Dear Karl
I hope you don't mind me writing to you but I'm confused about the best way to analyze my dataset and would really appreciate your advice.
I have 3 groups 2 drugs 1 placebo and 8 conditions. 15 subjects per group.
I (with the help of Cyril and Jan Glascher) have set up a flexible factorial design in spm5 with subjects, group and condition as factors.
I haven't been able to figure out how to do a main effect of group (with the 3 groups in) I have attached the design matrix to show what I have done.
Also I have now done comparisons between groups for each condition and have great results with a threshold of 0.001 unc and clusters corrected at 0.001 etc
so that is comparing drug 1 with placebo
then drug2 with placebo
and I compare drug 1 with drug 2
but is this the best way to do it? how do I really compare the 3 groups ?
also to extract the bold signal change i would like to do it for the clusters and have tried MARSbar but I'm getting errors
perhaps because its a flex design not sure
is there another way to extract the bold from perhaps a ROI identified from a condition activated by ALL-subjects (n=45)
so I have identified a cluster how do I analyse just this region between groups in spm5
or could i extract the data and then analyse this data in spss is this valid? would it get around the problem of comparing the 3 groups directly?
Thanks so much hope it made sense :)
Thanks
Ciara
Dr Ciara McCabe
Neurosciences Building
Department of Psychiatry
Warneford Hospital
University of Oxford
Warneford Lane
Oxford
Tel:01865 223778
mob:07966035631
E:[log in to unmask]
http://www.psychiatry.ox.ac.uk/research/researchunits/perl/perlmembers/ciaramccabe/
|