Skip to main content
. 2021 Dec 24;15:740143. doi: 10.3389/fninf.2021.740143

FIGURE 1.

FIGURE 1

(A) The edge information of individual networks is vectorized and stacked in a matrix which is provided as input to UMAP for low dimensional representation. (B) For the Fit-SNE version of t-SNE, dimension reduction using principal components analysis (PCA) is applied to the vectorized, stacked matrix during the initial embedding. Subsequently, the embedding is performed directly on the vectorized, stacked matrix.