Compute Connectivity Matrix from Atlas Parcellations

Extracts mean time-series for each atlas region from a 4D NeuroVec and computes pairwise correlations to produce a connectivity matrix.

Usage

atlas_connectivity(
  data_vol,
  atlas,
  method = c("pearson", "spearman", "partial"),
  threshold = NULL,
  stat_func = mean,
  ...
)

Arguments

data_vol

A NeuroVec (4D) containing the time-series data.

atlas

An atlas object defining the parcellation.

method

Character string specifying the correlation method: "pearson" (default), "spearman", or "partial". Partial correlation requires the corpcor package.

threshold

Optional numeric value. If supplied, entries with abs(r) < threshold are set to zero.

stat_func

Function used to summarise voxel values within each parcel (default: mean).

...

Additional arguments passed to reduce_atlas.

Value

A symmetric matrix of class c("atlas_connectivity", "matrix") with region labels as dimnames.

See also

reduce_atlas for the underlying extraction, as_igraph.atlas_connectivity for graph conversion

Examples

if (FALSE) { # \dontrun{
atlas <- get_schaefer_atlas(parcels = "100", networks = "7")
# data_vol is a 4D NeuroVec from an fMRI run
conn <- atlas_connectivity(data_vol, atlas)
conn_sparse <- atlas_connectivity(data_vol, atlas, threshold = 0.3)
} # }