Can anyone help with this?
If a correlation is significant in the whole brain analysis, then
shouldn't we expect that same correlation ought to also be signficant in
an ROI analysis? If not, could you explain why this might be?
Specifically, our fMRI data show a significant correlation between signal
in a localized area of PFC and one of our behavioral variables. However,
when we look at this area as a region of interest, we do not find a
significant correlation between the same behavioral variable and the
contrast values from this area (obtained using the VOI tool).
Further information: we took my contrast values from an ROI taken from the
whole brain one-sample t-test for this contrast. We have also tried taking
them from the image in which activation is correlated with my behavioral
variable. Both provide similar contrast values for this ROI and neither
set correlates with my behavioral variable.
Also, how might we calculate percent signal change from the contrast values?
Thanks,
Michael Snow
Cornell University
Department of Human Development
|