Exploring machine learning techniques for aerobic system analysis with applicabili...
| Grant number: | 18/19016-8 |
| Support type: | Scholarships abroad - Research Internship - Post-doctor |
| Effective date (Start): | March 15, 2019 |
| Effective date (End): | March 14, 2020 |
| Field of knowledge: | Biological Sciences - Physiology - Physiology of Effort |
| Principal Investigator: | Aparecida Maria Catai |
| Grantee: | Thomas Beltrame |
| Supervisor abroad: | Uwe Hoffmann |
| Home Institution: | Instituto de Computação (IC). Universidade Estadual de Campinas (UNICAMP). Campinas , SP, Brazil |
| Local de pesquisa : | German Sport University Cologne, Germany |
| Associated to the scholarship: | 17/09639-5 - Exploring machine learning techniques for aerobic system analysis with applicability for cardiorespiratory rehabilitation programs, BP.PD |
Abstract Efforts to better understand cardiorespiratory health are relevant for the future development of optimized physical activity programs, mainly those designed for chronic diseases that affect functional capacity and mortality. Cardiorespiratory health can be evaluated through the characterization of aerobic power or aerobic fitness, and each of these are related to different aspects of the aerobic system response. This research project aims to explore the associations between the ability that the aerobic system has to supply energy as fast as possible during exercise transitions (fitness) with its maximal capacity to supply energy during maximal exertion (power). It is hypothesized that a slower aerobic system response during exercise transitions leads to lower functional capacity. Data used in this current project will be originated from 40 healthy participants submitted to a constant, pseudo-random binary sequence, and cardiopulmonary exercise testing protocols. The functional capacity and aerobic fitness will be estimated based on maximal oxygen uptake and aerobic system dynamics analysis, respectively. Unique methods developed in Cologne, Germany, will be used to estimate muscular oxygen uptake dynamics based on heart rate and ventilatory variables. In addition, aerobic system analysis will be made by machine learning algorithms and by frequency and time domain methods. Therefore, this project will use advanced mathematical and computational methods to understand complex physiological responses to further understand the aerobic system and how it is related to function capacity (and therefore, mortality). Our results may show that simpler moderate exercise protocols might be as good as maximal exertion exercise protocols to obtain indexes related to cardiorespiratory health. In addition, novel data analysis methods will be developed to further explore and model the aerobic system response during dynamic exercise, increasing the specificity of future rehabilitation programs for functional capacity improvements. | |