Dear everyone,
I replicated a previous fmri study. Now I just finished data analysis for all of
12 subjects. I've got several questions. Hope somebody can help me out. Thanks.
First, let me give a brief introduction about the study. There were two runs in
the study, one was crossed hand(c), the other was uncrossed hand(u). in each run
there were two conditions, one was Compatible (C), the other was
Incompatible(I). So totally there were 4 conditions which are uC,uI, cC,cI.
For each condition (uC,uI,cC,cI contrast) in random effect analysis, I used
p<0.001 (uncorrect), I could see many activation. But for uI-uC, cI-cC,
(uI+cI)>(uC+cC), I could not see any activation with p<0.001(uncorrected), which
was used in original paper and got many activations. Even with p<0.005, no
activation either. Only for contrast (cC+cI)>(uC+uI), I could see several
activation with p<0.001. Only when I set up p<0.01 or p<0.05, many activation
could be seen.
So my question here is does this mean I failed to find something based on my
results?
I also found that half of subjects without head motion (between 0.5-1mm) and
half of them with obvious head motion for which I've discard corresponding
datapoints. I tried to compare the random effect results for both groups, I
found that I could see activations in all of 4 comparisons for the group without
head motion though activations were not as much as original paper. Meanwhile, I
also could see activations in 2 comparisons in the group with head motion, but
there was not activation in the other 2 comparisons of the group with head
motion. Does it mean I should do more strict head motion correction for the
group with head motion? Could anyone help me to anlyze it and suggest me what I
should do next? Thanks in advance!
Best regards,
--
Chunhong Shao
[log in to unmask]
Research Associate
Department of Psychological Sciences
Purdue University
|