Hi Rafael
An 'easy' way to do contrast is to make sure you sum up to 0 (although
is some cases a contrast is legal even if not summing up to 1 - a good
thing is illegal contrasts are now allowed within SPM). In your case
anyway [2 -1 -1] is fine, testing group 1 as stronger than gp 2 and 3
pooled together. For a full test you should use a F test [1 -1 0; 0 1
-1; 1 0 -1] revealing any differences between any groups.
For 'trajectories' or indeed to look at 'shapes' across groups another
easy way is simply to write what you want and remove the mean. So say
you want 1 2 3 the the contrast is [1 2 3] - mean([1 2 3]) which indeed
is [-1 0 1], but doing so you could test for other shapes like exp([1 2
3]) - mean(exp([1 2 3])) etc ..
good luck
cyril
> Dear SPM statisticians
>
> I compare fMRI activation maps between 3 Groups (full factorial design). First, I compared the levels in the Factor "Group" by post-hoc t-test (e.g. [1 -1 0]). Now, as my groups differ in age, I want to test developmental trajectories like "early development" or "late development" by contrasts like [1 -1 -1] or [-1 1 1].
> I am not sure if these contrasts are statistically legal, because I produce an inballance by comparing the average of two groups against one single group, probably violating some statistical rules? Maybe it is required to recover ballance by weighting the single group by [2 -1 -1]? Further, would you test a linear trayectory by [-1 0 1] or rather by [1 2 3]?
>
> Many thanks for any advise
>
> Rafael
>
>
--
Dr Cyril Pernet,
SBIRC fMRI Manager
Lecturer in cognitive neuroimaging
SFC Brain Imaging Research Center
Division of Clinical Neurosciences
University of Edinburgh
Western General Hospital
Crewe Road
Edinburgh
EH4 2XU
Scotland, UK
[log in to unmask]
tel: +44(0)1315373661
http://www.sbirc.ed.ac.uk/cyril
http://www.sinapse.ac.uk/
The University of Edinburgh is a charitable body, registered in
Scotland, with registration number SC005336.
|