Dear Anson,
I have a confusion here. I have
2 scans, one is acup, one is motor. I would like
to see their common activation sites. Should I use conjunctive
analysis.
Yes, this would be appropriate. Note that SPM99 only tests the
conjunction
null. This means you should make it clear that your inference is
based on
the minimum t-value (and does not necessarily mean both t-tests are
significant)
If so, I would like
to confirm the way I am doing it is right. I defined 4
contrast: acup rest>active; acup rest<active; motor rest>active
and motor
rest<active. My friend taught me, when I want to do a conjunctive
analysis, I
should choose acup rest<active and than press the CTRL key on the
keybroad, to
choose motor rest<active. Is it right? (I am using SPM99 in
WinXP)
Yes.
Or should I define
a new contrast, e.g. acup vs motor -1 1 -1 1?
No. This is the average of the rest>active over
rest and acupuncture.
(i.e. the main effect of active minus rest). The conjunction is a
multi-
variate test that requires both contrasts to have a t-statistic that is
greater than some minimum value
I hope this helps - Karl