What problem does it solve? Researchers who want to partition the brain into data-driven parcels rather than rely on a predefined atlas need a reproducible unsupervised route. This Skill provides model-level guidance for running hierarchical (agglomerative/Ward) clustering on neuroimaging features to produce parcel label maps and cluster summaries. ## Core Features & Use Cases - Data-driven parcellation: Partitions voxels, vertices, or ROI features into brain parcels from functional or structural similarity. - Multi-scale outputs: Exports parcel label maps, cluster size summaries, and optional dendrogram or merge information across scales. - Workflow delegation: Coordinates with fmri-skill and smri-skill for feature preparation and nilearn-tool for concrete masking, feature matrices, and parcel export. - Use Case: A researcher with preprocessed resting-state fMRI data wants a 200-parcel group-level parcellation; this Skill guides feature preparation, Ward clustering, and label map export. ## Quick Start Ask the agent to run hierarchical clustering parcellation on your preprocessed fMRI images with a group mask and a target of 200 parcels, exporting the label map to the output directory.
