Hi Jacqueline,
That is fine.
All the best,
Anderson
On Thu, 11 Oct 2018 at 13:34, Jacqueline Scholl <
[log in to unmask]> wrote:
> Hi Anderson,
> Thank you so much for your reply!
>
> I was thinking of testing with the one-sample ttest whether there is a
> (within subject) difference between eg the default mode network in two
> conditions (eg rest and rumination). For this i have made the difference of
> the dual regression cope outputs for the two conditions. Would a one-sample
> ttest not make sense in this case? (Or would it just not make sense to
> test eg the default mode network in rest per se, which i understand why
> that would be weird)
> Best wishes
> Jacquie
>
> On Thursday, October 11, 2018, Anderson M. Winkler <[log in to unmask]>
> wrote:
>
>> Hi Jacqueline,
>>
>> That's right. In general (but not for the dual regression), you could
>> still use the original design as that would accommodate the group
>> differences as nuisance.
>>
>> For the dual regression, however, it doesn't make much sense to do a
>> 1-sample t-test, because the ICs already have a non-zero average in the
>> locations of the networks they represent. Same applies to strictly positive
>> quantities, such as gray matter volume, fractional anisotropy, or cortical
>> thickness: it doesn't make sense to do a one-sample t-test to check whether
>> the mean is different than zero because it is. That is, there is no
>> negative gray matter volume, so there isn't really anything to be tested.
>>
>> All the best,
>>
>> Anderson
>>
>>
>> On Wed, 19 Sep 2018 at 12:54, Jacqueline Scholl <
>> [log in to unmask]> wrote:
>>
>>> I think i've figured out what the problem was: I tried to do one-sample
>>> t-tests from the design matrix above. If now instead I only use the design
>>> matrix for doing between-group t-tests and run the one-sample t-tests
>>> separately with the -1 option for randomise this work. I was just assuming
>>> it would also work in one large contrast design matrix to get randomise to
>>> do both (like FLAME does). but happy it's working now :)
>>>
>>> On Wed, Sep 19, 2018 at 9:51 AM Jacquelne Scholl <
>>> [log in to unmask]> wrote:
>>>
>>>> ps: I'm calling randomise like this:
>>>> randomise_parallel = -i ${newCompFile} -o ${resultsFile} -d
>>>> ${desListDir}/${ides}.mat -t ${desListDir}/{ides}.con -n 5000 -T -m
>>>> ${maskFile}
>>>>
>>>> and this is in a loop over the different (25) ICA components and
>>>> designs (30).
>>>>
>>>>
>>> ------------------------------
>>>
>>> To unsubscribe from the FSL list, click the following link:
>>> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1
>>>
>>
>> ------------------------------
>>
>> To unsubscribe from the FSL list, click the following link:
>> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1
>>
>
> ------------------------------
>
> To unsubscribe from the FSL list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1
>
########################################################################
To unsubscribe from the FSL list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1
