Key result
Linear and non-linear indicators of heart rate variability predicted salivary cortisol levels with a low mean absolute error of 7.43% using the Radial Basis Function regression algorithm.
Why the study?
The activation of the stress system is partly unconscious, and the study aimed to provide an algorithm for predicting the numerical value of salivary cortisol concentration using HRV indices.
Observational (n=601)
Effect estimate: MAE 7.43%
Heart rate variability indices can be used with machine learning algorithms, particularly Radial Basis Function, to accurately predict salivary cortisol levels as an objective measure of the stress system.
HRV-based cortisol prediction may enable non-invasive stress monitoring; leaves open clinical adoption pending validation.
Background and Aim:The stress system is one of the most important parts of maintaining living of beings.The indices of heart rate variation (HRV) and cortisol hormone are two outputs of stress system activity.The activation of the stress system is not necessarily in a consciousness state and part of it is in the unconscious.The aim of this study is to provide an algorithm for predicting the numerical value of the salivary cortisol concentration using HRV indices . Materials and Methods:The samples of this study included 601 healthy adult men (between 20 and 50 years old).The used algorithms were designed to predict the numerical value of salivary cortisol taken from 9:00 AM to 2:00 PM with HRV indicators.In each of the algorithms, a predicted value is compared with the actual value to determine which was more successful.Results: The results of this study showed that the frequency and non-linear indicators of HRV are able to predict the amount of salivary cortisol with use of Multi Layer Perceptron (MLP), XGBoost(XGB), Support Vector Machine (SVM) and Radial Basis Function (RBF) regression algorithms with the average absolute error, 7.78, 8.06, 8.37 and 7.43 percent respectively.Conclusion: In this study, it was found that a set of linear and non-linear indicators of HRV with high power can predict the amount of salivary cortisol in the best case with a low error percentage of 7.43 by the RBF algorithm, and instead of stress self-report that does not cover the physiological part.It can be a more accurate tool in the intelligent evaluation of the stress system.
No takes yet. Share an insight, caveat, or question.
Noruzi et al. (2024) conducted an observational in Healthy (n=601). Heart rate variability (HRV) indicators was evaluated on Prediction of salivary cortisol concentration (Mean Absolute Error) (MAE 7.43%). Linear and non-linear indicators of heart rate variability predicted salivary cortisol levels with a low mean absolute error of 7.43% using the Radial Basis Function regression algorithm.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: