discr.freq.obs2discr: Frequency Discriminability

Description Usage Arguments Value Author(s)

Description

A function that performs discriminability in the frequency domain.

Usage

1
2
discr.freq.obs2discr(signal, ids, tr, lc = NaN, hc = NaN,
  spectrum = "amp", rank = FALSE, font.size = 15, method = "H")

Arguments

signal

[[n]][nt, nroi] the signal for each of the n subjects, containing an array of nt observations for nroi rois.

ids

[n] the ids for each scan corresponding to the signal from above.

tr=NaN

[1] the repetition time of the dataset. NaN for none.

lc=NaN

[1] the lower cutoff for highpass filtering. NaN for none.

hc=NaN

[1] the high cutoff for low-pass filtering. NaN for none.

spec='amp'

the spectrum to work with. 'amp' for amplitude, 'pow' for power.

rank=FALSE

a boolean indicating whether to do unranked (FALSE) or ranked (TRUE) graphs, whereby all the edge-weights gare ranked before computing distances.

font.size=10

the default font size for the plot text.

method='H'

the method to use for the distance computation.

'F'

Frobenius norm between pairs of graphs

'H'

Hellinger distance between pairs of graphs

Value

d [1] the discriminability statistic for the data.

dist [n, n] the distance matrix associated with the data.

distplot a plot of the distance matrix.

kdeplot a plot of the density estimate of the intra vs inter subject distances.

combinedplot a multiplot showing the dist plot and the kde plot.

Author(s)

Eric Bridgeford


ebridge2/Discriminability documentation built on May 15, 2019, 7:48 p.m.