|
syam.gadde
BIAC Staff
USA
421 Posts |
 Posted - Apr 21 2005 : 09:28:02 AM
|
bxhreorient has been updated on golgi. This release fixes a bug where the labels for the spatial dimensions (x, y, z) were not reset in the output, causing potential problems for programs that depended on the dimension labels being ordered (x, y, z). Problems due to this bug, if any, are likely to be obvious.
At best, programs may have introduced an unnecessary permutation when reading this data (but the dimensions info, such as orientation, would remain consistent because it would be permuted along with the data). At worst, programs that don't check this info may incorrectly use permuted data.
The problem only occurs if the re-orientation involves more than a simple flipping of axes (i.e. RAS->LAS, LPS->LPI, RAS->LPI are all OK). Additionally, programs that use multiple input datasets are likely to fail in an obvious way if their dimension sizes don't match. An exception is coregistration and normalization, in that SPM doesn't really care that the dimensions of anatomical data are different from functional data or a template. If you co-registered, say, sagittal anatomical data to axial functional data using the standard golgi pre-processing scripts, please review your pre-processing output; in this case, the results should be obviously wrong.
In summary, we believe the bxhreorient bug should not introduce any subtle errors -- if this bug had an effect on your data, it should be obvious. If you have any questions about this issue, please reply to this forum post.
NOTE: the reorient function in the BIAC MATLAB tools is not affected by this bug. |
Edited by - syam.gadde on Apr 21 2005 09:31:08 AM |
|
bxhreorient updated on golgi
