How to avoid dicom conversion in CMTK

[Kaveh Kohan]

Hi,

I am new to CMTK. I have two questions and I would be thankful if you answer:

- I have rsFMRI and T1 data already converted to nii.gz, how can I bypass the docom conversion and the folder structure should be?

- In the CMTK website, you mentioned that at the end it is possible to get average time-course for each ROI. Is it possible to have time-courses sampled from cortex? I mean for each vertex on the cortex we have one time-series?


Thanks

[Qingyang Li]

Hi Kaveh,

Your first question: see here

Your second question. I don't think CMP does this for you. But you can take the preprocessed refMRI data and register it a surface space... this can be done by freesurfer cmd line.

Yag

------
Qingyang Li

Balance and focus.

[Kaveh Kohan]

Hi,

Thank you for you reply. I followed the instruction and it worked up
to the point that it wanted to register T1 to rs-FMRI. I need to
mention that I don't have DTI (DSI) data (just T1 and rs-fmri). It
seems that T1-2-fmri is failing, I should not need DTI for this step
should I? From the error message it seems that it wants to compute the
correlation.
Here is the list of steps I requested for:

Create Folder
DICOM Converter
Segmentation
Registration
Parcellation
Apply registration
Resting-state fMRI
Skip previously completed stages

Here is the error message I got:

INFO : 2012-04-01 13:22:46,492 : main : Stage
'cmp.stages.rsfmri.correlation' missing input, file not found: /home/
kayhan/test/myproject/control001/tp1/CMP/fs_output/HR__registered-TO-
b0/freesurferaparc/ROIv_HR_th.nii.gz
ERROR : 2012-04-01 13:22:46,492 : main : Required input file
missing for stage: 'cmp.stages.rsfmri.correlation'. Please run all
previous stages.
ERROR : 2012-04-01 13:22:46,492 : main : Exception occured:
Required input file missing for stage:
'cmp.stages.rsfmri.correlation'. Please run all previous stages.
ERROR : 2012-04-01 13:22:46,492 : main : Traceback (most
recent call last):
File "/usr/local/lib/python2.6/dist-packages/cmp/gui.py", line 66,
in object_run_changed
cmp.connectome.mapit(object)
File "/usr/local/lib/python2.6/dist-packages/cmp/connectome.py",
line 66, in mapit
raise Exception( msg )
Exception: Required input file missing for stage:
'cmp.stages.rsfmri.correlation'. Please run all previous stages.

On Mar 31, 8:22 pm, Qingyang Li <con...@gmail.com> wrote:
> Hi Kaveh,
>
> Your first question: see

> here<http://groups.google.com/group/cmtk-users/browse_thread/thread/4b9cda...>

[Qingyang Li]

OK, let's take a look at this line of error msg:

INFO     : 2012-04-01 13:22:46,492 : main       : Stage
'cmp.stages.rsfmri.correlation' missing input, file not found: /home/kayhan/test/myproject/control001/tp1/CMP/fs_output/HR__registered-TO-b0/freesurferaparc/ROIv_HR_th.nii.gz

It seems that CMP requires the parcellation mask ( which should be corregistered with the DTI (b0) image) for the connectivity analysis of rs-fMRI data. 

As fa as I can tell, you will have to have DTI/DWI data for rs-fMRI data analysis if you use CMP. 

Yang

------
Qingyang Li

Balance and focus.

[Kaveh Kohan]

Ok, if I can see how CMP runs the registration command (bbregister) on
a dataset with DTI (eg the one you provided CMP website), I can figure
it out. Is there any log file keeping track of commands called by CMP
(eg how it calls recon-all, how it call bregister, etc) ?

On Apr 1, 3:10 pm, Qingyang Li <con...@gmail.com> wrote:
> OK, let's take a look at this line of error msg:
>
> INFO     : 2012-04-01 13:22:46,492 : main       : Stage
> 'cmp.stages.rsfmri.correlation' missing input, file not found: /home/

> kayhan/test/myproject/control001/tp1/CMP/fs_output*/**HR__registered-TO-
> b0/freesurferaparc/ROIv_HR_th.nii.gz**
> *

[Qingyang Li]

yes, if you go to Project folder > subject > tp1 > LOG

There is a '.log' file. Everything is there...

------
Qingyang Li

Balance and focus.

[Alessandra Griffa]

Hi Kaveh,

despite the CMP was originally thought to process diffusion data, it does not require the DWI images to process rs-fMRI data.
The error you got was caused by a bug. You can download the fix from github, or if you prefer replace 'conf.get_cmp_tracto_mask_tob0()' with 'conf.get_cmp_tracto_mask()' in cmp/stages/rsfmri/correlation.py, line 236.

For the moment the CMP does not output the fMRI signal for each GM point, but the average fMRI signals for GM parcels.
Anyway if you choose the parcellation scheme 'lausanne2008' you can get the fMRI average signals from small, isosurface (~1.5 cm2) GM parcels.


Hope this could help you.
Regards,
Alessandra