fmri.cluster: Cluster thresholding.

Description

Detection of activated regions using cluster thresholding.

Usage

fmri.cluster(spm, alpha = 0.05, ncmin = 2, ncmax=ncmin, 
             minimum.signal = 0, verbose = FALSE)

Value

Object with class attributes "fmripvalue" and "fmridata"

pvalue: cluster based p-values for voxel that were detected for any cluster size, a value of 1 otherwise.
mask: mask of detected activations
weights: voxelsize ratio
dim: data dimension
hrf: expected BOLD response for contrast (single stimulus only)

Arguments

spm: fmrispm object
alpha: multiple test (over volume and cluster sizes) adjusted significance level used for thresholds.
ncmin: minimal cluster size used. An activation is detected if for any clustersize in nvmin:20 the size specific threshold is exceeded.
ncmax: maximal cluster size used. An activation is detected if for any clustersize in ncmin:ncmax the size specific threshold is exceeded.
minimum.signal: allows to specify a (positive) minimum value for detected signals. If minimum.signal >0 the thresholds are to conservative, this case needs further improvements.
verbose: intermediate diagnostics

Author

Joerg Polzehl polzehl@wias-berlin.de

Details

Approximate thresholds for the existence of a cluster with spm-values exceeding a 1-beta threshold k_{nc,na:ne} for cluster size nc are based on a simulation study under the hypothesis and adjusted for number of voxel in mask and spatial correlation. beta is chosen such that under the hypothesis the combined (over cluster sizes ncmin:ncmax) test has approximate significance level alpha.

Examples

Run this code

  if (FALSE) fmri.cluster(fmrispmobj)

Run the code above in your browser using DataLab