Logistic Random Effects Based Parcellation

Groupwise Structural Parcellation of the Whole Cortex: A Logistic Random Effects Model Based Approach

Installation

• clone repository
• python setup.py install

Command Line Tools

Please read the . More examples coming soon.

Averaging CIFTI connectivity files

The cifti_average tool computes the average connectivity from a group of connectivity matrices. If the in_logodds flag is present, then the matrices are transformed to the LogOdds space, averaged, and transformed back.

cifti_average -matrices CIFTI_FILE_1 CIFTI_FILE_2 ... -out CIFTI_OUT -in_logodds

Parcelling a CIFTI connectivity file

The cifti_parcellate tool parcellates a CIFTI connectivity file in a given DIRECTION. If the flag transform is present, the data is transformed to the LogOdds space as explained in Gallardo et al. (2017). If a constraint and a minimum_size are given, then the clustering is performed only between neighbors, until the min_size is reached.

cifti_parcellate -cifti CIFTI_FILE [-direction ROW/COLUMN] [-transform] [-constraints SURF/VOLUME-FILE] [-min_size N] -out DENDROGRAM_FILE.csv

Extracting a parcellation from the dendrogram

The extract_parcellation retrieves a parcellation with a defined number of parcels from a dendrogram. The output file can be both a CIFTI LABEL file (dlabel.nii) or a TXT file.

extract_parcellation -dendrogram DENDROGRAM_FILE.csv -parcels nparcels -out OUT_FILE.

Reference

Guillermo Gallardo, William Wells III, Rachid Deriche, Demian Wassermann, Groupwise structural parcellation of the whole cortex: a logistic random effects model based approach. 2017, in press. NeuroImage.