Christain:
For those codes you sent me for reshaping the IC data in to a 2D matrix, (#ICs x #voxels), I have some questions. First show your code below:
data = read_avw('test');
mask=read_avw('mask');
mask=reshape(mask,prod(size(mask)),1)' > 0;
data=reshape(data,prod(size(mask)),prod(size(data))/prod(size(mask)))';
data=data(:,mask);
My understanding is data is just a 4D matrix read from the IC results. But what is the mask and 'mask'? mask=read_avw('mask');, this seems like the 'mask' has already been somewhere in fsl folder. Or, if not, do I have to arbitrarily design a mask? Also, mask is read from 'mask' by read_avw, so it seems that mask is a 4D matrix as well, but why would we reshape the mask into a column vector in mask=reshape(mask,prod(size(mask)),1)' > 0? And in the next line: data=reshape(data,prod(size(mask)),prod(size(data))/prod(size(mask)))';, how can we make sure that prod(size(data))/prod(size(mask)) is an integer? Plus, I dont see any convolution in those steps, and I am just confused on the term of "mask".
Could you explain more about those codes?
Thanks
Xuelin
****************************************
Xuelin Cui
Department of Electrical Engineering
University of Hawaii-Manoa
Honolulu HI 96822
Tel: 1-808-349-0983
Email: [log in to unmask]
****************************************
----- Original Message -----
From: Christian Beckmann <[log in to unmask]>
Date: Wednesday, July 4, 2007 4:20 am
Subject: Re: [FSL] How to find the exact numeric value of those spatial IC maps in MELODIC?
To: [log in to unmask]
> Hi
>
> These scripts are wrapper functions, you will need a complete
> _working_ install of FSL on each computer you want to run these
> scripts on.
> hope this helps
> christian
>
>
>
>
> On 29 Jun 2007, at 01:41, Xuelin Cui wrote:
>
> > hi Christain:
> >
> > More over. I want to run my nifti data on a Mac computer, but do
>
> > the ICA data analysis on another Windows computer. Those matlab
> > scripts in $FSLDIR/etc/matlab I have already moved them in the
> > correct path on the Windows computer.
> >
> > But when I ran the read_avw( ) function, there was mistakes. I
> > check the source code, there is a line in read_avw file is:
> > system(['sh -c "FSLOUTPUTTYPE=NIFTI_PAIR; export FSLOUTPUTTYPE;
> > $FSLDIR/bin/avwmaths ', fname, ' ', tmpname,'"']);
> > and I think the matlab scripts is default for the case that both
>
> > FSL and these scripts are on the same computer. Plus, I saw
> there
> > is a file named "melodic_ICA.nii.gz" in my results folder. I am
> not
> > sure if we still need the command line above to generate another
>
> > nifiti file?
> >
> > My real question is: How can modify these script so that I can
> do
> > my ICA data analysis on a different computer? Can I generate
> nifti
> > format file of ICA results on my Mac(if needed), and doin the
> ICA
> > data analysis using my Windows?
> >
> > Thanks
> >
> > XC
> >
> > ****************************************
> > Xuelin Cui
> > Department of Electrical Engineering
> > University of Hawaii-Manoa
> > Honolulu HI 96822
> >
> > Tel: 1-808-349-0983
> > Email: [log in to unmask]
> > ****************************************
> >
> > ----- Original Message -----
> > From: Christian Beckmann <[log in to unmask]>
> > Date: Thursday, June 28, 2007 3:14 pm
> > Subject: Re: [FSL] How to find the exact numeric value of those
> > spatial IC maps in MELODIC?
> > To: [log in to unmask]
> >
> >> Hi,
> >>
> >> simply use read_avw (in $FSLDIR/etc/matlab) to load the entire 4d
> >> file into matlab. You can then check each voxels value in each map
> >>
> >> easily:
> >>
> >> data=read_avw('blah.ica/melodic_IC');
> >> data(20,20,10,3)
> >>
> >> ans = some value
> >>
> >> This then is the value of voxel 20,20,10 in IC map 3. If you want
> >> to
> >> mask out the background voxels then you can do the following
> >>
> >> data = read_avw('test');
> >> mask=read_avw('mask');
> >> mask=reshape(mask,prod(size(mask)),1)' > 0;
> >> data=reshape(data,prod(size(mask)),prod(size(data))/prod(size
> >> (mask)))';
> >> data=data(:,mask);
> >>
> >> data then is a 2d matrix (#ICs x #voxels). If you only want to
> >> find
> >> out about a small number of values you can simply load melodic_IC
> >> to
> >> fslview and click on the voxels of interest.
> >> hope this helps
> >> christian
> >>
> >>
> >> On 28 Jun 2007, at 01:45, Xuelin Cui wrote:
> >>
> >>> Dear all:
> >>>
> >>> As we know, ICA decomposes the fMRI data in to a time course
> >> matrix
> >>> and a spatial map matrix. Those time course values are easy to
> >> see
> >>> if we simply click the time course picture on the html file.
> >> Also,
> >>> we can see *.txt files, which is the time course data, in the
> >> folder.>
> >>> I just wonder are there a similar txt files for spatial map?
> >> What
> >>> if I am interested in the exact intensity of each spatial map? I
> >> am
> >>> think of using Matlab to read those spatial map, but the problem
> >> is
> >>> those IC maps are put together, and there are lots of marginal
> >> area
> >>> on the picture that is not needed. So, is there a way to find
> >> out
> >>> the exact value of each independent spatial maps?
> >>>
> >>> Thanks a lot
> >>>
> >>> Xuelin
> >>>
> >>> ****************************************
> >>> Xuelin Cui
> >>> Department of Electrical Engineering
> >>> University of Hawaii-Manoa
> >>> Honolulu HI 96822
> >>>
> >>> Tel: 1-808-349-0983
> >>> Email: [log in to unmask]
> >>> ****************************************
> >>
> >> ____
> >> Christian F. Beckmann
> >> University Research Lecturer
> >> Oxford University Centre for Functional MRI of the Brain (FMRIB)
> >> John Radcliffe Hospital, Headington, Oxford OX3 9DU, UK.
> >> [log in to unmask] http://www.fmrib.ox.ac.uk/~beckmann
> >> tel: +44 1865 222551 fax: +44 1865 222717
> >>
>
> ____
> Christian F. Beckmann
> University Research Lecturer
> Oxford University Centre for Functional MRI of the Brain (FMRIB)
> John Radcliffe Hospital, Headington, Oxford OX3 9DU, UK.
> [log in to unmask] http://www.fmrib.ox.ac.uk/~beckmann
> tel: +44 1865 222551 fax: +44 1865 222717
>
|