Hi,
You want to ignore the A-B-C variability, so the answer depends on
whether you still want to use (esimate) the cross-subject variability
- so yes probably FE at this highest level isn't what you want as it
will ignore all cross-input variabilities. So either your approach is
fine, or you could do FE across A,B,C separately for each subject at
an intermediate level analysis (and then simple ME using the outputs
of that across all subjects) at highest level, or your approach should
be fine too.
Cheers.
On 25 Nov 2008, at 13:34, Bradley Goodyear wrote:
> Thanks Steve.
> Yes, but in this case, A,B, and C are levels of task difficulty
> actually.
> Not just three sessions of the same thing.
> I was interested in mean task activity in the absence of the
> difficulty modulation.
> That's a little different than just using the fixed effects option
> isn't it?
>
> -Brad
>
>
> On Nov 25, 2008, at 1:48 AM, Steve Smith wrote:
>
>> Hi,
>>
>> A = a + b + c
>> B = a -b
>> C = a-c
>>
>> A+B+C = 3a
>>
>> So yes a [1 0 0 ] contrast gives you the mean of A,B,C.
>>
>> However if you just want to ignore cross-session variability you
>> could just use an all 1s single EV and use the fixed-effects option?
>>
>> Cheers.
>>
>>
>>
>>
>> On 24 Nov 2008, at 22:04, Brad Goodyear wrote:
>>
>>> Hi.
>>> I want to compute the average across three conditions for each
>>> subject by removing any
>>> differences between the conditions.
>>> For two conditions, I understand the EVs would be
>>>
>>> EV1 EV2
>>> 1 1
>>> 1 -1
>>>
>>> would it not, and I compute the contrast (1,0) for the average
>>> across the two conditions
>>> with any differences removed?
>>>
>>> For three conditions, is it
>>>
>>> EV1 EV2 (a) EV3 (b)
>>> 1 1 1
>>> 1 -1 0
>>> 1 0 -1
>>>
>>> and then compute (1,0,0) since conditions A+B+C = a+b-a-b = 0, as
>>> per the tripled t-test
>>> example?
>>> I then plan to average this (1,0,0) contrast across subjects to
>>> get the mean across
>>> condition with any differences between the conditions removed.
>>>
>>> -Brad
>>>
>>
>>
>> ---------------------------------------------------------------------------
>> 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
---------------------------------------------------------------------------
|