SEARCH

SEARCH BY CITATION

Keywords:

  • accelerometry;
  • energy expenditure;
  • indirect calorimetry;
  • physical activity

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Objective: To develop regression-based equations that estimate physical activity ratios [energy expenditure (EE) per minute/sleeping metabolic rate] for low-to-moderate intensity activities using total acceleration obtained by triaxial accelerometry.

Research Methods and Procedures: Twenty-one Japanese adults were fitted with a triaxial accelerometer while also in a whole-body human calorimeter for 22.5 hours. The protocol time was composed of sleep (8 hours), four structured activity periods totaling 4 hours (sitting, standing, housework, and walking on a treadmill at speeds of 71 and 95 m/min, 2 × 30 minutes for each activity), and residual time (10.5 hours). Acceleration data (milligausse) from the different periods and their relationship to physical activity ratio obtained from the human calorimeter allowed for the development of EE equations for each activity. The EE equations were validated on the residual times, and the percentage difference for the prediction errors was calculated as (predicted value − measured value)/measured value × 100.

Results: Using data from triaxial accelerations and the ratio of horizontal to vertical accelerations, there was relatively high accuracy in identifying the four different periods of activity. The predicted EE (882 ± 150 kcal/10.5 hours) was strongly correlated with the actual EE measured by human calorimetry (846 ± 146 kcal/10.5 hours, r = 0.94 p < 0.01), although the predicted EE was slightly higher than the measured EE.

Discussion: Triaxial accelerometry, when total, vertical, and horizontal accelerations are utilized, can effectively evaluate different types of activities and estimate EE for low-intensity physical activities associated with modern lifestyles.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Activity thermogenesis can be separated into two components: exercise-related activity thermogenesis and non-exercise activity thermogenesis (NEAT)1 (1). NEAT, composed mainly of the energy expenditure (EE) related to low-to-moderate intensity daily physical activity (PA), is likely to have greater individual variation than exercise-related activity thermogenesis and body size-dependent basal metabolic rate. Levine et al. (2) used inclinometers and triaxial accelerometers to reveal that obese participants were seated for 164 min/d more than and were upright for 152 min/d less than lean participants. Moreover, if the obese subjects had the same posture allocation as the lean subjects, they would have expended an additional 352 kcal/d. Therefore, NEAT has been highlighted recently for helping to prevent weight gain. However, there are currently few effective methods to objectively and noninvasively evaluate the type or quantity of low-intensity PA in free-living conditions.

Triaxial accelerometers that are small in size and minimally intrusive to normal subject movement can be useful devices for predicting PA EE (3). Previous studies demonstrated higher correlation coefficients between counts obtained with triaxial accelerometry and the EE measured by chamber in comparison with counts from uniaxial accelerometry (4, 5, 6). However, these previous studies researched moderate-intensity PA such as slow and brisk walking and jogging, not low-intensity lifestyle activities. Moreover, Bassett et al. (7) found that uniaxial waist-mounted accelerometers overestimated the EE of walking and underestimated the EE of all other activities. Thus, we hypothesized that methods for estimating EE would be improved by the development of equations for each daily lifestyle PA.

To accurately predict EE using equations for each activity, it is necessary to classify each daily lifestyle PA using triaxial accelerometry. There are currently no published data concerning the identification of body posture in free-living conditions using triaxial accelerometry, especially light-to-moderate intensity PA with upper body movement such as sweeping, mopping, and window washing, which is a relatively high-energy cost during daily living (4, 7). However, a previous study that evaluated standing balance using a triaxial accelerometer found that the accelerometer measurements, especially horizontal acceleration, were able to distinguish between the different test conditions and simultaneous force platform measurements (8). Concomitantly, it is speculated that household activities with upper body movement (e.g., cleaning and sweeping) may have larger horizontal acceleration than sitting and standing. We hypothesized that low-intensity PA in free-living conditions can be identified by using horizontal acceleration obtained from triaxial accelerometry.

Thus, the purpose of the present study was to develop regression-based equations that estimate EE from total acceleration, which was based on the defined thresholds of accelerations that can be used to delineate low-to-moderate intensity PA. Furthermore, we compared the ability to identify the type and quantity of the low-intensity PA and predicted EE using either triaxial acceleration or only vertical acceleration from a triaxial accelerometer.

Research Methods and Procedures

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Subjects

Twenty-one Japanese adults (8 men and 13 women) living in the Tokyo metropolitan area were recruited for the study (Table 1). All subjects were adults (≥20 years) and were without any chronic diseases that could affect EE or daily PA. All subjects received a verbal and written description of the study and gave their informed consent to participate before testing. The study protocol was approved by the Ethical Committee of the National Institute of Health and Nutrition.

Table 1. . Subject characteristics
 Men (n = 8)Women (n = 13)
Age (yrs)33 ± 1531 ± 10
Standing height (cm)171.2 ± 4.7161.0 ± 5.3
Body weight (kg)65.3 ± 4.155.8 ± 9.8
BMI (kg/m2)22.3 ± 2.021.5 ± 3.5
Fat (%)13.2 ± 3.723.3 ± 8.4
Fat-free mass (kg)52.0 ± 5.832.6 ± 5.4

Anthropometry

Body weight was measured on a digital balance to the nearest 0.1 kg, and height was measured on a stadiometer to the nearest 0.1 cm. BMI was calculated as the body weight in kilograms divided by the height in meters squared. Body composition was evaluated by the skinfold method at two skinfolds (triceps and subscapular) to the nearest 0.1 mm. The measurements were repeated until the difference between the two readings reached within 1 mm, and the mean value was used. Body density was assessed using the equations for Japanese (9), and the percentage of body fat was estimated using the equation of Brozek et al. (10). Body fat mass and the fat-free mass were calculated from body weight and percent of body fat.

Study Protocol

Subjects were fitted at the left hip with a triaxial accelerometer (AC-301, 51 × 77 × 15 mm, 87 grams; or AC-210, 48 × 67 × 16 mm, 57 grams; GMS, Tokyo, Japan) while also in the indirect human calorimeter (IHC) for 22.5 hours (from 6 pm to 4:30 pm the next day). The triaxial accelerometer obtained three-dimensional accelerations every 40 ms with a sensitivity of 2 milligausse (mG) and a band-pass filter of 0.3 to 100 Hz. The acceleration count was calculated as the average of the absolute values for acceleration in each direction for a given interval (1 minute). The subjects ate breakfast, lunch, and dinner at 8:15 am, 12:30 pm, and 6:30 pm, respectively. They went to bed at 11 pm and were gently awakened at 7 am. They were permitted to go to the toilet and were asked to return to bed immediately. The schedule included 8 sessions of standardized activities: 2 × 30 minutes sessions each of walking on a treadmill (95 m/min in the morning and 71 m/min in the afternoon), sitting, standing, and housework representative of typical activities in free-living conditions. Subjects were permitted to spend time freely in a sitting or standing position as long as posture was maintained and to rest periodically during the housework period. During the remaining time periods, subjects were only permitted to do light activities such as reading, writing, viewing television, dressing, and undressing. They were asked to refrain from sleeping and planned strenuous exercise except during the walking periods. Meals were given three times a day to provide the predicted basal metabolic rate (11) multiplied by the estimated PA level (1.5).

IHC

An open-circuit IHC was used to evaluate the EE of the four standardized activities totaling 4 hours, the sleeping time for 8 hours, and the residual time for 10.5 hours. Details of IHC have been reported previously (12, 13). Briefly, the respiratory chamber was an air-tight room (20,000 liters), equipped with a bed, desk, chair, television with video deck, compact disc player, telephone, toilet, sink, and treadmill. The temperature and relative humidity in the room were controlled at 25 °C and 55%, respectively. The O2 and CO2 concentrations of the air supply and exhaust were measured by mass spectrometry. For each experiment, the gas analyzer (ARCO-1000A-CH; Arco System, Inc., Kashiwa, Japan) was initially calibrated using a certified gas mixture and atmospheric air. The flow rate exhausted from the chamber was measured by pneumotachograph (FLB1; Arco System, Inc.). The flow meter was calibrated before each measurement, and the flow rate was maintained at ∼60 L/min. Oxygen uptake (Vo2) and carbon dioxide production (Vco2) were determined by the flow rate of exhaust from the chamber and the concentrations of the inlet and outlet air of the chamber, respectively (12). Values of Vo2 and Vco2 were expressed under the conditions of standard temperature and pressure and under dry conditions. EE was estimated from Vo2 and Vco2 using Weir's equation (14). The accuracy and precision of our IHC for measuring EE as determined by the alcohol combustion test was 99.8 ± 0.5% [mean ± standard deviation (SD)] in 6 hours and 99.4 ± 3.1% in 30 minutes. Sleeping metabolic rate was defined as the average EE over 8 hours of sleep. The PA ratio (PAR) was calculated as the EE during sitting, standing, housework, or walking periods divided by the sleeping metabolic rate.

Identification of the Types for PA

Minute-to-minute anterior-posterior (x-axis), mediolateral (y-axis), vertical (z-axis), and total (synthesized triaxes) accelerations were obtained from a triaxial accelerometer during four standardized periods (sitting, standing, housework, and walking on a treadmill, 2 × 30 minutes each activity). Twenty-eight of the 30 minutes of each structured period, which excluded the first and last minute of each session, were used for the analysis (i.e., 28 data points × two replicate sessions × 21 subjects = 1176 data points for four types of activity). One of the acceleration data for walking on a treadmill at 71 m/min was excluded for the analysis because the subject walked at a different speed. In addition, because the hip-fitted triaxial accelerometer could shift horizontally while the subject was in the IHC, anterior-posterior (x-axis) and mediolateral (y-axis) were synthesized as horizontal acceleration for the analysis. Optimal thresholds for classifying total acceleration and the ratio of vertical to horizontal acceleration into sitting, standing, housework, and walking were determined by receiver operating characteristic analysis, which is the standard approach to evaluate the sensitivity and specificity of test results. We adopted the acceleration for the highest product of sensitivity and specificity as optimal thresholds for each binary classification. Furthermore, the threshold of each activity was defined using only vertical acceleration.

Prediction and Validation of EE

The total accelerations from the different periods and the data's relationship to PAR obtained from the IHC allowed for the development of EE equations for four types of activity (sitting, standing, housework, and walking). The averaged value of minute-to-minute total acceleration for each activity was used for the analysis (i.e., one data point × 21 subjects = 21 data points for four types of activity), which corresponded to the 30-minute averaged PAR data obtained by IHC. The validation of the EE equations was tested on the residual time (10.5 hours). Initially, the minute-to-minute total acceleration for the residual time was classified into four types of activity using thresholds we developed. Subsequently, the PAR for each minute was predicted using a selected equation among four types of regression-based equations and/or constant value. The estimated EE for 1 minute was calculated as follows: the predicted PAR × the measured sleeping metabolic rate, which is a highly stable value in IHC. The estimated EE per 1 minute for the residual time (i.e., 630 minutes = 10.5 hours) was totaled. We investigated the validity of the equations by comparing the EE measured by IHC with the EE estimated using the developed equations. Similarly, in cases that only utilized vertical acceleration, the development and validation of equations were conducted.

Supplementary Experiment

To supplement the data of housework and walking, additional protocols that tested these activities were conducted using the same triaxial accelerometer and a portable gas analyzer (Metamax3B; CORTEX, Leipzig, Germany). Japanese adults (5 men and 7 women) 21 to 38 years old were recruited for the study. The measurement time was 4 minutes for housework (pull up weeds and sweep up) and 5 minutes for walking (walk in place and walk slowly). The relationship between the acceleration data (mG) from the different periods and PAR was tested.

Statistics

Statistical analyses were performed using SPSS for Windows (version 10.0; SPSS, Inc., Chicago, IL). All results are presented as the mean ± SD. The relationship between two variables was evaluated by Pearson's and Spearman's correlation. The percentage difference was calculated as follows: [(predicted value − measured value)/measured value] × 100. Agreement of EE between the predicted and measured values was further examined by plotting the difference in predicted values against the mean with limits of agreement (mean difference ± 2 SD of the differences, which gives an indication of the precision of the method), as suggested by Bland and Altman (15). Differences were regarded as significant when the probabilities were <0.05.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

The physical characteristics of the subjects are shown in Table 1. In general, the mean values were comparable with those obtained in the National Nutrition Survey, although a slightly larger variation was observed for body weight among women. Means and SD of total, horizontal, and vertical acceleration and the ratio of vertical-to-horizontal acceleration for structured activities are listed in Table 2. Only the vertical-to-horizontal acceleration ratio for walking exceeded 1.00. The resulting receiver operating characteristic curve characterized the performance of a binary classification by describing the trade-off between sensitivity and specificity over an entire range of possible thresholds (Table 3). The thresholds for sitting vs. standing and standing vs. housework were classified by total acceleration. Because it is possible to combine the total acceleration between housework and walking activities, the threshold for housework vs. walking was determined by the vertical-to-horizontal acceleration ratio and 30 mG or more of total acceleration. Sensitivities and specificities were >75% for each combination of two activities, except for specificity of sitting vs. standing. Moreover, when classifying PA by the threshold in the present study, the percentage of each classified PA was calculated during standardized periods of sitting (sitting, 75.3%; standing, 22.2%; housework, 2.5%; and walking, 0%), standing (sitting, 35.4%; standing, 43.5%; housework, 20.6%; and walking, 0.5%), housework (sitting, 8.2%; standing, 15.5%; housework, 72.4%; and walking, 3.8%), and walking (sitting, 0%; standing, 0.4%; housework, 5.1%; and walking, 94.4%). The same thresholds were also obtained by discriminant analysis. In contrast, when using vertical acceleration only, standing, housework, and walking activities were identified as accurately as total acceleration (sensitivity and specificity: standing vs. housework, 82% and 74%; housework vs. walking, 99% and 99%); however, it was not possible to distinguish between sitting and standing positions.

Table 2. . Minute-to-minute acceleration data for each activity
 Acceleration (mG) 
ActivityTotalHorizontalVerticalVertical/Horizontal
  1. mG, milligausse.

Sit6.1 ± 8.53.5 ± 6.70.7 ± 2.30.05 ± 0.17
Stand19.0 ± 20.813.0 ± 16.24.4 ± 7.30.18 ± 0.21
Housework52.8 ± 31.637.5 ± 23.618.7 ± 14.00.44 ± 0.22
Walk436.3 ± 107.7261.2 ± 62.7281.1 ± 87.31.08 ± 0.27
Table 3. . Threshold, sensitivity, and specificity (%) for each activity
 Acceleration (mG)   
ActivityTotalVertical/ horizontalSensitivity (%)Specificity (%) 
  1. mG, milligausse.

When using tri-axes acceleration:     
 Sit<7<0.75075.364.6Sit vs. stand
 Stand8 to 29<0.75078.976.3Stand vs. housework
 Housework>30<0.75095.994.5Housework vs. walk
 Walk>30>0.751   
When using vertical acceleration:     
 Sit<7   Sit vs. stand
 Stand<7 82.473.5Stand vs. housework
 Housework8 to 99 99.899.5Housework vs. walk
 Walk>100    

The averaged values of PAR were 1.38 ± 0.07 for sitting, 1.54 ± 0.18 for standing, 2.39 ± 0.27 for housework, and 4.34 ± 0.84 for walking, which corresponded to total acceleration values of 7.0 ± 2.9, 19.5 ± 14.7, 54.2 ± 14.6, and 426.0 ± 95.3 mG, respectively. Significant simple correlations were observed between PAR obtained by IHC and total acceleration obtained by triaxial accelerometry for standing, housework, and walking [R2 = 0.45 to 0.72, p < 0.01, standard error of estimation (SEE) = 0.05 to 0.32] (Table 4, Figure 1A). Because PAR for sitting was not associated with total acceleration, the averaged value of PAR (i.e., 1.3786) was used for predicting EE. Thresholds between the activities and three equations, or a constant value, for each kind of activity to predict EE were applied to the residual time for validation. There was a strong correlation between the measured and predicted EE (r = 0.94, p < 0.01) (Figure 2), although the predicted EE (882 ± 150 kcal/10.5 hours) was slightly higher than the EE measured by IHC (846 ± 146 kcal/10.5 hours; 4.4 ± 6.2% difference) (Figure 3). The same analyses were also performed using only vertical acceleration. Three EE equations (1, sitting and standing; 2, housework; 3, walking) were developed using only vertical acceleration (R2 = 0.51 to 0.64, p < 0.01, SEE = 0.02 to 0.29) (Table 4) but overestimated EE (p < 0.01) (981 ± 181 kcal/10.5 hours, 16.0 ± 10.0% difference).

Table 4. . Prediction equation for each activity
 PAR
ActivityModelR2SEE
  1. PAR, physical activity ratio; SEE, standard error of the estimate; AC, acceleration count; mG, milligausse.

When using tri-axes acceleration:   
 Sit1.3786  
 Stand0.0093AC (mG) + 1.35660.660.05
 Housework0.0123AC (mG) + 1.72080.450.18
 Walk0.0081AC (mG) + 0.92340.720.32
When using vertical acceleration   
 Sit0.0329AC (mG) + 1.38460.510.02
 Stand   
 Housework0.0333AC (mG) + 1.73160.600.13
 Walk0.0092AC (mG) + 1.84430.640.29
image

Figure 1. : Relationship between total acceleration and PAR. Original data (A) with additional protocol data (B).

Download figure to PowerPoint

image

Figure 2. : Relationship between measured and predicted 10.5-hour EE.

Download figure to PowerPoint

image

Figure 3. : Bland and Altman analysis. The differences between measured and predicted 10.5-hour EE are plotted against the measured and predicted mean 10.5-hour EE.

Download figure to PowerPoint

In the supplemental experiment, the average values of PAR and total acceleration were, respectively, 3.22 and 91.8 mG in men (n = 5), and 3.12 and 85.3 mG in women (n = 7) for pulling up weeds and 3.12 and 106.4 mG in men and 3.16 and 117.6 mG in women for sweeping up, which were categorized as housework in our study (Figure 1B, open triangle). Similarly, the PAR and total acceleration were, respectively, 2.90 and 170.2 mG in men, and 2.84 and 188.2 mG in women for walking in place and 3.21 and 202.1 mG in men and 2.84 and 218.1 mG in women for walking slowly, which were categorized as walking (Figure 1B, closed rhombus).

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

The major finding of this study is that we can accurately identify four different periods of activity (i.e., sitting, standing, housework, and walking) using total acceleration and the vertical-to-horizontal acceleration ratio obtained from a triaxial accelerometer under close-to-normal living conditions. When we used vertical accelerations only, it was not possible to distinguish between sitting and standing positions. In addition, the sensitivity and specificity between housework and walking using the vertical-to-horizontal acceleration ratio, which was our original method, was over 90%. A recent study found that the time allocated to sitting and standing was closely related to weight gain (2). Moreover, PA with upper body movement such as housework has a relatively high energy cost during daily living (4, 7). Therefore, the classification of daily lifestyle PA in our study could be a significant contribution to weight management, especially in the area of clinical practice.

Additionally, we found a high validation of predicting EE in low-intensity PA. Our results indicate that EE measured by chamber was closely correlated with EE estimated using the three equations and one constant value (percentage difference, 4.4%; correlation coefficient, 0.94; SEE, 61 kcal/10.5 hours). Although a previous study that estimated daily EE using triaxial accelerometry was limited, the percentage difference between EE measured by chamber and EE estimated by the developed non-linear model using Tritrac (triaxial accelerometer) was small (16). Moreover, Plasqui et al. (17) observed the relationship between total EE measured by the doubly labeled water technique for 15 consecutive days and the predicted EE using the equation of counts for Tracmor (triaxial accelerometer), age, weight, and height as parameters. These authors indicated that the correlation coefficient was 0.90, and SEE was 167 kcal/d between measured and predicted EE. Our study presents a novel method to objectively evaluate EE of low-intensity PA under close-to-normal living conditions using triaxial accelerometry that compares favorably with the previous study.

We believe that our highly accurate prediction of EE for low-intensity PA is due to the method used to develop each equation for standing, housework, and walking. A previous study reported that an equation based on the acceleration of walking underestimated EE of moderate-intensity lifestyle activities (7). Recently, Crouter et al. (18) found that the estimation of EE both in walking and lifestyle activity could be improved by the two regression lines. In the present study, if only one equation was developed from the relationship between total acceleration and PAR of all plots, including sitting, standing, housework, and walking [EE kcal = 0.0068 × acceleration count (mG) + 1.5509], the predicted EE of residual time (10.5 hours) would be overestimated (931 ± 155 kcal/10.5 hours, p < 0.01, 10.3 ± 5.2% difference). One possible explanation for this overestimation is that EE of static body posture such as sitting and standing may be overestimated by all plots included in the equation. Thus, the developed equations for each daily lifestyle PA are a novel method for predicting EE.

A previous study compared the ability to predict EE using uniaxial and triaxial accelerometry (7, 19). The results indicated that triaxial accelerometry had higher accuracy of estimating EE than uniaxial accelerometry. However, as Plasqui et al. (17) pointed out, because two devices from different manufacturers were used, no conclusions can be drawn regarding the possible benefits of triaxial vs. uniaxial accelerometry. When Plasqui et al. (17) initially observed the contributions of vertical and horizontal acceleration to total EE per day adjusted for weight, height, and age, vertical acceleration explained an additional 16% of the variation in total EE. Furthermore, because horizontal acceleration contributed another 5%, it was concluded that triaxial accelerometers are more suitable than uniaxial accelerometers for estimating daily life activities. Similarly, the present study also compared the ability to quantitate low-intensity PA using either triaxial acceleration or only vertical acceleration from a triaxial accelerometer. Our results demonstrate that EE equations developed using only vertical acceleration overestimated EE by 135 kcal/10.5 hours. Further analysis of our data shows that there is no difference in EE for sitting and standing between equations using triaxial acceleration and only vertical acceleration (triaxial, 681 kcal/10.5 hours vs. uniaxial, 672 kcal/10.5 hours, p = 0.06), whereas the equation using only vertical acceleration overestimated the EE of housework periods by 109 kcal/10.5 hours (triaxial, 195 kcal/10.5 hours vs. uniaxial, 304 kcal/10.5 hours, p < 0.01). Therefore, we conclude that a triaxial accelerometer has a higher ability to predict EE of low-intensity PA, especially when the activity includes a large variation in horizontal acceleration, such as housework. Additionally, the technique of using not only total acceleration but also the vertical-to-horizontal acceleration ratio can be emphasized as a merit of the three-dimensional accelerometer.

There are some limitations of this study. The first limitation concerns the validity of the equations developed by comparing the EE measured by IHC with the EE estimated using developed equations for the residual time (i.e., 630 minutes = 10.5 hours). It is noted that this approach tends to overestimate the validity of the methods developed. We need to test the prediction equations of the present study in free-living conditions using the doubly labeled water method. The second limitation was that total acceleration data from 100 to 250 mG were blank during the chamber stay, although the relationship between PAR and total acceleration allowed for the development of EE equations for each activity. However, the plots describing the relationship between PAR and total acceleration for housework and walking in the supplemental experiment were likely to be an extension of the regression line, explaining this relationship in both activities in the present study. The results indicate that either of the equations for housework and walking can be applied to the range of 100 to 250 mG for total acceleration. Another limitation is that we did not develop an equation for cycling, which is a very popular lifestyle PA. Future studies should apply to all types of lifestyle activities. Lastly, the reason for the slight overestimation of the EE/10.5 hours in the present study should be clarified.

In conclusion, we identified low-intensity PA with high accuracy using total acceleration and the vertical-to-horizontal acceleration ratio obtained from a triaxial accelerometer. Notably, the use of the vertical-to-horizontal acceleration ratio is a novel method. Due to the classification of low-intensity PA, it is possible to accurately predict EE using equations for each activity. We demonstrated that triaxial accelerometry, when the total, vertical, and horizontal accelerations are utilized, can effectively evaluate different types of activities and estimate EE for low-intensity physical activities associated with modern lifestyles. In combination with measured or a highly accurately predicted sleeping metabolic rate (20), EE in sedentary lifestyle can be obtained.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Heartfelt thanks are due to the subjects who participated in the study. We thank Hirokazu Osanai for effort in gathering the subjects. We thank the members of the National Institute of Health and Nutrition, especially Hiroko Kogure for help in data acquisition and analyses. This work was performed as part of the Surveys and Research on Energy Metabolism of Healthy Japanese by the National Institute of Health and Nutrition (Project Leader, I.T.) and was supported, in part, by the 20th Research Grant for Medical and Health Science from the Meiji Yasuda Life Foundation of Health and Welfare (to S.T.).

Footnotes
  • 1

    Nonstandard abbreviations: NEAT, non-exercise activity thermogenesis; EE, energy expenditure; PA, physical activity; IHC, indirect human calorimeter; mG, milligausse; Vo2, oxygen uptake; Vco2, carbon dioxide production; SD, standard deviation; PAR, PA ratio; SEE, standard error of estimation.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  • 1
    Levine, JA. (2004) Nonexercise activity thermogenesis (NEAT): environment and biology. Am J Physiol Endocrinol Metab. 286: 675685.
  • 2
    Levine, J. A., Lanningham-Foster, L. M., McCrady, S. K., et al (2005) Interindividual variation in posture allocation: possible role in human obesity. Science. 307: 584586.
  • 3
    Chen, K. Y., Bassett, D. R., Jr (2005) The technology of accelerometry-based activity monitors: current and future. Med Sci Sports Exerc. 37: S490S500.
  • 4
    Hendelman, D., Miller, K., Baggett, C., Debold, E., Freedson, P. (2000) Validity of accelerometry for the assessment of moderate intensity physical activity in the field. Med Sci Sports Exerc. 32: S442S449.
  • 5
    Levine, J. A., Baukol, P. A., Westerterp, KR. (2001) Validation of the Tracmor triaxial accelerometer system for walking. Med Sci Sports Exerc. 33: 15931597.
  • 6
    Rowlands, A. V., Thomas, P. W., Eston, R. G., Topping, R. (2004) Validation of the RT3 triaxial accelerometer for the assessment of physical activity. Med Sci Sports Exerc. 36: 518524.
  • 7
    Bassett, D. R., Jr, Ainsworth, B. E., Swartz, A. M., Strath, S. J., O'Brien, W. L., King, G. A. (2000) Validity of four motion sensors in measuring moderate intensity physical activity. Med Sci Sports Exerc. 32: S471S480.
  • 8
    Mayagoitia, R. E., Lotters, J. C., Veltink, P. H., Hermens, H. (2002) Standing balance evaluation using a triaxial accelerometer. Gait Posture. 16: 5559.
  • 9
    Nagamine, S., Suzuki, S. (1964) Anthropometry and body composition of Japanese young men and women. Hum Biol. 36: 815.
  • 10
    Brozek, J., Grande, F., Anderson, J. T., Keys, A. (1963) Densitometric analysis of body composition: revision of some quantitative assumptions. Ann N Y Acad Sci. 110: 113140.
  • 11
    Ministry of Health, Labour and Welfare, Japan (2005) Dietary Reference Intakes for Japanese, 2005. Daiichi Shuppan: Tokyo, Japan.
  • 12
    Futami, J., Tanaka, S., Yamamura, C., Oka, J., Ishikawa-Takata, K., Kashiwazaki, H. (2003) Measurement of energy expenditure by whole-body indirect human calorimeter: evaluation of validity and error factors. Nippon Eiyo Shokuryo Gakkaishi (J Jpn Soc Nutr Food Sci). 56: 229236.
  • 13
    Yamamura, C., Tanaka, S., Futami, J., Oka, J., Ishikawa-Takata, K., Kashiwazaki, H. (2003) Activity diary method for predicting energy expenditure as evaluated by a whole-body indirect human calorimeter. J Nutr Sci Vitaminol. 49: 262269.
  • 14
    Weir, JB. (1949) New methods for calculating metabolic rate with special reference to protein metabolism. J Physiol. 109: 19.
  • 15
    Bland, J. M., Altman, DG. (1986) Statistical methods for assessing agreement between two methods of clinical measurement. Lancet. 1: 307310.
  • 16
    Chen, K. Y., Sun, M. (1997) Improving energy expenditure estimation by using a triaxial accelerometer. J Appl Physiol. 83: 21122122.
  • 17
    Plasqui, G., Joosen, A. M., Kester, A. D., Goris, A. H., Westerterp, KR. (2005) Measuring free-living energy expenditure and physical activity with triaxial accelerometry. Obes Res. 13: 13631369.
  • 18
    Crouter, S. E., Clowers, K. G., Bassett, D. R., Jr (2006) A novel method for using accelerometer data to predict energy expenditure. J Appl Physiol. 100: 13241331.
  • 19
    Bouten, C. V., Westerterp, K. R., Verduin, M., Janssen, JD. (1994) Assessment of energy expenditure for physical activity using a triaxial accelerometer. Med Sci Sports Exerc. 26: 15161523.
  • 20
    Ganpule, A. A., Tanaka, S., Ishikawa-Takata, K., Tabata, I. (2007) Interindividual variability in sleeping metabolic rate in Japanese subjects. Eur J Clin Nutr. 61: 12561261.