Thanks very much. I thought the original question is going to be buried
without being answered. I've done Approach 1 and am holding to re-do it
with Approach 2 until I get some expert opinion on it. The second approach
definitely requires more labor. Are there any strong reasons against the
first approach (e.g. statistical validity, filter efficiency, motion
correction across runs, etc.)? I will go ahead and anlyze the data with
Approach 2 and see how much the end results differ.
P.S. Another message (Distributed Computing Once More) I sent on Aug. 30
did not get people involved to discuss. Hopefully FSL development team and
other experts on this issue can look into it to see if it is useful. One of
the advantages of FSL is that it enables easy batch processing (mostly
serially) with scripts. Batch parallel processing has to be able to run on
distributed/clustered system to gain full power.
On Sat, 4 Sep 2004 09:21:50 +0100, Stephen Smith <[log in to unmask]>
wrote:
>Hi, yes that all makes sense. Approach 2 is what you want and you're right
>to take out the constant-EV ev1 from your original email.
>
>Cheers.
>
>
>On Tue, 31 Aug 2004, Xun Liu wrote:
>
>> On Tue, 31 Aug 2004 21:17:33 +0100, X Liu <[log in to unmask]>
wrote:
>>
>> >Concatenate run/session or not?
>> >
>> >We have a design with 2 independent variables (IV1 and IV2), each with 3
>> >levels. IV1 is manipulated across blocks within a run and IV2 is
>> >manipulated across runs. There are two approaches I can think of to
anaylze
>> >the data for the main effects and interaction for this design. What is
the
>> >advantage and disadvantage of each approach, from the conceptual and
>> >practical perspectives? Is one more valid than the other from the
>> >statistics point of view? Thanks very much.
>> >
>> >
>> >Approach 2:
>> >Analyze each run separately and model just the 3 conditions/levels of
IV1
>> >(ev1, ev2, ev3). Then the main effects of IV1 can be set up as below.
>> >
>> >Contrast ev1 ev2 ev3
>> >mean (1) 1 1 1
>> >IV1 (2) 1 -1 0
>> > (3) 1 0 -1
>> > (4) 0 1 -1
>> >
>> >And then proceed to the group analysis with the EVs/contrasts setup as
>> >below for the three levels of IV2 (say for 5 subjects)
>> >
>> >Group ev1 ev2 ev3 ev4 ev5 ev6 ev7 ev8
>> >1 1 1 1 1 0 0 0 0
>> >1 1 1 1 0 1 0 0 0
>> >1 1 1 1 0 0 1 0 0
>> >1 1 1 1 0 0 0 1 0
>> >1 1 1 1 0 0 0 0 1
>> >1 1 -1 0 1 0 0 0 0
>> >1 1 -1 0 0 1 0 0 0
>> >1 1 -1 0 0 0 1 0 0
>> >1 1 -1 0 0 0 0 1 0
>> >1 1 -1 0 0 0 0 0 1
>> >1 1 0 -1 1 0 0 0 0
>> >1 1 0 -1 0 1 0 0 0
>> >1 1 0 -1 0 0 1 0 0
>> >1 1 0 -1 0 0 0 1 0
>> >1 1 0 -1 0 0 0 0 1
>> >
>> >Contrast ev1 ev2 ev3 ev4 ev5 ev6 ev7 ev8
>> >c1 1 0 0 0 0 0 0 0
>> >c2 0 1 0 0 0 0 0 0
>> >c3 0 0 1 0 0 0 0 0
>> >c4 0 1 -1 0 0 0 0 0
>> >
>> >Then the .gfeat folder should include 4 cope#.feat subfolders, one for
each
>> >of the contrasts from the first level. zstat2 to zstat4 of cope1.feat
will
>> >assess the main effect of IV2 (zstat1 is the overall grand mean of both
IVs
>> >again baseline). zstat1 of cope2.feat to cope4.feat will assess the main
>> >effect of IV1. zstat2 to zstat4 of cope2.feat to cope4.feat will assess
the
>> >interactions.
>>
>> Just found out an error. To correct myself, ev1 is a linear combination
of
>> ev4 to ev8. So take out ev1 and contrast 1 is just the combination of the
>> new ev3 to ev7 (weighted or not WRT to the other contrasts?). The
>> interpretation of the main effects and interaction above stands.
>>
>> Group ev1 ev2 ev3 ev4 ev5 ev6 ev7
>> 1 1 1 1 0 0 0 0
>> 1 1 1 0 1 0 0 0
>> 1 1 1 0 0 1 0 0
>> 1 1 1 0 0 0 1 0
>> 1 1 1 0 0 0 0 1
>> 1 -1 0 1 0 0 0 0
>> 1 -1 0 0 1 0 0 0
>> 1 -1 0 0 0 1 0 0
>> 1 -1 0 0 0 0 1 0
>> 1 -1 0 0 0 0 0 1
>> 1 0 -1 1 0 0 0 0
>> 1 0 -1 0 1 0 0 0
>> 1 0 -1 0 0 1 0 0
>> 1 0 -1 0 0 0 1 0
>> 1 0 -1 0 0 0 0 1
>>
>> Contrast ev1 ev2 ev3 ev4 ev5 ev6 ev7
>> c1 0 0 0.2 0.2 0.2 0.2 0.2
>> c2 1 0 0 0 0 0 0
>> c3 0 1 0 0 0 0 0
>> c4 1 -1 0 0 0 0 0
>>
>
> Stephen M. Smith DPhil
> Associate Director, FMRIB and Analysis Research Coordinator
>
> Oxford University Centre for Functional MRI of the Brain
> John Radcliffe Hospital, Headington, Oxford OX3 9DU, UK
> +44 (0) 1865 222726 (fax 222717)
>
> [log in to unmask] http://www.fmrib.ox.ac.uk/~steve
|