Measuring the replicability of fMRI results using a large database.

Github project for the paper: "The empirical replicability of task-based fMRI in function of sample size."

Introduction

In this project, we investigate the replicability of fMRI results. We measure the latter through the (un)conditional test-retest reliability, coherence and stability of fMRI group studies under increasing sample size. We use the a database with N = 1400 subjects from where we subsample independent replications.

Structure of project

• Sampling: contains the scripts which were used for sampling and generating the results.
• Analyses: contains the scripts to generate the pre-processed data, the analyses and finally the figures for the paper.

The actual database is separated from this repository due to privacy reasons. Pre-processed (and blinded) intermediate R files are saved in /2_Analyses/_IntData. To generate the figures, one then only needs this folder and the /2_Analyses/_Figures folder.

Contact

• Han Bossier : han.bossier@ugent.be

First Created: 26/06/2015