stableEventResponse: Estimate effect sizes per-voxel, per stimulus,...
In stnava/RKRNS: ANTsR Decoding

Description Usage Arguments Value Author(s) Examples

Use leave-a-run-out to summarize response variability across runs for a stimulus class. Returns a matrix of beta effect sizes per stimulus class where the beta effect size are defined as mean of the beta response across left out runs divided by standard deviation of the same.

1 2	stableEventResponse(boldmatrix, designmatrixIn, runIDs, hrf, verbose = F, polydegree = 4, baseshift = 0, timevals = NA)

`boldmatrix`	input raw bold data in time by space matrix
`runIDs`	numbers for the rows that should be treated together as runs
`hrf`	input hrf to use in regression
`polydegree`	number of polynomial predictors
`baseshift`	basis shift for design matrix
`designmatrix`	input design matrix - binary/impulse entries for event related design, blocks otherwise

returns a list of cross-validated effect sizes for different stimuli classes

Avants BB

fn<-paste(path.package("RKRNS"),"/extdata/111157_mocoref_masked.nii.gz",sep="") 
eximg<-antsImageRead(fn,3)
fn<-paste(path.package("RKRNS"),"/extdata/subaal.nii.gz",sep="") 
mask<-antsImageRead(fn,3)
bb<-simulateBOLD(option="henson",eximg=eximg,mask=mask)
boldImage<-bb$simbold
mat<-timeseries2matrix( bb$simbold, bb$mask )
runs<-bb$desmat$Run; 
hrf<-hemodynamicRF( 20, onsets=1, durations=1, rt=1,cc=0.1 )
stb<-stableEventResponse(mat,  bb$desmat[,1:4], runs,  hrf=hrf )
var1i<-antsImageClone( mask ) 
var1i[mask==1]<-stb[1,] # then write this out ...
# or run eigseg 
stb2<-stb
stb2[ stb < 6 ]<-0
ee<-eigSeg(mask, matrixToImages( stb2, mask) )
ImageMath(3,ee,'ClusterThresholdVariate',ee,mask,5)
# antsImageWrite ...