MRIPredict

Free R package to easily predict diagnosis from sMRI scans

Created by the Imaging of mood- and anxiety- related disorders (IMARD) group,
IDIBAPS, Hospital Clínic de Barcelona

MRIPredict is a software package for using sMRI data to easily predict diagnosis (e.g. whether the patient will respond to a treatment, or whether the patient has one or another disorder).

Note that it is based on the glmnet library

The model may include covariates, and the software conducts both cross-validation of the model and fitting for its use with new VBM data.

Work co-funded by the Instituto de Salud Carlos III - Subdirección General de Evaluación y Fomento de la Investigación and the European Regional Development Fund (FEDER). Project grants CP14/00041, PI14/00292, PI14/01148 and PI14/01151.

EASY AND COMPATIBLE

MRIPredict is aimed to be easy to use and user friendly. Both for R experts or those who prefer an easy to use interface-based application, to adapt best to every professional needs.

R package

R package and an easy to use interface-based shiny application.

INSTALLATION AND USAGE

devtools

install.packages("devtools")

devtools::install_github("alsolanes/mripredict")

download

install.packages("path/to/mripredict_1.01.tar.gz", repos = NULL, type="source")

COMPATIBLE WITH THE MAIN OS

MRIPredict, as an R package, is compatible with:

LINUX

MAC OS

WINDOWS

Download MRIPredict

The software is fully functional and may be already used within R. Currently, it has the following limitations, which will be addressed promptly. To run the software:

R code works but it is not a proper documented R package yet.
To load the package, install the package using the methods described in the section installation

library(mripredict)

After loading the package run ?mripredict in order to see examples on how to run the software.
To run the interface-based software, after loading the package, execute the instruction launchApp().

Download

DETAILS

Methodological details
The software is currently optimized for predicting binary or continuous variables, and conducting cox risk estimation survival analysis from VBM data, which must be registered to the MNI space (we also recommend smoothing). For each fold of the cross-validation as well as for the final fitting, the software creates a mask of the voxels with variance > 0, applies ComBat if the data is multisite and performs multiple imputation if there are missing values. Voxelwise applies a linear model to remove the effects of covariates, and includes the residuals into a lasso regression using the lambda that gives the minimum error in a cross-validation. The intercept of the lasso regression, as well as the MNI coordinates, the covariates coefficients and the lasso coefficient of the voxels with non-null lasso coefficients are saved.