Oxynet: A collective intelligence that detects ventilatory thresholds in cardiopulmonary exercise tests

The role of machine learning methods in physiological explorations of endurance trained athletes: a mini-review

Endurance-trained athletes require physiological explorations that have evolved throughout the history of exercise physiology with technological advances. From the use of the Douglas bag to measure gas exchange to the development of wearable connected devices, advances in physiological explorations have enabled us to move from the classic but still widely used cardiopulmonary exercise test (CPET) to the collection of data under real conditions on outdoor endurance or ultra-endurance events. However, such explorations are often costly, time-consuming, and complex, creating a need for efficient analysis methods. Machine Learning (ML) has emerged as a powerful tool in exercise physiology, offering solutions to these challenges. Given that exercise physiologists may be unfamiliar with ML, this mini-review provides a concise overview of its relevance to the field. It introduces key ML methods, highlights their ability to predict important physiological parameters (e.g., heart rate variability and exercise-induced hypoxemia), and discusses their strengths and limitations. Finally, it outlines future directions based on the challenges identified, serving as an initial reference for physiologists exploring the application of ML in endurance exercise.

Smart Devices for Health and Wellness Applied to Tele-Exercise: An Overview of New Trends and Technologies Such as IoT and AI

This descriptive article explores the use of smart devices for health and wellness in the context of telehealth, highlighting rapidly evolving technologies such as the Internet of Things (IoT) and Artificial Intelligence (AI). Key innovations, benefits, challenges, and opportunities related to the adoption of these technologies are outlined. The article provides a descriptive and accessible approach to understanding the evolution and impact of smart devices in the tele-exercise reality. Nowadays, technological advances provide solutions that were unthinkable just a few years ago. The habits of the general population have also changed over the past few years. Hence, there is a need to investigate this issue and draw the attention of the scientific community to this topic by describing the benefits and challenges associated with each topic. If individuals no longer go to exercise, the exercise must go to their homes instead.

Sequencing patterns of ventilatory indices in less trained adults

Submaximal ventilatory indices, i.e., point of optimal ventilatory efficiency (POE) and anaerobic threshold (AT), are valuable indicators to assess the metabolic and ventilatory response during cardiopulmonary exercise testing (CPET). The order in which the ventilatory indices occur (ventilatory indices sequencing pattern, VISP), may yield additional information for the interpretation of CPET results and for exercise intensity prescription. Therefore, we determined whether different VISP groups concerning POE and AT exist. Additionally, we analysed fat metabolism via the exercise intensity eliciting the highest fat oxidation rate (Fatmax) as a possible explanation for differences between VISP groups. 761 less trained adults (41-68 years) completed an incremental exercise test on a cycle ergometer until volitional exhaustion. The ventilatory indices were determined using automatic and visual detection methods, and Fatmax was determined using indirect calorimetry. Our study identified two VISP groups with a lower work rate at POE compared to AT in VISPPOE < AT but not in group VISPPOE = AT. Therefore, training prescription based on POE rather than AT would result in different exercise intensity recommendations in 66% of the study participants and consequently in unintended physiological adaptions. VISPPOE < AT participants were not different to VISPPOE = AT participants concerning VO2peak and Fatmax. However, participants exhibiting a difference in work rate (VISPPOE < AT) were characterized by a higher aerobic capacity at submaximal work rate compared to VISPPOE = AT. Thus, analysing VISP may help to gain new insights into the complex ventilatory and metabolic response to exercise. But a methodological framework still must be established.

Automatic Detection of Aerobic Threshold through Recurrence Quantification Analysis of Heart Rate Time Series

During exercise with increasing intensity, the human body transforms energy with mechanisms dependent upon actual requirements. Three phases of the body's energy utilization are recognized, characterized by different metabolic processes, and separated by two threshold points, called aerobic (AerT) and anaerobic threshold (AnT). These thresholds occur at determined values of exercise intensity(workload) and can change among individuals. They are considered indicators of exercise capacities and are useful in the personalization of physical activity plans. They are usually detected by ventilatory or metabolic variables and require expensive equipment and invasive measurements. Recently, particular attention has focused on AerT, which is a parameter especially useful in the overweight and obese population to determine the best amount of exercise intensity for weight loss and increasing physical fitness. The aim of study is to propose a new procedure to automatically identify AerT using the analysis of recurrences (RQA) relying only on Heart rate time series, acquired from a cohort of young athletes during a sub-maximal incremental exercise test (Cardiopulmonary Exercise Test, CPET) on a cycle ergometer. We found that the minima of determinism, an RQA feature calculated from the Recurrence Quantification by Epochs (RQE) approach, identify the time points where generic metabolic transitions occur. Among these transitions, a criterion based on the maximum convexity of the determinism minima allows to detect the first metabolic threshold. The ordinary least products regression analysis shows that values of the oxygen consumption VO₂, heart rate (HR), and Workload correspondent to the AerT estimated by RQA are strongly correlated with the one estimated by CPET (r > 0.64). Mean percentage differences are <2% for both HR and VO₂ and <11% for Workload. The Technical Error for HR at AerT is <8%; intraclass correlation coefficients values are moderate (≥0.66) for all variables at AerT. This system thus represents a useful method to detect AerT relying only on heart rate time series, and once validated for different activities, in future, can be easily implemented in applications acquiring data from portable heart rate monitors.

Machine Learning Models for the Automatic Detection of Exercise Thresholds in Cardiopulmonary Exercising Tests: From Regression to Generation to Explanation

The cardiopulmonary exercise test (CPET) constitutes a gold standard for the assessment of an individual's cardiovascular fitness. A trend is emerging for the development of new machine-learning techniques applied to the automatic process of CPET data. Some of these focus on the precise task of detecting the exercise thresholds, which represent important physiological parameters. Three are the major challenges tackled by this contribution: (A) regression (i.e., the process of correctly identifying the exercise intensity domains and their crossing points); (B) generation (i.e., the process of artificially creating a CPET data file ex-novo); and (C) explanation (i.e., proving an interpretable explanation about the output of the machine learning model). The following methods were used for each challenge: (A) a convolutional neural network adapted for multi-variable time series; (B) a conditional generative adversarial neural network; and (C) visual explanations and calculations of model decisions have been conducted using cooperative game theory (Shapley's values). The results for the regression, generation, and explanatory techniques for AI-assisted CPET interpretation are presented here in a unique framework for the first time: (A) machine learning techniques reported an expert-level accuracy in the classification of exercise intensity domains; (B) experts are not able to substantially differentiate between a real vs an artificially generated CPET; and (C) Shapley's values can provide an explanation about the choices of the algorithms in terms of ventilatory variables. With the aim to increase their technology-readiness level, all the models discussed in this contribution have been incorporated into a free-to-use Python package called pyoxynet (ver. 12.1). This contribution should therefore be of interest to major players operating in the CPET device market and engineering.

Detecting Metabolic Thresholds from Nonlinear Analysis of Heart Rate Time Series: A Review

Heart rate time series are widely used to characterize physiological states and athletic performance. Among the main indicators of metabolic and physiological states, the detection of metabolic thresholds is an important tool in establishing training protocols in both sport and clinical fields. This paper reviews the most common methods, applied to heart rate (HR) time series, aiming to detect metabolic thresholds. These methodologies have been largely used to assess energy metabolism and to identify the appropriate intensity of physical exercise which can reduce body weight and improve physical fitness. Specifically, we focused on the main nonlinear signal evaluation methods using HR to identify metabolic thresholds with the purpose of identifying a method which can represent a useful tool for the real-time settings of wearable devices in sport activities. While the advantages and disadvantages of each method, and the possible applications, are presented, this review confirms that the nonlinear analysis of HR time series represents a solid, robust and noninvasive approach to assess metabolic thresholds.

Enhancing instantaneous oxygen uptake estimation by non-linear model using cardio-pulmonary physiological and motion signals

Oxygen uptake (VO₂) is an important parameter in sports medicine, health assessment and clinical treatment. At present, more and more wearable devices are used in daily life, clinical treatment and health care. The parameters obtained by wearables have great research potential and application prospect. In this paper, an instantaneous VO₂ estimation model based on XGBoost was proposed and verified by using data obtained from a medical-grade wearable device (Beijing SensEcho) at different posture and activity levels. Furthermore, physiological characteristics extracted from single-lead electrocardiogram, thoracic and abdominal respiration signal and tri-axial acceleration signal were studied to optimize the model. There were 29 healthy volunteers recruited for the study to collect data while stationary (lying, sitting, standing), walking, Bruce treadmill test and recuperating with SensEcho and the gas analyzer (Metalyzer 3B). The results show that the VO₂ values estimated by the proposed model are in good agreement with the true values measured by the gas analyzer (R² = 0.94 ± 0.03, n = 72,235), and the mean absolute error (MAE) is 1.83 ± 0.59 ml/kg/min. Compared with the estimation method using a separate heart rate as input, our method reduced MAE by 54.70%. At the same time, other factors affecting the performance of the model were studied, including the influence of different input signals, gender and movement intensity, which provided more enlightenment for the estimation of VO₂. The results show that the proposed model based on cardio-pulmonary physiological signals as inputs can effectively improve the accuracy of instantaneous VO₂ estimation in various scenarios of activities and was robust between different motion modes and state. The VO₂ estimation method proposed in this paper has the potential to be used in daily life covering the scenario of stationary, walking and maximal exercise.

Using Machine Learning to Identify Organ System Specific Limitations to Exercise via Cardiopulmonary Exercise Testing

Cardiopulmonary Exer cise Testing (CPET) is a unique physiologic medical test used to evaluate human response to progressive maximal exercise stress. Depending on the degree and type of deviation from the normal physiologic response, CPET can help identify a patient's specific limitations to exercise to guide clinical care without the need for other expensive and invasive diagnostic tests. However, given the amount and complexity of data obtained from CPET, interpretation and visualization of test results is challenging. CPET data currently require dedicated training and significant experience for proper clinician interpretation. To make CPET more accessible to clinicians, we investigated a simplified data interpretation and visualization tool using machine learning algorithms. The visualization shows three types of limitations (cardiac, pulmonary and others); values are defined based on the results of three independent random forest classifiers. To display the models' scores and make them interpretable to the clinicians, an interactive dashboard with the scores and interpretability plots was developed. This machine learning platform has the potential to augment existing diagnostic procedures and provide a tool to make CPET more accessible to clinicians.

Oxygen uptake efficiency slope in healthy normal weight young males: an applicable framework for calculation and interpretation

Background

The oxygen uptake efficiency slope (OUES) is considered a reliable indicator of cardiorespiratory fitness in young and clinical populations who cannot achieve maximal effort during a graded exercise test. However, OUES accuracy depends on the data points used for its calculation and it is still not clear if the submaximal OUES can accurately assess CRF in healthy young males.

Objective

We investigated the association between peak oxygen uptake and peak and submaximal OUES, and the agreement between submaximal OUES and peak OUES in male adolescents and young adults.

Methods

In this cross-sectional, observational study, fifty normal weight healthy participants (age 14-22 years, peak oxygen uptake 43.8 ± 7.3 mL·min^-1·kg^-1) performed a graded exercise test on a cycle ergometer and pulmonary gas exchange was assessed using breath-by-breath analysis. Peak oxygen uptake, and oxygen consumption at the aerobic and at the anaerobic threshold were determined as the 30-s average of the oxygen consumption values. Peak OUES (up to peak) and submaximal OUES (up to the aerobic and anaerobic thresholds) were calculated from the logarithmic relation between oxygen consumption and pulmonary ventilation.

Results

Very strong correlations were observed between peak oxygen uptake and peak OUES (r = 0.80-0.88) while fair-to-very strong correlations were observed between the peak oxygen uptake and the two submaximal OUES (r = 0.32-0.81). The level of agreement between peak OUES and OUES up to the anaerobic threshold (r = 0.89-0.93; Typical percentage error 6%; Intraclass correlation coefficient = 0.89-0.93) was greater than the one between the peak oxygen uptake with OUES up to the aerobic threshold (r = 0.39-0.56; Typical percentage error 15%; Intraclass correlation coefficient = 0.38-0.56).

Conclusions

. The peak OUES is a better indicator of aerobic fitness than the OUES up to the anaerobic threshold in healthy, young males. The OUES up to the anaerobic threshold is a valid alternative to peak OUES.

A Reappraisal of Ventilatory Thresholds in Wheelchair Athletes With a Spinal Cord Injury: Do They Really Exist?

The ventilatory threshold (VT) separates low- from moderate-intensity exercise, the respiratory compensation point (RCP) moderate- from high-intensity exercise. Both concepts assume breakpoints in respiratory data. However, the objective determination of the VT and RCP using breakpoint models during upper-body modality exercise in wheelchair athletes with spinal cord injury (SCI) has received little attention. Therefore, the aim of this study was to compare the fit of breakpoint models (i.e., two linear regression lines) with continuous no-breakpoint models (i.e., exponential curve/second-order polynomial) to respiratory data obtained during a graded wheelchair exercise test to exhaustion. These fits were compared employing adjusted R², and blocked bootstrapping was used to derive estimates of a median and 95% confidence intervals (CI). V̇O₂-V̇CO₂ and V̇E/V̇O₂-time data were assessed for the determination of the VT, and V̇CO₂-V̇E and V̇E/V̇CO₂-time data for the determination of the RCP. Data of 9 wheelchair athletes with tetraplegia and 8 with paraplegia were evaluated. On an overall group-level, there was an overlap in the adjusted R² median ± 95% CI between the breakpoint and the no-breakpoint models for determining the VT (V̇O₂-V̇CO₂: 0.991 ± 0.003 vs. 0.990 ± 0.003; V̇E/V̇O₂-time: 0.792 ± 0.101 vs. 0.782 ± 0.104, respectively) and RCP (V̇E-V̇CO₂: 0.984 ± 0.004 vs. 0.984 ± 0.004; V̇E/V̇CO₂-time: 0.729 ± 0.064 vs. 0.691 ± 0.063, respectively), indicating similar model fit. We offer two lines of reasoning: (1) breakpoints in these respiratory data exist but are too subtle to result in a significant difference in adjusted R² between the investigated breakpoint and no-breakpoint models; (2) breakpoints do not exist, as has been argued previously.

A Learn-to-Rank Approach for Predicting Road Cycling Race Outcomes

Professional road cycling is a very competitive sport, and many factors influence the outcome of the race. These factors can be internal (e.g., psychological preparedness, physiological profile of the rider, and the preparedness or fitness of the rider) or external (e.g., the weather or strategy of the team) to the rider, or even completely unpredictable (e.g., crashes or mechanical failure). This variety makes perfectly predicting the outcome of a certain race an impossible task and the sport even more interesting. Nonetheless, before each race, journalists, ex-pro cyclists, websites and cycling fans try to predict the possible top 3, 5, or 10 riders. In this article, we use easily accessible data on road cycling from the past 20 years and the Machine Learning technique Learn-to-Rank (LtR) to predict the top 10 contenders for 1-day road cycling races. We accomplish this by mapping a relevancy weight to the finishing place in the first 10 positions. We assess the performance of this approach on 2018, 2019, and 2021 editions of six spring classic 1-day races. In the end, we compare the output of the framework with a mass fan prediction on the Normalized Discounted Cumulative Gain (NDCG) metric and the number of correct top 10 guesses. We found that our model, on average, has slightly higher performance on both metrics than the mass fan prediction. We also analyze which variables of our model have the most influence on the prediction of each race. This approach can give interesting insights to fans before a race but can also be helpful to sports coaches to predict how a rider might perform compared to other riders outside of the team.

Recurrence quantification analysis of heart rate variability to detect both ventilatory thresholds

Aims of this study were: to verify if Recurrence Quantification Analysis (RQA) of Heart Rate Variability (HRV) time series could determine both ventilatory thresholds in individuals with different fitness levels, and to assess the validity of RQA method compared to gas-exchange method (GE). The two thresholds were estimated in thirty young individuals during incremental exercise on cycle-ergometer: Heart rate (HR), Oxygen consumption (VO₂) and Workload were measured by the two methods (RQA and GE). Repeated measures ANOVA was used to assess main effects of methods and methods-by-groups interaction effects for HR, VO₂ and Workload at aerobic (AerT) and anaerobic (AnT) thresholds. Validity of RQA at both thresholds was assessed for HR, VO₂ and Workload by Ordinary Least Products (OLP) regression, Typical Percentage Error (TE), Intraclass Correlation Coefficients (ICC) and the Bland Altman plots. No methods-by-groups interaction effects were detected for HR, VO₂ and Workload at AerT and AnT. The OLP analysis showed that at both thresholds RQA and GE methods had very strong correlations (r >0.8) in all variables (HR, VO₂ and Workload). Slope and intercept values always included the 1 and the 0, respectively. At AerT the TE ranged from 4.02% (5.48 bpm) to 10.47% (8.53 Watts) (HR and Workload, respectively) and in all variables ICC values were excellent (≥0.85). At AnT the TE ranged from 2.53% (3.98 bpm) to 6.64% (7.81 Watts) (HR and Workload, respectively) and in all variables ICC values were excellent (≥0.90). Therefore, RQA of HRV time series is a new valid approach to determine both ventilatory thresholds in individuals with different physical fitness levels, it can be used when gas analysis is not possible or not convenient.

Detection of the Anaerobic Threshold in Endurance Sports: Validation of a New Method Using Correlation Properties of Heart Rate Variability

Past attempts to define an anaerobic threshold (AnT) have relied upon gas exchange kinetics, lactate testing and field-based evaluations. DFA a1, an index of heart rate (HR) variability (HRV) fractal correlation properties, has been shown to decrease with exercise intensity. The intent of this study is to investigate whether the AnT derived from gas exchange is associated with the transition from a correlated to uncorrelated random HRV pattern signified by a DFA a1 value of 0.5. HRV and gas exchange data were obtained from 15 participants during an incremental treadmill run. Comparison of the HR reached at the second ventilatory threshold (VT2) was made to the HR reached at a DFA a1 value of 0.5 (HRVT2). Based on Bland-Altman analysis and linear regression, there was strong agreement between VT2 and HRVT2 measured by HR (r = 0.78, p < 0.001). Mean VT2 was reached at a HR of 174 (±12) bpm compared to mean HRVT2 at a HR of 171 (±16) bpm. In summary, the HR associated with a DFA a1 value of 0.5 on an incremental treadmill ramp was closely related to that of the HR at the VT2 derived from gas exchange analysis. A distinct numerical value of DFA a1 representing an uncorrelated, random interbeat pattern appears to be associated with the VT2 and shows potential as a noninvasive marker for training intensity distribution and performance status.