Hi,
Sorry for the previous email, the rank deficient problem didn't come from the EB but the covariate that I had just added. This covariate had been demeaned before. Do you know how to correct this problem. My design is the same as example in fslwiki:
EV1: run (1 for 1st scan, -1 for the 2nd)
EV2: group x run (1 for group 1, -1 for group 2)
EV3,...: mean subject
And the demeaned covariate.
Many thanks,
Best regard,
Kim TD
On Mon, 28 Aug 2017 15:30:22 +0100, Kim TRAN <[log in to unmask]> wrote:
>Dear Mark,
>
>Thank you so much for your advise. As each subject has 2 time scans, should I add Exchange Block (EB) into model? I tried with EB and this made the design rank deficient.
>
>Many thanks,
>
>Best regard,
>
>Kim TD
>
>On Mon, 28 Aug 2017 14:18:56 +0000, Mark Jenkinson <[log in to unmask]> wrote:
>
>>Hi,
>>
>>Yes, you can add a covariate to this model. Just make sure that:
>>1 - it is demeaned
>>2 - it is not the same as one of the factors or a combination of them (as this would make the design rank deficient)
>>
>>If your confound varies with subject then the second condition should be met, in which case just make sure that you demean the values before entering them into the EV.
>>
>>All the best,
>> Mark
>>
>>
>>> On 28 Aug 2017, at 10:58, Kim TRAN <[log in to unmask]> wrote:
>>>
>>> Dear FSL expert,
>>>
>>> I want to perform an ANOVA model for 2-groups, 2-levels per subject (2-way Mixed Effect ANOVA) following fslwiki GLM. Otherwise, I also doubt about some confounding factors, could I add them to model as covariate? I really appreciate any advice you can provide.
>>>
>>> Best regard,
>>>
>>> Kim TD
|