Prediction of the Wingate anaerobic mechanical power outputs from a maximal incremental cardiopulmonary exercise stress test using machine-learning approach

A computational predictor of the anaerobic mechanical power outputs from a clinical exercise stress test

We previously were able to predict the anaerobic mechanical power outputs using features taken from a maximal incremental cardiopulmonary exercise stress test (CPET). Since a standard aerobic exercise stress test (with electrocardiogram and blood pressure measurements) has no gas exchange measurement and is more popular than CPET, our goal, in the current paper, was to investigate whether features taken from a clinical exercise stress test (GXT), either submaximal or maximal, can predict the anaerobic mechanical power outputs to the same level as we found with CPET variables. We have used data taken from young healthy subjects undergoing CPET aerobic test and the Wingate anaerobic test, and developed a computational predictive algorithm, based on greedy heuristic multiple linear regression, which enabled the prediction of the anaerobic mechanical power outputs from a corresponding GXT measures (exercise test time, treadmill speed and slope). We found that for submaximal GXT of 85% age predicted HRmax, a combination of 3 and 4 variables produced a correlation of r = 0.93 and r = 0.92 with % error equal to 15 ± 3 and 16 ± 3 on the validation set between real and predicted values of the peak and mean anaerobic mechanical power outputs (p < 0.001), respectively. For maximal GXT (100% of age predicted HRmax), a combination of 4 and 2 variables produced a correlation of r = 0.92 and r = 0.94 with % error equal to 12 ± 2 and 14 ± 3 on the validation set between real and predicted values of the peak and mean anaerobic mechanical power outputs (p < 0.001), respectively. The newly developed model allows to accurately predict the anaerobic mechanical power outputs from a standard, submaximal and maximal GXT. Nevertheless, in the current study the subjects were healthy, normal individuals and therefore the assessment of additional subjects is desirable for the development of a test applicable to other populations.

Automatic Cardiopulmonary Endurance Assessment: A Machine Learning Approach Based on GA-XGBOOST

OBJECTIVE

Among various assessment paradigms, the cardiopulmonary exercise test (CPET) provides rich evidence as part of the cardiopulmonary endurance (CPE) assessment. However, methods and strategies for interpreting CPET results are not in agreement. The purpose of this study is to validate the possibility of using machine learning to evaluate CPET data for automatically classifying the CPE level of workers in high-latitude areas.

METHODS

A total of 120 eligible workers were selected for this cardiopulmonary exercise experiment, and the physiological data and completion of the experiment were recorded in the simulated high-latitude workplace, within which 84 sets of data were used for XGBOOST model training and36 were used for the model validation. The model performance was compared with Support Vector Machine and Random Forest. Furthermore, hyperparameter optimization was applied to the XGBOOST model by using a genetic algorithm.

RESULTS

The model was verified by the method of tenfold cross validation; the correct rate was 0.861, with a Micro-F1 Score of 0.864. Compared with RF and SVM, all data achieved a better performance.

CONCLUSION

With a relatively small number of training samples, the GA-XGBOOST model fits well with the training set data, which can effectively evaluate the CPE level of subjects, and is expected to provide automatic CPE evaluation for selecting, training, and protecting the working population in plateau areas.

Utilizing Machine Learning Techniques to Predict the Efficacy of Aerobic Exercise Intervention on Young Hypertensive Patients Based on Cardiopulmonary Exercise Testing

Recently, the incidence of hypertension has significantly increased among young adults. While aerobic exercise intervention (AEI) has long been recognized as an effective treatment, individual differences in response to AEI can seriously influence clinicians' decisions. In particular, only a few studies have been conducted to predict the efficacy of AEI on lowering blood pressure (BP) in young hypertensive patients. As such, this paper aims to explore the implications of various cardiopulmonary metabolic indicators in the field by mining patients' cardiopulmonary exercise testing (CPET) data before making treatment plans. CPET data are collected "breath by breath" by using an oxygenation analyzer attached to a mask and then divided into four phases: resting, warm-up, exercise, and recovery. To mitigate the effects of redundant information and noise in the CPET data, a sparse representation classifier based on analytic dictionary learning was designed to accurately predict the individual responsiveness to AEI. Importantly, the experimental results showed that the model presented herein performed better than the baseline method based on BP change and traditional machine learning models. Furthermore, the data from the exercise phase were found to produce the best predictions compared with the data from other phases. This study paves the way towards the customization of personalized aerobic exercise programs for young hypertensive patients.