No, I'm pretty sure we are using the same version of SPM!
(at least, SPM2_released_with_updates in the present context, though I would
recommend SPM5, as in previous email).
Perhaps the confusion is over the precise SPM2 menu option you chose:
one-way ANOVA
one-way ANOVA (with constant)
one-way ANOVA (within-subjects)
The former models condition effects only. The second adds a single constant
term. The third adds a column for each subject (see my previous email to
Bernhard). (I didn't choose these terms!). In the context of 2nd-level
designs for fMRI data, which is appropriate depends on the nature of the
1st-level contrasts used to generate the data (ie the contrast images). If
you want to be safe, you could always add subject effects (ie choose the
third option) and model all conditions, though this will not generalise to
ANOVAs with a between-subject factor.
Or perhaps I am not explaining things very well?! Please (and seriously) -
if someone else wants to try to explain ANOVAs and SPM - please do!
Rik
--------------------------------------------------------
DR RICHARD HENSON
MRC Cognition & Brain Sciences Unit
15 Chaucer Road, Cambridge,
CB2 2EF England
EMAIL: [log in to unmask]
URL: http://www.mrc-cbu.cam.ac.uk/~rik.henson
TEL +44 (0)1223 355 294 x522
FAX +44 (0)1223 359 062
MOB +44 (0)794 1377 345
--------------------------------------------------------
>-----Original Message-----
>From: SPM (Statistical Parametric Mapping)
>[mailto:[log in to unmask]] On Behalf Of FATIMA HUSAIN
>Sent: 25 May 2006 21:18
>To: [log in to unmask]
>Subject: Re: [SPM] ANOVA design matrix
>
>
>My experiment details seem similar to yours.
>For what it is worth, my ANOVA did not look anything like the
>Rik Henson
>one either. It looked like yours. I checked with other people
>and they got
>similar to ours. I tried different ways of setting up the ANOVA and
>could't get the Henson_Penny version.
>Perhaps Henson used a different version of SPM?
>
>Can't help with the "not uniquely specified" problem. Didn't get that.
>
>Fatima T. Husain, Ph.D.
>NIDCD/NIH
>Bldg. 10/ Room 8S235-D
>Phone: 301-594-7758
>
>On Thu, 25 May 2006, Bernhard Staresina wrote:
>
>> Dear SPM experts,
>>
>> I'm trying to set up a within subjects ANOVA in SPM2, with
>beta values for certain conditions as
>> dependent measures. I have 16 subjects and three conditions
>of interest. Correspondingly, I enter 16
>> subjects and 3 conditions at the prompts, and then load up
>the 3 beta images of interest, subject
>> after subject. However, when I look at the resulting design
>matrix (attached as 'anova_matrix'), it
>> doesn't look like the example provided by Rik Henson and
>Will Penny (attached as
>> 'r_henson_example'), and importantly it says that the betas
>are not uniquely specified. My question is
>> whether I did something wrong with the setup and/or whether
>I should be worried about the 'not
>> uniquely specified' statement.
>>
>> Thanks a lot for your help!
>>
>> Cheers,
>> Bernhard
>>
>>
>>
>>
>>
>
|