Hi Xiaoying,
Happy to answer all your questions, but can I ask you to post the same question to the icatb listserv instead of spm (not also many of the questions can be addressed by searching the archives)? We are now definitely off the topic of SPM and don't want to annoy the SPM list members. ;-) The website to join (if you aren't already) is: https://lists.sourceforge.net/lists/listinfo/icatb-discuss.
Thanks,
Vince
> -----Original Message-----
> From: Xiaoying Fan [mailto:[log in to unmask]]
> Sent: Thursday, March 15, 2012 12:59 PM
> To: [log in to unmask]; Vince Calhoun
> Subject: Re: group ICA problem
>
> Dear Dr. Vince Calhoun,
>
> Thank you so much for your quick reply. It seems I don't have good choice on that bad
> data.
> Do you mind if I ask a couple of more questions about ICA?
>
> 1. For the multi-session multi-subject ICA, how does GIFT deal with the multi-session
> for one subject? Concatenate all sessions together or average them first like GLM does
> somehow? How will it effect the 2nd level analysis( group ICA)?
>
> 2. Is the variance normalization GIFT uses same as FSL does for ICA? Could you
> please address a little more about intensity normalization that GIFT uses?
>
> 3. Do I need to do the slice timing correction for ICA? Or the principle is same as
> GLM?
>
> I am sorry for so many questions. Hope you don't mind replying.
>
> Best Regards,
>
> Xiaoying
|