An Adjusted Likelihood Ratio Test for Separability in Unbalanced Multivariate Repeated Measures Data

TL;DR

This paper introduces an adjusted likelihood ratio test for assessing two-factor separability in unbalanced multivariate repeated measures data, especially where within-subject correlations decay exponentially, with applications to medical imaging.

Contribution

The paper presents a novel adjusted likelihood ratio test tailored for unbalanced data with exponential correlation decay, generalizable to various matrix structures.

Findings

Simulation study confirms the test's accuracy.

Application to medical imaging demonstrates practical utility.

Method outperforms existing tests in unbalanced settings.

Abstract

We propose an adjusted likelihood ratio test of two-factor separability (Kronecker product structure) for unbalanced multivariate repeated measures data. Here we address the particular case where the within subject correlation is believed to decrease exponentially in both dimensions (e.g., temporal and spatial dimensions). However, the test can be easily generalized to factor specific matrices of any structure. A simulation study is conducted to assess the inference accuracy of the proposed test. Longitudinal medical imaging data concerning schizophrenia and caudate morphology illustrates the methodology.