Multi-session Data¶

Multi-session subject data (i.e., a data set with multiple runs) can be processed by CBPtools in full when sessions are defined in the configuration file. Data for each session will be processed separately until the connectivity step, after which the connectivity matrices for each subject will be averaged across all the sessions. When multi-session data is being used, the (optional) PCA transformation will be performed after the sessions have been averaged, whereas the other transformations (Fisher’s Z transform and cubic transform) will be applied before averaging. The averaged connectivity matrices are then used for the remainder of the procedure.

Multi-session data can be defined in the configuration file as follows:

modality: rsfmri

data:
    session: [sess1, sess2, sess3]
    time_series: /path/to/data_set/{participant_id}/{session}/time_series.nii.gz
    confounds:
        columns: ['constant', 'wm.linear', 'csf.linear', 'motion-*']
        delimiter: \t
        file: /path/to/data_set/{participant_id}/{session}/confounds.tsv
    masks:
        seed: /path/to/seed_mask.nii.gz
        target: /path/to/target_mask.nii.gz
        space: standard

...

Note how the {session} wildcard is now used to reference the data set files. Assuming two subjects with participant IDs ‘sub-001’ and ‘sub-002’, and the sessions defined as in the code-block above, the time_series files will become:

/path/to/data_set/sub-001/sess1/time_series.nii.gz
/path/to/data_set/sub-001/sess2/time_series.nii.gz
/path/to/data_set/sub-001/sess3/time_series.nii.gz
/path/to/data_set/sub-002/sess1/time_series.nii.gz
/path/to/data_set/sub-002/sess2/time_series.nii.gz
/path/to/data_set/sub-002/sess3/time_series.nii.gz

All wildcards can be used multiple times and don’t have to indicate a folder. As long as the structure of the data set is consistent, there should be no problem. For example, the time-series can also be referenced as:

time_series: /path/to/data_set/xyz-{participant_id}a/session_{session}/sw{participant_id}_{session}.nii.gz

For participant ‘sub-001’ and session sess1 this will become: /path/to/data_set/xyz-sub-001a/session_sess1/swsub-001_sess1.nii.gz

Note

This example is by no means an endorsement for storing your data in such a way.