Description
Clusterwise Independent Component Analysis.
Description
Clustering multi-subject resting state functional Magnetic Resonance Imaging data. This methods enables the clustering of subjects based on multi-subject resting state functional Magnetic Resonance Imaging data. Objects are clustered based on similarities and differences in cluster-specific estimated components obtained by Independent Component Analysis.
README.md
Clusterwise Independent Component Analysis R package
Version notes
Version of CICA on CRAN notes:
- CRAN v0.1.0: CICA version with ALS random start procedure
Version of CICA on GitHub:
Download the development version of CICA using the devtools package: devtools::install_github('jeffreydurieux/CICA')
This version contains:
R v0.1.0: CICA version with ALS random start procedure
R v0.2.1: CICA with (pseudo-) rational start options
- v0.2.0: modified RV matrix computations (computeRVmat()). A (dis) similarity matrix is computed between a list of input matrices. This is based on the two-step clustering procedure from Durieux & Wilderjans (2019).
- v0.2.0: FindRationalStarts() function. This function applies the two-step procedure using several hierarchical clustering methods in order to find rational starts for the ALS algorithm for CICA. Cluster perturbation options are also included. This function returns an object of class
rstarts. This object can be passed to the CICA main function. - v0.2.0: These options are also directly included in the CICA main function.
- v0.2.1: Update of example data. Added a single example data set from the simulation design of Durieux & Wilderjans (2019). It contains 60 subjects and original cluster specific components and the true simulated clustering is added.
R v0.3.0 CICA version with multiple CICA models
R v1.0.0 CICA version with all working functionalities. This version is also available on CRAN. This package version includes the papayar archived files that were made by John Muschelli.
R v1.1.0 CICA version with a fast EVD based estimation procedure. This results in an equal (or similar) clustering. Use the final clustering to seed the CICA (using method = 'fastICA') to extract independent components.