 |
 |
you can I'm sure do single subject in fsl as well for dr, but not for spatially constrained. It's basically like a weighted seed approach. In my experience it will not cleanly handle noise as well for subjects that weren't included in the original ica or if you include only a subset of the components, so you might consider some additional filtering in that case. My two cents. Vdc ------Original Message------ From: John Fredy To: Vince D. Calhoun Cc: SPM List Subject: Re: [SPM] default mode network Sent: May 18, 2012 4:38 PM Thanks, I always work with fsl and I have very single cases Best On Fri, May 18, 2012 at 5:31 PM, Vince Calhoun <[log in to unmask]> wrote: Hi John, From: John Fredy [mailto:[log in to unmask]] Sent: Friday, May 18, 2012 4:27 PM To: Vince Calhoun Cc: [log in to unmask] Subject: Re: [SPM] default mode network Dear Vince, can you please explain me with more detail the last point?,How I can run ICA spatially constrained?, Or, I can download the templates as volumes and run a multiple subject concatenation and then run dual regression? Well, if you use GIFT you can use the included spatially constrained ICA algorithm [1], and select the template you want and the data you want to analyze. You can do this for a single subject or a group of subjects. Likewise if you want to use the STR/DR approach this is provided as either a back-reconstruction option or as a utility (drop down on front screen) where it will ask you for the single subject fmri data and the template images. Regards, Vince [1] Q. Lin, J. Liu, Y. Zheng, H. Liang, and V. D. Calhoun, "Semi-blind Spatial ICA of fMRI Using Spatial Constraints," Hum. Brain Map., vol. 31, 2010. ----------------- Thanks in advance On Fri, May 18, 2012 at 9:22 AM, Vince Calhoun <[log in to unmask]> wrote: Hi Syu-Jyun, There are multiple ways to do this. You can just run single-subject ICA and likely you will find one or more DMN components, one will likely have maximal values in posterior cingulate regions, so you can just select one or more of these. You can also take templates from another study (e.g. http://mialab.mrn.org/data) or from an atlas and run either spatiotemporal regression/dual regression or spatially constrained ICA to get a single component out. I prefer the latter since it does a better job of handing the subject specific variability, and it gives you what you want, but any of the above will ‘work’. ;-) Regards, Vince From: SPM (Statistical Parametric Mapping) [mailto:[log in to unmask]] On Behalf Of Peng Syu-Jyun Sent: Friday, May 18, _____________________________________ Vince D. Calhoun, Ph.D. Chief Executive Officer and Director Image Analysis & MR Research The Mind Research Network Professor Depts. of Electrical and Computer Engineering (primary), Neurosciences, Psychiatry, and Computer Science University of New Mexico