Hi,<div><br></div><div>Yes, that makes sense now. Thanks for the time and thought you put into it!</div><div><br></div><div>Best,</div><div>Mike<br><br><div class="gmail_quote">On Mon, Nov 7, 2011 at 6:20 PM, Yaroslav Halchenko <span dir="ltr"><<a href="mailto:debian@onerussian.com">debian@onerussian.com</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;"><div class="im"><br>
On Mon, 07 Nov 2011, Mike E. Klein wrote:<br>
</div><div class="im">> - My reason for wanting to do this is because of a relative paucity of<br>
> examples (9 per run, 3 categories, 9 runs), which I plan to reduce in<br>
> number further by some averaging. It seems (to me) that the zscoring would<br>
> be more accurate when considering thousands of voxels (per example) as<br>
> opposed to several voxels (per run). <br>
<br>
</div>oy... full answer and pros/cons/against/for would take a while... there<br>
are papers which were published and such "preprocessing" was in-place...<br>
I even debated with one of the authors but never got to submit some kind<br>
of cririque... in case of Francisco's -- I guess I have missed that<br>
particular piece or it was in respect only for classification (which is<br>
as I said is ok)<br>
<br>
<br>
really coarse reasoning from me: the problem with zscoring (or<br>
even just plain demeaning) across voxels is that it leaks information<br>
among voxels... e.g. consider the most obvious simplified example<br>
where you have 2 voxels, 1 of which is informative, and one is not.<br>
<br>
after demeaning (let stay to the ground) -- they both become informative<br>
in respect to the condition of interest, so if you were to judge among<br>
"who is informative" -- you would be misguided<br>
<br>
in case of full brain, where majority of voxels is not informative and<br>
only few (unless you have really simple contrast/paradigm) are<br>
informative, zscoring shouldn't be as detremental, but fact would remain<br>
valid -- mean/std might be stimuli-dependent thus you would leak<br>
that information into each voxel (possibly even removing effects from<br>
informative voxels). So in case of distributed effects having a large<br>
"mass" -- they will be reflected as diagnostic information in the mean<br>
of the volumes, and then would be introduced into every voxel in the<br>
volume.<br>
<br>
altogether -- to be on a safe side I would not do anything like that ;)<br>
<br>
zscoring -- I would just zscore across run before extracting/meaning<br>
any samples you want to use for classification...<br>
<br>
hope it makes sense...<br>
<div class="im HOEnZb"><br>
> - I guess I don't see why zscoring this way would render a searchlight<br>
> invalid. I'm looking at the Pereira paper clip that was referenced<br>
> recently in the listserv: "In the example study, we normalized each<br>
> example (row) to have mean 0 and standard deviation 1. The idea in this<br>
> case is to reduce the effect of large, image-wide signal changes. Another<br>
> possibility would be to normalize each feature (column) to have mean 0 and<br>
> standard deviation 1, either across the entire experiment or within<br>
> examples coming from the same run." <br>
</div><div class="HOEnZb"><div class="h5">--<br>
=------------------------------------------------------------------=<br>
Keep in touch <a href="http://www.onerussian.com" target="_blank">www.onerussian.com</a><br>
Yaroslav Halchenko <a href="http://www.ohloh.net/accounts/yarikoptic" target="_blank">www.ohloh.net/accounts/yarikoptic</a><br>
<br>
_______________________________________________<br>
Pkg-ExpPsy-PyMVPA mailing list<br>
<a href="mailto:Pkg-ExpPsy-PyMVPA@lists.alioth.debian.org">Pkg-ExpPsy-PyMVPA@lists.alioth.debian.org</a><br>
<a href="http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-exppsy-pymvpa" target="_blank">http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-exppsy-pymvpa</a></div></div></blockquote></div><br></div>