Dear Mind-Tsung,
SPM will not let you calculate the single contrast because the columns in
the design matrix for a paired t-test are not uniquely specified. You can
test the effect with the contrast:
[ones(1,n)./n,0,1]
where n is the number of subjects.
A better approach (in terms of the way the variance is calculated) may be to
run a one-sample t-test on "high-temperature", save the thresholded results,
convert them to a binary inclusion mask (use ImCalc) and then apply this
mask to the paired t-test using WFU PickAtlas.
This will test for regions where high > low AND high > 0.
Hope this helps,
Paul Wright
On Thu, 4 Jun 2009 06:17:29 +0100, Ming-Tsung Tseng <[log in to unmask]>
wrote:
>Hi SPMers,
>
>(Here I paste my question again for more clear description)
>
>In my experiment, 14 subjects received 2 fMRI scannings (low-temperature vs.
>high-temperature stimuli), and I want to investigate the difference in brain
>activations for the 2 stimulus temperatures.
>
>The contrast ˇ§(high temperature) minus (low temperature)ˇ¨ means that which
>brain regions show significantly higher responses for the high-temperature
>stimulus than those for the low-temperature stimulus.
>
>I used paired t-test in SPM5 to perform this analysis. The procedures are as
>follows:
>
>(1).14 pairs of con*.img files (low temperature vs. high temperature) for each
>subject were established
>(2).The JPEG of the design matrix is attached
>(3).The contrast was 0 0 0 0 0 0 0 0 0 0 0 0 0 0 -1 1.
>(4).Then I got the activation map
>
>However, I want to get another map.....
>
>The contrast ˇ§(high temperature) minus (low temperature), with the (high
>temperature) as the inclusive maskˇ¨ means that which high temperature-
>activated brain regions show significantly higher responses for the high-
>temperature stimulus than those for the low-temperature stimulus
>(Is my understanding correct?).
>
>Again, I used paired t-test in SPM5 to perform this analysis. However, how
>can I ˇ§maskˇ¨ the ˇ§(high temperature)ˇ¨? Because it seems not allowed to
>establish the contrast ˇ§0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1ˇ¨.
>
>Can I inclusively mask ˇ§(high temperature) minus (low temperature)ˇ¨
>with ˇ§(high temperature)ˇ¨ in the paired t-test analysis in SPM5 2nd-level
>analysis?
>
>I'd appreciate for any help.
>
>Thanks
>
>Dr. Tseng
>
>
>
|