Hi,

I recently downloaded the 2.2.2 version of MedInria and tried to analyze DTI data with
- a single sagittal 4D image containing a series of diffusion weighted images
- a text file containing the list of gradients in the order x, y, z

What are the convention for the x, y, z axis ?
Is it x (Right-Left of the first image), y (Top-Down of the first image), z (First to last of the image) ?

When I load image in Analyze format with ImageJ software, I have remarked the image appears differently comparing with MedInria : it is flipped (Top-Down).

Thanks in advance for your help.

Hi,

First things first, about the orientation of the image. I do not know how ImageJ handles image orientation but medInria is using the orientation matrix information to orient the displayed image in space. That matrix tells how to go from the acquired image space to the real life coordinates.

This takes me to the second part. By default, medinria considers that the axes correspond to the image axes. If your 4D image was acquired in the axial plane, this is probably the axes you are mentioning. Otherwise, it may be different. Another option is not to write the file yourself but to convert the dicom files using, say, dcm2nii and use the bvec file output by the conversion as the gradient file.

Hope this helps, don't hesitate if this is not clear.

Thanks for your quick reply.

I don't understand where you find the matrix information.

For MR machine, the z axis is along the horizontal.
So for axial images, the first image is in plane xy plane (x righr to left).
Do you suppose this convention for the three columns xyz of the gradient file ?
Do the images have to be in axial rather than sagittal ?

Philippe