This class represents a three-dimensional brain image divided into N disjoint
partitions or clusters. It extends the SparseNeuroVol class
to provide efficient storage and manipulation of clustered neuroimaging data.
Construct a ClusteredNeuroVol instance
Arguments
- mask
an instance of class
LogicalNeuroVol- clusters
a vector of clusters ids with length equal to number of nonzero voxels in mask
mask- label_map
an optional
listthat maps from cluster id to a cluster label, e.g. (1 -> "FFA", 2 -> "PPA")- label
an optional
characterstring used to label of the volume
Details
The ClusteredNeuroVol class is designed for efficient representation and manipulation of brain images with distinct, non-overlapping regions or clusters. It combines the memory efficiency of sparse representations with additional structures for managing cluster information.
The use case of ClusteredNeuroVol is to store volumetric data that has been clustered into discrete sets of voxels,
each of which has an associated id. For example, this class can be used to represent parcellated neuroimaging volumes.
Slots
maskA
LogicalNeuroVolobject representing the logical mask indicating the spatial domain of the set of clusters.clustersAn integer vector representing the cluster number for each voxel in the mask.
label_mapA named list where each element represents a cluster and its name.
cluster_mapAn
environmentobject that maps from cluster id to the set of 1D spatial indices belonging to that cluster.
Methods
This class inherits methods from the SparseNeuroVol class.
Additional methods specific to cluster operations may be available.
Usage
ClusteredNeuroVol objects are particularly useful for:
Representing parcellated brain images
Storing results of clustering algorithms applied to neuroimaging data
Efficient manipulation and analysis of region-based neuroimaging data
See also
SparseNeuroVol-class for the parent sparse volume class.
LogicalNeuroVol-class for the mask representation.
Examples
if (FALSE) { # \dontrun{
# Create a simple clustered brain volume
dim <- c(10L, 10L, 10L)
mask_data <- array(rep(c(TRUE, FALSE), 500), dim)
mask <- new("LogicalNeuroVol", .Data = mask_data,
space = NeuroSpace(dim = dim, origin = c(0,0,0), spacing = c(1,1,1)))
clusters <- as.integer(runif(sum(mask_data)) * 5)
label_map <- list("Cluster1" = 1, "Cluster2" = 2, "Cluster3" = 3,
"Cluster4" = 4, "Cluster5" = 5)
cluster_map <- new.env()
for (i in 1:5) {
cluster_map[[as.character(i)]] <- which(clusters == i)
}
clustered_vol <- new("ClusteredNeuroVol",
mask = mask,
clusters = clusters,
label_map = label_map,
cluster_map = cluster_map)
} # }
# Create a simple space and volume
space <- NeuroSpace(c(16, 16, 16), spacing = c(1, 1, 1))
vol_data <- array(rnorm(16^3), dim = c(16, 16, 16))
vol <- NeuroVol(vol_data, space)
# Create a binary mask (e.g., values > 0)
mask_data <- vol_data > 0
mask_vol <- LogicalNeuroVol(mask_data, space)
# Get coordinates of masked voxels
mask_idx <- which(mask_data)
coords <- index_to_coord(mask_vol, mask_idx)
# Cluster the coordinates into 10 groups
set.seed(123) # for reproducibility
kmeans_result <- kmeans(coords, centers = 10)
# Create the clustered volume
clustered_vol <- ClusteredNeuroVol(mask_vol, kmeans_result$cluster)
# Print information about the clusters
print(clustered_vol)
#>
#> ClusteredNeuroVol
#> ============================================================
#>
#> > Type: Clustered Volume
#> > Dimensions: 16 x 16 x 16
#> > Spacing: 1 x 1 x 1 mm
#> > Origin: 0 x 0 x 0 mm
#> > Orientation: Left-to-Right Posterior-to-Anterior Inferior-to-Superior
#>
#> Cluster Information
#> ----------------------------------------
#> > Total Clusters: 10
#> > Active Voxels: 2033 (49.6% of volume)
#>
#> Region Labels
#> ----------------------------------------
#> > Clus_1 [1]
#> > Clus_2 [2]
#> > Clus_3 [3]
#> > Clus_4 [4]
#> > Clus_5 [5]
#> ... and 5 more regions
#>