Version: 0.2
Date: 2018-07-27
Title: All-Resolution Inference
Author: Livio Finos, Jelle Goeman, Wouter Weeda, Jonathan Rosenblatt, Aldo Solari
Maintainer: Livio Finos <livio.finos@unipd.it>
Description: It performs All-Resolutions Inference (ARI) on functional Magnetic Resonance Image (fMRI) data. As a main feature, it estimates lower bounds for the proportion of active voxels in a set of clusters as, for example, given by a cluster-wise analysis. The method is described in Rosenblatt, Finos, Weeda, Solari, Goeman (2018) <doi:10.1016/j.neuroimage.2018.07.060>.
License: GPL-2 | GPL-3 [expanded from: GPL (≥ 2)]
RoxygenNote: 6.0.1
Suggests: knitr, rmarkdown
Imports: hommel, RNifti, plyr
VignetteBuilder: knitr
NeedsCompilation: no
Packaged: 2018-08-01 11:27:29 UTC; livio
Repository: CRAN
Date/Publication: 2018-08-01 12:20:02 UTC

All-Resolutions Inference

Description

It performs All-Resolutions Inference on fMRI data. As a main feature, it estimates lower bounds for the proportion of active voxels in a set of clusters as, for example, given by a cluster-wise analysis.

Author(s)

all of us

Examples

pvalue_name <- system.file("extdata", "pvalue.nii.gz", package="ARIbrain")
cluster_name <- system.file("extdata", "cluster_th_3.2.nii.gz", package="ARIbrain")
zstat_name <- system.file("extdata", "zstat.nii.gz", package="ARIbrain")
mask_name <- system.file("extdata", "mask.nii.gz", package="ARIbrain")

ARI(Pmap = pvalue_name, clusters= cluster_name, 
    mask=mask_name, Statmap = zstat_name)

Valid Circular Inference (ARI) for Brain Imaging

Description

Valid Circular Inference (ARI) for Brain Imaging

Usage

ARI(Pmap, clusters, mask = NULL, alpha = 0.05, Statmap = function(ix)
  -qnorm(Pmap[ix]), summary_stat = c("max", "center-of-mass"),
  silent = FALSE)

Arguments

Pmap

3D array of p-values or a (character) nifti file name.

clusters

3D array of cluster ids (0 when voxel does not belong to any cluster) or a (character) nifti file name.

mask

3D array of locicals (i.e. TRUE/FALSE in/out of the brain). Alternatively it may be a (character) nifti file name. If mask=NULL, it is assumed that non of the voxels have to be excluded.

alpha

Significance level. alpha=.05 by default.

Statmap

Statistics (usually t-values) on which the summaries are based. Can be either a 3D array, a (character) nifti file name or a function with argument ix used in the function to select the voxels belonging to a given cluster. By default Statmap = function(ix) -qnorm(Pmap[ix]) which convert the p-values in one-sided z-score.

summary_stat

Choose among =c("max", "center-of-mass").

silent

FALSE by default.

Value

A matrix reporting Size, FalseNull, TrueNull, ActiveProp and other statistics for each cluster.

Examples

pvalue_name <- system.file("extdata", "pvalue.nii.gz", package="ARIbrain")
cluster_name <- system.file("extdata", "cluster_th_3.2.nii.gz", package="ARIbrain")
zstat_name <- system.file("extdata", "zstat.nii.gz", package="ARIbrain")
mask_name <- system.file("extdata", "mask.nii.gz", package="ARIbrain")

print(mask_name)
print(pvalue_name)
print(cluster_name)
print(zstat_name)

ARI(Pmap = pvalue_name, clusters= cluster_name, 
    mask=mask_name, Statmap = zstat_name)

cluster_threshold

Description

Get spatially-connected clusters starting from a 3D map of logical values

Usage

cluster_threshold(map, max_dist = sqrt(3))

Arguments

map

3D map of logical values. TRUE if the voxel it to be clustered (e.g. it is supra-threshold).

max_dist

maximum distance allowed to in the same cluster. By default: max_dist=sqrt(3) i.e. comprises all the voxels up to the corners souranding the target voxel. A value such as max_dist=sqrt(2) excludes the corners.

Value

a 3D map (same size of map) with integer values identifying the cluster and 0 elsewhere.

Examples

## Not run: 
Tmap = RNifti::readNifti(system.file("extdata", "zstat.nii.gz", package="ARIbrain"))
clstr=cluster_threshold(Tmap>3.2)
table(clstr)

## End(Not run)