On Fri, Feb 25, 2011 at 7:02 PM, Salem Boussida
<[log in to unmask]> wrote:
> Dear Donald,
>
> Thank you for your help. I am still confused about some details:
>
>
> "MCLAREN, Donald" <[log in to unmask]> a écrit :
>
>> Salem,
>>
>> This isn't a mixed effects design as you only have one group and only
>> one condition. What you want to do is to use the con_001 images in a
>> one-sample t-test. DO NOT USE the con_002 images in the model. I
>> suspect that might have been what happened.
>
> I am still confused about "mixed effects analysis"and random effects
> analysis. In prevoius discussions from SPMlist users and the notes from
> Terry Oakes page
> (http://psyphz.psych.wisc.edu/~oakes/spm/spm_random_effects.html), it is
> mentioned that "Random-Effects" analysis is also referred to as a "Mixed
> Effects" analysis, since it considers both within- and between-subject
> variance. Also, in SPM manual, the one-sample-t-test is considered as a
> "random effects analysis".
> Any comments about this?
> Which test (in SPM) should I use to consider both fixed and random effects?
The whole discussion of fixed effects and random effects are best left
to the field of statistician. If you want the details of why it could
be considered either, see Ch.12 in Human Brain Function.
It is better to think of the analysis as within-subject factor and
within-subject errors OR between-subject factors and between errors.
In group comparison or comparing the group against zero, you have
between subject errors (1 sample t-test, 2-sample t-test, ANOVA).
Paired t-tests and repeated measures ANOVAs have within-subject error
terms. And then, you can have models with both within-subject and
between-subject errors.
>
>
>
>>
>> Your design (SPM.xX.X)hould have a size of [7 1]; load SPM.mat and
>> then use size(SPM.xX.X) to find the answer.
>
> The "SPM.xX.X" command gives me : SPM.xX.X =
> 1
> 1
> 1
> 1
> 1
> 1
> 1
> Is it right?
Yes.
>
>>
>> A t-contrast of 1 will find the regions that are activated. A
>> t-contrast of -1 will find regions that are deactivated. An F-contrast
>> is simply the square of the 2 t-tests and does not have a direction.
>>
> Thank you again for your help.
>
> Best regards,
>
> Salem
>
>
>> Best Regards, Donald McLaren
>> =================
>> D.G. McLaren, Ph.D.
>> Postdoctoral Research Fellow, GRECC, Bedford VA
>> Research Fellow, Department of Neurology, Massachusetts General
>> Hospital and Harvard Medical School
>> Office: (773) 406-2464
>> =====================
>> This e-mail contains CONFIDENTIAL INFORMATION which may contain
>> PROTECTED HEALTHCARE INFORMATION and may also be LEGALLY PRIVILEGED
>> and which is intended only for the use of the individual or entity
>> named above. If the reader of the e-mail is not the intended recipient
>> or the employee or agent responsible for delivering it to the intended
>> recipient, you are hereby notified that you are in possession of
>> confidential and privileged information. Any unauthorized use,
>> disclosure, copying or the taking of any action in reliance on the
>> contents of this information is strictly prohibited and may be
>> unlawful. If you have received this e-mail unintentionally, please
>> immediately notify the sender via telephone at (773) 406-2464 or
>> email.
>>
>>
>>
>> On Fri, Feb 25, 2011 at 2:33 PM, SALEM BOUSSIDA
>> <[log in to unmask]> wrote:
>>>
>>> Dear SPMers,
>>>
>>> I have a group of 7 subjects (rats) and I would like to perform a group
>>> analysis (mixed effects).
>>> I have done the 1st level analysis (fixed effects) for each subject and
>>> I
>>> have two con_001 and con_002 images, wich are respectively the contrast
>>> image for a positive effect (activation) and for a negative effect
>>> (deactivation) of one condition (current stimulation) .
>>> Then, I have done a one-sample-t-test : I feed all the con_001 of each
>>> subject into the 2nd level analysis , then I entered a T contrast [1]
>>> and
>>> SPM displayed the thresholded T-statistic image. But when I entered a F
>>> contrast [1], I have got this error:
>>> "
>>>
>>> ??? Inf computed by model function, fitting cannot continue.
>>> Try using or tightening upper and lower bounds on coefficients.
>>>
>>> "
>>>
>>> Based on the procedure described in SPM8 manuel concerning the "face
>>> group
>>> data", It is mentionned that I have to enter a F contrast.
>>> I am confused by this and I have some questions:
>>> (1) which contrast should I enter, F or T contrast? and what is the
>>> differenece between them?
>>> (2) Is the one-sample-t-test, the appropriate one to run mixed effects
>>> analysis?
>>> (3) Do I have to run two separate "one-sample-t-test" for
>>> positive(activation) and negative-deactivation) effects?
>>>
>>>
>>> Thank you for your help.
>>>
>>> Best regards,
>>>
>>> Salem
>>>
>>
>
>
>
>
>
|