Hello!
Yes the -e option is really only needed for repeated-measures,
although the numbering is the other way around ( 1 is for the 1st
subject, 2 for the 2nd and so on).
many Regards
Matthew
>
> Just to understand that correctly: So I only would need the -e option when
> I
> am doing repeated-measures?
> And it would be 1 for the first measuer and 2 for the same subject's 2nd
> measure etc.?
>
> Thanks again!
> Markus
>
>
> 2010/1/7 Matthew Webster <[log in to unmask]>
>
>> Hello,
>> As long as you have independent data, all different subjects etc ,
>> then there is no need to include an exchangeability-blocks file (
>> it can be called anything, design.grp is just an example name ). It
>> looks like your design below is fine - you don't need to generate a
>> design.grp file ).
>>
>> Many Regards
>> Matthew
>> > Dear Experts,
>> >
>> > I came across the message below from Tom Nichols, because I was
>> searching
>> > for a reason why randomise did not work properly when I included a
>> > design.grp file (made with Glm) with the -e option.
>> >
>> > 1) Is this still true? What exactly has to be edited? Is the extension
>> > .grp
>> > or .exb?
>> >
>> > 2) I am trying to find different imaging correlates for two different
>> > regressors.
>> > My design is as follows:
>> >
>> > EV1 = con_regr1
>> > EV2 = pat_regr1
>> > EV3 = con_regr2
>> > EV4 = pat_regr2
>> > EV5 = con
>> > EV6 = pat
>> >
>> > con 2.3 0 30 0 1 0
>> > con 4.3 0 39 0 1 0
>> > con 1.1 0 28 0 1 0
>> > pat 0 4.5 0 10 0 1
>> > pat 0 6.5 0 20 0 1
>> > pat 0 7.8 0 17 0 1
>> >
>> > contrast will be [1 -1 0 0 0 0 ] and [0 0 1 -1 0 0 ] with F-test.
>> >
>> > Does this sound reasonable? And my question: Do I nead to include a
>> > desing.grp file with 1 for con and 2 for pat?
>> >
>> > Thanks very much!
>> > Markus
>> >
>> >
>> >
>> >
>> > 2008/10/1 Thomas Nichols <[log in to unmask]>
>> >
>> >> Xiaochu,
>> >>
>> >> Glm provides a group file expected by Feat; randomise expects an
>> >> exchangeability block file. Unfortunately, you have to hand-edit the
>> >> exchangeability block file to denote blocks of data coming from the
>> same
>> >> subjects (i.e. the repeated measures). In your case, you can copy
>> your
>> >> designg.grp into, say, design.exb, and then have the following after
>> the
>> >> line "/Matrix"
>> >> 1
>> >> 2
>> >> 3
>> >> ...
>> >> 24
>> >>
>> >> 25
>> >> 1
>> >> 2
>> >> 3
>> >> ...
>> >> 25
>> >> This tells randomise that you don't have independent data, and that
>> the
>> >> 1st
>> >> and 26th observation are from the same subject, and the 2nd and 27th
>> are
>> >> same, etc. Then supply this .exb file to randomise's -e option (or
>> -g
>> >> if
>> >> before FSL 2.1).
>> >>
>> >> That should get you sorted.
>> >>
>> >> -Tom
>> >>
>> >>
>> >>
>> >> On Wed, Oct 1, 2008 at 2:55 PM, Zhang, Xiaochu (NIH/NIDA) [F] <
>> >> [log in to unmask]> wrote:
>> >>
>> >>>
>> >>> Thank you so much Tom!
>> >>> Something should be wrong. I need to find where is wrong.
>> >>> My design.grp is all "1". The attachment is the parameter files I
>> used
>> >>> for
>> >>> paired t-test. They are created by "Glm" program automatically.
>> >>> For save the email size, I will send the files for 25 participants
>> >>> (paired t test). My scans order is "1 2 3.....25 1 2 3.....25".
>> >>> Could you please do me a favor and check these files for me? If it
>> is
>> >>> possible, could you please do me a favor and send your files to me?
>> >>>
>> >>> For threshold, now I knew it was "one-tail" but not "two-tail". We
>> are
>> >>> consistent each other in this field.
>> >>>
>> >>> Xiaochu Zhang PhD
>> >>>
>> >>> Visiting Research Fellow
>> >>>
>> >>> Neuroimaging Research Branch
>> >>>
>> >>> National Institute on Drug Abuse - IRP
>> >>>
>> >>> Biomedical Research Center
>> >>>
>> >>> 251 Bayview Blvd.
>> >>>
>> >>> Suite 200 (NIDA)
>> >>>
>> >>> Baltimore MD
>> >>>
>> >>> 21224
>> >>>
>> >>> Tel: 443-740-2619
>> >>>
>> >>>
>> >>
>> >>
>> >>
>> >> --
>> >> ____________________________________________
>> >> Thomas Nichols, PhD
>> >> Director, Modelling & Genetics
>> >> GlaxoSmithKline Clinical Imaging Centre
>> >>
>> >> Senior Research Fellow
>> >> Oxford University FMRIB Centre
>> >>
>> >
>>
>
>
>
> --
>
>
> Dr. med. Markus Gschwind, M.D.
> Laboratory for Neurology and Imaging of Cognition - Dept of Neurosciences
> University Medical Center (CMU)
> 1, Michel-Servet
> CH-1211 GENEVA - Switzerland
>
> Tel 0041 (0) 22 379 5324
> Fax 0041 (0) 22 379 5402
> email: [log in to unmask]
> http://labnic.unige.ch
>
|