A novel group-fused sparse partial correlation method for simultaneous estimation of functional networks in group comparisons

Group-fused multiple-graphical lasso cobined with stability selection (GMGLASS) is a software toolbox that can be employed to simultaneously estimate both individual- and group-level functional networks from 2 groups.
The software:
(1) Reads time series from 2 groups of subjects;
(2) Randomly subsamples data 100 times to estimate stability with stability
selection;
(3) Group-fused multiple-graphical lasso is applied with two hyper-parameters: alpha and beta;
(4) Appropriate ranges of hyper-parameters are chosen for achieving stability selection;
(5) Both individual- and group-level functional networks can be estimated.

The method is described in the following paper:
Xiaoyun Liang, David N. Vaughan, Alan Connelly, Fernando Calamante. A novel group-fused sparse partial correlation method for simultaneous estimation of functional networks.in group comparison studies. Brain Topography, 12/2017; DOI:10.1007/s10548-017-0615-6.
Download Now Download Now
OR See All Files

Specifications

License:GNU General Public License (GPL)
Domain:MR

Recent Activity

Show: 
File Activity

gmglass: GMGLASS1.01 release

Github link posted by Xiaoyun Liang on Jan 10

File Activity

gmglass: GMGLASS1.01 release

GMGLASS.zip posted by Xiaoyun Liang on Jan 10

Forum Activity

open-discussion forum

Github link posted by Xiaoyun Liang on Jan 9

File Activity

gmglass: GMGLASS1.0 release

GMGLASS.zip posted by Xiaoyun Liang on Jan 9

Forum Activity

open-discussion forum

Welcome to Open-Discussion posted by Xiaoyun Liang on Jan 9