Hi all,
we are currently looking for a way to (region-wise) quantify the amount of entropy in our resting state time series. I’ve never done that before and wonder what measures are commonly used to quantify entropy in rs-fmri time series? Moreover, I am a little bit confused about which measure to take as established entropy measures usually work on discrete vectors (something like [1,0,0,1,1,…]) and not on continuous variables (where each value in the set is a rational number and can theoretically range from -∞ to +∞)?
I came across antropy and PyInform which offer a variety of measures, is there some best practice wich measure is typically used for rs-fmri?
from nilearn import datasets
from nilearn.maskers import NiftiLabelsMasker
# load one nifti file
run = datasets.fetch_development_fmri(n_subjects=1)['func'][0]
# load an atlas
atlas = datasets.fetch_atlas_harvard_oxford('cort-maxprob-thr25-2mm')
# Instantiate the masker with label image and label values
masker = NiftiLabelsMasker(labels_img=atlas.maps,
labels=atlas.labels,
standardize=True)
# get signals as time x region matrix
signals = masker.fit_transform(run)
# Now compute entropy for each atlas-region over time series . . .