 |
 |
Hi Thang,While the degrees of freedom will make some difference, I wouldn’t think it would be that dramatic. Are you testing the conjunction null or global null? This will make a larger impact and changes the interpretation.Best,Donald--Hi Jacek,Thank you for your response. The degrees of freedom may be the issue. There was significant more activation. One with barely anything at voxel p = .05 FWE and the other with clusters of 1000 voxels. Anyway, I ended up following the steps from an old post on here.https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind04&L=SPM&P=R187333
This gave me results that I would have expected.
Dear Thang,
Could you provide some pictures? How much "more" activations (actually
overlap not activations per se) do you observe in your conjunction? I'm
not sure but it may be related to having two times more degrees of
freedom in your join model with both A>B and C>D in your model, so each
map resulting from it's contrast has more "power"? Therefore their
conjunction shows overlap in regions not detected by one sample Ts for
each of your conditions separately?
You can check the differences with projected overlaps of A>B from one
sample T and A>B from your "conjunction 2nd level"
Don't quote me on that though, there are much smarter people here so
let's wait for their advice :)
Best regards,
Jacek
W dniu 02.10.2018 o 21:04, Thang Le pisze:
> Hi fellow SPM users,
>
> I created a conjunction analysis for (A>B) AND (C>D) at the second
> level. However, I noticed that the results showed way more activations
> for this conjunction analysis than I would if I were to do an overlap
> analysis using the one-sample t-tests of (A>B) and (C>D) individually.
> Anyone has any ideas why this might be the case? Is this because of
> the difference between the one-sample t-test model when I do each
> contrast separately and the factorial design used in the conjunction?
> If so, how should I go about correcting the conjunction?
>
> Thanks everyone in advance!
>
> tl
--
Laboratory of Brain Imaging (LOBI)
Neurobiology Center
The Nencki Institute of Experimental Biology
Pasteur 3, 02-093 Warsaw, Poland
Tel.: +48 22 5892 551
mail: [log in to unmask]
website: http://lobi.nencki.gov.pl/
Best Regards,Donald McLaren, PhD