centralveins: Central Vein Detection

In neuroconductor-devel/lesiontools: Statistical and Computational Tools for Brain Lesion Analysis

Description

This function obtains the probability that each lesion in a subject's deep white-matter has a central vein.

Usage

centralveins(epi, t1, flair, probmap = NULL, binmap = NULL,
  parallel = F, cores = 2, skullstripped = F, biascorrected = F,
  c3d = F)

Arguments

epi	a T2*-EPI volume of class nifti.
t1	a T1-weighted volume of class nifti.
flair	a T2-FLAIR volume of class nifti.
probmap	an image of class nifti, containing the probability that each voxel is a lesion voxel. If a probability map is not included, the MIMoSA model will be applied (Valcarcel et al., 2017).
binmap	a nifti mask in which voxels are classified as either lesion voxels or not lesion voxels. Note that mask should be in the same space as the probmap volume.
parallel	is a logical value that indicates whether the user's computer is Linux or Unix (i.e. macOS), and should run the code in parallel.
cores	if parallel = TRUE, cores is an integer value that indicates how many cores the function should be run on.
skullstripped	a logical value reflecting whether or not the images have already been skull-stripped.
biascorrected	a logical value reflecting whether or not the images have already been bias-corrected.
c3d	a logical value reflecting whether or not the Convert3D imaging toolbox is installed.

Value

A list containing candidate.lesions (a nifti file with labeled lesions evaluated for CVS), cvs.probmap (a nifti file in which candidate lesions are labeled with their CVS probability), and cvs.biomarker (a numeric value representing the average CVS probability of a subject's lesions).

Examples

## Not run: 
library(neurobase)
epi <- readnii('path/to/epi')
flair <- readnii('path/to/flair')
t1 <- readnii('path/to/t1')
cvs <- centralveins(epi = epi, t1 = t1, flair = flair
                         parallel = TRUE, cores = 4, c3d = T) 
## End(Not run)

neuroconductor-devel/lesiontools documentation built on May 15, 2019, 3:16 p.m.