Generalized Additive Models for location, scale and shape in analyzing the number of cortical lesions in patients with multiple sclerosis

Authors

Keywords:

Count data, Poisson inverse Gaussian distribution, generalized additive models, GAMLSS

Abstract

Multiple sclerosis (MS) is a chronic autoimmune inflammatory disease that causes demyelination and neurodegeneration in the central nervous system. There is a growing interest in investigating variables associated with the number of cortical lesions in MS. In this context, the aim of this paper is to analyze the relationship between the number of cortical lesions with age, graduation EDSS (expanded disability status scale) and disease duration in patients with MS. Therefore, we analyzed 30 patients with MS. The analysis was conducted using the generalized additive models for location, scale and shape (GAMLSS). Four probability distributions were tested for the response variable. The results showed that the Poisson inverse Gaussian was the most suitable distribution for the data analysis, where it was possible to model the mean and dispersion parameters as functions of some covariates. Overall, the results indicate that patients with higher neurological deficits (represented by higher value in EDSS graduation), younger and presenting longer disease duration are the ones that showed the most cortical lesions. 

Published

08-01-2017

How to Cite

Petterle, R. R., & Formighieri, L. (2017). Generalized Additive Models for location, scale and shape in analyzing the number of cortical lesions in patients with multiple sclerosis. Sigmae, 5(1), 37–44. Retrieved from https://publicacoes.unifal-mg.edu.br/revistas/index.php/sigmae/article/view/508

Issue

Section

Applied Statistics