sdm-help-list > How to determine if someone used Brett transf
Showing 1-2 of 2 posts
Display:
Results per page:
Oct 31, 2013  09:10 PM | Nobody
How to determine if someone used Brett transf
We are finding it difficult to replicate some of the findings of the research on the SDM website. We think there are two reasons for that:

1. We are unsure of when authors are using Matthew Brett's conversion versus when authors are reporting correct Talairach space (which influences whether we create mni2tal.txt or tal.txt files).
For example, in the Lansley 2013 meta-analysis, Audoin has two articles (2006 and 2010). The exact wording about the conversion process is very similar, yet one is coded as mni2tal.txt and the other is coded as tal.txt

Here is what the articles say about the conversion:

The Audoin 2010 article says "Coordinates of significant clusters in the MNI space were transformed into Talairach coordinates using a nonlinear transformation to locate these clusters" (This was coded as Standard space: Talairach)

The Audoin 2006 article says: "MNI coordinates were converted to the Talairach space [34] using a non-linear transformation (www.mrccbu.cam.ac.uk/imaging/mnispace.html)." (This was coded as Standard space: mni2tal transformation)

Although not really the most reliable method, we thought of using the table provided in brainmap.org as a rule of thumb for determining what software uses what type of transformation. Is that a good solution?


Is there a reliable method to determine if the authors of the primary articles used/did not use Brett transformation?
Nov 4, 2013  09:11 PM | Nobody
RE: How to determine if someone used Brett transf
Sometimes you may not know whether an article used Brett's conversion or not, giving room to some uncertainty.

However, note that coordinates in Audoin 2006 can be assumed to be in mni2tal because the article specifically states the Brett's conversion website. Conversely, Audoin 2010 does not state the conversion method and is a relatively recent article, for what we don't know whether it used a newer conversion method.

In any case, SDM uses relatively large kernels (e.g. 20mm), for what the effects of specifying tal instead of mni2tal or vice versa should be relatively subtle.

Regarding the table, please note that you don't need to transform the coordinates in SDM. The software will conduct the required transformations.