On Mon, Jan 14, 2013 at 4:51 AM, Cong <[log in to unmask]> wrote:
> Hi,
>
> I just ran my first PPI analysis using FEAT and was unsure how to interpret the results. I followed the steps outlined on the FMRIB site and made a GLM using PSYCH/PHYS/PPI EVs then set up a positive and negative contrast for the interaction effect (0,0,1; 0,0,-1). There were no significant clusters found for the positive contrast, but there were for the negative contrast. I've been searching the mailing list for clues on how to interpret this and only came across this post "https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind1101&L=FSL&D=0&P=237986".
>
> So a negative interaction means there is decreased connectivity between the seed region and structures activated as a result of the PSYCH EV?
>>>> The PSYCH EV is A-B, so if you were to flip the tasks, you'd get a positive effect. All a negative contrast means is that you one condition is greater than the other.
Does this mean that as activity in the seed region increases there is
a decrease in activity elicited by the task EV (or vice versa)?
>>>> PPI represents a difference in the connectivity of the tasks. Since you only have a difference, you can't really interpret what it means. Its like a BOLD contrast of A-B, you can't say if A or B are both positiive, both negative, or one positive and one negative from the this contrast. You'd need information about A and B alone. In terms of A versus baseline, you'd need another PPI to look at that. Then you could interpret A relative to baseline, but you'd need to look at the baseline connectivity to interpret if A is increasing in absolute terms. It may be sufficient to say the connectivity is greater during the task than baseline though.
>
> The earlier post I found also mentioned that there hasn't been any reports in the literature of decreased connectivity. So I'm just wondering if I've skipped over some information regarding PPI analysis and this contrast shouldn't have been run in the first place.
>>> This is related to what an individual calls A and what an individual calls B. Negative or decreased connectivity occurs during rest, so it must also be able to occur in a task setting too.
>
> Also my analysis only has one task EV, unlike the scenario in the previous post.
Hope this helps. I also recommend reading the paper on Generalized PPI.
A generalized form of context-dependent psychophysiological
interactions (gPPI): a comparison to standard approaches.
McLaren DG, Ries ML, Xu G, Johnson SC.
Neuroimage. 2012 Jul 16;61(4):1277-86. doi:
10.1016/j.neuroimage.2012.03.068. Epub 2012 Mar 30.
>
> Thanks for your help,
>
> Cong
|