Hi Carlotta,
It is entirely up to you whether you want to correlate against each
individual behavioural variable, or against the mean across all of
them. It depends on what questions you want to ask, and whether it
makes sense or not to calcualte the mean across your behavioural
variables.
Paul
On 29/03/2019, Carlotta Fabris <[log in to unmask]> wrote:
> Thank you!
>
> This example was really easy to understand and I got what you were trying
> to tell me, so now I will apply it to my specific experiment.
>
> Just one question: suppose that in the experiment I scanned the patients
> while ingesting different types of cheese. Now if I want to see the
> correlation between mean activation of the ingestion of cheese in general
> and preferential scores for all the different cheese types, should I do a
> mean of the preferential scores of each cheese type?
>
> Thanks again, you are always very clear and kind!
> Carlotta.
>
>
> Il Ven 29 Mar 2019, 14:31 Matthew Webster <[log in to unmask]>
> ha scritto:
>
>> Hi Carlotta,
>> It’s unclear to me what _specific_ hypothesis is being tested here.
>> Without knowing this, it is very difficult to answer the question.
>>
>> The following simplified example shows one way in which a data-set could
>> be analysed ( with some reference to the JoM article):
>>
>> Imagine an experiment in which participants are asked to state their
>> preference for a particular foodstuff ( e.g. cheese ) on a scale of 1 to
>> 4.
>> The participants are scanned while ingesting a number of foods including
>> cheese. A potential contrast for a first-level FEAT might be activation
>> while ingesting cheese ( assuming some kind of block design ). A
>> higher-level analysis could combine these contrasts to generate a group
>> ROI
>> ( cluster-mask ) for cheese-related activation. This mask could then be
>> applied to the subject-specific first-level contrasts, using the
>> standard-space versions of the copes, to obtain each subject’s mean
>> response in that ROI. A correlation could then be calculated ( e.g. in
>> MATLAB ) between preference scores and mean activation.
>>
>> This is obviously a very simple example, but the concepts should be
>> expandable to your specific analysis,
>> Hope this helps,
>> Kind Regards,
>> Matthew
>>
>> --------------------------------
>> Dr Matthew Webster
>> FMRIB Centre
>> John Radcliffe Hospital
>> University of Oxford
>>
>> On 29 Mar 2019, at 08:48, Carlotta Fabris <[log in to unmask]>
>> wrote:
>>
>> Thank you so much Paul!
>>
>> My DVs represent the answer that the partecipants gave for each stimuli.
>> They are on a scale from 1 to 4.
>> If I do this in the group level analysis, what is the best approach to
>> follow?
>>
>> Thanks,
>> Carlotta.
>>
>> Il Gio 28 Mar 2019, 16:52 paul mccarthy <[log in to unmask]> ha
>> scritto:
>>
>>> Hi Carlotta,
>>>
>>> > I was interested in calculating the correlation of my DVs with the
>>> > peaks
>>> > of the third level analyses I did, so I won't have the values for each
>>> > individual, but for the different groups. Should I use the second
>>> > level
>>> > analyses to calculate the scalar values for each subject?
>>>
>>> You should do this at whatever level is relevant to the question you
>>> want to ask.
>>>
>>> > Another thing I wanted to ask is, I have 250 DVs for each subject,
>>> should I
>>> > do a mean of all of them to find the correlation of these with the
>>> scalar
>>> > values?
>>>
>>> I don't know - what do your variables represent?
>>>
>>> Paul
>>>
>>>
>>>
>>> On 27/03/2019, Carlotta Fabris <[log in to unmask]> wrote:
>>> > Thank you Paul!
>>> >
>>> > I was interested in calculating the correlation of my DVs with the
>>> peaks of
>>> > the third level analyses I did, so I won't have the values for each
>>> > individual, but for the different groups. Should I use the second
>>> > level
>>> > analyses to calculate the scalar values for each subject?
>>> >
>>> > Another thing I wanted to ask is, I have 250 DVs for each subject,
>>> should I
>>> > do a mean of all of them to find the correlation of these with the
>>> scalar
>>> > values?
>>> >
>>> > Thanks,
>>> > Carlotta.
>>> >
>>> >
>>> > Il Mer 27 Mar 2019, 15:59 paul mccarthy <[log in to unmask]> ha
>>> > scritto:
>>> >
>>> >> Hi Carlotta,
>>> >>
>>> >> There are many different ways to create ROI masks, such as manually
>>> >> drawing, using a discrete or probabilistic atlas, or defining a
>>> >> spheres areound areas of peak activation. You need to decide which
>>> >> areas of the brain you are interested in, and then create a binary
>>> >> mask image with 1s in those areas, and 0s everywhere else.
>>> >>
>>> >> After you've calculated a scalar value, derived from COPE values
>>> >> within your region, for each individual in your study, you can then
>>> >> correlate them with your other measures as you were already trying to
>>> >> do - using MATLAB, SPSS, or similar.
>>> >>
>>> >> Cheers,
>>> >>
>>> >> Paul
>>> >>
>>> >> On 27/03/2019, Carlotta Fabris <[log in to unmask]> wrote:
>>> >> > Hi Paul!
>>> >> >
>>> >> > I am sorry to bother you, but could you explain me something more
>>> about
>>> >> what
>>> >> > should I do to create the mask and also what I should do after the
>>> >> > calculation of the regional metrics?
>>> >> >
>>> >> > Thank you a lot,
>>> >> > Carlotta.
>>> >> >
>>> >> >
>>> ########################################################################
>>> >> >
>>> >> > 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
>>>
>>
>> ------------------------------
>>
>> 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
|