regpostsim: Estimation of voxel activations

In cudaBayesreg: CUDA Parallel Implementation of a Bayesian Multilevel Model for fMRI Data Analysis

Description

regpostsim estimates voxel activation and plots the posterior distribution of a regression coefficient.

Usage

regpostsim(pmeans, vreg, plot=T)

Arguments

pmeans	Posterior Means of Coefficients as processed by pmeans.hcoef()
vreg	regression variable to process
plot	plot the histogram, (default=T)

Details

Estimate the active and non-active voxels based on the highest posterior density (HPD) of the coefficients simulated by the multilevel method. Plot the histogram of the posterior distribution of regression coefficient vreg

See Also

cudaMultireg.slice, pmeans.hcoef, plot.hcoef.post, post.simul.hist, post.simul.betadraw, post.ppm, post.tseries

Examples

## Not run: 
slicedata <- read.fmrislice(fbase="fmri", slice=3, swap=FALSE)
ymaskdata <- premask(slicedata)
fsave <- paste(tempdir(),"/simultest1",fileext = ".sav", sep="")
answ <- readline("Run MCMC simulation first ? ")
run <- FALSE
if (substr(answ, 1, 1) == "y") { run <- TRUE }
if(run) {
	out <- cudaMultireg.slice(slicedata, ymaskdata, R=2000, keep=5, nu.e=3,
   fsave=fsave, zprior=FALSE)
} else {
	load(file=fsave)
	cat("loaded",fsave,"\n")
}
##
pmeans <- pmeans.hcoef(out$betadraw) 
px <- regpostsim(pmeans, vreg=2)
pm2 <- pmeans[,vreg]
spma <- px$spma # active voxels
spmn <- px$spmn # non-active voxels

## End(Not run)

cudaBayesreg documentation built on May 29, 2017, 6:19 p.m.