Closed jcohenadad closed 2 years ago
@jcohenadad I tried to find any persitent information about the specifications for AcquisitionDuration
on https://neurostars.org/search?q=Acquisition%20duration%20tag%3Abids and https://github.com/bids-standard/bids-specification/search?q=acquisitionduration.
@jcohenadad I will start populating AcquisitionDuration
in JSON metadata for the T2star scans.
@jcohenadad I will start populating AcquisitionDuration in JSON metadata for the T2star scans.
thank you. Just make sure you document where you got the information so other people can double check
As per the documentation for AcquisitionDuration
: https://bids-specification.readthedocs.io/en/stable/04-modality-specific-files/01-magnetic-resonance-imaging-data.html#other-recommended-metadata
I ran a script to see if AcquisitionDuration
, RepetitionTime
, VolumeTiming
, SliceTiming
are available in the JSON metadata for *_T2star.json
. I couldn't find AcquisitionDuration
for any of the scans. I will manually check for the data that we have in DCM format.
Thanks @alexfoias ! It would be nice to put your script under scripts/ in this repository as it could be useful for others
The other thing we need to address is that AcquisitionDuration is mutually exclusive with TR, which is a problem because we will be violating the BIDS convention. So we need to address this somehow. I've opened an issue here: https://neurostars.org/t/why-should-acquisitionduration-and-repetitiontime-be-mutually-exclusive/21153
Added the script here: https://github.com/spine-generic/data-multi-subject/pull/112
@alexfoias could you please also display the matrix size (dim x,y), that will be informative.
@jcohenadad sure, I will add this info.
I added AcquisitionMatrixPE
& ReconMatrixPE
from the json sidecar. Is this what you referred as the matrix size (dim x,y) ? Otherwise I can use nibabel to get the info from the T2star.nii.gz.
@jcohenadad I compiled the AcqDuration
manually comparing the protocol PDF with the TR
and AcqMatrix
of each scan.
I have a few issues with:
TR 0.625, AcqMatrix 320
instead of TR 0.6, AcqMatrix 448
(indicated in the protocol). I have looked in the DCM from tokyo that have this matrix and I found 242.5 in the DCM tag 0018,9073. I applied this value to all scans that have similar TR
and AcqMatrix
.Here is a table without VolumeTiming
, SliceTiming
, but with AcquisitionDurationManual
.
Thank you @alexfoias!
GE scans: I cannot find the duration in the protocol or in the DCM. As per their documentation, GE uses DICOM tag: 0019,105A. This tag does not exist in our DCMs.
I found this XML file from a session I did in Tokyo, if you look at the ME-GRE section, it says:
<ses:property name="scanTime" value="288000"/>
Now we just hope that all GE used the same parameters 😅 (which could be verified)
Philips scans: Some have TR 0.625, AcqMatrix 320 instead of TR 0.6, AcqMatrix 448 (indicated in the protocol). I have looked in the DCM from tokyo that have this matrix and I found 242.5 in the DCM tag 0018,9073. I applied this value to all scans that have similar TR and AcqMatrix.
good!
Some subjects don't have AcqMatrix in their json. What should I do in these cases ?
You can assume it is the same as the other sites (per vendor).
I assume that scanTime
is in miliseconds.
Okay. I will open a PR with the changes in AcquisitionDuration.
I assume that scanTime is in miliseconds.
fair assumption 😅
I pushed the changes. I will double check the values tomorrow morning and ping you when ready to review. I appended AcquisitionDuration
at the end of the jsons.
See https://github.com/sct-pipeline/gm-challenge-data/issues/25 for context