mni.compute.FDR: Compute the q value at every vertex

Description Usage Arguments Details Value See Also

View source: R/vertex_glim.R

Description

Uses the False Discovery Rate as described by Genovese et. al. to compute both the q-values and a t-threshold for a corresponding set of t or p values.

Usage

1
2
mni.compute.FDR(t.stats=NULL, p.values=NULL, filename=NULL,
                column.name=NULL, df=Inf, fdr=0.05, plot.fdr=FALSE) 

Arguments

t.stats

An array of t-statistics.

p.values

An array of p-values.

filename

Quoted filename of previous results

column.name

The column of the file specified by filename to retrieve

df

The degrees of freedom.

fdr

The threshold to test for.

plot.fdr

Whether to produce a plot of the FDR procedure.

Details

This function computes the False Discovery Rate as described by the NeuroImage paper by Genovese et. al. It can handle three possible inputs: an array of p values, and array of t statistics, or a filename of vertstats results (as generated by mni.write.vertex.stats). In the latter case, the filename should be quoted and accompanied by a column name (specified by the column.name argument, also quoted). If you are unsure about the possible column names in that file, leave that argument out and mni.compute.FDR will report an error but also give a list of the possible column names in that file.

Value

Returns both the threshold as well as a list of q-values corresponding to the t or p values.

See Also

mni.read.glim.file. mni.build.data.table. mni.vertex.statistics. mni.mean.statistics. mni.write.vertex.stats.


BIC-MNI/mni.cortical.statistics documentation built on May 5, 2019, 10:25 a.m.