Application of cross-sectional time series modeling for the prediction of energy expenditure from heart rate and accelerometry

Are Activity Wrist-Worn Devices Accurate for Determining Heart Rate during Intense Exercise?

The market for wrist-worn devices is growing at previously unheard-of speeds. A consequence of their fast commercialization is a lack of adequate studies testing their accuracy on varied populations and pursuits. To provide an understanding of wearable sensors for sports medicine, the present study examined heart rate (HR) measurements of four popular wrist-worn devices, the (Fitbit Charge (FB), Apple Watch (AW), Tomtom runner Cardio (TT), and Samsung G2 (G2)), and compared them with gold standard measurements derived by continuous electrocardiogram examination (ECG). Eight athletes participated in a comparative study undergoing maximal stress testing on a cycle ergometer or a treadmill. We analyzed 1,286 simultaneous HR data pairs between the tested devices and the ECG. The four devices were reasonably accurate at the lowest activity level. However, at higher levels of exercise intensity the FB and G2 tended to underestimate HR values during intense physical effort, while the TT and AW devices were fairly reliable. Our results suggest that HR estimations should be considered cautiously at specific intensities. Indeed, an effective intervention is required to register accurate HR readings at high-intensity levels (above 150 bpm). It is important to consider that even though none of these devices are certified or sold as medical or safety devices, researchers must nonetheless evaluate wrist-worn wearable technology in order to fully understand how HR affects psychological and physical health, especially under conditions of more intense exercise.

Estimation of Heart Rate and Energy Expenditure Using a Smart Bracelet during Different Exercise Intensities: A Reliability and Validity Study

Background. With wrist-worn wearables becoming increasingly available, it is important to understand their reliability and validity in different conditions. The primary objective of this study was to examine the reliability and validity of the Lexin Mio smart bracelet in measuring heart rate (HR) and energy expenditure (EE) in people with different physical activity levels exercising at different intensities. Methods. A total of 65 participants completed one maximal oxygen uptake test and two running exercise tests wearing the Mio smart bracelet, the Polar H10 HR band, and a gas-analysis system. Results. In terms of HR measurement reliability, the Mio smart bracelet showed good reliability in a left versus right test and good test−retest reliability (p > 0.05; mean absolute percentage error (MAPE) < 10%; intraclass correlation coefficient (ICC) > 0.4). For EE measurement, the Mio smart bracelet showed good reliability in a left versus right test, good test−retest reliability on the right (p > 0.05; MAPE > 10%; ICC > 0.4), and low test−retest reliability on the left (p > 0.05; MAPE > 10%; ICC < 0.4). Regarding validity, the Mio smart bracelet showed good validity for HR measurement (p > 0.05; MAPE < 10%; ICC > 0.4) and low validity for EE measurement (p < 0.05; MAPE > 10%; ICC < 0.4). Conclusion. The Lexin Mio smart bracelet showed good reliability and validity for HR measurement among people with different physical activity levels exercising at various exercise intensities in a laboratory setting. However, the smart bracelet showed good reliability and low validity for the estimation of EE.

Integrating Multiple Inputs Into an Artificial Pancreas System: Narrative Literature Review

Background

Type 1 diabetes (T1D) is a chronic autoimmune disease in which a deficiency in insulin production impairs the glucose homeostasis of the body. Continuous subcutaneous infusion of insulin is a commonly used treatment method. Artificial pancreas systems (APS) use continuous glucose level monitoring and continuous subcutaneous infusion of insulin in a closed-loop mode incorporating a controller (or control algorithm). However, the operation of APS is challenging because of complexities arising during meals, exercise, stress, sleep, illnesses, glucose sensing and insulin action delays, and the cognitive burden. To overcome these challenges, options to augment APS through integration of additional inputs, creating multi-input APS (MAPS), are being investigated.

Objective

The aim of this survey is to identify and analyze input data, control architectures, and validation methods of MAPS to better understand the complexities and current state of such systems. This is expected to be valuable in developing improved systems to enhance the quality of life of people with T1D.

Methods

A literature survey was conducted using the Scopus, PubMed, and IEEE Xplore databases for the period January 1, 2005, to February 10, 2020. On the basis of the search criteria, 1092 articles were initially shortlisted, of which 11 (1.01%) were selected for an in-depth narrative analysis. In addition, 6 clinical studies associated with the selected studies were also analyzed.

Results

Signals such as heart rate, accelerometer readings, energy expenditure, and galvanic skin response captured by wearable devices were the most frequently used additional inputs. The use of invasive (blood or other body fluid analytes) inputs such as lactate and adrenaline were also simulated. These inputs were incorporated to switch the mode of the controller through activity detection, directly incorporated for decision-making and for the development of intermediate modules for the controller. The validation of the MAPS was carried out through the use of simulators based on different physiological models and clinical trials.

Conclusions

The integration of additional physiological signals with continuous glucose level monitoring has the potential to optimize glucose control in people with T1D through addressing the identified limitations of APS. Most of the identified additional inputs are related to wearable devices. The rapid growth in wearable technologies can be seen as a key motivator regarding MAPS. However, it is important to further evaluate the practical complexities and psychosocial aspects associated with such systems in real life.

Validity of Wrist-Worn photoplethysmography devices to measure heart rate: A systematic review and meta-analysis

Heart rate (HR), when combined with accelerometry, can dramatically improve estimates of energy expenditure and sleep. Advancements in technology, via the development and introduction of small, low-cost photoplethysmography devices embedded within wrist-worn consumer wearables, have made the collection of heart rate (HR) under free-living conditions more feasible. This systematic review and meta-analysis compared the validity of wrist-worn HR estimates to a criterion measure of HR (electrocardiography ECG or chest strap). Searches of PubMed/Medline, Web of Science, EBSCOhost, PsycINFO, and EMBASE resulted in a total of 44 articles representing 738 effect sizes across 15 different brands. Multi-level random effects meta-analyses resulted in a small mean difference (beats per min, bpm) of -0.40 bpm (95 confidence interval (CI) -1.64 to 0.83) during sleep, -0.01 bpm (-0.02 to 0.00) during rest, -0.51 bpm (-1.60 to 0.58) during treadmill activities (walking to running), while the mean difference was larger during resistance training (-7.26 bpm, -10.46 to -4.07) and cycling (-4.55 bpm, -7.24 to -1.87). Mean difference increased by 3 bpm (2.5 to 3.5) per 10 bpm increase of HR for resistance training. Wrist-worn devices that measure HR demonstrate acceptable validity compared to a criterion measure of HR for most common activities.

Adaptive Control of an Artificial Pancreas Using Model Identification, Adaptive Postprandial Insulin Delivery, and Heart Rate and Accelerometry as Control Inputs

BACKGROUND

People with type 1 diabetes (T1D) have varying sensitivities to insulin and also varying responses to meals and exercise. We introduce a new adaptive run-to-run model predictive control (MPC) algorithm that can be used to help people with T1D better manage their glucose levels using an artificial pancreas (AP). The algorithm adapts to individuals' different insulin sensitivities, glycemic response to meals, and adjustment during exercise as a continuous input during free-living conditions.

METHODS

A new insulin sensitivity adaptation (ISA) algorithm is presented that updates each patient's insulin sensitivity during nonmeal periods to reduce the error between the actual glucose levels and the process model. We further demonstrate how an adaptive learning postprandial hypoglycemia prevention algorithm (ALPHA) presented in the previous work can complement the ISA algorithm, and the algorithm can adapt in several days. We further show that if physical activity is incorporated as a continuous input (heart rate and accelerometry), performance is improved. The contribution of this work is the description of the ISA algorithm and the evaluation of how ISA, ALPHA, and incorporation of exercise metrics as a continuous input can impact glycemic control.

RESULTS

Incorporating ALPHA, ISA, and physical activity into the MPC improved glycemic outcome measures. The adaptive learning postprandial hypoglycemia prevention algorithm combined with ISA significantly reduced time spent in hypoglycemia by 71.7% and the total number of rescue carbs by 67.8% to 0.37% events/day/patient. Insulin sensitivity adaptation significantly reduced model-actual mismatch by 12.2% compared to an AP without ISA. Incorporating physical activity as a continuous input modestly improved time in the range 70 to 180 mg/dL during high physical activity days from 84.4% to 84.9% and reduced the percentage time in hypoglycemia by 23.8% from 2.1% to 1.6%.

CONCLUSION

Adapting postprandial insulin delivery, insulin sensitivity, and adapting to physical exercise in an MPC-based AP systems can improve glycemic outcomes.

Accuracy of the Multisensory Wristwatch Polar Vantage's Estimation of Energy Expenditure in Various Activities: Instrument Validation Study

Background

Sport watches and fitness trackers provide a feasible way of obtaining energy expenditure (EE) estimations in daily life as well as during exercise. However, today’s popular wrist-worn technologies show only poor-to-moderate EE accuracy. Recently, the invention of optical heart rate measurement and the further development of accelerometers in wrist units have opened up the possibility of measuring EE.

Objective

This study aimed to validate the new multisensory wristwatch Polar Vantage and its EE estimation in healthy individuals during low-to-high-intensity activities against indirect calorimetry.

Methods

Overall, 30 volunteers (15 females; mean age 29.5 [SD 5.1] years; mean height 1.7 [SD 0.8] m; mean weight 67.5 [SD 8.7] kg; mean maximal oxygen uptake 53.4 [SD 6.8] mL/min·kg) performed 7 activities—ranging in intensity from sitting to playing floorball—in a semistructured indoor environment for 10 min each, with 2-min breaks in between. These activities were performed while wearing the Polar Vantage M wristwatch and the MetaMax 3B spirometer.

Results

After EE estimation, a mean (SD) of 69.1 (42.7) kcal and 71.4 (37.8) kcal per 10-min activity were reported for the MetaMax 3B and the Polar Vantage, respectively, with a strong correlation of r=0.892 (P<.001). The systematic bias was 2.3 kcal (3.3%), with 37.8 kcal limits of agreement. The lowest mean absolute percentage errors were reported during the sitting and reading activities (9.1%), and the highest error rates during household chores (31.4%). On average, 59.5% of the mean EE values obtained by the Polar Vantage were within ±20% of accuracy when compared with the MetaMax 3B. The activity intensity quantified by perceived exertion (odds ratio [OR] 2.028; P<.001) and wrist circumference (OR −1.533; P=.03) predicted 29% of the error rates within the Polar Vantage.

Conclusions

The Polar Vantage has a statistically moderate-to-good accuracy in EE estimation that is activity dependent. During sitting and reading activities, the EE estimation is very good, whereas during nonsteady activities that require wrist and arm movement, the EE accuracy is only moderate. However, compared with other available wrist-worn EE monitors, the Polar Vantage can be recommended, as it performs among the best.

A statistical virtual patient population for the glucoregulatory system in type 1 diabetes with integrated exercise model

Purpose

We introduce two validated single (SH) and dual hormone (DH) mathematical models that represent an in-silico virtual patient population (VPP) for type 1 diabetes (T1D). The VPP can be used to evaluate automated insulin and glucagon delivery algorithms, so-called artificial pancreas (AP) algorithms that are currently being used to help people with T1D better manage their glucose levels. We present validation results comparing these virtual patients with true clinical patients undergoing AP control and demonstrate that the virtual patients behave similarly to people with T1D.

Methods

A single hormone virtual patient population (SH-VPP) was created that is comprised of eight differential equations that describe insulin kinetics, insulin dynamics and carbohydrate absorption. The parameters in this model that represent insulin sensitivity were statistically sampled from a normal distribution to create a population of virtual patients with different levels of insulin sensitivity. A dual hormone virtual patient population (DH-VPP) extended this SH-VPP by incorporating additional equations to represent glucagon kinetics and glucagon dynamics. The DH-VPP is comprised of thirteen differential equations and a parameter representing glucagon sensitivity, which was statistically sampled from a normal distribution to create virtual patients with different levels of glucagon sensitivity. We evaluated the SH-VPP and DH-VPP on a clinical data set of 20 people with T1D who participated in a 3.5-day outpatient AP study. Twenty virtual patients were matched with the 20 clinical patients by total daily insulin requirements and body weight. The identical meals given during the AP study were given to the virtual patients and the identical AP control algorithm that was used to control the glucose of the virtual patients was used on the clinical patients. We compared percent time in target range (70–180 mg/dL), time in hypoglycemia (<70 mg/dL) and time in hyperglycemia (>180 mg/dL) for both the virtual patients and the actual patients.

Results

The subjects in the SH-VPP performed similarly vs. the actual patients (time in range: 78.1 ± 5.1% vs. 74.3 ± 8.1%, p = 0.11; time in hypoglycemia: 3.4 ± 1.3% vs. 2.8 ± 1.7%, p = 0.23). The subjects in the DH-VPP also performed similarly vs. the actual patients (time in range: 75.6 ± 5.5% vs. 71.9 ± 10.9%, p = 0.13; time in hypoglycemia: 0.9 ± 0.8% vs. 1.3 ± 1%, p = 0.19). While the VPPs tended to over-estimate the time in range relative to actual patients, the difference was not statistically significant.

Conclusions

We have verified that a SH-VPP and a DH-VPP performed comparably with actual patients undergoing AP control using an identical control algorithm. The SH-VPP and DH-VPP may be used as a simulator for pre-evaluation of T1D control algorithms.

Accuracy of Wrist-Worn Activity Monitors During Common Daily Physical Activities and Types of Structured Exercise: Evaluation Study

BACKGROUND

Wrist-worn activity monitors are often used to monitor heart rate (HR) and energy expenditure (EE) in a variety of settings including more recently in medical applications. The use of real-time physiological signals to inform medical systems including drug delivery systems and decision support systems will depend on the accuracy of the signals being measured, including accuracy of HR and EE. Prior studies assessed accuracy of wearables only during steady-state aerobic exercise.

OBJECTIVE

The objective of this study was to validate the accuracy of both HR and EE for 2 common wrist-worn devices during a variety of dynamic activities that represent various physical activities associated with daily living including structured exercise.

METHODS

We assessed the accuracy of both HR and EE for two common wrist-worn devices (Fitbit Charge 2 and Garmin vívosmart HR+) during dynamic activities. Over a 2-day period, 20 healthy adults (age: mean 27.5 [SD 6.0] years; body mass index: mean 22.5 [SD 2.3] kg/m²; 11 females) performed a maximal oxygen uptake test, free-weight resistance circuit, interval training session, and activities of daily living. Validity was assessed using an HR chest strap (Polar) and portable indirect calorimetry (Cosmed). Accuracy of the commercial wearables versus research-grade standards was determined using Bland-Altman analysis, correlational analysis, and error bias.

RESULTS

Fitbit and Garmin were reasonably accurate at measuring HR but with an overall negative bias. There was more error observed during high-intensity activities when there was a lack of repetitive wrist motion and when the exercise mode indicator was not used. The Garmin estimated HR with a mean relative error (RE, %) of -3.3% (SD 16.7), whereas Fitbit estimated HR with an RE of -4.7% (SD 19.6) across all activities. The highest error was observed during high-intensity intervals on bike (Fitbit: -11.4% [SD 35.7]; Garmin: -14.3% [SD 20.5]) and lowest error during high-intensity intervals on treadmill (Fitbit: -1.7% [SD 11.5]; Garmin: -0.5% [SD 9.4]). Fitbit and Garmin EE estimates differed significantly, with Garmin having less negative bias (Fitbit: -19.3% [SD 28.9], Garmin: -1.6% [SD 30.6], P<.001) across all activities, and with both correlating poorly with indirect calorimetry measures.

CONCLUSIONS

Two common wrist-worn devices (Fitbit Charge 2 and Garmin vívosmart HR+) show good HR accuracy, with a small negative bias, and reasonable EE estimates during low to moderate-intensity exercise and during a variety of common daily activities and exercise. Accuracy was compromised markedly when the activity indicator was not used on the watch or when activities involving less wrist motion such as cycle ergometry were done.

Energy expenditure estimation from respiration variables

The aim of this study was to develop and cross-validate two models to estimate total energy expenditure (TEE) based on respiration variables in healthy subjects during daily physical activities. Ninety-nine male and female subjects systematically varying in age (18-60 years) and body mass index (BMI; 17-36 kg*m-2) completed eleven aerobic activities with a portable spirometer as the criterion measure. Two models were developed using linear regression analyses with the data from 67 randomly selected subjects (50.0% female, 39.9 ± 11.8 years, 25.1 ± 5.2 kg*m-2). The models were cross-validated with the other 32 subjects (49% female, 40.4 ± 10.7 years, 24.7 ± 4.6 kg*m-2) by applying equivalence testing and Bland-and-Altman analyses. Model 1, estimating TEE based solely on respiratory volume, respiratory rate, and age, was significantly equivalent to the measured TEE with a systematic bias of 0.06 kJ*min-1 (0.22%) and limits of agreement of ±6.83 kJ*min-1. Model 1 was as accurate in estimating TEE as Model 2, which incorporated further information on activity categories, heart rate, sex, and BMI. The results demonstrated that respiration variables and age can be used to accurately determine daily TEE for different types of aerobic activities in healthy adults across a broad range of ages and body sizes.

An original piecewise model for computing energy expenditure from accelerometer and heart rate signals

OBJECTIVE

Activity energy expenditure (EE) plays an important role in healthcare, therefore, accurate EE measures are required. Currently available reference EE acquisition methods, such as doubly labeled water and indirect calorimetry, are complex, expensive, uncomfortable, and/or difficult to apply on real time. To overcome these drawbacks, the goal of this paper is to propose a model for computing EE in real time (minute-by-minute) from heart rate and accelerometer signals.

APPROACH

The proposed model, which consists of an original branched model, uses heart rate signals for computing EE on moderate to vigorous physical activities and a linear combination of heart rate and counts per minute for computing EE on light to moderate physical activities. Model parameters were estimated from a given data set composed of 53 subjects performing 25 different physical activities (light-, moderate- and vigorous-intensity), and validated using leave-one-subject-out. A different database (semi-controlled in-city circuit), was used in order to validate the versatility of the proposed model. Comparisons are done versus linear and nonlinear models, which are also used for computing EE from accelerometer and/or HR signals.

MAIN RESULTS

The proposed piecewise model leads to more accurate EE estimations ([Formula: see text], [Formula: see text] and [Formula: see text] J kg^-1 min^-1 and [Formula: see text], [Formula: see text], and [Formula: see text] J kg^-1 min^-1 on each validation database).

SIGNIFICANCE

This original approach, which is more conformable and less expensive than the reference methods, allows accurate EE estimations, in real time (minute-by-minute), during a large variety of physical activities. Therefore, this model may be used on applications such as computing the time that a given subject spent on light-intensity physical activities and on moderate to vigorous physical activities (binary classification accuracy of 0.8155).

Use of Computer and Mobile Technologies in the Treatment of Depression

Major depression (MDD) is a common and disabling disorder. Research has shown that most people with MDD receive either no treatment or inadequate treatment. Computer and mobile technologies may offer solutions for the delivery of therapies to untreated or inadequately treated individuals with MDD. The authors review currently available technologies and research aimed at relieving symptoms of MDD. These technologies include computer-assisted cognitive-behavior therapy (CCBT), web-based self-help, Internet self-help support groups, mobile psychotherapeutic interventions (i.e., mobile applications or apps), technology enhanced exercise, and biosensing technology.

Functional data analysis of sleeping energy expenditure

Adequate sleep is crucial during childhood for metabolic health, and physical and cognitive development. Inadequate sleep can disrupt metabolic homeostasis and alter sleeping energy expenditure (SEE). Functional data analysis methods were applied to SEE data to elucidate the population structure of SEE and to discriminate SEE between obese and non-obese children. Minute-by-minute SEE in 109 children, ages 5-18, was measured in room respiration calorimeters. A smoothing spline method was applied to the calorimetric data to extract the true smoothing function for each subject. Functional principal component analysis was used to capture the important modes of variation of the functional data and to identify differences in SEE patterns. Combinations of functional principal component analysis and classifier algorithm were used to classify SEE. Smoothing effectively removed instrumentation noise inherent in the room calorimeter data, providing more accurate data for analysis of the dynamics of SEE. SEE exhibited declining but subtly undulating patterns throughout the night. Mean SEE was markedly higher in obese than non-obese children, as expected due to their greater body mass. SEE was higher among the obese than non-obese children (p<0.01); however, the weight-adjusted mean SEE was not statistically different (p>0.1, after post hoc testing). Functional principal component scores for the first two components explained 77.8% of the variance in SEE and also differed between groups (p = 0.037). Logistic regression, support vector machine or random forest classification methods were able to distinguish weight-adjusted SEE between obese and non-obese participants with good classification rates (62-64%). Our results implicate other factors, yet to be uncovered, that affect the weight-adjusted SEE of obese and non-obese children. Functional data analysis revealed differences in the structure of SEE between obese and non-obese children that may contribute to disruption of metabolic homeostasis.

Incorporating an Exercise Detection, Grading, and Hormone Dosing Algorithm Into the Artificial Pancreas Using Accelerometry and Heart Rate

In this article, we present several important contributions necessary for enabling an artificial endocrine pancreas (AP) system to better respond to exercise events. First, we show how exercise can be automatically detected using body-worn accelerometer and heart rate sensors. During a 22 hour overnight inpatient study, 13 subjects with type 1 diabetes wearing a Zephyr accelerometer and heart rate monitor underwent 45 minutes of mild aerobic treadmill exercise while controlling their glucose levels using sensor-augmented pump therapy. We used the accelerometer and heart rate as inputs into a validated regression model. Using this model, we were able to detect the exercise event with a sensitivity of 97.2% and a specificity of 99.5%. Second, from this same study, we show how patients' glucose declined during the exercise event and we present results from in silico modeling that demonstrate how including an exercise model in the glucoregulatory model improves the estimation of the drop in glucose during exercise. Last, we present an exercise dosing adjustment algorithm and describe parameter tuning and performance using an in silico glucoregulatory model during an exercise event.

The current state of physical activity assessment tools

Physical activity (PA) is a behavior that involves bodily movements resulting in energy expenditure. When assessing PA, the goal is to identify the frequency, duration, intensity, and types of behaviors performed during a period of time. Self-report measures of PA include administration of questionnaires and completion of detailed diaries and/or brief logs. Direct measures include motion sensors such as accelerometers, pedometers, heart-rate monitors, and multiple-sensor devices. The PA assessment period can range from a few hours to a lifetime depending on the tools used. Considerations when selecting a PA tool should include the literacy requirements of a tool, the purpose for assessing PA, the recall or time period to measure, the validity evidence of an assessment tool for the populations measured, and the generalizability of the results to diverse populations.

Comparison of total energy expenditure assessed by two devices in controlled and free-living conditions

The objective of this study was to evaluate the validity of total energy expenditure (TEE) provided by Actiheart and Armband. Normal-weight adult volunteers wore both devices either for 17 hours in a calorimetric chamber (CC, n = 49) or for 10 days in free-living conditions (FLC) outside the laboratory (n = 41). The two devices and indirect calorimetry or doubly labelled water, respectively, were used to estimate TEE in the CC group and FLC group. In the CC, the relative value of TEE error was not significant (p > 0.05) for Actiheart but significantly different from zero for Armband, showing TEE underestimation (-4.9%, p < 0.0001). However, the mean absolute values of errors were significantly different between Actiheart and Armband: 8.6% and 6.7%, respectively (p = 0.05). Armband was more accurate for estimating TEE during sleeping, rest, recovery periods and sitting-standing. Actiheart provided better estimation during step and walking. In FLC, no significant error in relative value was detected. Nevertheless, Armband produced smaller errors in absolute value than Actiheart (8.6% vs. 12.8%). The distributions of differences were more scattered around the means, suggesting a higher inter-individual variability in TEE estimated by Actiheart than by Armband. Our results show that both monitors are appropriate for estimating TEE. Armband is more effective than Actiheart at the individual level for daily light-intensity activities.

Modeling energy expenditure in children and adolescents using quantile regression

Advanced mathematical models have the potential to capture the complex metabolic and physiological processes that result in energy expenditure (EE). Study objective is to apply quantile regression (QR) to predict EE and determine quantile-dependent variation in covariate effects in nonobese and obese children. First, QR models will be developed to predict minute-by-minute awake EE at different quantile levels based on heart rate (HR) and physical activity (PA) accelerometry counts, and child characteristics of age, sex, weight, and height. Second, the QR models will be used to evaluate the covariate effects of weight, PA, and HR across the conditional EE distribution. QR and ordinary least squares (OLS) regressions are estimated in 109 children, aged 5-18 yr. QR modeling of EE outperformed OLS regression for both nonobese and obese populations. Average prediction errors for QR compared with OLS were not only smaller at the median τ = 0.5 (18.6 vs. 21.4%), but also substantially smaller at the tails of the distribution (10.2 vs. 39.2% at τ = 0.1 and 8.7 vs. 19.8% at τ = 0.9). Covariate effects of weight, PA, and HR on EE for the nonobese and obese children differed across quantiles (P < 0.05). The associations (linear and quadratic) between PA and HR with EE were stronger for the obese than nonobese population (P < 0.05). In conclusion, QR provided more accurate predictions of EE compared with conventional OLS regression, especially at the tails of the distribution, and revealed substantially different covariate effects of weight, PA, and HR on EE in nonobese and obese children.

Support vector machines classifiers of physical activities in preschoolers

The goal of this study is to develop, test, and compare multinomial logistic regression (MLR) and support vector machines (SVM) in classifying preschool-aged children physical activity data acquired from an accelerometer. In this study, 69 children aged 3-5 years old were asked to participate in a supervised protocol of physical activities while wearing a triaxial accelerometer. Accelerometer counts, steps, and position were obtained from the device. We applied K-means clustering to determine the number of natural groupings presented by the data. We used MLR and SVM to classify the six activity types. Using direct observation as the criterion method, the 10-fold cross-validation (CV) error rate was used to compare MLR and SVM classifiers, with and without sleep. Altogether, 58 classification models based on combinations of the accelerometer output variables were developed. In general, the SVM classifiers have a smaller 10-fold CV error rate than their MLR counterparts. Including sleep, a SVM classifier provided the best performance with a 10-fold CV error rate of 24.70%. Without sleep, a SVM classifier-based triaxial accelerometer counts, vector magnitude, steps, position, and 1- and 2-min lag and lead values achieved a 10-fold CV error rate of 20.16% and an overall classification error rate of 15.56%. SVM supersedes the classical classifier MLR in categorizing physical activities in preschool-aged children. Using accelerometer data, SVM can be used to correctly classify physical activities typical of preschool-aged children with an acceptable classification error rate.

Computational methods for estimating energy expenditure in human physical activities

Accurate and reliable methods for assessing human physical activity (PA) energy expenditure (PAEE) are informative and essential for understanding individual behaviors and quantifying the effect of PA on disease, for PA surveillance, and for examining determinants of PA in different populations. This article reviews recent advances in the estimation of PAEE in three interrelated areas: 1) types of sensors worn by human subjects, 2) features extracted from the measured sensor signals, and 3) modeling techniques to estimate the PAEE using these features. The review illustrates three directions in the PAEE studies and provides recommendations for future research, with the aim to produce valid, reliable, and accurate assessment of PAEE from wearable sensors.

Cross-sectional time series and multivariate adaptive regression splines models using accelerometry and heart rate predict energy expenditure of preschoolers

Prediction equations of energy expenditure (EE) using accelerometers and miniaturized heart rate (HR) monitors have been developed in older children and adults but not in preschool-aged children. Because the relationships between accelerometer counts (ACs), HR, and EE are confounded by growth and maturation, age-specific EE prediction equations are required. We used advanced technology (fast-response room calorimetry, Actiheart and Actigraph accelerometers, and miniaturized HR monitors) and sophisticated mathematical modeling [cross-sectional time series (CSTS) and multivariate adaptive regression splines (MARS)] to develop models for the prediction of minute-by-minute EE in 69 preschool-aged children. CSTS and MARS models were developed by using participant characteristics (gender, age, weight, height), Actiheart (HR+AC_x) or ActiGraph parameters (AC_x, AC_y, AC_z, steps, posture) [x, y, and z represent the directional axes of the accelerometers], and their significant 1- and 2-min lag and lead values, and significant interactions. Relative to EE measured by calorimetry, mean percentage errors predicting awake EE (-1.1 ± 8.7%, 0.3 ± 6.9%, and -0.2 ± 6.9%) with CSTS models were slightly higher than with MARS models (-0.7 ± 6.0%, 0.3 ± 4.8%, and -0.6 ± 4.6%) for Actiheart, ActiGraph, and ActiGraph+HR devices, respectively. Predicted awake EE values were within ±10% for 81-87% of individuals for CSTS models and for 91-98% of individuals for MARS models. Concordance correlation coefficients were 0.936, 0.931, and 0.943 for CSTS EE models and 0.946, 0.948, and 0.940 for MARS EE models for Actiheart, ActiGraph, and ActiGraph+HR devices, respectively. CSTS and MARS models should prove useful in capturing the complex dynamics of EE and movement that are characteristic of preschool-aged children.

Assessing physical activity using wearable monitors: measures of physical activity

BACKGROUND

Physical activity may be defined broadly as "all bodily actions produced by the contraction of skeletal muscle that increase energy expenditure above basal level." Physical activity is a complex construct that can be classified into major categories qualitatively, quantitatively, or contextually. The quantitative assessment of physical activity using wearable monitors is grounded in the measurement of energy expenditure. Six main categories of wearable monitors are currently available to investigators: pedometers, load transducers/foot-contact monitors, accelerometers, HR monitors, combined accelerometer and HR monitors, and multiple sensor systems.

BEST PRACTICES

Currently available monitors are capable of measuring total physical activity as well as components of physical activity that play important roles in human health. The selection of wearable monitors for measuring physical activity will depend on the physical activity component of interest, study objectives, characteristics of the target population, and study feasibility in terms of cost and logistics.

FUTURE DIRECTIONS

Future development of sensors and analytical techniques for assessing physical activity should focus on the dynamic ranges of sensors, comparability for sensor output across manufacturers, and the application of advanced modeling techniques to predict energy expenditure and classify physical activities. New approaches for qualitatively classifying physical activity should be validated using direct observation or recording. New sensors and methods for quantitatively assessing physical activity should be validated in laboratory and free-living populations using criterion methods of calorimetry or doubly labeled water.

Measurement of human energy expenditure, with particular reference to field studies: an historical perspective

Over the years, techniques for the study of human movement have ranged in complexity and precision from direct observation of the subject through activity diaries, questionnaires, and recordings of body movement, to the measurement of physiological responses, studies of metabolism and indirect and direct calorimetry. This article reviews developments in each of these domains. Particular reference is made to their impact upon the continuing search for valid field estimates of activity patterns and energy expenditures, as required by the applied physiologist, ergonomist, sports scientist, nutritionist and epidemiologist. Early observers sought to improve productivity in demanding employment. Direct observation and filming of workers were supplemented by monitoring of heart rates, ventilation and oxygen consumption. Such methods still find application in ergonomics and sport, but many investigators are now interested in relationships between habitual physical activity and chronic disease. Even sophisticated questionnaires still do not provide valid information on the absolute energy expenditures associated with good health. Emphasis has thus shifted to use of sophisticated pedometer/accelerometers, sometimes combining their output with GPS and other data. Some modern pedometer/accelerometers perform well in the laboratory, but show substantial systematic errors relative to laboratory reference criteria such as the metabolism of doubly labeled water when assessing the varied activities of daily life. The challenge remains to develop activity monitors that are sufficiently inexpensive for field use, yet meet required accuracy standards. Possibly, measurements of oxygen consumption by portable respirometers may soon satisfy part of this need, although a need for valid longer term monitoring will remain.

A new method to estimate energy expenditure from abdominal and rib cage distances

The aim of this paper is to validate a new method of energy expenditure (EE) estimation stemming solely from the measurement of rib cage, abdominal and chest wall distances. We set out to prove that the variations of these distances, measured by two pairs of electromagnetic coils, lead to the estimation of the ventilation (VE) and the EE. Eleven subjects were recruited to take part in this study (27.6 ± 5.4 years; 73.7 ± 9.7 kg). Each subject participated in two tests. The objective of Test 1 was to determine the individual and group equations between the VE and EE during light to moderate activities while Test 2 compared the two pairs of electromagnetic coils with the indirect calorimetry so as to estimate EE in upright sitting and standing positions and during walking exercises. During Test 2, we compared EE measured by indirect calorimetry (EE(IC-Val-REF)) with EE estimated by the two pairs of electromagnetic coils through the application of: (1) the individual equation (EE(mag-Val-INDIV)) and (2) the group equation (EE(mag-val-GROUP)). The results show that there is no significant difference between EE(IC-Val-REF) and EE(mag-Val-INDIV) and between EE(IC-Val-REF) and EE(mag-val-GROUP) for each activity. Furthermore, the mean difference seems to show that the estimation of EE is better with the group equation. In conclusion, on the proven basis of this study we are able to validate this new method which permits the estimation of EE from abdominal and rib cage distances. This study also highlights the advantage of using a group equation to the estimate EE.

Validation of the ActiGraph two-regression model for predicting energy expenditure

PURPOSE

The purpose of this study was to validate a two-regression model for predicting energy expenditure (EE) from ActiGraph GT1M accelerometer-generated activity counts using a whole-room indirect calorimeter and the doubly labeled water (DLW) technique. We also investigated if a low-pass filter (LPF) approach would improve the model's accuracy in the minute-to-minute EE prediction.

METHODS

Thirty-four healthy volunteers (age = 20-67 yr, body mass index = 19.3-52.1 kg.m) spent approximately 24 h in a room calorimeter while wearing a GT1M monitor and performed structured and self-selected activities followed by overnight sleep. The EE predicted by the models and expressed in metabolic equivalents (MET-minutes) during waking times was compared with the room calorimeter-measured EE. A subset of volunteers (n = 22) completed a 14-d DLW protocol in free living while wearing an ActiGraph. The average daily EE predicted by the models (MET-minutes) was compared with the DLW.

RESULTS

Compared with the room calorimeter, the two-regression model overpredicted EE by 10.2% +/- 11.4% (1282 +/- 125 and 1174 +/- 152 MET.min, P < 0.001) and time spent in moderate physical activity (PA) by 36.9 +/- 46.0 min while underestimating the time spent in light PA by -48.3 +/- 55.0 min (P < 0.05). The LPF reduced the squared and mean absolute error in the EE prediction (P < 0.05) but not the prediction error in time spent in moderate or light PA (both P > 0.05). The EE measured by DLW (2108 +/- 358 MET.min.d) and predicted by both filtered and unfiltered models (2104 +/- 218 and 2192 +/- 228 MET.min.d, respectively) were similar (P > 0.05).

CONCLUSIONS

The two-regression model with LPF showed good agreement with total EE measured using room calorimeter and DLW. However, the individual variability in assessing time spent in sedentary, low, and moderate PA intensities and related EE remains significant.

Estimation of free-living energy expenditure using a novel activity monitor designed to minimize obtrusiveness

The aim of this study was to investigate the ability of a novel activity monitor designed to be minimally obtrusive in predicting free-living energy expenditure. Subjects were 18 men and 12 women (age: 41 +/- 11 years, BMI: 24.4 +/- 3 kg/m(2)). The habitual physical activity was monitored for 14 days using a DirectLife triaxial accelerometer for movement registration (Tracmor(D)) (Philips New Wellness Solutions, Lifestyle Incubator, the Netherlands). Tracmor(D) output was expressed as activity counts per day (Cnts/d). Simultaneously, total energy expenditure (TEE) was measured in free living conditions using doubly labeled water (DLW). Activity energy expenditure (AEE) and the physical activity level (PAL) were determined from TEE and sleeping metabolic rate (SMR). A multiple-linear regression model predicted 76% of the variance in TEE, using as independent variables SMR (partial-r(2) = 0.55, P < 0.001), and Cnts/d (partial r(2) = 0.21, P < 0.001). The s.e. of TEE estimates was 0.9 MJ/day or 7.4% of the average TEE. A model based on body mass (partial-r(2) = 0.31, P < 0.001) and Cnts/d (partial-r(2) = 0.23, P < 0.001) predicted 54% of the variance in TEE. Cnts/d were significantly and positively associated with AEE (r = 0.54, P < 0.01), PAL (r = 0.68, P < 0.001), and AEE corrected by body mass (r = 0.71, P < 0.001). This study showed that the Tracmor(D) is a highly accurate instrument for predicting free-living energy expenditure. The miniaturized design did not harm the ability of the instrument in measuring physical activity and in determining outcome parameters of physical activity such as TEE, AEE, and PAL.

Validation of cross-sectional time series and multivariate adaptive regression splines models for the prediction of energy expenditure in children and adolescents using doubly labeled water

Accurate, nonintrusive, and inexpensive techniques are needed to measure energy expenditure (EE) in free-living populations. Our primary aim in this study was to validate cross-sectional time series (CSTS) and multivariate adaptive regression splines (MARS) models based on observable participant characteristics, heart rate (HR), and accelerometer counts (AC) for prediction of minute-by-minute EE, and hence 24-h total EE (TEE), against a 7-d doubly labeled water (DLW) method in children and adolescents. Our secondary aim was to demonstrate the utility of CSTS and MARS to predict awake EE, sleep EE, and activity EE (AEE) from 7-d HR and AC records, because these shorter periods are not verifiable by DLW, which provides an estimate of the individual's mean TEE over a 7-d interval. CSTS and MARS models were validated in 60 normal-weight and overweight participants (ages 5-18 y). The Actiheart monitor was used to simultaneously measure HR and AC. For prediction of TEE, mean absolute errors were 10.7 +/- 307 kcal/d and 18.7 +/- 252 kcal/d for CSTS and MARS models, respectively, relative to DLW. Corresponding root mean square error values were 305 and 251 kcal/d for CSTS and MARS models, respectively. Bland-Altman plots indicated that the predicted values were in good agreement with the DLW-derived TEE values. Validation of CSTS and MARS models based on participant characteristics, HR monitoring, and accelerometry for the prediction of minute-by-minute EE, and hence 24-h TEE, against the DLW method indicated no systematic bias and acceptable limits of agreement for pediatric groups and individuals under free-living conditions.

Multivariate adaptive regression splines models for the prediction of energy expenditure in children and adolescents

Advanced mathematical models have the potential to capture the complex metabolic and physiological processes that result in heat production or energy expenditure (EE). Multivariate adaptive regression splines (MARS) is a nonparametric method that estimates complex nonlinear relationships by a series of spline functions of the independent predictors. The specific aim of this study is to construct MARS models based on heart rate (HR) and accelerometer counts (AC) to accurately predict EE, and hence 24-h total EE (TEE), in children and adolescents. Secondarily, MARS models will be developed to predict awake EE, sleep EE, and activity EE also from HR and AC. MARS models were developed in 109 and validated in 61 normal-weight and overweight children (ages 5–18 yr) against the criterion method of 24-h room respiration calorimetry. Actiheart monitor was used to measure HR and AC. MARS models were based on linear combinations of 23–28 basis functions that use subject characteristics (age, sex, weight, height, minimal HR, and sitting HR), HR and AC, 1- and 2-min lag and lead values of HR and AC, and appropriate interaction terms. For the 24-h, awake, sleep, and activity EE models, mean percent errors were −2.5 ± 7.5, −2.6 ± 7.8, −0.3 ± 8.9, and −11.9 ± 17.9%, and root mean square error values were 168, 138, 40, and 122 kcal, respectively, in the validation cohort. Bland-Altman plots indicated that the predicted values were in good agreement with the observed TEE, and that there was no bias with increasing TEE. Prediction errors for 24-h TEE were not statistically associated with age, sex, weight, height, or body mass index. MARS models developed for the prediction of EE from HR monitoring and accelerometry were demonstrated to be valid in an independent cohort of children and adolescents, but require further validation in independent, free-living populations.

Distributed lag and spline modeling for predicting energy expenditure from accelerometry in youth

Movement sensing using accelerometers is commonly used for the measurement of physical activity (PA) and estimating energy expenditure (EE) under free-living conditions. The major limitation of this approach is lack of accuracy and precision in estimating EE, especially in low-intensity activities. Thus the objective of this study was to investigate benefits of a distributed lag spline (DLS) modeling approach for the prediction of total daily EE (TEE) and EE in sedentary (1.0-1.5 metabolic equivalents; MET), light (1.5-3.0 MET), and moderate/vigorous (> or = 3.0 MET) intensity activities in 10- to 17-year-old youth (n = 76). We also explored feasibility of the DLS modeling approach to predict physical activity EE (PAEE) and METs. Movement was measured by Actigraph accelerometers placed on the hip, wrist, and ankle. With whole-room indirect calorimeter as the reference standard, prediction models (Hip, Wrist, Ankle, Hip+Wrist, Hip+Wrist+Ankle) for TEE, PAEE, and MET were developed and validated using the fivefold cross-validation method. The TEE predictions by these DLS models were not significantly different from the room calorimeter measurements (all P > 0.05). The Hip+Wrist+Ankle predicted TEE better than other models and reduced prediction errors in moderate/vigorous PA for TEE, MET, and PAEE (all P < 0.001). The Hip+Wrist reduced prediction errors for the PAEE and MET at sedentary PA (P = 0.020 and 0.021) compared with the Hip. Models that included Wrist correctly classified time spent at light PA better than other models. The means and standard deviations of the prediction errors for the Hip+Wrist+Ankle and Hip were 0.4 +/- 144.0 and 1.5 +/- 164.7 kcal for the TEE, 0.0 +/- 84.2 and 1.3 +/- 104.7 kcal for the PAEE, and -1.1 +/- 97.6 and -0.1 +/- 108.6 MET min for the MET models. We conclude that the DLS approach for accelerometer data improves detailed EE prediction in youth.

An artificial neural network to estimate physical activity energy expenditure and identify physical activity type from an accelerometer

The aim of this investigation was to develop and test two artificial neural networks (ANN) to apply to physical activity data collected with a commonly used uniaxial accelerometer. The first ANN model estimated physical activity metabolic equivalents (METs), and the second ANN identified activity type. Subjects (n = 24 men and 24 women, mean age = 35 yr) completed a menu of activities that included sedentary, light, moderate, and vigorous intensities, and each activity was performed for 10 min. There were three different activity menus, and 20 participants completed each menu. Oxygen consumption (in ml x kg(-1) x min(-1)) was measured continuously, and the average of minutes 4-9 was used to represent the oxygen cost of each activity. To calculate METs, activity oxygen consumption was divided by 3.5 ml x kg(-1) x min(-1) (1 MET). Accelerometer data were collected second by second using the Actigraph model 7164. For the analysis, we used the distribution of counts (10th, 25th, 50th, 75th, and 90th percentiles of a minute's second-by-second counts) and temporal dynamics of counts (lag, one autocorrelation) as the accelerometer feature inputs to the ANN. To examine model performance, we used the leave-one-out cross-validation technique. The ANN prediction of METs root-mean-squared error was 1.22 METs (confidence interval: 1.14-1.30). For the prediction of activity type, the ANN correctly classified activity type 88.8% of the time (confidence interval: 86.4-91.2%). Activity types were low-level activities, locomotion, vigorous sports, and household activities/other activities. This novel approach of applying ANNs for processing Actigraph accelerometer data is promising and shows that we can successfully estimate activity METs and identify activity type using ANN analytic procedures.

Estimating activity-related energy expenditure under sedentary conditions using a tri-axial seismic accelerometer

Activity-related energy expenditure (AEE) is difficult to quantify, especially under sedentary conditions. Here, a model was developed using the detected type of physical activity (PA) and movement intensity (MI), based on a tri-axial seismic accelerometer (DynaPort MiniMod; McRoberts B.V., The Hague, the Netherlands), with energy expenditure for PA as a reference. The relation between AEE (J/min/kg), MI, and the type of PA was determined for standardized PAs as performed in a laboratory including: lying, sitting, standing, and walking. AEE (J/min/kg) was calculated from total energy expenditure (TEE) and sleeping metabolic rate (SMR) as assessed with indirect calorimetry ((TEEx0.9)-SMR). Subsequently, the model was validated over 23-h intervals in a respiration chamber. Subjects were 15 healthy women (age: 22+/-2 years; BMI: 24.0+/-4.0 kg/m2). Predicted AEE in the chamber was significantly related to measured AEE both within (r2=0.81+/-0.06, P<0.00001) and between (r2=0.70, P<0.001) subjects. The explained variation in AEE by the model was higher than the explained variation by MI alone. This shows that a tri-axial seismic accelerometer is a valid tool for estimating AEE under sedentary conditions.

Assessment of physical activity in youth

Despite much progress with physical activity assessment, the limitations concerning the accurate measurement of physical activity are often amplified in young people due to the cognitive, physiological, and biomechanical changes that occur during natural growth as well as a more intermittent pattern of habitual physical activity in youth compared with adults. This mini-review describes and compares methods to assess habitual physical activity in youth and discusses main issues regarding the use and interpretation of data collected with these techniques. Self-report instruments and movement sensing are currently the most frequently used methods for the assessment of physical activity in epidemiological research; others include heart rate monitoring and multisensor systems. Habitual energy expenditure can be estimated from these input measures with varying degree of uncertainty. Nonlinear modeling techniques, using accelerometry perhaps in combination with physiological parameters like heart rate or temperature, have the greatest potential for increasing the prediction accuracy of habitual physical activity energy expenditure. Although multisensor systems may be more accurate, this must be balanced against feasibility, a balance that shifts with technological and scientific advances and should be considered at the beginning of every new study.