ICASSO
ICASSO
toolbox is used in GIFT to determine the reliability of ICA algorithm. ICA
algorithm is run several times to determine the algorithmic reliability or
stability. Reliable estimates correspond to tight clusters and un-reliable ones
do not point to any cluster. Figure 1 will open when you click on
Utilities drop down box and select
"ICASSO". The parameters in the figure are as follows:
- Select Mode - Options available are 'RandInit', 'Bootstrap' and 'both'. The
explanation of each option is given below:
- RandInit - Algorithm starts with different initial values.
- Bootstrap - Bootstrap technique is used.
- Both - Uses both 'RandInit' and 'Bootstrap' options.
- Select number of ICA runs - Number of times ICA algorithm will be run.
ICASSO results are written to a MAT file with suffix '_icasso_results.mat'. This
MAT file contains the following variables:
- iq - Stability index.
- sR - Variable containing information about similarity measure, clustering
and projection.
- A - Mixing matrix.
- W - Un-mixing matrix.
- S - Source signal.
We use centroid of the cluster for each component instead of the average of the
individual ICA runs as it is more stable. After the ICASSO step is completed,
subsequent group ICA analysis steps like Back Reconstruction, Scaling Components
and Group Stats are run.
Figure 1: GUI for running ICASSO.
