zscore_img {neurobase}

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
)

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?

Value

Array of object of class nifti

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")