neurobase (version 1.32.4)

zscore_img: Get Z-score over a margin of an img

Description

Standardizes an image either by the axial, sagittal, or coronal slice or whole image

Usage

zscore_img(
  img,
  mask = NULL,
  margin = NULL,
  centrality = c("mean", "median", "trimmed_mean"),
  variability = c("sd", "iqrdiff", "mad", "maddiff", "iqr", "trimmed_sd"),
  trim = 0.2,
  remove.na = TRUE,
  remove.nan = TRUE,
  remove.inf = TRUE,
  remove.val = 0,
  remask = TRUE
)

Value

Array of object of class nifti

Arguments

img

character path of image or an object of class nifti

mask

character path of mask or an object of class nifti

margin

Margin of image to z-score over (NULL - whole brain, 3-Axial, 2-Sagittal, 1-Coronal)

centrality

(character) Measure to center the data, either mean or median

variability

(character) Measure to scale the data

trim

if centrality is trimmed_mean or variability is trimmed_sd, then the amount of trimming

remove.na

(logical) change NAs to remove.val

remove.nan

(logical) change NaN to remove.val

remove.inf

(logical) change Inf to remove.val

remove.val

(logical) value to put the NA/NaN/Inf

remask

(logical) Should the image be remasked after normalizing?

See Also

aperm

Examples

dim = c(100, 30, 5)
img = array(rnorm(prod(dim), mean=4, sd=4), 
dim=dim)

truth2 = img
for (i in 1:dim(img)[2]) {
truth2[,i,] = (truth2[,i,]- mean(truth2[,i,]))/sd(truth2[,i,])
}

truth1 = img
for (i in 1:dim(img)[1]) {
truth1[i,,] = (truth1[i,,]- mean(truth1[i,,]))/sd(truth1[i,,])
}

truth3 = img
for (i in 1:dim(img)[3]) {
truth3[,,i] = (truth3[,,i]- mean(truth3[,,i]))/sd(truth3[,,i])
}
try3 = zscore_img(img, margin=3)
stopifnot(all.equal(try3, truth3))
try2 = zscore_img(img, margin=2)
stopifnot(all.equal(try2, truth2))
try1 = zscore_img(img, margin=1)
stopifnot(all.equal(try1, truth1))
  
z = zscore_img(img, margin=NULL)
ztrim = zscore_img(img, margin=NULL, 
centrality = "trimmed_mean", variability = "trimmed_sd")

z = zscore_img(img, centrality = "median", variability = "iqr")
z = zscore_img(img, centrality = "median", variability = "iqrdiff")
z = zscore_img(img, centrality = "median", variability = "maddiff")