Dear Olga,
It won't be an issue that your con images are from SPM2.
To reduce the amount of information you're inputting to your 2nd level
design perhaps you could create a contrast at the first level for
each subject: after ttt - before ttt. This would remove your 'scan'
factor. You now don't need to worry that subject X's overall response
may be larger than subject Y's as you have relative values for each
subject - ie. you now don't need to have subject as a factor either.
So you should be able to do all you need with a two-sample t-test: group
1 versus group 2. Scans for each are after ttt - before ttt.
Best,
Will.
Olga Prilipko wrote:
> Dear Will,
>
> I am writing to you directly as I have seen that you have answered
> several questions concerning the full factorial/flexible factorial
> designs in SPM5 on the SPM email list and thus was wondering if you
> could help me with a problem I have encountered in SPM5:
>
> when specifying a full factorial (2x2) design or a flexible factorial
> with 3 factors (subject, group and scan - I have 28 subjects divided
> in 2 groups of 14 and who have all been scanned 2 times (before and
> after ttt), I get an error at the estimation level (after the
> hyperparameter estimation passes successfully): "please check your
> data, there are no significant voxels" This is quite disturbing since
> the con.img used are working fine in other types of analysis and also
> in a flexible factorial design if the subject factor is not
> specified. The con.img I am using have been created in SPM2 - could
> this be a problem? Any ideas on why this is happening and how to
> solve this would be of great help.
>
> Thank you very much,
>
> Olga Prilipko, MD Postdoctoral fellow Stanford Sleep Clinic
>
>
--
William D. Penny
Wellcome Trust Centre for Neuroimaging
University College London
12 Queen Square
London WC1N 3BG
Tel: 020 7833 7475
FAX: 020 7813 1420
Email: [log in to unmask]
URL: http://www.fil.ion.ucl.ac.uk/~wpenny/
|