Doug Greve a écrit :

How do you know it is left-right swapped? What tool are you using to visualize it? Do you have a fiducial?

Hello, the surface reconstruction (visualised with scuba) gave a left hemisphere that should be the right one (I am using icbm data that have also been processed by other softwares so I could compare). Now thinking about it I realise that the problem could be other than the file conversion. Does recon_all expect a particular convention (i.e. neurological/radiological) ? Thanks, Olivier

Olivier Coulon wrote:

...

Hi everyone,

...

...

but then, when I do mri_convert -it analyze -i subject -o subject.mgz, the volume is shifted to one side (in addition to left/right swapped). If you look at the coronal sections I'm attaching, you see that a piece of ear the right ear is appearing at the left.

I also have this problem of left/right inversion and I cannot figure out why this happens. Do we have to specify a --out_orientation string ?

Thanks,

Olivier

Thanks for the answer. The problem comes from my original analyze file, not from the conversion or recon_all.

Olivier

Doug Greve a écrit :

It does not require neurological/radiological. It only requres that the geometry in the input file be accurate. Is it? You can test it by viewing it in tkmedit.

doug

Olivier Coulon wrote:

...

Doug Greve a écrit :

...

How do you know it is left-right swapped? What tool are you using to visualize it? Do you have a fiducial?

Hello, the surface reconstruction (visualised with scuba) gave a left hemisphere that should be the right one (I am using icbm data that have also been processed by other softwares so I could compare). Now thinking about it I realise that the problem could be other than the file conversion. Does recon_all expect a particular convention (i.e. neurological/radiological) ? Thanks, Olivier

...

Olivier Coulon wrote:

...

Hi everyone,

...

...

but then, when I do mri_convert -it analyze -i subject -o subject.mgz, the volume is shifted to one side (in addition to left/right swapped). If you look at the coronal sections I'm attaching, you see that a piece of ear the right ear is appearing at the left.

I also have this problem of left/right inversion and I cannot figure out why this happens. Do we have to specify a --out_orientation string ?

Thanks,

Olivier