Skip to content


Folders and files

Last commit message
Last commit date

Latest commit


Repository files navigation



Data analysis code for the myconnectome project

The goal of this project is to demonstrate reproducible analysis for a large and complex dataset. The package provides a set of scripts that will implement all of the statistical analyses and some of the preprocessing steps for the data from the MyConnectome study, as reported by Poldrack et al. (submitted). Code is not currently included for the cortical parcellation steps that were performed at Washington University, which is available [here] (

Users who wish to use this package in a turnkey fashion should try the Myconnectome-VM which will automatically set up a virtual machine that will complete the full statistical analysis workflow.


The code shared here requires a number of dependencies:


If you don't already have a scientific Python distribution installed, I would recommend Anaconda. Beyond the standard scientific Python stack, a number of additional packages are required, which you can install using the following set of commands (assuming that you have Anaconda already installed):

pip install setuptools

conda install --yes pip numpy scipy nose traits networkx

conda install --yes dateutil ipython-notebook matplotlib

conda install --yes statsmodels boto pandas scikit-learn

pip install nibabel

pip install gtf_to_genes

pip install suds

pip install mygene


The packages used here require R 3.0 or greater; I would recommend installing the latest version from CRAN. The R scripts in the package will attempt to automatically install all necessary packages, so you shouldn't need to install any additional packages yourself.

Connectome Workbench

Some of the processing operations require the [Connectome Workbench] (, which can also be used to visualize the surface-based results.


Some of the resting fMRI processing operations using MATLAB with the Brain Connectivity Toolbox. If you wish for those to be completed on your machine, you must have MATLAB installed with the BCT in your MATLAB path. If MATLAB is not installed, then those results will be downloaded directly from our archive.

Getting started

To install the package, first clone it to your local machine:

git clone

Then run the setup script:

cd myconnectome python install

Once this is installed, you need to set some environment variables:

  • MYCONNECTOME_DIR: this is the directory where all of the results will be put
  • WORKBENCH_BIN_DIR: this is the directory where the Connectome Workbench binary files (e.g. wb_command) are located
  • DAVID_EMAIL: This optional setting specifies an email address that has been registered for the [DAVID web service] ( If is it not specified, then the results will be downloaded directly from our archive.