Modelagem bayesiana da frequência cardíaca com cargas crescentes de trabalho

• Ano de defesa: 2017
• Autor(a) principal: Silveira, Sílvio de Castro
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Lavras; Programa de Pós-Graduação em Estatística e Experimentação Agropecuária; UFLA; brasil; Departamento de Ciências Exatas
• Palavras-chave em Português: Frequência cardíaca – Medição – Métodos estatísticos; Teoria bayesiana de decisão estatística; Esportes – Aspectos fisiológicos; Heart rate – Measurement – Statistical methods; Bayesian statistical decision theory; Sports – Physiological aspects; Probabilidade e Estatística Aplicadas
• Link de acesso: http://repositorio.ufla.br/jspui/handle/1/28257

Considering heart rate as one of the most used physiological variables in physical exercise assessment and prescription programs, many studies are conducted to analyze its behavior in progressive tests because it represents a noninvasive alternative for the identification of metabolic transition thresholds. Linear model adjustments are still widely used to represent the heart rate curve as a function of increasing loads, even though there are indications of nonlinear behavior at the extremes of the heart rate curve at increasing loads. The objective of this study was to evaluate and compare, through the Bayesian procedure, Logistic model and Boltzmann's Sigmoidal model for the heart rate curve in increasing loads of healthy people, to verify if, with the improvement in the parameter estimation, there is some modification of the association of the transition points of the heart rate curve with the lactate thresholds. More specifically, we expect the heart rate inflection point to be associated with the first lactate threshold and the heart rate deflection point at the second lactate threshold. We used the data of the average of 16 individuals, tested in mechanical braking cycle ergometer, initial load of 0 kpm.min-1 and increment of 90kpm.min-1 every minute until exhaustion. The chains for the models were generated using OpenBugs software that uses the MCMC method. The convergence of the chains was verified through a package BOA of software R. The graphs, estimates and interval HPD were also made using software R. The models were compared by the Bayes factor and the results indicated that the Boltzmann sigmoidal model is superior to the Logistic model, there was an association between the heart rate inflexion point and the first lactate threshold, and there was an association between the heart rate deflection point and the second lactate threshold only to Boltzmann sigmoidal model.