Check your import results using the following tools. The outputs from running the following executables will be in the _proc folder to be viewed using any visualization software that supports the viewing of RGB format files. You will find the following executable in the ./TORTOISE_V3.1.3/DIFFCALC/DIFFCALCV313/ folder:
Linear fitting should be performed to output a DT.nii
Usage: EstimateTensorWLLS full_path_to_listfile full_path_to_mask_file (optional) dti_bval_cutoff (optional. all volumes will be used)
Non symmetric heuristic maps You can check the non symm heuristic map to determine if the data has been imported correctly by checking the non symm directionally encoded color maps.
Usage: ComputeNSDECMap full_path_to_tensor_image
Tensor glyphs can also be viewed to determine if the import is correct
Usage: ComputeGlyphMaps full_path_to_tensor_NIFTI_file axis (optional) slice_number (optional) full_path_to_output (optional)
Axis: axial, sagittal or coronal. Default axial. slice_number: default center slice
Gyphs can be useful in detecting if the data has been imported correctly. It is sometimes not sufficient to look at just the axial slice as the difference maybe more evident in the coronal and/or sagittal slices. Notice in the example below, the glyphs in the axial slice appears correct in both imports but the difference is seen in the coronal and sagittal slices.
Incorrect (example when flip x and flip y are flipped)
It is essential to check the bmatrix file of the imported data. If it is a new study or if there has been a software upgrade at the scanner, it may affect the header information of the data. During such instances, it is advisable to ensure there has not been any change in the interpretation of the gradient file provided in the acquisition and how this information is stored in the header. The following quality check is advice to be performed periodically on a study, especially if it lasts many years.
- open the imported raw bmtxt file in an excel sheet
- add the elements in the first, fourth and sixth column of the data. The values of the sum should be in equal in order of the bvalues used in your protocol for the study. If you notice discrepancy in the order or the values, it maybe related to the values read from the header. In such instances, it is advised to provide the gradient file and maximum bvalue used in the protocol and re-check the new imported bmatrix. If the problem persists, please consult with TORTOISE team for further suggestions on import.
- Bmatrix check is crucial in blip up and blip down acquisition and to use DRBUDDI correction. Please refer to the following blip up blip down acquisition tips provided by Dr.Joelle Sarrlls, to avoid bmatrix issues between up and down data.