A co-alteration parceling of the cingulate cortex

Abstract

The cingulate cortex is known to be a complex structure, involved in several cognitive and emotional functions, as well as being altered by a variety of brain disorders. This heterogeneity is reflected in the multiple parceling models proposed in the literature. At the present, sub-regions of the cingulate cortex had been identified taking into account functional and structural connectivity, as well as cytological and electrochemical properties. In the present work, we propose an innovative node-wise parceling approach based on meta-analytic Bayesian co-alteration. To this aim, 193 case-control voxel-based morphometry experiments were analyzed, and the Patel's κ index was used to assess probability of morphometric co-alteration between nodes placed in the cingulate cortex and in the rest of the brain. Hierarchical clustering was then applied to identify nodes in the cingulate cortex exhibiting a similar pattern of whole-brain co-alteration. The obtained dendrogram highlighted a robust fronto-parietal cluster compatible with the default mode network, and being supported by the interplay between the retrosplenial cortex and the anterior and posterior cingulate cortex, rarely described in the literature. This ensemble was further confirmed by the analysis of functional patterns. Leveraging on co-alteration to investigate cortical organization could, therefore, allow to combine multimodal information, resolving conflicting results sometimes coming from the separate use of singular modalities. Crucially, this provides a valuable way to understand the pathological brain using data driven, whole-brain informed and context-specific evidence in a way not yet explored in the field.

Keywords: Bayesian statistic; Cingulate cortex; Hierarchical clustering; Morphometric co-alteration network; Retrosplenial cortex.

Fig. 1

Graphical representation of the pipeline used. a The MCN is obtained on the basis of the ALE map. Root nodes (i.e. localized in the cingulate cortex) are colored in white, while nodes in the rest of the brain are colored in gray; b only first neighbors of the root nodes are retained, and edges between couples of non-root nodes are removed; c the root nodes × first neighbor nodes matrix is built. Values are based on the Patel’s κ, representing the likelihood of co-alteration between each couple of nodes. This matrix is used as input for the hierarchical clustering of the root nodes (i.e. computed between rows); d clustering results are visualized on the brain. In this case, colours refers to c = 3, as marked with the dashed line on the dendrogram in (c). For the sake of clarity, the visualization is based on synthetic and simplified data rather than on the real one used for the analyses

Fig. 2

Results of the hierarchical clustering of the 30 root nodes, obtained using Pearson correlation and WPGMA. Axial views are in neurological convention (left is left). C = 5 is not shown since only the orange posterior node changed with respect to c = 4. Colors are coherent between the dendrogram and the visualization of the nodes

Fig. 3

Evaluation of the role of local co-alteration. Left: hierarchical clustering result at c = 3, obtained after excluding edges between root nodes in the cingulate-cortex. Right: details of local co-alteration between root nodes. Edges’ color from blue to red represents increasing Patel’s κ values (thresholded at 0.5 for visualization purpose). Nodes’ colors are based on the clusters obtained for the whole MCN (including both local and global co-alteration), and hence it is coherent with Fig. 1. Nodes with no edges surviving the imposed κ threshold were not shown

Fig. 4

Results of the MDS. Nodes in the obtained 2D spatial distribution were colored based on the results of hierarchical clustering at different cardinalities, as obtained for whole-brain co-alteration. Colors are coherent with the sagittal views, which were also shown in Fig. 2. C = 5 is not shown since only the orange central node changed with respect to c = 4

Fig. 5

Results of the hierarchical clustering (Pearson correlation and WPGMA) of the SVC maps originating from each of the 30 root nodes. Axial views are in neurological convention (left is left). C = 4 is not shown since only the yellow middle node changed with respect to c = 3. Colors are coherent between the dendrogram and the visualization of the nodes

Fig. 6

A comparison of the two clusters solution based on co-alteration (left) or SVC (right) at level c = 2. Axial views are in neurological convention (left is left)

Fig. 7

Results of the functional network decomposition for the fronto-parietal cluster. Top: Each line represents the thresholding on a different percentile of the Patel’s κ values distribution. Values express the percentage of nodes located in a network with respect to the total number of nodes. VIS visual, SM sensory-motor, dATT dorsal attentional, vATT ventral attentional, LIMB limbic, fr-par fronto-parietal, DMN default mode network. Bottom: visualization of the nodes with at least one edge above the threshold of the 90th percentile (i.e. data behind the blue line in the radar chart). Root nodes are colored in yellow. Axial view is in neurological convention (left is left)