PRoNTo: Pattern Recognition for Neuroimaging Toolbox

Abstract

In the past years, mass univariate statistical analyses of neuroimaging data have been complemented by the use of multivariate pattern analyses, especially based on machine learning models. While these allow an increased sensitivity for the detection of spatially distributed effects compared to univariate techniques, they lack an established and accessible software framework. The goal of this work was to build a toolbox comprising all the necessary functionalities for multivariate analyses of neuroimaging data, based on machine learning models. The “Pattern Recognition for Neuroimaging Toolbox” (PRoNTo) is open-source, cross-platform, MATLAB-based and SPM compatible, therefore being suitable for both cognitive and clinical neuroscience research. In addition, it is designed to facilitate novel contributions from developers, aiming to improve the interaction between the neuroimaging and machine learning communities. Here, we introduce PRoNTo by presenting examples of possible research questions that can be addressed with the machine learning framework implemented in PRoNTo, and cannot be easily investigated with mass univariate statistical analysis.

Regional volumes and spatial volumetric distribution of gray matter in the gender dysphoric brain

“…In our study, we used the analysis pipeline for pattern recognition analyses provided by PRoNTo (http://www.mlnl.cs.ucl.ac.uk/pronto). This pipeline can be used to automatically search for regularities in the data and train a classifier function that models the relation between spatial signal patterns and experimental factors (in our case, group) based on a training dataset (Schrouff et al, 2013). This classifier can then be used to predict the group a new image belongs to using the spatial distribution of the signal within the image, and compute the accuracy with which groups can be discriminated from one another based on whole-brain spatial signal patterns.…”

Brain functional plasticity associated with the emergence of expertise in extreme language control

“…These difference maps were labeled as coming from the trained participants or the control participants, and submitted to a multivariate classification procedure using the PRoNTo toolbox (Schrouff et al, 2013), running in Matlab. A linear support vector machine (SVM) was trained to classify the difference maps as originating either from the interpreters or the controls, using a 35-fold leave one subject out procedure.…”

Localizing and Comparing Weight Maps Generated from Linear Kernel Machine Learning Models

“…More specifically, as in [9], the older subjects (from 60 to 90 years old) were selected for regression based on the scalar momentum features resulting from the normalization [10], [11].…”

“…In both cases, a leaveone-subject-out cross-validation was performed to compute model performance, its significance being assessed by a nonparametric testing using 1000 random permutations of the training labels. All machine learning based modelling steps have been performed in PRoNTo [9] 3 .…”