files = {‘/data/user/home/khwood/SPSS2_Osprey_processed_MRS/sub-SPSS214931/ses-01/mrs/sub-SPSS214931_ses-01_acq-rtVS_press/meas_MID00084_FID46817_RightVS_svs_se_30_TXRX_TR200.dat’};
files_ref = {‘/data/user/home/khwood/SPSS2_Osprey_processed_MRS/sub-SPSS214931/ses-01/mrs/sub-SPSS214931_ses-01_acq-rtVS_press-ref/meas_MID00085_FID46818_RightVS_svs_se_30_TXRX_NWS_TR2000.dat’};
I’m using Matlab version R2019a
Can you give a later version a shot, please? I am running 2022 here.
Wow, thank you!! Using 2022 I got all the way to Quantify in the GUI! Then received this error:
Do I need to create the “rtVS_stat.csv” file BEFORE running Osprey? I thought this was an output file created by Matlab upon analysis…
I created an empty .csv file, ran the analysis again, and made it to the end of the GUI. However, the rtVS_stat.csv file remained empty…
Just comment out the csv line in the job file - it’s not going to complain. It’s an optional thing in which you can record if your participants belong to certain groups - just helps OspreyOverview create grouped plots. You don’t even need OspreyOverview for further analysis, it’s simply a tool to help with exploring the results.
Good morning @admin!
Thank you for the prompt reply and guidance! Commenting out the csv line worked and I have successfully made it through processing 2 participants’ vmPFC and rtVS data.
However, I am still unable to analyze our ACC data (that was acquired with a slightly different protocol). From what I understand, the protocol is different such that when the data are loaded in a viewer we are looking at a single TR rather than the averaged time course.
I’ve uploaded the DICOMS from the WS and NWS ACC data hoping you will see the parameters you need in the DICOM header, found here.
My current issue: Once I reach the “Model Data” portion of the GUI I receive this error:
Any help or guidance you can provide is greatly appreciated!
–Kimberly
As long as you put all DICOMs for each acquisition in a separate folder, Osprey should automatically find all transients. Seems like it worked for the case you’re showing (you’re clearly seeing all the transients loaded, the frequency trace in the upper right, and the averaged spectrum in the bottom right)
The error is related to specifying the basis set. Do you have any line in your job file that starts with opts.fit.basisSetFile? If your sequence and TE aren’t too exotic (and I believe they aren’t - it’s still PRESS and TE30 right? You never shared the ACC protocol), Osprey should automatically pick the right basis set.
You have massive out-of-voxel artefacts (likely with some kind of lipid contamination as well) in the area between 1 and 2 ppm. This will cause big modeling problems, and you will need to do some work finding the best voxel geometry parameters (this requires a lot of piloting and trial and error). See this thread for some initial guidance (“Avoiding lipid contamination and out-of-voxel echoes”). Do you have a screenshot of your voxel placement? It might already help to move the voxel away from the edge of the brain a little, change the angulation, or flip some of the slice-selective gradients so that you don’t excite the lipids as much.
I did not put any DICOMs for the acquisition in any folders, only the TWIX.dat file.
I do not have any line in my job file that starts with “opts.fit.basisSetFile”. Unfortunately, I cannot seem to get my hands on a copy of the protocol from the console, it’s not in the study for some reason… perhaps because it’s a WIP protocol? Sending the DICOMs for your reference was my technologist’s recommended solution…
Question: If I add the “basisSetFile” command will it resolve my issue with the ACC data analysis?
Unfortunately, all data for the study have already been collected, it looks like troubleshooting the best voxel geometry is not an option…Also, no screenshots were taken at the time of data acquisition. I do, however, have a visualization from Osprey of the voxel placement.
I’m not sure I understand what you are saying - didn’t you say you uploaded DICOM data from the WS and NWS ACC data above?
We’re going to need the protocol (or at least the main settings, i.e. sequence and TE) to figure out where the basis set file error comes from (and also to choose the right basis set which would then, yes, be set in the basisSetFile option).
Right, that’s not good - if that kind of artefact is present in all spectra, the analysis will be a little hairy, I’d suggest limiting the fit range to begin at 1.6 ppm. Certainly not optimal.
My apologies! I completely misunderstood your #1. Yes, the DICOM data for the WS and NWS ACC data are in separate folders in the Google Drive I shared.
My hope is you are able to see the main settings (sequence, TE) in the DICOM header I shared for the ACC data. I’ve just learned I need to use our basis set for all analyses. Copy in the Google Drive I shared (hoping this will also provide parameter insight?)
Thank you for the guidance on setting the fit range to begin at 1.6 ppm! Agree, it is not optimal…
This is STEAM data (TE = 6.5 ms), so you definitely cannot use a TE = 30 ms PRESS basis set to fit them. It’s an entirely different pulse sequence.
I’ll have to see if I can dig up a basis set from anywhere. It definitely explains the error (Osprey cannot find an appropriate basis set in its own library).
I added a steam basis set to the Google drive I shared. Can you tell me if it is the appropriate file to use for the ACC data?
Many thanks!!
–Kimberly
Yeah, that one looks right (TE/TM = 6.5/10 ms). I’d give it a pop (use the 'LCModel' option to fit). Where did you get that from?
Wonderful!! Thank you greatly for this feedback and guidance. I tried it and it worked! The PRESS basis set came from LC Model…
And the STEAM basis set?
Glad it works, we just need to make sure that the basis set really matches the sequence as closely as possible. All the ones on the LCModel website are from an unknown source (we don’t even know the manufacturer).
The STEAM basis set was created by a developer team and I did not have permission to share. Please remove it from your database and computer. My apologies for any inconvenience or confusion.
Many thanks!
–Kimberly
