[FieldTrip] :Questions about source interpolate and source reconstruction for specific MNI coordinate

[陈骏浩]

Dear Fieldtripers,
 
I tried to conduct source reconstruction and extract time series from specific MNI coordinate on MEG data, but I got some problems. I sorted my questions as follow:
 
(1 In the process of constructing sourcemodel, firstly I conducted ‘ft_volumerealign’ (cfg.coordsys='4d') on subject MRI to define the position of LPA, RPA and nasion according to 4d/bti coordinate system. Then I conducted:
 
cfg = [];
 
cfg.output = ‘brain’;
 
segmentedmri=ft_volumesegment(cfg,mri);
 
I checked the segmentedmri but I saw the brain stem and part of spinal cord were also comprised in it (Fig.1 and Fig.2). I think this result may affect source reconstruction. So, is this result normal? Or did I do something incorrectly?
 
 
 
(2 Next I used template-sourcemodel and subject-MRI to construct sourcemodel in MNI space. After I got subject headmodel and sourcemodel, I conducted ‘ft_prepare_leadfield’ and then ‘ft_sourceanalysis’. When I conducted ‘ft_soureinterpolate’ to interpolate this source on subject MRI and plot it, I saw some activation outside the brain, especially in brain stem and spinal cord. But the most important thing is when I interpolate the source on template-MRI (spm8-T1.nii), the functional image and anatomical image do not match, the sagittal plane of functional image is plotted on the coronal plane of anatomical image, the transverse plane of functional image is plotted on the sagittal plane of anatomical image (Fig. 3). But if i interpolated the source on subject MRI, then use 'ft_volumenormalise', they will be match. I have not conducted ‘ft_volumelookup’ and ‘ft_volumereslice’ in the process, is this the reason? Or did I do wrong step in the process?
 
 
 
(3 I want to extract time series form specific ROI in MNI space according to MNI coordinate. And I prepare to do this by:
 
       norm = ft_volumenormalise([],mri); %normalise subject MRI to MNI space
 
       mnipos = [x y z]; %define ROI position in MNI space
 
       posback=ft_warp_apply(norm.params,mnipos,'sn2individual');
 
 btipos= ft_warp_apply(pinv(norm.initial),posback); % position in individual coordinates
 
Then conducted ‘ft_prepare_leadfield’ again for ROI only, and conducted ‘ft_sourceanalysis’. Can this process work?
 
 
 
Thank you very much for your time and consideration. Looking forward to your reply.
 
 
 
Best regards,
Chen
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20190206/9070448a/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Fig. 1.png
Type: application/octet-stream
Size: 265329 bytes
Desc: not available
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20190206/9070448a/attachment-0003.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Fig. 2.png
Type: application/octet-stream
Size: 22899 bytes
Desc: not available
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20190206/9070448a/attachment-0004.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Fig. 3.png
Type: application/octet-stream
Size: 183732 bytes
Desc: not available
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20190206/9070448a/attachment-0005.obj>