Parag Kumar Mital

Performing a region of interest analysis on MRI requires knowing where the regions are in your subject data. Typically, this has been done using hand-drawn masks in a 3d viewer. However, recent research has made the process mostly automatic and the open-source community has implemented everything you will need to automatically create labeled volumetric regions of interest [1-3]. With FreeSurfer 5.3, we have the option of performing cortical parcellation using 4 different atlases:

Destrieux atlas: aparc.a2009s
Desikan-Killiany atlas: aparc
Mindboggle: aparc.DKTatlas40
Brodman areas: BA and BA.thresh

We’ll first use freesurfer’s recon-all tool to perform a cortical reconstruction of our anatomical scans. Download freesurfer and register your copy. You’ll be sent an e-mail with a license. Follow the instructions and create the license file “.license” inside your freesurfer home directory (check the environment variable, FREESURFER_HOME, e.g., “$ echo $FREESURFER_HOME"). Then run the script, “$FREESURFER_HOME/FreeSurferEnv.sh” to setup necessary paths.

Next make sure you have set the environment variable for SUBJECTS_DIR to where you’d like your analysis to go (e.g., “$ export SUBJECTS_DIR=/some/directory“). For our example, we’ll keep this to a directory called “freesurfer” in our home directory, “~/”. Each subject we analyze will have its own folder insider SUBJECTS_DIR (i.e., “~/freesurfer”, for our example), with the name specified by recon-all’s “subjid” parameter. We’ll come back to that if it doesn’t make sense yet.

We’re ready to run recon-all now. It can take up to 12 hours on a new machine. Since we have a few subjects, we used our computing cluster to perform the analysis. The cluster we use requires us to submit jobs to a queue using qsub. This bash script file will loop through all our subjects and run recon-all on each of them using a new “job” submitted through qsub. Our subject’s anatomical data is stored in a Nifti file format, “.nii”, in the directory ~/anatomical.

# set the environment variable to where our subject data will be stored
export SUBJECTS_DIR=~/freesurfer

# these are the names of the subjects for which we have high-res scans for
declare -a arr=("15jul13rr" "16jul13ad" "17jul13bs" "18jul13ys")

# we are going to use this temporary directory while processing data
mkdir /global/scratch/pkm

# loop through our subjects
for i in "${arr[@]}"
do
   echo "$i"

   # copy the scan to the temp directory
   datasetlocation=/global/scratch/pkm/"$i"_anat.nii
   cp ../anatomical/"$i"_anat.nii $datasetlocation

   # subject a job with the current subject's file
   qsub -v subject=$i,location=$datasetlocation run_freesurfer.pbs
done

This pbs file is our job script which takes one subject’s mri data and runs recon-all on it.
#!/bin/bash -l

# Name your job (used in the PBS output file names)
#PBS -N $subject

# request the queue (enter the possible names, if omitted, serial is the default)
#PBS -q default

# request 1 node and  request 1 processor per node
#PBS -l nodes=1:ppn=1

# Specify how much time you think the job will run
#PBS -l walltime=36:00:00

# By default, PBS scripts execute in your home directory, not the
# directory from which they were submitted. The following line
# places you in the directory from which the job was submitted.
cd $PBS_O_WORKDIR

# Set the environment variable again as this is on another machine
export SUBJECTS_DIR=~/freesurfer

# Filename for logging
filename="$LOGID"_`date +%Y%m%d-%H%M%S`_freesurfer.log

# Run recon-all
recon-all -i $location -subjid $subject -all &> logs/$filename

If you don’t have a cluster, the line to perform is simply:
recon-all -i YOUR_NIFTI_DATA.nii -subjid SOME_IDENTIFIER -all

where YOUR_NIFTI_DATA may be something like ’15julrr_anat.nii’, and SOME_IDENTIFIER could be ’15julrr’. This identifier is the directory which will be created inside SUBJECTS_HOME.

Next we’ll make use of AFNI’s toolset (free download here). The first is @SUMA_Make_Spec_FS, which will prepare analysis for suma and store the results inside the subject’s surf directory in its own folder, SUMA. This process takes about 15-30 minutes. We then use @SUMA_AlignToExperiment to re-align our data and labels.
Finally, we are ready to extract VOIs. We’ll use whereami to do this. We first find what the name of our region is within our atlas code:
$ whereami -show_atlases
$ whereami -show_atlas_code -atlas ATLAS_CODE

where ATLAS_CODE is an atlas identifier from the output of show_atlases. If we use, “aparc.a2009s+aseg_rank”, then we are using the Destrieux’s Atlas (see here), and we can see the codes for our atlas like so:
$ whereami -show_atlas_code -atlas aparc.a2009s+aseg_rank
++ Input coordinates orientation set by default rules to RAI

Atlas aparc.a2009s+aseg_rank,      194 regions
----------- Begin regions for aparc.a2009s+aseg_rank atlas-----------
l:Left-Cerebellum-White-Matter:4
l:Left-Hippocampus:13
r:Right-choroid-plexus:34
u:ctx_lh_G_cingul-Post-ventral:55
u:ctx_lh_G_rectus:76
u:ctx_lh_S_front_inf:97
u:ctx_lh_S_oc_sup_and_transversal:103
u:ctx_rh_G_and_S_transv_frontopol:124
u:ctx_rh_G_pariet_inf-Supramar:145
u:ctx_rh_S_circular_insula_ant:166
u:ctx_rh_S_precentral-inf-part:187
l:Left-Cerebellum-Cortex:5
l:Left-Amygdala:14
u:5th-Ventricle:35
u:ctx_lh_G_cuneus:56
u:ctx_lh_G_subcallosal:77
u:ctx_lh_S_front_middle:98
u:ctx_lh_S_occipital_ant:104
u:ctx_rh_G_and_S_cingul-Ant:125
u:ctx_rh_G_parietal_sup:146
u:ctx_rh_S_circular_insula_inf:167
u:ctx_rh_S_precentral-sup-part:188
l:Left-Thalamus-Proper:6
u:CSF:15
u:WM-hypointensities:36
u:ctx_lh_G_front_inf-Opercular:57
u:ctx_lh_G_temp_sup-G_T_transv:78
u:ctx_lh_S_front_sup:99
u:ctx_lh_S_oc-temp_lat:105
u:ctx_rh_G_and_S_cingul-Mid-Ant:126
u:ctx_rh_G_postcentral:147
u:ctx_rh_S_circular_insula_sup:168
u:ctx_rh_S_suborbital:189
l:Left-Caudate:7
l:Left-Accumbens-area:16
u:non-WM-hypointensities:37
u:ctx_lh_G_front_inf-Orbital:58
u:ctx_lh_G_temp_sup-Lateral:79
u:ctx_lh_S_oc-temp_med_and_Lingual:106
u:ctx_rh_G_and_S_cingul-Mid-Post:127
u:ctx_rh_G_precentral:148
u:ctx_rh_S_collat_transv_ant:169
u:ctx_rh_S_subparietal:190
l:Left-Putamen:8
l:Left-VentralDC:17
u:Optic-Chiasm:38
u:ctx_lh_G_front_inf-Triangul:59
u:ctx_lh_G_temp_sup-Plan_polar:80
u:ctx_lh_S_orbital_lateral:107
u:ctx_rh_G_cingul-Post-dorsal:128
u:ctx_rh_G_precuneus:149
u:ctx_rh_S_collat_transv_post:170
u:ctx_rh_S_temporal_inf:191
l:Left-Pallidum:9
l:Left-vessel:18
u:CC_Posterior:39
u:ctx_lh_G_front_middle:60
u:ctx_lh_G_temp_sup-Plan_tempo:81
u:ctx_lh_S_orbital_med-olfact:108
u:ctx_rh_G_cingul-Post-ventral:129
u:ctx_rh_G_rectus:150
u:ctx_rh_S_front_inf:171
u:ctx_rh_S_temporal_sup:192
l:Left-choroid-plexus:19
u:CC_Mid_Posterior:40
u:ctx_lh_G_front_sup:61
u:ctx_lh_G_temporal_inf:82
u:ctx_lh_S_orbital-H_Shaped:109
u:ctx_rh_G_cuneus:130
u:ctx_rh_G_subcallosal:151
u:ctx_rh_S_front_middle:172
u:ctx_rh_S_temporal_transverse:193
r:Right-Cerebral-White-Matter:20
u:CC_Central:41
u:ctx_lh_G_Ins_lg_and_S_cent_ins:62
u:ctx_lh_G_temporal_middle:83
u:ctx_lh_S_parieto_occipital:110
u:ctx_rh_G_front_inf-Opercular:131
u:ctx_rh_G_temp_sup-G_T_transv:152
u:ctx_rh_S_front_sup:173
r:Right-Lateral-Ventricle:21
u:CC_Mid_Anterior:42
u:ctx_lh_G_insular_short:63
u:ctx_lh_Lat_Fis-ant-Horizont:84
u:ctx_lh_S_pericallosal:111
u:ctx_rh_G_front_inf-Orbital:132
u:ctx_rh_G_temp_sup-Lateral:153
u:ctx_rh_S_interm_prim-Jensen:174
r:Right-Inf-Lat-Vent:22
u:CC_Anterior:43
u:ctx_lh_G_occipital_middle:64
u:ctx_lh_Lat_Fis-ant-Vertical:85
u:ctx_lh_S_postcentral:112
u:ctx_rh_G_front_inf-Triangul:133
u:ctx_rh_G_temp_sup-Plan_polar:154
u:ctx_rh_S_intrapariet_and_P_trans:175
r:Right-Cerebellum-White-Matter:23
u:ctx-lh-unknown:44
u:ctx_lh_G_occipital_sup:65
u:ctx_lh_Lat_Fis-post:86
u:ctx_lh_S_precentral-inf-part:113
u:ctx_rh_G_front_middle:134
u:ctx_rh_G_temp_sup-Plan_tempo:155
u:ctx_rh_S_oc_middle_and_Lunatus:176
r:Right-Cerebellum-Cortex:24
u:ctx-rh-unknown:45
u:ctx_lh_G_oc-temp_lat-fusifor:66
u:ctx_lh_Pole_occipital:87
u:ctx_lh_S_precentral-sup-part:114
u:ctx_rh_G_front_sup:135
u:ctx_rh_G_temporal_inf:156
u:ctx_rh_S_oc_sup_and_transversal:177
r:Right-Thalamus-Proper:25
u:ctx_lh_G_and_S_frontomargin:46
u:ctx_lh_G_oc-temp_med-Lingual:67
u:ctx_lh_Pole_temporal:88
u:ctx_lh_S_suborbital:115
u:ctx_rh_G_Ins_lg_and_S_cent_ins:136
u:ctx_rh_G_temporal_middle:157
u:ctx_rh_S_occipital_ant:178
r:Right-Caudate:26
u:ctx_lh_G_and_S_occipital_inf:47
u:ctx_lh_G_oc-temp_med-Parahip:68
u:ctx_lh_S_calcarine:89
u:ctx_lh_S_subparietal:116
u:ctx_rh_G_insular_short:137
u:ctx_rh_Lat_Fis-ant-Horizont:158
u:ctx_rh_S_oc-temp_lat:179
r:Right-Putamen:27
u:ctx_lh_G_and_S_paracentral:48
u:ctx_lh_G_orbital:69
u:ctx_lh_S_central:90
u:ctx_lh_S_temporal_inf:117
u:ctx_rh_G_occipital_middle:138
u:ctx_rh_Lat_Fis-ant-Vertical:159
u:ctx_rh_S_oc-temp_med_and_Lingual:180
r:Right-Pallidum:28
u:ctx_lh_G_and_S_subcentral:49
u:ctx_lh_G_pariet_inf-Angular:70
u:ctx_lh_S_cingul-Marginalis:91
u:ctx_lh_S_temporal_sup:118
u:ctx_rh_G_occipital_sup:139
u:ctx_rh_Lat_Fis-post:160
u:ctx_rh_S_orbital_lateral:181
r:Right-Hippocampus:29
u:ctx_lh_G_and_S_transv_frontopol:50
u:ctx_lh_G_pariet_inf-Supramar:71
u:ctx_lh_S_circular_insula_ant:92
u:ctx_lh_S_temporal_transverse:119
u:ctx_rh_G_oc-temp_lat-fusifor:140
u:ctx_rh_Pole_occipital:161
u:ctx_rh_S_orbital_med-olfact:182
u:Unknown:0
r:Right-Amygdala:30
u:ctx_lh_G_and_S_cingul-Ant:51
u:ctx_lh_G_parietal_sup:72
u:ctx_lh_S_circular_insula_inf:93
u:ctx_rh_G_and_S_frontomargin:120
u:ctx_rh_G_oc-temp_med-Lingual:141
u:ctx_rh_Pole_temporal:162
u:ctx_rh_S_orbital-H_Shaped:183
l:Left-Cerebral-White-Matter:1
u:3rd-Ventricle:10
r:Right-Accumbens-area:31
u:ctx_lh_G_and_S_cingul-Mid-Ant:52
u:ctx_lh_G_postcentral:73
u:ctx_lh_S_circular_insula_sup:94
u:ctx_lh_S_interm_prim-Jensen:100
u:ctx_rh_G_and_S_occipital_inf:121
u:ctx_rh_G_oc-temp_med-Parahip:142
u:ctx_rh_S_calcarine:163
u:ctx_rh_S_parieto_occipital:184
l:Left-Lateral-Ventricle:2
u:4th-Ventricle:11
r:Right-VentralDC:32
u:ctx_lh_G_and_S_cingul-Mid-Post:53
u:ctx_lh_G_precentral:74
u:ctx_lh_S_collat_transv_ant:95
u:ctx_lh_S_intrapariet_and_P_trans:101
u:ctx_rh_G_and_S_paracentral:122
u:ctx_rh_G_orbital:143
u:ctx_rh_S_central:164
u:ctx_rh_S_pericallosal:185
l:Left-Inf-Lat-Vent:3
u:Brain-Stem:12
r:Right-vessel:33
u:ctx_lh_G_cingul-Post-dorsal:54
u:ctx_lh_G_precuneus:75
u:ctx_lh_S_collat_transv_post:96
u:ctx_lh_S_oc_middle_and_Lunatus:102
u:ctx_rh_G_and_S_subcentral:123
u:ctx_rh_G_pariet_inf-Angular:144
u:ctx_rh_S_cingul-Marginalis:165
u:ctx_rh_S_postcentral:186
----------- End regions for aparc.a2009s+aseg_rank atlas --------------

To extract one of these regions as a mask, we can use whereami:
$ whereami -atlas aparc.a2009s+aseg_rank_Alnd_Exp
           -mask_atlas_region aparc.a2009s+aseg_rank_Alnd_Exp::ctx_lh_G_temporal_middle
           -prefix lh.mtg.al+orig

More reading:

http://brainybehavior.com/neuroimaging/2010/05/converting-cortical-labels-from-freesurfer-to-volumetric-masks/

http://blog.cogneurostats.com/?p=268

http://blog.cogneurostats.com/?p=217

http://neuroimage.usc.edu/brainstorm/Tutorials/LabelFreeSurfer
References:

[1]. Destrieux, C., Fischl, B., Dale, A., & Halgren, E. (2010). Automatic parcellation of human cortical gyri and sulci using standard anatomical nomenclature. Neuroimage, 53(1), 1–15. doi:10.1016/j.neuroimage.2010.06.010.Automatic
[2]. Fischl et al. (2004). Automatically Parcellating the Human Cerebral Cortex. Cerebral Cortex, 14:11-22.
[3]. Desikan et al. (2006). An automated labeling system for subdividing the human cerebral cortex on MRI scans into gyral based regions of interest. NeuroImage, 31(3):968-80.
Notes:

Many thanks to Beau Sievers and Carolyn Parkinson for their help in this process.