Table of Contents

• Analyzing the FIAC data using SPM and Phiwave
  • Download and unpack the data
  • Setting up the analysis
  • tsdiffana review
  • Subject, session exclusion
  • Fieldmap undistort
  • Coregistration
  • Analysis stream

Analyzing the FIAC data using SPM and Phiwave

FIAC information here:

http://www.madic.org/fiac/how_to_participate.html

http://www.madic.org/fiac/MRIsequences.htm

Download and unpack the data

Download the data, and unpack it into a directory. On my machine the directory was /home/imagers/FIAC, and unpacking gave me, for the second subject, a subdirectory fiac1, with sub-subdirectories like this:

|-- fiac1
|   |-- anat1
|   |-- fieldmap2
|   |-- fonc1
|   |-- fonc2
|   |-- fonc3
|   `-- fonc4

Make sure you have also downloaded the onsets. For your convenience I have attached an archive of FIAC onsets that unpacks into the directory structure above.

Setting up the analysis

1. The unwarp toolbox. This should be in the matlab path ahead of SPM
2. The FieldMap toolbox. You should replace the FieldMap.m file in that toolbox with the edited FieldMap.m version.
3. MarsBaR. You need MarsBaR in order to use Phiwave (below) - at least version v0.38.2
4. Phiwave - version 3.3.
5. GroovyBatch SPM2 metabatch scripts.
6. The FIAC GroovyBatch scripts.

tsdiffana review

http://www.mrc-cbu.cam.ac.uk/Imaging/Common/downloads/SPMUtils/tsdiffana.tar.gz

I first ran tsdiffana on each of the datasets, in fact using the groovy_diff batch script, from GroovyBatch.

See the tsdiffana plots in the attachment table at the bottom of this page. The following table summarizes (s2 means session 2).

The following table gives the output from the tsdiffana run for all the subjects; each is a pdf file or about 340k

Subject, session exclusion

Omit subjects 0 and 11 (no fieldmaps)

Omit subject 5 (no anatomical)

Omit session 3 for subject 10 (asleep)

From tsdiffana analysis: omit subject 8, subject 12 session 1.

If we had been using subject 11, we would have excluded session 3, due to tsdiffana spikes.

Fieldmap undistort

http://www.fil.ion.ucl.ac.uk/spm/toolbox/fieldmap/

In order to get the fieldmap undistort working with the batch system, I had to use a modified version of the main FieldMap script. Please download this to replace the original if you want to use the batch files here.

Coregistration

The FIAC images all have .mat files, which the pvconv conversion program generates automatically from the scanner data, to give the position of the image in terms of the magnet isocentre. This means that the anatomical, fieldmap etc images are already coregistered using the information in the .mat files.

After a little testing, it seemed that the automated SPM coregistration was giving a less successful coregistration than the pvconv scanner data, so I did not use coregistration for matching the anatomical to the EPI, or the fieldmap routines.

Analysis stream

Analysis stream common to SPM and Phiwave

1. tsdiffana (and subject / session exclusion)

2. slice timing

3. analyze fieldmaps

4. realign / unwarp + undistort using fieldmaps

5. segment structural

6. normalize segmented structural to gray template

7. write EPIs normalized

For SPM only

1. Smooth 5mm

2. Set up SPM model

3. Run SPM contrasts

For Phiwave only

1. Import SPM model

2. Convert to use unsmoothed images

3. Run model

4. Denoise contrasts

Last Refreshed: Thu Jan 12 12:42:27 GMT 2006