Skip to contents

Build a parcel-level Laplacian

make_parcel_laplacian.Rd

Constructs a parcel adjacency-based Laplacian either from a supplied adjacency matrix, parcel centroids (k–NN), or a label volume. Normalization and spectral scaling mirror the voxel-level Laplacian.

Usage

make_parcel_laplacian(
  adjacency = NULL,
  parcel_coords = NULL,
  parc_vol = NULL,
  k = 6,
  sigma = 2.5,
  normalized = TRUE
)

Arguments

adjacency

Optional symmetric adjacency/weight matrix (P x P).

parcel_coords

Optional P x 3 matrix of parcel centroids.

parc_vol

Optional NeuroVol of parcel labels (used to compute centroids).

k, sigma

KNN neighbors and Gaussian bandwidth.

normalized

Logical; return normalized Laplacian.

Value

Sparse symmetric Matrix (P x P) Laplacian.

See also

make_laplacian() for voxel-level Laplacian construction

Other spatial metrics: build_spatial_metric(), build_spatial_metric_parcel(), compute_tsnr(), compute_tsnr_parcel(), make_laplacian()

Examples

# \donttest{
Lp <- make_parcel_laplacian(parc_vol = parc, k = 6, sigma = 2)
#> Error: object 'parc' not found
# }