questions > Visualization trouble .nii with Mricron
Showing 1-4 of 4 posts
Display:
Results per page:
Feb 24, 2020  04:02 AM | Luca Cuffaro - University of East London, NeuroRehabiliation Unit
Visualization trouble .nii with Mricron
Dear all,

as you can see in the attached files, I have this problem with the visualization of the MRI images, so I cannot go on.
All images are from stroke patients (MRI machine is a Siemens Avanto 1.5T) and they are stored in .dcom format all together in a single folder, so I use MicroDicom or MRIconvert to perform a selection of those I want to convert.
I have tried with also T1w or adding T2 or other series, but the mail problem is that I don't have T1w for all the three plans. I can use T1w Axial and Sagittal, but I have to use a T2 Coronal, because it is the only coronal I have.
I have used SPM12, MicroGL and MRIcron with dcm2niix, but I achieve the same results: several .nii files ("TEST 1__20170319151600_2.nii", "TEST 1__20170319151600_3.nii", "TEST 1__20170319151600_4.nii", etc) and none does't lead to a correct visualization of all three plans of images in a unique file. I have read several papers and istructions from Rorden and Friston, I have checked the manual of SPM12, a lot about spatial normalization, but I blocked at the beginning. 
I think I am doing something wrong, for sure because I am a neurologist, i don't have a profound expertise in neuroengineering knowledge and coding, but I am trying to understand.

Thanks in advance for your time and for your suggestions!
Luca
(also manual, book, paper, everything could help is sincerely welcome)
Attachment: 1..jpg
Feb 24, 2020  05:02 AM | Chris Rorden
RE: Visualization trouble .nii with Mricron
It sounds like images were acquired with anisotropic resolution. To confirm, choose Window/Information and look at the "Spacing" column for the Spatial dimensions, I suspect you will see that the image spacing is of higher resolution in plane (I Space, J Space) than the between slice resolution (K Space). If this is the case, this reflect properties of your image acquisition. Rather than acquire three orthogonal series (Coronal, Sagittal, Axial) with thick slices, I would suggest acquiring a single 3D image with isotropic resolution. I suspect this will give you similar SNR with similar or lower can time.

If this is the case, you might want to choose the option "Reorient images when loading" which will interpolate your data to isotropic. Alternatively, if you have a modern computer, I would suggest using MRIcroGL instead of MRIcron, which will stretch dimensions by default. Whichever method you use, images will exhibit poor resolution across slices, reflecting the inherent properties of your acquisition.
Feb 24, 2020  07:02 AM | Luca Cuffaro - University of East London, NeuroRehabiliation Unit
RE: Visualization trouble .nii with Mricron
Dear Chris,

many thanks for your reply and your time.

I have checked as you suggested and there is a discrepancy between plane and slice resolution in different combinations all among image series that I converyed to .nii. So, do you suggest to "convert" a single 3D image of those acquired (for example? just a good axial T1?), or was your suggestion to modify acquisition proprierties for next patients? 

Furthermore, I have look for  "Reorient images when loading", to interpolate to isotropic, but you can see attached what option I have.

I have already asked to the servicedesk to install MRIcroGL. I will try as soon as possible, I have to wait for university's technicians to install something on my pc. 

As conclusion, is your opinion that I have to to accept that these images will show anyway a poor resolution. Am I right? 

Many thanks, again
Attachment: 6.jpg
Feb 24, 2020  08:02 AM | Chris Rorden
RE: Visualization trouble .nii with Mricron
1.) The "Reorient Images when loading" checkbox is in your Preferences. For the Windows operating system you choose Help/Preferences to change this setting.
 2.) I would recommend using a 3D sequence for future acquisitions, e.g. for Siemens this would be called "MP-RAGE". The 3D sequence can be set up to acquire high resolution isotropic voxels in a short scan time.
Attachment: preferences.png