fmri_findclusters

Description

FMRI_FINDCLUSTERS Finds and label connected clusters in a binary brain map. Identifies all connected clusters of suprathreshold voxels in a binary brain map, returns a labelled map and a summary of cluster sizes. Clusters are defined using 18-connectivity (face and edge connected voxels).

Usage

[clusters, n_clusters, sizes] = fmri_findclusters(data)
[clusters, n_clusters, sizes] = fmri_findclusters(data, min_size)

Inputs

DATA: Binary brain map, vectorized (228453 x 1) or volume (91 x 109 x 91). Non-zero voxels define clusters.
MIN_SIZE: Minimum cluster size in voxels. Clusters smaller than this are removed from the output. Default: 1 (keep all).

Outputs

CLUSTERS: Integer-labelled map (same format as DATA). Each connected cluster has a unique integer label 1..n_clusters, sorted by descending cluster size. Zero = background.
N_CLUSTERS: Number of clusters found (after applying MIN_SIZE).
SIZES: Vector of cluster sizes in voxels (descending order), one entry per cluster.

Examples

[cl, n, sz] = fmri_findclusters(thresh_map);
fprintf('%d clusters found; largest = %d voxels\n', n, sz(1))
% Only keep clusters > 50 voxels
[cl, n, sz] = fmri_findclusters(thresh_map, 50);
fmri_showslices_int(cl, 1, 3)
% Get binary map of largest cluster only
biggest = double(cl == 1);