mris_ca_label - For a single subject, produces an annotation file, in which each cortical surface vertex is assigned a neuroanatomical label.

This automatic procedure employs data from a previously-prepared atlas file. An atlas file is created from a training set, capturing region data manually drawn by neuroanatomists combined with statistics on variability correlated to geometric information derived from the cortical model (sulcus and curvature). Besides the atlases provided with FreeSurfer, new ones can be prepared using mris_ca_train).


mris_ca_label [options] <subject> <hemi> <canonsurf> <classifier> <outputfile>


Positional Arguments


-sdir, -orig, -long, -novar, -nbrs, -f, -v, -w, -r, --help, --version


subject id


hemisphere: rh or lh


canonical surface file. Ie: the name of the spherical surface file which describes the registration of a subject's vertices to the reference "average" surface. Example: sphere.reg


specify classifier array input file


annotated surface output file. Ie: an annotation file listing the region label for each vertex. See LabelsClutsAnnotationFiles

Required Flagged Arguments

subject hemi canonsurf classifier outputfile

Optional Flagged Arguments

-sdir <subject dir>

specify a subjects directory (default=$SUBJECTS_DIR)

-orig <filename>

specify filename of original surface (default=smoothwm)


refines the initial labeling read-in from -r (default: disabled)

-r <filename>

file containing precomputed parcellation


sets all covariance matrices to the identify (default: disabled)

-nbrs <number>

neighborhood size (default=2)

-f <number>

applies mode filter <number> times before writing output (default=10)

-t <filename>

specify parcellation table input file (default: none)

-p <filename>

output a file containing label probability for each vertex (default: none)

-v <number>

diagnostic level (default=0)

-w <number> <filename>

writes-out snapshots of gibbs process every <number> iterations to <filename> (default=disabled)


print help info


print version info



labeled cortical surface model


mris_ca_label \
 -orig white \
 -novar \
 -t $SUBJECTS_DIR/scripts/colortable_final.txt \
 my_subject \
 lh \
 sphere.reg \
 $SUBJECTS_DIR/average/lh.rahul.gcs \

In this example, mris_ca_label take sphere.reg as the canonical surface input file, lh.rahul.gcs as the classifier array input file, and writes the annotated surface info to lh.raparc.annot. The file colortable_final.txt is embedded into the output file, so that tksurfer (and other utilities) can read it in.



See Also

mris_sample_parc, mris_ca_train, SurfaceRegAndTemplates


CorticalParcellation, FreeSurfer, FsFast


Automatically Parcellating the Human Cerebral Cortex, Fischl et al., (2004). Cerebral Cortex, 14:11-22.

Reporting Bugs

Report bugs to <>



mris_ca_label (last edited 2013-11-11 12:40:07 by MaritzaEbling)