Data fusion of body-worn accelerometers and heart rate to predict VO2max during submaximal running

Heart Rate Variability Based Estimation of Maximal Oxygen Uptake in Athletes Using Supervised Regression Models

Wearable Heart Rate monitors are used in sports to provide physiological insights into athletes' well-being and performance. Their unobtrusive nature and ability to provide reliable heart rate measurements facilitate the estimation of cardiorespiratory fitness of athletes, as quantified by maximum consumption of oxygen uptake. Previous studies have employed data-driven models which use heart rate information to estimate the cardiorespiratory fitness of athletes. This signifies the physiological relevance of heart rate and heart rate variability for the estimation of maximal oxygen uptake. In this work, the heart rate variability features that were extracted from both exercise and recovery segments were fed to three different Machine Learning models to estimate maximal oxygen uptake of 856 athletes performing Graded Exercise Testing. A total of 101 features from exercise and 30 features from recovery segments were given as input to three feature selection methods to avoid overfitting of the models and to obtain relevant features. This resulted in the increase of model's accuracy by 5.7% for exercise and 4.3% for recovery. Further, post-modelling analysis was performed to remove the deviant points in two cases, initially in both training and testing and then only in training set, using k-Nearest Neighbour. In the former case, the removal of deviant points led to a reduction of 19.3% and 18.0% in overall estimation error for exercise and recovery, respectively. In the latter case, which mimicked the real-world scenario, the average R value of the models was observed to be 0.72 and 0.70 for exercise and recovery, respectively. From the above experimental approach, the utility of heart rate variability to estimate maximal oxygen uptake of large population of athletes was validated. Additionally, the proposed work contributes to the utility of cardiorespiratory fitness assessment of athletes through wearable heart rate monitors.

Encoding Cardiopulmonary Exercise Testing Time Series as Images for Classification using Convolutional Neural Network

Exercise testing has been available for more than a half-century and is a remarkably versatile tool for diagnostic and prognostic information of patients for a range of diseases, especially cardiovascular and pulmonary. With rapid advancements in technology, wearables, and learning algorithm in the last decade, its scope has evolved. Specifically, Cardiopulmonary exercise testing (CPX) is one of the most commonly used laboratory tests for objective evaluation of exercise capacity and performance levels in patients. CPX provides a non-invasive, integrative assessment of the pulmonary, cardiovascular, and skeletal muscle systems involving the measurement of gas exchanges. However, its assessment is challenging, requiring the individual to process multiple time series data points, leading to simplification to peak values and slopes. But this simplification can discard the valuable trend information present in these time series. In this work, we encode the time series as images using the Gramian Angular Field and Markov Transition Field and use it with a convolutional neural network and attention pooling approach for the classification of heart failure and metabolic syndrome patients. Using GradCAMs, we highlight the discriminative features identified by the model. Clinical relevance- The proposed framework can process multivariate exercise testing time-series data and accurately predict cardiovascular diseases. Interpretable Grad-CAMs can be obtained to explain the prediction.

Greater Breast Support Is Associated With Reduced Oxygen Consumption and Greater Running Economy During a Treadmill Running Task

Introduction

Breast pain is a major barrier to running for women. While breast support through the use of sports bras reduces breast-related discomfort, the effect of breast support on running performance is less understood. Therefore, the purpose of the current study was to evaluate the effect of greater breast support on oxygen consumption and running economy during a treadmill running task.

Methods

Fifteen female recreational runners performed a 10-min treadmill running task at their preferred running speed in each of two sports bra conditions: low support and high support. Participants ran on an instrumented treadmill (1,200 Hz, Bertec) while indirect calorimetry was performed using a metabolic measurement system (100 Hz, TrueOne, ParvoMedics). Average VO₂ (absolute and relative) from the third to 10th minutes was used to evaluate oxygen consumption. Running economy was calculated as the distance traveled per liter of oxygen consumed. Paired samples t-tests were used to compare mean oxygen consumption and running economy values between breast support conditions. Correlation analysis was performed to evaluate the relationship between breast size and change in running performance.

Results

Greater breast support was associated with reductions in absolute (p < 0.001) and relative oxygen consumption (p < 0.001; LOW: 30.9 ± 7.1 ml/kg/min; HIGH: 28.7 ± 6.7 ml/kg/min). Greater breast support was associated with increases in running economy (p < 0.001; LOW: 88.6 ± 29.1 m/L O₂; HIGH: 95.2 ± 31.1 m/L O₂). No changes in temporospatial characteristics of running were observed including cadence (p = 0.149), step length (p = 0.300) or ground contact time (p = 0.151). Strong positive linear correlations were observed between the change in running performance metrics and breast size (Oxygen Consumption: p < 0.001, r = 0.770; Relative Oxygen Consumption: p < 0.001, r = 0769; Running Economy: p < 0.001, r = 0.807).

Conclusions

Greater breast support was associated with reduced oxygen consumption and increased running economy. These findings demonstrate that greater breast support is not only associated with improved comfort but also improved running performance.

An Experimental Feasibility Study Evaluating the Adequacy of a Sportswear-Type Wearable for Recording Exercise Intensity

Sportswear-type wearables with integrated inertial sensors and electrocardiogram (ECG) electrodes have been commercially developed. We evaluated the feasibility of using a sportswear-type wearable with integrated inertial sensors and electrocardiogram (ECG) electrodes for evaluating exercise intensity within a controlled laboratory setting. Six male college athletes were asked to wear a sportswear-type wearable while performing a treadmill test that reached up to 20 km/h. The magnitude of the filtered tri-axial acceleration signal, recorded by the inertial sensor, was used to calculate the acceleration index. The R-R intervals of the ECG were used to determine heart rate; the external validity of the heart rate was then evaluated according to oxygen uptake, which is the gold standard for physiological exercise intensity. Single regression analysis between treadmill speed and the acceleration index in each participant showed that the slope of the regression line was significantly greater than zero with a high coefficient of determination (walking, 0.95; jogging, 0.96; running, 0.90). Another single regression analysis between heart rate and oxygen uptake showed that the slope of the regression line was significantly greater than zero, with a high coefficient of determination (0.96). Together, these results indicate that the sportswear-type wearable evaluated in this study is a feasible technology for evaluating physical and physiological exercise intensity across a wide range of physical activities and sport performances.

Is This the Real Life, or Is This Just Laboratory? A Scoping Review of IMU-Based Running Gait Analysis

Inertial measurement units (IMUs) can be used to monitor running biomechanics in real-world settings, but IMUs are often used within a laboratory. The purpose of this scoping review was to describe how IMUs are used to record running biomechanics in both laboratory and real-world conditions. We included peer-reviewed journal articles that used IMUs to assess gait quality during running. We extracted data on running conditions (indoor/outdoor, surface, speed, and distance), device type and location, metrics, participants, and purpose and study design. A total of 231 studies were included. Most (72%) studies were conducted indoors; and in 67% of all studies, the analyzed distance was only one step or stride or <200 m. The most common device type and location combination was a triaxial accelerometer on the shank (18% of device and location combinations). The most common analyzed metric was vertical/axial magnitude, which was reported in 64% of all studies. Most studies (56%) included recreational runners. For the past 20 years, studies using IMUs to record running biomechanics have mainly been conducted indoors, on a treadmill, at prescribed speeds, and over small distances. We suggest that future studies should move out of the lab to less controlled and more real-world environments.

Assessment of Peak Oxygen Uptake with a Smartwatch and its Usefulness for Training of Runners

Peak oxygen uptake (˙VO _2peak ) is an important factor contributing to running performance. Wearable technology may allow the assessment of ˙VO _2peak more frequently and on a larger scale. We aim to i) validate the ˙VO _2peak assessed by a smartwatch (Garmin Forerunner 245), and ii) discuss how this parameter may assist to evaluate and guide training procedures. A total of 23 runners (12 female, 11 male; ˙VO _2peak : 48.6±6.8 ml∙min ⁻¹ ∙kg ⁻¹ ) visited the laboratory twice to determine their ˙VO _2peak during a treadmill ramp test. Between laboratory visits, participants wore a smartwatch and performed three outdoor runs to obtain ˙VO _2peak values provided by the smartwatch. The ˙VO _2peak obtained by the criterion measure ranged from 38 to 61 ml∙min ⁻¹ ∙kg ⁻¹ . The mean absolute percentage error (MAPE) between the smartwatch and the criterion ˙VO _2peak was 5.7%. The criterion measure revealed a coefficient of variation of 4.0% over the VO2peak range from 38–61 ml∙min ⁻¹ ∙kg ⁻¹ . MAPE between the smartwatch and criterion measure was 7.1, 4.1 and −6.2% when analyzing ˙VO _2peak ranging from 39–45 ml∙min ⁻¹ ∙kg ⁻¹ , 45–55 ml∙min ⁻¹ ∙kg ⁻¹ or 55–61 ml∙min ⁻¹ ∙kg ⁻¹ , respectively.

Recommendations for marathon runners: on the application of recommender systems and machine learning to support recreational marathon runners

Every year millions of people, from all walks of life, spend months training to run a traditional marathon. For some it is about becoming fit enough to complete the gruelling 26.2 mile (42.2 km) distance. For others, it is about improving their fitness, to achieve a new personal-best finish-time. In this paper, we argue that the complexities of training for a marathon, combined with the availability of real-time activity data, provide a unique and worthwhile opportunity for machine learning and for recommender systems techniques to support runners as they train, race, and recover. We present a number of case studies-a mix of original research plus some recent results-to highlight what can be achieved using the type of activity data that is routinely collected by the current generation of mobile fitness apps, smart watches, and wearable sensors.

A New Fitness Test of Estimating VO2max in Well-Trained Rowing Athletes

Background

This study was designed to investigate the validity of maximal oxygen consumption (VO_2max) estimation through the Firstbeat fitness test (FFT) method when using submaximal rowing and running programs for well-trained athletes.

Methods

Well-trained flatwater rowers (n = 45, 19.8 ± 3.0 years, 184 ± 8.7 cm, 76 ± 12.9 kg, and 58.7 ± 6.0 mL⋅kg^–1⋅min^–1) and paddlers (n = 45, 19.0 ± 2.5 years, 180 ± 7.7 cm, 74 ± 9.4 kg, and 59.9 ± 4.8 mL⋅kg^–1⋅min^–1) completed the FFT and maximal graded exercise test (GXT) programs of rowing and running, respectively. The estimated VO_2max was calculated using the FFT system, and the measured VO_2max was obtained from the GXT programs. Differences between the estimated and measured VO_2max values were analyzed to assess the accuracy and agreement of the predictions. Equations from the previous study were also used to predict the VO_2max in the submaximal programs to compare the accuracy of prediction with the FFT method.

Results

The FFT method was in good agreement with the measured VO_2max in both groups based on the intraclass correlation coefficients (>0.8). Additionally, the FFT method had considerable accuracy in VO_2max estimation as the mean absolute percentage error (≤5.0%) and mean absolute error (<3.0 mL⋅kg^–1⋅min^–1) were fairly low. Furthermore, the FFT method seemed more accurate in the estimation of VO_2max than previously reported equations, especially in the rowing test program.

Conclusion

This study revealed that the FFT method provides a considerably accurate estimation of VO_2max in well-trained athletes.

Sacral acceleration can predict whole-body kinetics and stride kinematics across running speeds

Background

Stress fractures are injuries caused by repetitive loading during activities such as running. The application of advanced analytical methods such as machine learning to data from multiple wearable sensors has allowed for predictions of biomechanical variables associated with running-related injuries like stress fractures. However, it is unclear if data from a single wearable sensor can accurately estimate variables that characterize external loading during running such as peak vertical ground reaction force (vGRF), vertical impulse, and ground contact time. Predicting these biomechanical variables with a single wearable sensor could allow researchers, clinicians, and coaches to longitudinally monitor biomechanical running-related injury risk factors without expensive force-measuring equipment.

Purpose

We quantified the accuracy of applying quantile regression forest (QRF) and linear regression (LR) models to sacral-mounted accelerometer data to predict peak vGRF, vertical impulse, and ground contact time across a range of running speeds.

Methods

Thirty-seven collegiate cross country runners (24 females, 13 males) ran on a force-measuring treadmill at 3.8-5.4 m/s while wearing an accelerometer clipped posteriorly to the waistband of their running shorts. We cross-validated QRF and LR models by training them on acceleration data, running speed, step frequency, and body mass as predictor variables. Trained models were then used to predict peak vGRF, vertical impulse, and contact time. We compared predicted values to those calculated from a force-measuring treadmill on a subset of data (n = 9) withheld during model training. We quantified prediction accuracy by calculating the root mean square error (RMSE) and mean absolute percentage error (MAPE).

Results

The QRF model predicted peak vGRF with a RMSE of 0.150 body weights (BW) and MAPE of 4.27  ±  2.85%, predicted vertical impulse with a RMSE of 0.004 BW*s and MAPE of 0.80  ±  0.91%, and predicted contact time with a RMSE of 0.011 s and MAPE of 4.68  ±  3.00%. The LR model predicted peak vGRF with a RMSE of 0.139 BW and MAPE of 4.04  ±  2.57%, predicted vertical impulse with a RMSE of 0.002 BW*s and MAPE of 0.50  ±  0.42%, and predicted contact time with a RMSE of 0.008 s and MAPE of 3.50  ±  2.27%. There were no statistically significant differences between QRF and LR model prediction MAPE for peak vGRF (p = 0.549) or vertical impulse (p = 0.073), but the LR model's MAPE for contact time was significantly lower than the QRF model's MAPE (p = 0.0497).

Conclusions

Our findings indicate that the QRF and LR models can accurately predict peak vGRF, vertical impulse, and contact time (MAPE < 5%) from a single sacral-mounted accelerometer across a range of running speeds. These findings may be beneficial for researchers, clinicians, or coaches seeking to monitor running-related injury risk factors without force-measuring equipment.

Factors associated with validity of consumer-oriented wearable physical activity trackers: a meta-analysis

The purposes of this study were to examine (1) the strength of the criterion validity evidence of various consumer-oriented wearable physical activity trackers, (2) the influence of brands of consumer-oriented wearable physical activity on validity evidence and (3) factors that may contribute to differences in the strength of the criterion validity evidence. A total of 589 articles were identified through four databases. Pairs of researchers reviewed the articles to determine eligibility. A total of 29 studies with 96 validity coefficients were included in the meta-analysis. Five different moderators, including the brands of physical activity trackers, placement of devices, type of activities (ambulatory vs. lifestyle activities), population, and release year, were analysed to examine which factors impact the validity evidence. The summarised validity coefficient between activity trackers and energy expenditure ranged from r = .41 to r = .91. Moderator analyses revealed that the brand, placement of the device, and population significantly impact the magnitude of the validity evidence, while the type of activity and release year of the devices do not. Device brand, population, andplacement are each factor that significantly affects the validity coefficientsbetween consumer-oriented wearable physical activity trackers. Efforts should be made to improve the accuracy of these devices to maintain the credibility of the research and the trust of consumers.

Estimation of Instantaneous Oxygen Uptake During Exercise and Daily Activities Using a Wearable Cardio-Electromechanical and Environmental Sensor

OBJECTIVE

To estimate instantaneous oxygen uptake VO₂ with a small, low-cost wearable sensor during exercise and daily activities in order to enable monitoring of energy expenditure (EE) in uncontrolled settings. We aim to do so using a combination of seismocardiogram (SCG), electrocardiogram (ECG) and atmospheric pressure (AP) signals obtained from a minimally obtrusive wearable device.

METHODS

In this study, subjects performed a treadmill protocol in a controlled environment and an outside walking protocol in an uncontrolled environment. During testing, the COSMED K5 metabolic system collected gold standard breath-by-breath (BxB) data and a custom-built wearable patch placed on the mid-sternum collected SCG, ECG and AP signals. We extracted features from these signals to estimate the BxB VO₂ data obtained from the COSMED system.

RESULTS

In estimating instantaneous VO₂, we achieved our best results on the treadmill protocol using a combination of SCG (frequency) and AP features (RMSE of 3.68 ± 0.98 ml/kg/min and R² of 0.77). For the outside protocol, we achieved our best results using a combination of SCG (frequency), ECG and AP features (RMSE of 4.3 ± 1.47 ml/kg/min and R² of 0.64). In estimating VO₂ consumed over one minute intervals during the protocols, our median percentage error was 15.8[Formula: see text] for the treadmill protocol and 20.5[Formula: see text] for the outside protocol.

CONCLUSION

SCG, ECG and AP signals from a small wearable patch can enable accurate estimation of instantaneous VO₂ in both controlled and uncontrolled settings. SCG signals capturing variation in cardio-mechanical processes, AP signals, and state of the art machine learning models contribute significantly to the accurate estimation of instantaneous VO₂.

SIGNIFICANCE

Accurate estimation of VO₂ with a low cost, minimally obtrusive wearable patch can enable the monitoring of VO₂ and EE in everyday settings and make the many applications of these measurements more accessible to the general public.

Prediction of VO2max based on a 3-kilometer running test for water sports athletes

BACKGROUND

No studies have reported the 3-kilometer running test (3KRT) intending to predict VO2max for water sports athletes. Therefore, the purpose of this study was to develop a new model to predict the maximal aerobic capacity (VO2max) for water sports athletes based on 3KRT.

METHODS

One hundred and two water sports athletes completed two sessions of experiments consisting of a maximal graded exercise test (GXT) and a 3KRT. Multiple linear regression was applied to predict VO2max value based on the performance and physiological responses of 3KRT, along with participants' anthropometric and demographic variables. The predicted residual error sum of square (PRESS) and error terms (constant error and total error) were calculated to further evaluate the predictive accuracy.

RESULTS

Two significant prediction models based on elapsed exercise time (T3KRT), post-exercise heart rate (PHR3KRT), body mass, and gender were proposed. One model including PHR3KRT was identified (VO2max=120.77-0.028×T3KRT [second]-0.11×PHR3KRT [bpm]-0.334×body mass [kg]+8.70×gender [1: male, 0: female]), with an adjusted R2 of 0.723. Another model excluding PHR3KRT was also identified (VO2max=103.65-0.034×T3KRT [second]-0.317×Body mass [kg] + 7.89×gender [1: male, 0: female]), with an adjusted R2 of 0.713. Both models were further validated by the result of PRESS statistics.

CONCLUSIONS

This endurance 3-kilometer running test accurately predicted VO2max value for water sports athletes (rowers, canoeists, and kayakers), and the model excluding PHR3KRT would be easier to use.

Using Biosensors and Digital Biomarkers to Assess Response to Cardiac Rehabilitation: Observational Study

BACKGROUND

Cardiac rehabilitation (CR) is known for its beneficial effects on functional capacity and is a key component within current cardiovascular disease management strategies. In addition, a larger increase in functional capacity is accompanied by better clinical outcomes. However, not all patients respond in a similar way to CR. Therefore, a patient-tailored approach to CR could open up the possibility to achieve an optimal increase in functional capacity in every patient. Before treatment can be optimized, the differences in response of patients in terms of cardiac adaptation to exercise should first be understood. In addition, digital biomarkers to steer CR need to be identified.

OBJECTIVE

The aim of the study was to investigate the difference in cardiac response between patients characterized by a clear improvement in functional capacity and patients showing only a minor improvement following CR therapy.

METHODS

A total of 129 patients in CR performed a 6-minute walking test (6MWT) at baseline and during four consecutive short-term follow-up tests while being equipped with a wearable electrocardiogram (ECG) device. The 6MWTs were used to evaluate functional capacity. Patients were divided into high- and low-response groups, based on the improvement in functional capacity during the CR program. Commonly used heart rate parameters and cardiac digital biomarkers representative of the heart rate behavior during the 6MWT and their evolution over time were investigated.

RESULTS

All participating patients improved in functional capacity throughout the CR program (P<.001). The heart rate parameters, which are commonly used in practice, evolved differently for both groups throughout CR. The peak heart rate (HRpeak) from patients in the high-response group increased significantly throughout CR, while no change was observed in the low-response group (F4,92=8.321, P<.001). Similar results were obtained for the recovery heart rate (HRrec) values, which increased significantly over time during every minute of recuperation, for the high-response group (HRrec1: P<.001, HRrec2: P<.001, HRrec3: P<.001, HRrec4: P<.001, and HRrec5: P=.02). The other digital biomarkers showed that the evolution of heart rate behavior during a standardized activity test differed throughout CR between both groups. These digital biomarkers, derived from the continuous measurements, contribute to more in-depth insight into the progression of patients' cardiac responses.

CONCLUSIONS

This study showed that when using wearable sensor technology, the differences in response of patients to CR can be characterized by means of commonly used heart rate parameters and digital biomarkers that are representative of cardiac response to exercise. These digital biomarkers, derived by innovative analysis techniques, allow for more in-depth insights into the cardiac response of cardiac patients during standardized activity. These results open up the possibility to optimized and more patient-tailored treatment strategies and to potentially improve CR outcome.

A Machine Learning Approach to Estimate Hip and Knee Joint Loading Using a Mobile Phone-Embedded IMU

Hip osteoarthritis patients exhibit changes in kinematics and kinetics that affect joint loading. Monitoring this load can provide valuable information to clinicians. For example, a patient's joint loading measured across different activities can be used to determine the amount of exercise that the patient needs to complete each day. Unfortunately, current methods for measuring joint loading require a lab environment which most clinicians do not have access to. This study explores employing machine learning to construct a model that can estimate joint loading based on sensor data obtained solely from a mobile phone. In order to learn such a model, we collected a dataset from 10 patients with hip osteoarthritis who performed multiple repetitions of nine different exercises. During each repetition, we simultaneously recorded 3D motion capture data, ground reaction force data, and the inertial measurement unit data from a mobile phone attached to the patient's hip. The 3D motion and ground reaction force data were used to compute the ground truth joint loading using musculoskeletal modeling. Our goal is to estimate the ground truth loading value using only the data captured by the sensors of the mobile phone. We propose a machine learning pipeline for learning such a model based on the recordings of a phone's accelerometer and gyroscope. When evaluated for an unseen patient, the proposed pipeline achieves a mean absolute error of 29% for the left hip and 36% for the right hip. While our approach is a step in the direction of using a minimal number of sensors to estimate joint loading outside the lab, developing a tool that is accurate enough to be applicable in a clinical context still remains an open challenge. It may be necessary to use sensors at more than one location in order to obtain better estimates.