[pymvpa] Pre-processing the BOLD images (no qform_code or sform_code)

Xixi Wang xixi.wang577 at gmail.com
Thu Feb 19 21:04:27 UTC 2015

Hi PyMVPA users,

I'm trying to perform a surface based analysis using the Haxby dataset and
I've been stuck with the preprocessing of the fMRI dataset.

According to the tutorial page (
http://dev.pymvpa.org/datadb/tutorial_data.html#datadb-tutorial-data), the
4D fMRI time-series have been skull-stripped and motion corrected and the
preprocessed data is the bold.nii.gz file.

I tried doing similar preprocessing steps on all six raw bold data using
either A) the following command:
$ bet bold_raw bold_brain -F -f 0.5 -g 0
$ mcflirt -in bold_brain -out bold_mc -plots

or B) pymvpa2-prep-fmri.py bold_raw

I then checked the raw bold data and the output file (bold_mc.nii.gz) using
$fslhd and fslview and found out that the qform_code and sform_code were
both 0, and the volumes were incorrectly oriented (the axial and coronal
views should be swapped and the sagittal view was in wrong orientation).
However, for the preprocessed tutorial data bold.nii.gz, the header
information qform_code and sform_code are both 1 and the qto_xyz and
sto_xyz matrices are different from my preprocessed data. I was wondering
whether I'm missing some key preprocessing steps which might lead to my
incorrect results here.

Any help would be greatly appreciated!

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.alioth.debian.org/pipermail/pkg-exppsy-pymvpa/attachments/20150219/970193d6/attachment.html>

More information about the Pkg-ExpPsy-PyMVPA mailing list