A comparison of various MRI feature types for characterizing whole brain anatomical differences using linear pattern recognition methods

Gemma C Monté-Rubio,Carles Falcón,Edith Pomarol-Clotet,John Ashburner

doi:10.1016/j.neuroimage.2018.05.065

Abstract

There is a widespread interest in applying pattern recognition methods to anatomical neuroimaging data, but so far, there has been relatively little investigation into how best to derive image features in order to make the most accurate predictions. In this work, a Gaussian Process machine learning approach was used for predicting age, gender and body mass index (BMI) of subjects in the IXI dataset, as well as age, gender and diagnostic status using the ABIDE and COBRE datasets. MRI data were segmented and aligned using SPM12, and a variety of feature representations were derived from this preprocessing. We compared classification and regression accuracy using the different sorts of features, and with various degrees of spatial smoothing. Results suggested that feature sets that did not ignore the implicit background tissue class, tended to result in better overall performance, whereas some of the most commonly used feature sets performed relatively poorly.

Highlights

A common goal of neuroimaging research involves identifying morphometric alterations associated with particular diseases
Assessment using test information showed that body mass index (BMI) can best be predicted from unmodulated features
Yokum et al, (2012) showed that obese participants had lower total GM and WM volume than lean and overweight participants, but BMI correlated with higher WM volumes in the middle temporal gyrus, fusiform gyrus, parahippocampal gyrus, Rolandic operculum, and dorsal striatum

Summary

Introduction

A common goal of neuroimaging research involves identifying morphometric alterations associated with particular diseases. Thousands of studies have involved applying the Voxel-Based Morphometry (VBM) technique (Wright et al, 1995; Ashburner and Friston 2000, 2001) for comparing brain anatomies. With VBM, the aim is to test a hypothesis at each voxel using multiple linear regression (“mass-univariate statistics”). Multiple linear regression is a special case of the general linear model, which is a framework that encompasses multivariate approaches (such as MANOVA and MANCOVA) that deal with multiple independent and dependent variables. There are thousands or even millions of dependent variables, so many recent developments have been based on pattern recognition and other machine learning approaches that provide principled ways of dealing with the “curse of dimensionality”

Objectives

Methods

Results

Conclusion