Multimodal Image Analysis in Alzheimer's Disease via Statistical Modelling of Non-local Intensity Correlations.

Marco Lorenzi,Sebastien Ourselin,Jonathan M Schott,Sjoerd B Vos,Ivor J Simpson,Marc Modat,Alex F Mendelson,M Jorge Cardoso

doi:10.1038/srep22161

Abstract

The joint analysis of brain atrophy measured with magnetic resonance imaging (MRI) and hypometabolism measured with positron emission tomography with fluorodeoxyglucose (FDG-PET) is of primary importance in developing models of pathological changes in Alzheimer’s disease (AD). Most of the current multimodal analyses in AD assume a local (spatially overlapping) relationship between MR and FDG-PET intensities. However, it is well known that atrophy and hypometabolism are prominent in different anatomical areas. The aim of this work is to describe the relationship between atrophy and hypometabolism by means of a data-driven statistical model of non-overlapping intensity correlations. For this purpose, FDG-PET and MRI signals are jointly analyzed through a computationally tractable formulation of partial least squares regression (PLSR). The PLSR model is estimated and validated on a large clinical cohort of 1049 individuals from the ADNI dataset. Results show that the proposed non-local analysis outperforms classical local approaches in terms of predictive accuracy while providing a plausible description of disease dynamics: early AD is characterised by non-overlapping temporal atrophy and temporo-parietal hypometabolism, while the later disease stages show overlapping brain atrophy and hypometabolism spread in temporal, parietal and cortical areas.

Highlights

The joint analysis of brain atrophy measured with magnetic resonance imaging (MRI) and hypometabolism measured with positron emission tomography with fluorodeoxyglucose (FDG-PET) is of primary importance in developing models of pathological changes in Alzheimer’s disease (AD)
We notice that partial least squares regression (PLSR) generally provides a better fit, while the local patch based method leads to larger estimation errors in parietal and temporal areas
We have investigated the problem of multimodal analysis of biomedical images in AD, by comparing two different modelling hypothesis based on state-of-art techniques, PLSR and patch-based local correlation, to promote non-local correlation analysis approaches with respect to localized ones in describing multimodal correlation patterns in AD

Summary

Introduction

The joint analysis of brain atrophy measured with magnetic resonance imaging (MRI) and hypometabolism measured with positron emission tomography with fluorodeoxyglucose (FDG-PET) is of primary importance in developing models of pathological changes in Alzheimer’s disease (AD). The multimodal analysis of anatomical and physiological images is of primary importance in developing comprehensive models of biological processes and pathologies, and increasing the statistical power of current imaging biomarkers Already, both brain atrophy, measured in magnetic resonance images (MRIs), and hypometabolism, quantified by positron emission tomography with fluorodeoxyglucose radiotracers (FDG-PET), are among the primary diagnostic biomarkers of Alzheimer’s disease (AD). Several techniques have been proposed for modelling non-overlapping signal correlations in the field of functional MRI analysis Both independent component analysis (ICA) or partial least square (PLS) approaches have been successfully applied to the joint analysis of functional activation in the brain and covariates drawn from genetic, clinical, or imaging data[5,6,7]. This enables them to model the potentially significant interactions between voxels located in completely different areas of a single image, or between voxels in images of different modalities

Objectives

Results

Conclusion