknown. Therefore, purpose the current study was to evaluate the validity of the Axivity accelerometers in relation to establish algorithm in the measurement of physical activity intensity in boys and girls. Methods: A total of 143 participants recruited from a local school in Oxfordshire and 119 of them (84 boys and 35 girls, age (12.71 ± 0.46)) completed the shuttle run test while wearing an AX3 accelerometer. Signal Vector Magnitude (gravity-subtracted) (SVMgs) and Metabolic equivalent (MET) values were calculated for both of boys and girls. Results: The study result showed that girls had significantly higher SVMgs value (P< 0.05) while according to the standard equations, girls have lower VO2max and MET in each shuttle run lap. Conclusion: This study result suggest that different algorithms might be necessary to measure physical activity in different genders.

produce enough energy for conducting physical activity. 9,10 Maximal oxygen consumption (VO2max) is considered by the World Health Organization (WHO) to be the best cardiorespiratory fitness indicator. 11 There are different ways for physical and cardiorespiratory fitness assessment in adolescents, but 20meter Shuttle Run Test is the most popular one 12, 13 and frequently being used in cardiorespiratory fitness studies. 14 For example a recent research (Fit to study) is using the test for evaluating cardio fitness in 100 secondary schools in UK (https://www.fit-to-study.org/). The shuttle run, is included in a broad number of physical fitness test batteries. One of the most frequently-used versions is the PACER (Progressive Aerobic Cardiovascular Endurance Run), the standard CRF test for the FITNESSGRAM battery. 15 Based on the evidence, physical activity level could be another indicator of physical fitness Also MVPA can advance the strength, cardiorespiratory fitness, body composition and reduce the cardiovascular risk factors in adolescents. [5][6][7] There are several studies which explore different objective and subjective physical activity assessment methods in adolescents and children. 6,7,16,17,18 Accelerometers are a popular choice of instrument for the objective measurement of physical activity intensity. Accelerometers are portable devices which categorise the physical activity level base on the Signal Vector Magnitude (SVM) in three different Axis (vertical(Y), horizontal right-left (X) and horizontal front-back axis (Z)). 19 However, there is one study which used the accelerometers (Geneactiv) for the objective measurement of physical activity in adolescents during shuttle run test 16 and two recent studies which measured validity of the Axivity device wearing at thigh and back 20, 21 but to date there is no evidence to evaluate accuracy of the wrist worn Axivity algorithms using the (SVM) during a vigorous activity and comparing results between boys and girls.
Given the likelihood that shuttle run test and accelerometers will continue to be used by researchers as a tool to measure physical fitness and physical activity level in adolescents, aim of the current study was to evaluate the extent of the strength of relationship of SVM algorithms in measurement of physical activity intensity alongside of evaluating the SVM and cardio-fitness differences in boys and girls.

Materials And Methods Design
A Cross sectional study design.

Population
All of the participants were recruited from a local school in Oxfordshire. Ethical approval granted by Oxford Brookes University Ethical Advisory Committee. Permission was gained from each school's head teachers to recruit participants, and opt-out consent was collected from each participant's parent or legal guardian.

Measurements
Anthropometric measures were taken which included height (measured using a portable stadiometer, Seca UK), waist circumference (measured mid-way between the lower rib margin and the iliac crest using anthropometry tape), and body weight and composition (measured using a Tanita Body Composition Analyser, model: BC-418 MA, Tanita, UK).

Cardiorespiratory fitness measure
Shuttle run test used for the cardiorespiratory fitness evaluation. Total of 119 participants completed shuttle run test which according to the FITNESSGRAM (PACER), the test consists in running back and forth between two lines 20 meters apart, with running speed determined by audio signals from a prerecorded music CD. 15 The running speed increases at the end of each one-minute stage. The running speed is 8.0 km.h-1 for the first stage, 9.0 km.h-1 for the second stage, and thereafter increases by 0.5 km.h-1 each minute. The test ends when the subjects twice fail to reach the lines at the time indicated by the audio signals, demonstrating an inability to keep the required pace. 15

AX3
All the participants wore a wrist-band accelerometer during whole shuttle run test. AX3 accelerometer was attached, using a watch strap positioned on the dorsal aspect of participants' non-dominant hand. The AX3 (axivity, UK) which was used in this study as a measuring device is a continuous logging accelerometer that has been designed to monitor physical activity intensity and duration. It is a triaxial, ±16g acceleration sensor housed in a small (23 x 32.5 x 7.6 mm), light weight (11 g) and encased splash-proof. The sampling frequency of the AX3 ranges from 12.5 Hz to 3200 Hz and the battery can last up to 14 days when recording data at 100 Hz.

Data processing and statistical analyses
There are several methods and equations to estimate VO2max during SR test in adolescents. 22 One of these methods is Multiple Linear Regression (MLR) model which is a stepwise method. [23][24][25][26] Another method is Artificial Neural Networks (ANN) models which has been validated by Ruiz et al in (2008 and2009) for VO2max assessment in adolescents. 27,28 Also this method is widely being used in different studies. 26,[29][30][31][32] Raw 100 Hz triaxial Ax3 data was analysed in a bespoke LabVIEW programme (National Instruments, In addition, a bespoke programme was used to calculated the estimated VO2max for each participant using three different methodologies considering participants' sex, age, height, weight, body mass index (BMI) and Pacer laps according to Mahar et al 2011 22 , Multiple Linear Regression (MLR) [23][24][25][26] and using artificial neural networking (ANN) methodologies 27,28 To achieve the MET values associated with each level of shuttle run, VO2 values were converted to METs using age specific values.The MET values were calculated using the standardized MET formula for adolescents: 1 MET = 5.92 ml kg -1 min -1 (8-12 boys/8-11 girls) and 4.85 ml kg -1 min -1 (13-15 boys/12-14 girls). 35 The Shapiro-Wilk test confirmed that all data were normally distributed. Also SVMgs average calculated for each shuttle run level and compared between genders using univariate variance test. In the event of a significant ANOVA result, Bonferroni -corrected post hoc comparisons were undertaken to determine where the significant differences occurred. Differences between the BMI groups were tested using independent t-tests. P < 0.05 was considered significant and all tests were 2-sided. All statistical analyses were performed using SPSS version 25 (IBM SPSS Statistics). Data are displayed as mean (± SD) and mean (95 % CI) in the text and tables.

Discussion
We observed that the SVM gs was significantly higher in girls than boys with girls having a higher level of physical activity intensity and higher MET during the shuttle run test with standard formulas (Mahar et al, 2011; MLR and ANN). MET is an indicator of Physical activity intensity 36 and sedentary behaviour is ranged (1.0-1.5 METs) 35,36 and light-intensity is (1.6-2.9 METs) 37, 39 , moderate-intensity (3-5.9 METs), and vigorous-intensity (≥6 METs) activities. 40 Therefore we found that shuttle run test started in the vigorous physical activity level in both genders. Also based on the standard formula results, girls to have a lower MET level during all the shuttle run laps. There are limited studies which have compared the MET in boys and girls but this result is broadly consistent with other studies which comparing the energy expenditure and METs in male and females during different activities. For example a study result by Zunzer et al (2013) showed that male golf players had significantly higher energy expenditure and MET rather than female participants. 41 Based on the evidence and current study MET result, we expected girls to have a lower physical activity intensity during per shuttle run lap but accelerometer data showed that female participants had a significantly higher SVM gs rather than boys per shuttle run level. It might be because of the running economy or lower coordination level in girls. A study result by Telford et al, (2016) showed that girls had 44% lower eye-hand coordination during physical activity 42 and it could be a reason of the accelerometers SVM gs over estimating. Present study is the first research which compared the AX3, SVMgs results in adolescent boys and girls. Therefore further researches are required to confirm this result but base on the current study result, the AX3 accelerometer algorithm might need re-assessment to have the more accurate physical activity measurement for this age group in boys and girls.
Current study provided an evidence of cardio-fitness in boys and girls during shuttle run test. In the present study girls completed 7 levels and boys finished 9 levels of shuttle run laps. It could indicate that girls have a lower level of cardio-fitness rather than boys. This result is broadly consistent with other studies. For example a systematic review result by Tomkinson et al (2016) in 1142026 children age (9-17) during shuttle run test showed that boys had a better performance than girls in all the age groups. 43 Furthermore a study by Lintu et al (2015) in 9-11 years old children demonstrated that boys had a better cardio respiratory fitness rather than girls. 44 Also another study result in 7244 students (9-17.9) by Ramírez-Vélez et al, (2017) showed that boys had a higher shuttle and VO2peak in all of the age groups rather than girls. 45 This gender differences in cardio-fitness level could be because of the lower level of physical activity in girls rather than boys. A study result by Telford et al (2016) in 276 boys and 279 girls from 29 schools showed that girls 19% less active than boys and their cardio respiratory fitness was 18% less than boys. 42 This lower level of physical activity could have a negative impact on girls' health in adulthood. Therefore designing different interventions to increase physical activity level in girls could improve the cardio-fitness in them.
The limitations of this study include having relatively small sample size and not measuring the actual VO2max and resting metabolic rate in participants. Study strengths however include the novel assessment and comparison of Axivity SVM data during an observed free living style activity in boy and girl adolescents.

Conclusion
Current study presents important and novel finding that when using the Ax3 accelerometer raw data and comparing the result between male and female participants. Current study result showed that the SVM result was significantly different between genders and gone higher intensity levels for same shuttle lap in girls. This suggests that a different accelerometer algorithm might be required for accurate measurement of physical activity intensity in boys and girls. Further research is now required with more participants and different age groups of adolescents to confirm this result alongside actual measures of energy expenditure.

Declarations
Ethics approval and consent to participate Ethical approval granted by Oxford Brookes University Ethical Advisory Committee. Permission was gained from each school's head teachers to recruit participants, and opt-out consent was collected from each participant's parent or legal guardian.

Consent for publication
Not applicable

Availability of data and materials
The data that support the findings of this study are available from Oxford Brookes University but restrictions apply to the availability of these data, which were used under license for the current study, and so are not publicly available. Data are however available from the authors upon reasonable request and with permission of Oxford Brookes University.
Competing interests "The authors declare that they have no competing interests" Funding This work is supported by Elizabeth Casson Trust (ECT), EEF, NIHR Oxford Biomedical Centre (BRC) and the CLEAR trust.

Role of funding source
Authors declare that funding sources had no involvement in study design, collection, analysis and interpretation of data, in the writing of the report and the decision to submit the article for publication.    Figure 1 Gender differences in SVM during shuttle run among adolescents

Supplementary Files
This is a list of supplementary files associated with this preprint. Click to download.