Quoting shruthi chakrapani <[log in to unmask]>:
> Hello All
> I am new to SPM and I am trying to analyse some of the data set using SPM5
> which were already alaysed in SPM99.I am following the same pipeline used
> with SPM99.I have 2 Groups with 2 conditions i.e
> Cont grp 1: has 12 subjects with 2 conditions and Meth grp 2: has 10
> subjects with 2 conditons.
> I am trying to specify the parameters for the Flexible factorial design and
> I am a little bit confused while doing this.Do you guys have an example with
> multiple condition and grp analysis like what you have in the SPM5 manual
> for Subject and condition(Verbal Fluency PET data).
> In SPM99 and SPM2 they ask questions like ,How many subjects? How many
> groups? How many conditions? and so on.If there is a standard example for
> group analysis in SPM5 that would be really helpful.
> Meanwhile here is what I have done.
> 3 factors: Subj,
> Cond
> Grp
> Under subject: (GRP 1)
> Subject 1
> Scan : Input file -Subj1,Cond1,Grp1
> Condition : 1 1 1
> Subject 2
> Scan : Input file -Subj1,cond2,grp1
> Condition : 1 2 1
> -
> -
> -
> Subject n (GRP 2)
> Scan : Input file -Subjn,cond 1 ,grp 2
> Condition : n 1 2
> Subject n+1
> Scan : Input file -Subjn+1,cond 2,grp2
> Condition : n+1 2 2
> -
> -
> -
> - so on for all the subjects.
>
>
> Does this look correct as to how I am specifying the data and the
> conditions.Any help would be greatly appreciated.
>
> Thanks in advance
> Shruthi
Hi Shruthi,
I was able to make the SPM5 design matrix look almost exactly like
the SPM99 and SPM2 analyses of the same data, once by entering "all
scans" at once with a 44 row factor-specifying matrix, and also by
entering the scans subject by subject. The only difference is that the
22 subjects-by-group interaction columns are listed as "conditions" in
SPM5 instead of the "block effects" of earlier SPM versions.
Unfortunately, either way I set up the design, estimation hangs near
the end with an error message that "matrix dimensions" must match. All
scans are the same size and worked fine in SPM99 and SPM2. They can be
viewed in SPM5, so I thought they could be analyzed. Can anyone on the
list shed light on what is going wrong?
Thanks,
|