 |
 |
Hi, I am currently using FSL FDT diffusion to process my diffusion data. In order to obtain the diffusion gradients directions bvecs, I used dcm2nii (mricron). However I have seen that the gradients directions in the bvecs file do not correspond to the gradients directions I have using directly the MATLAB command 'dicominfo' on my dicoms and then looking at the dicom attribute 'DiffusionGradientOrientation'. I think that the components of the diffusion gradients directions in the dicom files (DiffusionGradientOrientation attribute) is expressed in the patient frame ( the xaxis is increasing toward the Left hemisphere, the yaxis is increasing toward the posterior side and the z-axis is increasing toward the head of the patient) whereas the diffusion gradient directions from dcm2nii are in the image frame. Consequently, I have used the values of the direction cosines of the dicom ImageOrientationPatient attribute and created a matrix M in order to express the dicom diffusion gradient directions in the image frame (rather than in the patient frame). M = a1 b1 c1 a2 b2 c2 a3 b3 c3 where a1, a2 and a3 corresponds respectively to the 1st, 2nd and 3rd element of the ImageOrientationPatient attribute b1, b2 and b3 corresponds respectively to the 4th, 5th and 6th element of the ImageOrientationPatient attribute and c1 c2 and c3 are obtained using the vector cross product of ( a1 a2 a3) with (b1 b2 b3). then I deduced the relation between the gradient directions in the patient frame and the gradient directions in the image frame: gradient direction in the image frame = inverse(M)*gradient direction in the patient frame The gradients directions in the image frame I finally got are the same than the gradients given by dcm2nii HOWEVER I still have a sign problem. Indeed the sign of the first and of the 3rd component of the gradients directions in the image frame I have calculated do not have the same sign than the gradients directions I have got using dcm2nii. Do you know why? and what is the method used by dcm2nii to obtain the diffusion gradients directions from the dicom files? Thanks for your help Best regards L.Vanquin