{"id":761,"date":"2026-10-20T14:29:26","date_gmt":"2026-10-20T14:29:26","guid":{"rendered":"https:\/\/ni.cmu.edu\/computing\/?post_type=ht_kb&p=761"},"modified":"2026-10-20T14:40:23","modified_gmt":"2026-10-20T14:40:23","slug":"shared-neuroimaging-resources-guide-templates-atlases","status":"publish","type":"ht_kb","link":"https:\/\/ni.cmu.edu\/computing\/knowledge-base\/shared-neuroimaging-resources-guide-templates-atlases\/","title":{"rendered":"Shared Neuroimaging Resources Guide: Templates & Atlases"},"content":{"rendered":"

This guide provides instructions for accessing shared neuroimaging reference templates and atlases stored on the cluster at \/lab_data\/hawk\/resources\/neuroimaging\/<\/code>.<\/p>\n

1. Available Resources & Directory Paths<\/h2>\n\n\n\n\n\n\n\n
Resource<\/strong><\/td>\nFile Path \/ Location<\/strong><\/td>\nDescription<\/strong><\/td>\n<\/tr>\n<\/thead>\n
TemplateFlow Cache<\/b><\/span><\/td>\n\/lab_data\/hawk\/resources\/neuroimaging\/templateflow<\/code><\/span><\/td>\nCentral offline cache containing MNI152NLin2009cAsym<\/code> and standard templates.<\/span><\/td>\n<\/tr>\n
Glasser Atlas (Volumetric)<\/b><\/span><\/td>\n\/lab_data\/hawk\/resources\/neuroimaging\/atlases\/glasser\/<\/code><\/span><\/td>\nHCP-MMP1.0 converted to 3D NIfTI volume in MNI space (.nii.gz<\/code> + labels).<\/span><\/td>\n<\/tr>\n
Glasser Atlas (Surface)<\/b><\/span><\/td>\n\/lab_data\/hawk\/resources\/neuroimaging\/atlases\/glasser\/HCP-MMP1_32k_fs_LR.dlabel.nii<\/code><\/span><\/td>\nNative CIFTI surface parcellation for fsLR_32k<\/code> mesh analysis.<\/span><\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

2. Environment Setup<\/h2>\n

You should route processing tools (such as fMRIPrep<\/b> or QSIPrep<\/b>) to use the shared TemplateFlow directory.<\/p>\n

Add these environment variables to your ~\/.bashrc<\/code> file or load them inside your Slurm submission scripts:<\/p>\n

\n
\n
\n
\n

Bash<\/span><\/p>\n

<\/div>\n<\/div>\n
# Point local Python pipelines to the shared TemplateFlow directory<\/span>\r\nexport<\/span> TEMPLATEFLOW_HOME=\/lab_data\/hawk\/resources\/neuroimaging\/templateflow\r\n\r\n# Bind environment variable for Singularity\/Apptainer containers<\/span>\r\nexport<\/span> APPTAINERENV_TEMPLATEFLOW_HOME=\/opt\/templateflow\r\nexport<\/span> SINGULARITYENV_TEMPLATEFLOW_HOME=\/opt\/templateflow\r\n<\/code><\/pre>\n<\/div>\n<\/div>\n<\/div>\n
3. Preprocessing with fMRIPrep (Slurm Submission)<\/h2>\n
When running containerized pipelines via Apptainer\/Singularity, bind the shared template directory into \/opt\/templateflow<\/code> inside the container.<\/p>\n
\n
\n
\n
\n
Bash<\/span><\/p>\n
<\/div>\n<\/div>\n
#!\/bin\/bash<\/span>\r\n#SBATCH --job-name=fmriprep_job<\/span>\r\n#SBATCH --nodes=1<\/span>\r\n#SBATCH --cpus-per-task=8<\/span>\r\n#SBATCH --mem=32G<\/span>\r\n#SBATCH --time=12:00:00<\/span>\r\n#SBATCH --output=logs\/fmriprep_%j.log<\/span>\r\n\r\n# 1. Load required modules (adjust according to system module names)<\/span>\r\nmodule load apptainer\r\n\r\n# 2. Define environment variables<\/span>\r\nexport<\/span> TEMPLATEFLOW_HOME=\/lab_data\/hawk\/resources\/neuroimaging\/templateflow\r\nexport<\/span> APPTAINERENV_TEMPLATEFLOW_HOME=\/opt\/templateflow\r\n\r\n# 3. Execute fMRIPrep using shared templates<\/span>\r\napptainer run --cleanenv \\\r\n  -B \/lab_data\/hawk:\/lab_data\/hawk \\\r\n  -B \/lab_data\/hawk\/resources\/neuroimaging\/templateflow:\/opt\/templateflow \\\r\n  \/lab_data\/hawk\/software\/fmriprep.sif \\\r\n  \/lab_data\/hawk\/data\/bids \\\r\n  \/lab_data\/hawk\/data\/derivatives \\\r\n  participant \\\r\n  --participant-label 01 \\\r\n  --output-spaces MNI152NLin2009cAsym:res-2 \\\r\n  --notrack\r\n<\/code><\/pre>\n<\/div>\n<\/div>\n<\/div>\n
4. Extracting Brain Signals with the Glasser Atlas<\/h2>\n
Option A: 3D Volumetric NIfTI Signal Extraction (Python & Nilearn)<\/h3>\n
Use Nilearn<\/b> to regress out confounds (motion, white matter, CSF) and extract timeseries across the 360 Glasser cortical regions:<\/p>\n
\n
\n
\n
\n
Python<\/span><\/p>\n
<\/div>\n<\/div>\n
import<\/span> numpy as<\/span> np\r\nimport<\/span> pandas as<\/span> pd\r\nfrom<\/span> nilearn.maskers import<\/span> NiftiLabelsMasker\r\nfrom<\/span> nilearn.connectome import<\/span> ConnectivityMeasure\r\n\r\n# Paths<\/span>\r\nfunc_file = \"\/lab_data\/hawk\/data\/derivatives\/fmriprep\/sub-01\/func\/sub-01_task-rest_space-MNI152NLin2009cAsym_res-2_desc-preproc_bold.nii.gz\"<\/span>\r\nconfounds_file = \"\/lab_data\/hawk\/data\/derivatives\/fmriprep\/sub-01\/func\/sub-01_task-rest_desc-confounds_timeseries.tsv\"<\/span>\r\nglasser_atlas = \"\/lab_data\/hawk\/resources\/neuroimaging\/atlases\/glasser\/HCP-MMP1_on_MNI152NLin2009cAsym.nii.gz\"<\/span>\r\n\r\n# Configure Masker<\/span>\r\nmasker = NiftiLabelsMasker(\r\n    labels_img=glasser_atlas,\r\n    standardize=\"zscore_sample\"<\/span>,\r\n    memory=\"nilearn_cache\"<\/span>,\r\n    verbose=0<\/span>\r\n)\r\n\r\n# Extract confounds and transform BOLD data<\/span>\r\nconfounds_df = pd.read_csv(confounds_file, sep='\\t'<\/span>)\r\nselected_confounds = confounds_df[['trans_x'<\/span>, 'trans_y'<\/span>, 'trans_z'<\/span>, 'rot_x'<\/span>, 'rot_y'<\/span>, 'rot_z'<\/span>, 'csf'<\/span>, 'white_matter'<\/span>]]\r\n\r\n# Time-series shape: (n_volumes, 360)<\/span>\r\ntime_series = masker.fit_transform(func_file, confounds=selected_confounds)\r\n\r\n# Compute 360x360 Functional Connectivity correlation matrix<\/span>\r\ncorrelation_measure = ConnectivityMeasure(kind=\"correlation\"<\/span>)\r\nfc_matrix = correlation_measure.fit_transform([time_series])[0<\/span>]\r\n\r\n# Save output<\/span>\r\nnp.save(\"sub-01_glasser_fc_matrix.npy\"<\/span>, fc_matrix)\r\n<\/code><\/pre>\n<\/div>\n<\/div>\n<\/div>\n
Option B: Surface\/CIFTI Parcellation (Connectome Workbench)<\/h3>\n
If your output data is in CIFTI space (.dtseries.nii<\/code>), use Connectome Workbench (wb_command<\/code>):<\/p>\n
\n
\n
\n
\n
Bash<\/span><\/p>\n
<\/div>\n<\/div>\n
wb_command -cifti-parcellate \\\r\n  \/lab_data\/hawk\/data\/derivatives\/fmriprep\/sub-01\/func\/sub-01_task-rest_space-fsLR_den-32k_bold.dtseries.nii \\\r\n  \/lab_data\/hawk\/resources\/neuroimaging\/atlases\/glasser\/HCP-MMP1_32k_fs_LR.dlabel.nii \\\r\n  COLUMN \\\r\n  sub-01_task-rest_glasser_parcellated.ptseries.nii\r\n<\/code><\/pre>\n<\/div>\n<\/div>\n<\/div>\n
5. Support & Troubleshooting<\/h2>\n
    \n
  • \n
    Missing Template Errors:<\/b> If your pipeline attempts to connect to the internet to download a template, ensure TEMPLATEFLOW_HOME<\/code> and APPTAINERENV_TEMPLATEFLOW_HOME<\/code> are correctly exported in your submission script.<\/p>\n<\/li>\n
  • \n
    Permission Issues:<\/b> All files under \/lab_data\/hawk\/resources\/neuroimaging\/<\/code> are configured with read-only permissions (755<\/code>) for all cluster users.<\/p>\n<\/li>\n<\/ul>\n","protected":false},"excerpt":{"rendered":"
    This guide provides instructions for accessing shared neuroimaging reference templates and atlases stored on the cluster at \/lab_data\/hawk\/resources\/neuroimaging\/. 1. Available Resources & Directory Paths Resource File Path \/ Location Description TemplateFlow Cache \/lab_data\/hawk\/resources\/neuroimaging\/templateflow Central offline cache containing MNI152NLin2009cAsym and standard templates. Glasser Atlas (Volumetric) \/lab_data\/hawk\/resources\/neuroimaging\/atlases\/glasser\/ HCP-MMP1.0 converted to 3D NIfTI…<\/p>\n","protected":false},"author":1,"comment_status":"closed","ping_status":"closed","template":"","format":"standard","meta":{"footnotes":""},"ht-kb-category":[21],"ht-kb-tag":[],"class_list":["post-761","ht_kb","type-ht_kb","status-publish","format-standard","hentry","ht_kb_category-data"],"jetpack_sharing_enabled":true,"_links":{"self":[{"href":"https:\/\/ni.cmu.edu\/computing\/wp-json\/wp\/v2\/ht-kb\/761","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/ni.cmu.edu\/computing\/wp-json\/wp\/v2\/ht-kb"}],"about":[{"href":"https:\/\/ni.cmu.edu\/computing\/wp-json\/wp\/v2\/types\/ht_kb"}],"author":[{"embeddable":true,"href":"https:\/\/ni.cmu.edu\/computing\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/ni.cmu.edu\/computing\/wp-json\/wp\/v2\/comments?post=761"}],"version-history":[{"count":4,"href":"https:\/\/ni.cmu.edu\/computing\/wp-json\/wp\/v2\/ht-kb\/761\/revisions"}],"predecessor-version":[{"id":769,"href":"https:\/\/ni.cmu.edu\/computing\/wp-json\/wp\/v2\/ht-kb\/761\/revisions\/769"}],"wp:attachment":[{"href":"https:\/\/ni.cmu.edu\/computing\/wp-json\/wp\/v2\/media?parent=761"}],"wp:term":[{"taxonomy":"ht_kb_category","embeddable":true,"href":"https:\/\/ni.cmu.edu\/computing\/wp-json\/wp\/v2\/ht-kb-category?post=761"},{"taxonomy":"ht_kb_tag","embeddable":true,"href":"https:\/\/ni.cmu.edu\/computing\/wp-json\/wp\/v2\/ht-kb-tag?post=761"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}