help
help > RE: Naming convention preprocessing COND
Apr 29, 2016 09:04 PM | Alfonso Nieto-Castanon - Boston University
RE: Naming convention preprocessing COND
Hi Jalmar,
The association between condition names (the strings defined in the Setup.Conditions tab) and condition numbers (the numbers in the *_Condition###* files) is logged in the files named _list_conditions.txt in the same folder.
Regarding CONN format of these files (before converting these to nifti), typically CONN datafiles are stored as two files: 1) a .mat file (containing only meta-data); and 2) an associated .matc file (containing the actual data/timeseries). Whenever possible CONN uses "softlinks" (links to the main *Condition000.matc data file) to avoid redundancies (avoid having two data files containing the exact same info). In the most typical case you would have simply one .matc file (*Condition000.matc) containing the actual timeseries data, and then multiple .mat files (*Condition001.mat ... *ConditionN.mat) simply defining different subsets of that file. In your case both Condition001 and Condition003 are of this sort, if I understand correctly, they will have .mat files (metadata) but no .matc files (actual data), since their data is actually stored in the Condition000.matc file. If one of your conditions uses condition-specific filters then the timeseries associated with that condition will not be the same as those stored in the *Condition000.matc file (because of the additional filtering) so in those cases CONN will not use "softlinks" and instead will create both .matc and .mat files associated with that condition (e.g. your *Condition002* files). So my best guess, without actually looking at the _list_conditions.txt file, would be that Condition001 corresponds to your original condition (and the actual data for that is stored in the Condition000.matc file), Condition002 corresponds to your different-band-pass condition (and the actual data for that is stored in the Condition002.matc file), and Condition003 corresponds perhaps to a new (perhaps renamed?) main condition -without filtering- (and the actual data for that is again stored in the Condition000.matc file).
Let me know if this clarifies
Best
Alfonso
Originally posted by Jalmar Teeuw:
The association between condition names (the strings defined in the Setup.Conditions tab) and condition numbers (the numbers in the *_Condition###* files) is logged in the files named _list_conditions.txt in the same folder.
Regarding CONN format of these files (before converting these to nifti), typically CONN datafiles are stored as two files: 1) a .mat file (containing only meta-data); and 2) an associated .matc file (containing the actual data/timeseries). Whenever possible CONN uses "softlinks" (links to the main *Condition000.matc data file) to avoid redundancies (avoid having two data files containing the exact same info). In the most typical case you would have simply one .matc file (*Condition000.matc) containing the actual timeseries data, and then multiple .mat files (*Condition001.mat ... *ConditionN.mat) simply defining different subsets of that file. In your case both Condition001 and Condition003 are of this sort, if I understand correctly, they will have .mat files (metadata) but no .matc files (actual data), since their data is actually stored in the Condition000.matc file. If one of your conditions uses condition-specific filters then the timeseries associated with that condition will not be the same as those stored in the *Condition000.matc file (because of the additional filtering) so in those cases CONN will not use "softlinks" and instead will create both .matc and .mat files associated with that condition (e.g. your *Condition002* files). So my best guess, without actually looking at the _list_conditions.txt file, would be that Condition001 corresponds to your original condition (and the actual data for that is stored in the Condition000.matc file), Condition002 corresponds to your different-band-pass condition (and the actual data for that is stored in the Condition002.matc file), and Condition003 corresponds perhaps to a new (perhaps renamed?) main condition -without filtering- (and the actual data for that is again stored in the Condition000.matc file).
Let me know if this clarifies
Best
Alfonso
Originally posted by Jalmar Teeuw:
Hi Alfonso,
Could you help me out by explaining the naming convention used by CONN to store results for different conditions?
I had previously processed my data with a single condition, this provided me with a DATA_SubjectXXX_Condition001.mat file that had a 'softlink' variable pointing to the DATA_SubjectXX_Condition000.matc file containing the actual data.
I added a second condition to process the data with a different bandpass filter, and now I see several more files appearing that don't make much sense to me.
There are the two DATA_SubjectXXX_Condition001.mat and DATA_SubjectXX_Condition000.matc files as described above. Then there is a DATA_SubjectXXX_Condition002.mat but the 'softlink' variable is empty (i.e. '[]'), and there is no DATA_SubjectXXX_Condition001.matc file. Next, there is a DATA_SubjectXXX_Condition003.mat file, but the 'softlink' variable point to the DATA_SubjectXX_Condition000.matc file (same file as used by DATA_SubjectXXX_Condition001.mat). And finally, there is a DATA_SubjectXXX_Condition002.matc file with actual data, but it is not referred to by any of the other mat files.
CONN does generate a niftiDATA_SubjectXXX_Condition000.nii and niftiDATA_SubjectXXX_Condition002.nii files, and the contents of these files are distinct by MD% hash, so I assume the data for the two conditions are there. Which files belong to which condition?
Kind regards,
Jalmar
Could you help me out by explaining the naming convention used by CONN to store results for different conditions?
I had previously processed my data with a single condition, this provided me with a DATA_SubjectXXX_Condition001.mat file that had a 'softlink' variable pointing to the DATA_SubjectXX_Condition000.matc file containing the actual data.
I added a second condition to process the data with a different bandpass filter, and now I see several more files appearing that don't make much sense to me.
There are the two DATA_SubjectXXX_Condition001.mat and DATA_SubjectXX_Condition000.matc files as described above. Then there is a DATA_SubjectXXX_Condition002.mat but the 'softlink' variable is empty (i.e. '[]'), and there is no DATA_SubjectXXX_Condition001.matc file. Next, there is a DATA_SubjectXXX_Condition003.mat file, but the 'softlink' variable point to the DATA_SubjectXX_Condition000.matc file (same file as used by DATA_SubjectXXX_Condition001.mat). And finally, there is a DATA_SubjectXXX_Condition002.matc file with actual data, but it is not referred to by any of the other mat files.
CONN does generate a niftiDATA_SubjectXXX_Condition000.nii and niftiDATA_SubjectXXX_Condition002.nii files, and the contents of these files are distinct by MD% hash, so I assume the data for the two conditions are there. Which files belong to which condition?
Kind regards,
Jalmar
Threaded View
| Title | Author | Date |
|---|---|---|
| Jalmar Teeuw | Apr 29, 2016 | |
| Alfonso Nieto-Castanon | Apr 29, 2016 | |