I am not sure that there is a simple answer to this, work with ICA and ASL has been less common to date. My instinct would be to start with the difference data as this will remove a lot of sources of variance that you are not interested in. However, I probably wouldn’t do much other preprocessing.
Michael
> On 3 Jan 2016, at 20:51, gabriel robert <[log in to unmask]> wrote:
>
> Dear FSL experts,
>
> I would like to run some independent component analysis with Melodic on resting state ASL data. what kind of preprocessing would you recommend and what type of data shall I start Melodic on (ie shall I start with the pairs or with the difference between the pairs of images) ?
>
> Many thanks for your help,
>
> Gabriel
|