<div dir="ltr">Thanks for the answer, Yaroslav.<br><div style="visibility: hidden; display: inline;" id="avg_ls_inline_popup"></div><style type="text/css">#avg_ls_inline_popup { position:absolute; z-index:9999; padding: 0px 0px; margin-left: 0px; margin-top: 0px; width: 240px; overflow: hidden; word-wrap: break-word; color: black; font-size: 10px; text-align: left; line-height: 13px;}</style>To be on the safe side, what is the equation for " z-scoring against baseline" you mentioned? I guess that you subtract the baseline condition instead of the average. But how you calculate the std?<br>
<br>Do you mind name some "reasonable
scaling approaches"? In one of your papers you also tested " scaled percentage change relative to baseline".<br><br>Thanks,<br>Vadim<br><br><br><div style="visibility: hidden; display: inline;" id="avg_ls_inline_popup">
</div><style type="text/css">#avg_ls_inline_popup { position:absolute; z-index:9999; padding: 0px 0px; margin-left: 0px; margin-top: 0px; width: 240px; overflow: hidden; word-wrap: break-word; color: black; font-size: 10px; text-align: left; line-height: 13px;}</style><br>
<div class="gmail_quote">On Sat, Jul 31, 2010 at 4:10 AM, Yaroslav Halchenko <span dir="ltr"><<a href="mailto:debian@onerussian.com">debian@onerussian.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
primary reason for zscoring is usually to bring data into a reasonable<br>
range of values for a classifier to be sensitive... usually it implies<br>
that values are primarily around 0 with values from -10 to 10 or so with<br>
the bulk from -2 to 2... z-scoring against baseline also has direct<br>
interpretation of cause<br>
<br>
in your case, if they are already betas from GLM -- they might be fine<br>
already... iirc betas are though not scaled by the level of unexplained<br>
variance (i.e. noise) so their values might be somewhat wild, so for<br>
classification it might still be worth z-scoring or any other reasonable<br>
scaling approach<br>
<br>
for correlation analysis situation is quite different though, especially<br>
if doing gross z-scoring (instead of z-scoring against baseline). e.g. if<br>
you within your patterns have some prominent preference, for instance,<br>
anterior higher than posterior for one category, z-scoring might<br>
damage that effect heavily or obliterate it entirely.<br>
<br>
Summary: z-score or not depends on what you want from your analysis and<br>
what is your hypothesis ;-)<br>
<div><div></div><div class="h5"><br>
On Fri, 30 Jul 2010, Vadim Axel wrote:<br>
<br>
> Hi,<br>
> When I run classification on raw fMRI data I always make a z-score<br>
> (subtract the mean and divide by std) for each voxel / scan.<br>
> Now I tried some sort of correlation analysis, while I run a<br>
> correlation for beta images (similar to Haxby 2001). So, the z-score is<br>
> going to be across beta_per_condition values (several values only). The<br>
> question is: does is make sense to zcore the beta results?<br>
> I ran some simulations while I add a noise to each voxel. So, I know<br>
> that there is a real correlation in my data. Whereas without zscore the<br>
> results indeed show high correlation, after zscoring procedure the<br>
> correlation is virtually disappeared. I also tried just to subtract the<br>
> mean across conditions, but the correlations still do not look good.<br>
> What is the correct way to do run this analysis?<br>
</div></div>--<br>
.-.<br>
=------------------------------ /v\ ----------------------------=<br>
Keep in touch // \\ (yoh@|www.)<a href="http://onerussian.com" target="_blank">onerussian.com</a><br>
Yaroslav Halchenko /( )\ ICQ#: 60653192<br>
Linux User ^^-^^ [175555]<br>
<br>
<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/mailman/listinfo/pkg-exppsy-pymvpa" target="_blank">http://lists.alioth.debian.org/mailman/listinfo/pkg-exppsy-pymvpa</a><br>
</blockquote></div><br></div>