A multivariate finite mixture latent trajectory model with application to dementia studies

Dongbing Lai, Huiping Xu, Daniel Koller, Tatiana Foroud, Sujuan Gao

Research output: Contribution to journalArticle

3 Scopus citations

Abstract

Dementia patients exhibit considerable heterogeneity in individual trajectories of cognitive decline, with some patients showing rapid decline following diagnoses while others exhibiting slower decline or remaining stable for several years. Dementia studies often collect longitudinal measures of multiple neuropsychological tests aimed to measure patients’ decline across a number of cognitive domains. We propose a multivariate finite mixture latent trajectory model to identify distinct longitudinal patterns of cognitive decline simultaneously in multiple cognitive domains, each of which is measured by multiple neuropsychological tests. EM algorithm is used for parameter estimation and posterior probabilities are used to predict latent class membership. We present results of a simulation study demonstrating adequate performance of our proposed approach and apply our model to the Uniform Data Set from the National Alzheimer's Coordinating Center to identify cognitive decline patterns among dementia patients.

Original languageEnglish (US)
Pages (from-to)1-21
Number of pages21
JournalJournal of Applied Statistics
DOIs
StateAccepted/In press - Feb 17 2016

    Fingerprint

Keywords

  • cognitive decline
  • dementia
  • Multivariate finite mixture latent trajectory

ASJC Scopus subject areas

  • Statistics and Probability
  • Statistics, Probability and Uncertainty

Cite this