Longitudinal Canonical Correlation Analysis

TL;DR

This paper introduces a method called Longitudinal Canonical Correlation Analysis (LCCA) for analyzing correlations between two high-dimensional longitudinal datasets, especially when data are sampled irregularly, and demonstrates its effectiveness through simulations and real Alzheimer’s data.

Contribution

The paper develops a novel LCCA method that models trajectories with random effects and handles irregular sampling, advancing correlation analysis in longitudinal high-dimensional data.

Findings

LCCA accurately recovers correlation patterns in simulations.

LCCA identifies meaningful brain change profiles in Alzheimer's data.

Method outperforms existing approaches in irregular sampling scenarios.

Abstract

This paper considers canonical correlation analysis for two longitudinal variables that are possibly sampled at different time resolutions with irregular grids. We modeled trajectories of the multivariate variables using random effects and found the most correlated sets of linear combinations in the latent space. Our numerical simulations showed that the longitudinal canonical correlation analysis effectively recovers underlying correlation patterns between two high-dimensional longitudinal data sets. We applied the proposed LCCA to data from the Alzheimer's Disease Neuroimaging Initiative and identified the longitudinal profiles of morphological brain changes and amyloid cumulation.