cor_method = "pearson" (fully supported across weighted/unweighted settings, pairwise/listwise missingness, and optional p-value pruning)cor_method = "polychoric" (currently supported for unweighted correlation estimation; unsupported option combinations transparently fall back to Pearson correlation)missing_cor = c("pairwise","listwise") to explicitly control missing-data handling during correlation estimation (default: "pairwise").weights can now be supplied either as:
nrow(df), ordf (e.g., weights = "psweights").weights is supplied as a column name and node_vars = NULL, the weights column is automatically excluded from the response-node variables used for estimation.remove_nonsignificant_method, enabling Benjamini–Hochberg FDR control ("fdr") as an alternative to raw p-value pruning ("default"). sign_threshold is interpreted as the raw p-value cutoff under "default" and as the target FDR level under "fdr".EBICglasso and EBICglasso_arglist are retained for backwards compatibility but are deprecated and ignored as of 2.3.0. This feature is not statistically appropriate for the rank-deficient one-hot encoded ResIN correlation matrix.ResIN_boots_prepped and ResIN_boots_executed now have clearer print/summary behavior.ResIN_boots_execute() is more robust (handles failed iterations gracefully) and improves progress-bar formatting.ResIN_boots_extract() now supports multi-quantity extraction and improved summarizing behavior, including consistent data-frame output for summaries.as.igraph.ResIN(), as.qgraph.ResIN(), as.gephi.ResIN()as.network.ResIN() (statnet/network ecosystem), with as.network() re-exported for convenienceas.networkx.ResIN() exports NetworkX compatible edge & node CSV tables for Python workflows (loadable via pandas + networkx).as.graphsjl.ResIN() exports Graphs.jl compatible edge & node CSV tables for Julia workflows (with integer vertex IDs suitable for Graphs.jl), while preserving node/edge metadata as table columns.ResIN_to_igraph(), ResIN_to_qgraph(), ResIN_to_gephi()) are retained for compatibility and delegate to the new conversion methods where appropriate.Enhanced the optional multimodal output (multimodal = TRUE) to return tidy, analysis-ready tables in addition to graph objects:
ResIN_edgelist_multimodal: participant–response edge list (one row per observed response node selection).ResIN_nodeframe_multimodal: vertex table containing node identifiers, node type (participant vs response node), layout coordinates, and available metadata.Improved multimodal plotting:
plot_whichstat).plot_whichstat = "cluster", the legend title includes the clustering algorithm and reports participant/response counts per detected cluster.Added flexible edge overlay controls for multimodal plots via multimodal_edge_overlay:
"none": no edges drawn"multimodal": draw participant–response edges only"ResIN": draw the original ResIN (response–response) edges as an overlay, with line width optionally mapped via plot_edgestat"both": draw ResIN edges as a base layer and multimodal edges on top
ResIN edge overlays are mapped to the multimodal layout coordinates to ensure visual alignment with the multimodal node geometry.Added adjecency matrix retaining negative edges as part of aux_objects