Content area
Full Text
Neuroinform (2014) 12:471486 DOI 10.1007/s12021-014-9223-8
SOFTWARE ORIGINAL ARTICLE
MANIAA Pattern Classification Toolbox for Neuroimaging Data
Dominik Grotegerd & Ronny Redlich &
Jorge R. C. Almeida & Mona Riemenschneider &
Harald Kugel & Volker Arolt & Udo Dannlowski
Published online: 28 March 2014# Springer Science+Business Media New York 2014
Abstract Conventional univariate statistics are common and widespread in neuroimaging research. However, functional and structural MRI data reveal a multivariate nature, since neighboring voxels are highly correlated and different localized brain regions activate interdependently. Multivariate pattern classification techniques are capable of overcoming shortcomings of univariate statistics. A rising interest in such approaches on neuroimaging data leads to an increasing demand of appropriate software and tools in this field. Here, we introduce and release MANIAMachine learning Application for NeuroImaging Analyses. MANIA is a Matlab based software toolbox enabling easy pattern classification of neuroimaging data and offering a broad assortment of machine learning algorithms and feature selection methods. Between groups classifications are the main scope of this software, for instance the differentiation between patients
and controls. A special emphasis was placed on an intuitive and easy to use graphical user interface allowing quick implementation and guidance also for clinically oriented researchers. MANIA is free and open source, published under GPL3 license. This work will give an overview regarding the functionality and the modular software architecture as well as a comparison between other software packages.
Keywords Pattern classification . Neuroimaging software . Machine learning . fMRI . MVPA
Introduction
In the past, the overwhelming majority of fMRI data analysis strategies were based on mass univariate statistics (i.e. general linear model). These mass univariate analyses allow for statistically investigating regional differences of brain activation between different experimental conditions and/or different groups of participants and have been the state of the art approach in this field. Despite the fact that voxels are spatial correlated, these methods treat voxels as being independent from all others. Hence, they ignore the multivariate properties of the data (e.g. neighboring voxels are highly correlated). A possible answer to this drawback are multivariate machine learning techniques, which are already very well established in other research fields, for instance in drug research (Dybowski et al. 2011) and cancer research (Heider et al. 2010).
Recently, such multivariate techniques have become...