Tutorial : Pipeline

Running a pipeline

This is a step by step guide to setting up and running a pipelined analysis. For this tutorial we will be using a slightly modified version of the FBIRN Phase I travelling data set. The tutorial is based on an SPM analysis, with a few non-SPM things thrown into the mix.

Requirements

1. Release 0.1 of nipype and it’s dependencies have been installed
2. FSL and matlab are available and callable from the command line
3. SPM 5/8 is installed and callable in matlab
4. 2.5 GB of space

Step 0.

Download and extract the Pipeline tutorial data (429MB).

(checksum: f91b81050e1262f0508d35135c2369f5)

Step 1.

Ensure that all programs are available by calling bet, matlab and then which spm within matlab to ensure you have spm5/8 in your matlab path.

Step 2.

You can now run the tutorial by typing python tutorial_script.py within the nipype-tutorial directory. This will run a full first level analysis on two subjects following by a 1-sample t-test on their first level results. The next section goes through each section of the tutorial script and describes what it is doing.

Tutorials

In nipype, a pipeline is represented as an acyclic, directed data-flow graph, where each node of the graph is a process (e.g., realignment, smoothing) and the connections between the nodes control how data flows between the processes.

• Using SPM for analysis
• Using FreeSurfer for smoothing