Hi,
So each subject has 4 scans to combine together to get the 'double-
differential' contrast - you should feed all 4 first-level outputs
(which have all been run separately in first-level FEATs) into the
second-level FE analysis. You should be able to do something like
model=
1 0 0 0
0 1 0 0
0 0 1 0
0 0 0 1
contrast=
-1 1 1 -1
etc.
Then just feed this into a 3rd level group mean or difference between
the two groups of subjects.
Cheers.
On 18 Aug 2009, at 17:46, Tal Herbsman wrote:
> Thanks Steve,
> I read what I wrote below and I noticed that I worded it a little
> ambiguously so I just
> wanted to clarify:
>
> 1. the initial fixed effects analysis is run on the difference
> contrast COPE
> image for POST-PRE. so essentially I am asking is there a group
> difference
> among the difference of differences (active condition(post - pre) -
> placebo
> (post - pre)).
>
> 2. Not sure how to design it all in one matrix so instead I am
> running the
> lower level COPEs in seperate FEAT matrices, e.g.:
>
> FEAT run 1: subj1 POST - subj1 PRE (run 2n times for n subjects)
> FEAT run 2: ACTIVE difference COPE - PLACEBO difference cope for each
> subject (run n times)
> FEAT run 3: unpaired t test group difference of results of FEAT run
> 2. (run
> once)
>
> Thanks
> Tal.
>
>
> On Tue, 18 Aug 2009 08:34:11 +0200, Steve Smith
> <[log in to unmask]> wrote:
>
>> Hi,
>>
>> Yes - that looks right.
>>
>> Cheers.
>>
>>
>> On 17 Aug 2009, at 18:20, Tal Herbsman wrote:
>>
>>> I have the following experimental paradigm:
>>>
>>>
>>> 1. patient and healthy control groups
>>> 2. each subject has two separate scanning sessions on 2 days
>>> 3. on each day subject has fMRI task followed by a)active substance
>>> OR b)placebo
>>> administration and repeat fMRI task
>>>
>>> I wanted to confirm proper design matrix construction to test the
>>> question of whether
>>> activation deltas (post - pre substance administration) differ among
>>> group A and B
>>>
>>> My thought is to use a variation of the Multi-Session & Multi-
>>> Subject (Repeated Measures
>>> - Three Level Analysis) scenario on the FEAT webpage but instead of
>>> running the initial
>>> fixed effects analysis on subject mean COPEs, run it instead on
>>> COPEs generated by a
>>> fixed effect analysis of before and after condition for each
>>> subject.
>>>
>>> Not sure if this is correct or how to design the matrix. Any advice
>>> would be much
>>> appreciated.
>>>
>>> Tal.
>>>
>>
>>
>> ---------------------------------------------------------------------------
>> Stephen M. Smith, Professor of Biomedical Engineering
>> Associate Director, Oxford University FMRIB Centre
>>
>> FMRIB, JR Hospital, Headington, Oxford OX3 9DU, UK
>> +44 (0) 1865 222726 (fax 222717)
>> [log in to unmask] http://www.fmrib.ox.ac.uk/~steve
>> ---------------------------------------------------------------------------
>
---------------------------------------------------------------------------
Stephen M. Smith, Professor of Biomedical Engineering
Associate Director, Oxford University FMRIB Centre
FMRIB, JR Hospital, Headington, Oxford OX3 9DU, UK
+44 (0) 1865 222726 (fax 222717)
[log in to unmask] http://www.fmrib.ox.ac.uk/~steve
---------------------------------------------------------------------------
|