Fig. 1: Thalamocortical structural connectivity gradients.
From: A multimodal characterization of low-dimensional thalamocortical structural connectivity patterns
A Normalized group-level structural connectivity matrix (N = 50), resulting from probabilistic tractography computation between thalamic seed-voxels and cortical parcels (i.e., TC). B Affinity matrix derived from group-level structural connectivity matrix using a normalized angle similarity kernel. C Decomposition of affinity matrix into ten gradient components using diffusion map embedding. For each component, the corresponding explained variance is displayed. D Gradient loadings of component 1 (G1sc) projected onto the thalamus (axial planes). The red lines in the glass brain indicate the position of each respective axial slice of the displayed thalamus. E Gradient loadings of component 2 (G2sc) projected onto the thalamus. Slice positions are congruent to (D). F Decoding of G1sc and G2sc based on THOMAS atlas. (i) 2D space framed by G1sc and G2sc. Each data point represents a thalamic voxel, color-coded by the thalamic subnucleus to which it belongs. (ii) Raincloud plots display the gradient loadings of G1sc and G2sc per nucleus and are ordered by median, respectively. The boxes represent the interquartile range (25th to 75th percentile), lines depict medians, and whiskers are defined by values 1.5 times the interquartile range. All results are presented for the left hemisphere; however, they were similarly replicated in the right thalamus (Supplementary Fig. 3). AV anterior ventral nucleus, VA ventral anterior nucleus, VLa ventral lateral anterior nucleus, VLP ventral lateral posterior nucleus, VPL ventral posterior lateral nucleus, Pul pulvinar nucleus, MGN medial geniculate nucleus, CM centromedian nucleus, MD mediodorsal nucleus, Hb Habenular nucleus, MTT Mammillothalamic tract.
