Hi
In ICA in general the amount that an individual volume contributes to a given component is encoded by the corresponding entry in the unmixing matrix. You could then simply average the volumes per subject to get a surrogate measure of subject-wise contribution...
hth
Christian
On 20 Apr 2010, at 21:11, Renat Yakupov wrote:
> Hi Christian,
>
> Thank you for the reply. I was wondering if you could also help me with the
> following.
> Lets say we have a group of subjects and after analysis we end up with
> several default mode networks. Is there any way to find how much is
> contributed by each subject to each of those components? As I understand,
> each component can be represented as a linear combination of components from
> each subjects. Is there a way to look at those? Thank you.
>
> Regards,
> Renat.
>
> -----Original Message-----
> From: FSL - FMRIB's Software Library [mailto:[log in to unmask]] On Behalf
> Of Christian F. Beckmann
> Sent: Wednesday, April 07, 2010 1:39 AM
> To: [log in to unmask]
> Subject: Re: [FSL] how to quantify ICA analysis results
>
> Hi
>
> The Z-converted stats images are all contained in melodic_IC, i.e. these are
> maps where the 'background noise' should be scaled close to unit standard
> deviation so that every value in each map can be understood as a Z-score
> relative to the background noise distribution.
> hth
> Christian
>
> On 6 Apr 2010, at 22:26, Renat Yakupov wrote:
>
>> Hello,
>>
>> I am analyzing resting state data for 2 groups of subjects. I ran the
> analysis till the end and got nice maps with t-scores.
>> My question is how and where can I get the actual numbers that were used
> to produce those t-scores? Are there files for each subjects that have some
> kind of a resting state BOLD signal amplitude, just like you would have for
> a simple block-design paradigm? Thank you.
>>
>> Regards,
>> Renat.
|