Joint graphical models combined with stability selection (JGMSS)is a software toolbox that can be employed to robustly estimate both individual- and group-level sparse networks. The software:
(1) Reads time series from a group of subjects;
(2) Subsample data 100 times to estimate stability (i.e. stability
selection);
(3) Group graphical-lasso constraints are applied, including 2
regularization parameters;
(4) Ranges of regularization parameters are chosen to implement stability
selection;
(5) The alternating direction method of multipliers (ADMM) approach is
employed to solve the problem.

The Method is described in the following paper:
Xiaoyun Liang, Alan Connelly, Fernando Calamante. A novel joint sparse partial correlation method for estimating group functional networks. Human Brain Mapping 12/2015; DOI:10.1002/hbm.23092.