Want to know more about SICE ( Sparse Inverse Covariance Estimation)

Hii all,

I would like to know more about SICE method. In GraphVar I am using following procedure for creating connectivity matrices from time series.

Using time series of fmri images I want to create connectivity matrix of particular SICE target density (0.25-0.45) with an increment of 0.05.  I would like to know more about SICE target density. Does it means, say 0.25 target density means 25 percent of your connections are retained or not?.

Pls help me to clear this doubt.

Thanks,

Mini