Hi,

Only the second option will be forming a binary mask for fslmeants, as when flirt performs a transformation it will use trilinear interpolation and hence the mask will no longer be binary.

All the best,

Mark

On 7 Nov 2016, at 14:17, Ramesh Babu <[log in to unmask]> wrote:

Dear Mark

Thank you very much for your suggestions. I have included -init highres2example_func.mat command in flirt call. I got time series data by following two steps given below. Both the steps are giving different time series data. Please see the attached file. I would like to confirm which step should I follow? step-1 or 2?

Step-1 [fslmaths (erode-threshold-binarize) > transform > extract time series]

fslmaths N5_T1_brain_pve_0.nii.gz -ero -thr .9 -bin csf_m -odt float

flirt -in csf_m.nii.gz -ref filtered_func_data.nii.gz -out csf_m_fl -applyxfm -init highres2example_func.mat

fslmeants -i filtered_func_data.nii.gz -o csf_m.txt -m csf_m_fl.nii.gz

Step-2 [transform > fslmaths (erode-threshold-binarize) >  extract time series]

flirt -in N5_T1_brain_pve_0.nii.gz -ref filtered_func_data.nii.gz -out csf_fl -applyxfm -init highres2example_func.mat

fslmaths csf_fl.nii.gz -ero -thr .9 -bin csf_m_fl_m -odt float

fslmeants -i filtered_func_data.nii.gz -o csf_ts.txt -m csf_fl_m.nii.gz

Thanks

Ramesh

On Mon, Nov 7, 2016 at 4:18 PM, Mark Jenkinson <[log in to unmask]> wrote:

Hi,

This looks generally fine to me, except that I think you've left off "-init highres2example_func.mat" in the flirt call (as otherwise the output, csf_m_fl, won't be registered to your filtered_func_data).  Did you look at that?

I would also threshold and binarise the output from flirt before using it as a mask in fslmeants.

All the best,

Mark

On 5 Nov 2016, at 02:02, Ramesh Babu <[log in to unmask]> wrote:

Dear Mark,

I got the time series data by following the steps given below.

fslmaths N5_T1_brain_pve_0.nii.gz -ero -thr .9 -bin csf_m -odt float
flirt -in csf_m.nii.gz -ref filtered_func_data.nii.gz -out csf_m_fl -applyxfm
fslmeants -i filtered_func_data.nii.gz -o csf_m.txt -m csf_m_fl.nii.gz

Please see the attached file and give your suggestion.

Thanks
Ramesh

On Fri, Nov 4, 2016 at 8:35 PM, Ramesh Babu <[log in to unmask]> wrote:

Dear Mark,

Thank you for your prompt reply.

I am going to use it into regression model. What are the steps should I follow to extract time series?

Thanks

Ramesh

On Fri, Nov 4, 2016 at 3:34 AM, Mark Jenkinson <[log in to unmask]> wrote:

Hi,

The answer to this depends on what you are going to do with the extracted time series.

If you are going to enter it into a regression then you should extract it from the same data that is going into the regression (so after whatever preprocessing is done to that data).

If you are using it for some other purpose (e.g., as a QC measure for your data) then you might need to extract it from something else like the raw data.  It all depends on your application.

All the best,

Mark

On 3 Nov 2016, at 09:26, Ramesh Babu <[log in to unmask]> wrote:

Dear Expert,

By using FAST segmentation tool I obtained segmented GM, WM and CSF. I want extract time series of WM, and CSF by using fsl meants. 

I would like to know Is there any steps before extracting time series from these segmented images?

To extract time series should I use raw BOLD sequence or filtered..image produced by FEAT preprocessing steps?

Please provide the steps to extract time series from WM and CSF.

Thanks in advance

Ramesh

<csf_m.txt><csf_mask.png>

<step-1_csf.txt><step-2_csf.txt>