Introduction to nipy

We first had a look at the range of nipy projects at http://nipy.org.

As part of nipy history, I (MB) mentioned Keith Worsley and Jonathan Taylor. Nipy largely started life as BrainStat which was Jonathan’s Python port of Keith’s fmristat package, written in MATLAB.

I mentioned the use of virtualenv to make isolated Python environments for installing new packages, and showed the install command for the libraries in the notebooks:

pip install numpy scipy matplotlib ipython[notebook] # general libraries
pip install nibabel nipy dipy  # imaging libraries

We started off looking at the NIPY diagnostics:

• nipy diagnostics.

These are based on the earlier tsdiffana code, written in MATLAB.

Next we talked about NIPY image slicing:

• nipy image slicing.

This brought us to a discussion of the image affine.

Then we did a brief run through of:

• combined slice timing and realignment;
• nipy image registration;
• dipy image registration using algorithms patterned after the ANTS toolbox;
• nipy statistical modeling including AR(1) modeling.