An R package for the analysis of Functional Connectivity matrices, lesional maps, or disconnection maps through elastic NETs.
The package is currently available through GitHub. The installation requires the R package devtools.
# install.packages("devtools")
devtools::install_github("EBlini/FCnet")The analysis of (Functional Connectivity) neuroimaging data can be daunting due to the very high dimensionality of the features involved. In time, several approaches to the problem have been devised. FCnet allows one to easily implement a three steps procedure consisting of:
1) Feature reduction: the functional connectivity matrices (or volumes with lesion/disconnection mapping) are first summarized through data reduction techniques such as Principal Component Analysis or Independent Components Analysis.
2) Robust regression: the reduced matrix of Weights is then entered into a robust regression model (with either ridge or LASSO penalty). The model is crossvalidated internally by means of Leave-One-Out (possibly nested) crossvalidation.
3) Back-projection: models’ coefficients can be back-projected onto the original space, in order to rank the most predictive edges of a matrix or voxels.
For useful references, see: Siegel et al., 2016; Salvalaggio et al., 2020; Calesella et al., 2020.
An overview of the package is available here.