Dear TORTOISE group,
Thank you for an updated software version. I have used TORTOISE long time ago when it had only the GUI version.
Here is the question. It is written in the documentation that "if your data contains intermediate b values (for ex: 200, 500, 1000) then please provide a gradient text file that is scaled to the maximum bvalue". I have combined two separate datasets of diffusion-weighted images with b=300 and b=1000. In both datasets the gradient vector directions have unit norm. Does one have to normalize the gradient table to maximum bvalue in this case? Or does the scaling relate to multi-shell acquisition with non-unit gradient vectors?
Thank you in advance for answers.
Sorry for the delayed response.
1) We recommend importing the dicom data first without a gradient file (both in cases with or without intermediate bvalues). The imported data should be tensor fit (quick linear fit) and then generate the directionally encoded color (DEC) map/ glyphs maps to see if they resulting maps are correct.
2) If after step 1, the imported data looks incorrect (please follow the checks prescribed under 'check import results' in the software guide), then the gradient information recorded in the dicom header is incorrect and it is being read incorrectly during TORTOISE import routine. If this is the case, you can then provide the gradient text file, that was used in the acquisition, while importing the data. Advice is again re-check the imported data using the tensor fitting routine just to ensure information is being read and saved correctly after import.
3) If you are following step 2 and you have data that have two datasets with b300 and b1000, you can import them separately using their associated gradient text files and then combine the resulting '.list' files using CombineListFiles command from DIFFPREP folder. This results in a combined list file, nii and bmxt file comprising of information from b300 and b1000.
4) If you are following step 2 and you have a folder consisting of multi shell data (all in one), then you will need to provide the gradient text file that is scaled to the square root of the maximum b value.