Covariate-Aware Longitudinal Modelling for Neurodegenerative Diseases - Inria - Institut national de recherche en sciences et technologies du numérique
Poster Communications Year : 2022

Covariate-Aware Longitudinal Modelling for Neurodegenerative Diseases

Abstract

Longitudinal modelling is of pivotal interest for the study of neurodegenerative diseases. The Disease Course Mapping is a multivariate Bayesian mixed-effect progression model that is able to recover the course of a disease from a cohort with multimodal longitudinal observations (imaging variables, cognitive and clinical scores) and to extract interpretable parameters to describe each patient. It has been validated on multiple diseases and on multiple applications settings (cohort study, trial enrichment, data simulation, ...) However, its statistical formulation relies only on time-dependent observations. It thus fail to integrate time-independent information (gender, education levels, genetic factors, ...) in its modelling, even though such covariates are known to modulate clinical disease courses. We propose a mixed-effect formulation that captures the influence of such covariates over the dynamic of the disease.
Fichier principal
Vignette du fichier
poster_final.pdf (800.34 Ko) Télécharger le fichier
Origin Files produced by the author(s)

Dates and versions

hal-03832976 , version 1 (28-10-2022)

Licence

Identifiers

  • HAL Id : hal-03832976 , version 1

Cite

Nemo Fournier, Stanley Durrleman​. Covariate-Aware Longitudinal Modelling for Neurodegenerative Diseases. 43rd Annual Conference of the International Society for Clinical Biostatistics (ISCB), Aug 2022, Newcastle, United Kingdom. . ⟨hal-03832976⟩
66 View
50 Download

Share

More