The Human Thalamus Is an Integrative Hub for Functional Brain Networks

Abstract

The thalamus is globally connected with distributed cortical regions, yet the functional significance of this extensive thalamocortical connectivity remains largely unknown. By performing graph-theoretic analyses on thalamocortical functional connectivity data collected from human participants, we found that most thalamic subdivisions display network properties that are capable of integrating multimodal information across diverse cortical functional networks. From a meta-analysis of a large dataset of functional brain-imaging experiments, we further found that the thalamus is involved in multiple cognitive functions. Finally, we found that focal thalamic lesions in humans have widespread distal effects, disrupting the modular organization of cortical functional networks. This converging evidence suggests that the human thalamus is a critical hub region that could integrate diverse information being processed throughout the cerebral cortex as well as maintain the modular structure of cortical functional networks.SIGNIFICANCE STATEMENT The thalamus is traditionally viewed as a passive relay station of information from sensory organs or subcortical structures to the cortex. However, the thalamus has extensive connections with the entire cerebral cortex, which can also serve to integrate information processing between cortical regions. In this study, we demonstrate that multiple thalamic subdivisions display network properties that are capable of integrating information across multiple functional brain networks. Moreover, the thalamus is engaged by tasks requiring multiple cognitive functions. These findings support the idea that the thalamus is involved in integrating information across cortical networks.

Keywords: brain networks; diaschisis; functional connectivity; graph theory; thalamus.

Figure 1.

Thalamus-mediating corticocortical communication for functional brain networks. A, As a provincial hub, the thalamus is connected with cortical regions that belong to the same cortical functional network (represented in solid blue circles). A provincial hub will have a high WMD z-score. B, As a connector hub, the thalamus is connected with cortical regions in multiple cortical functional networks (one network colored in blue, and the other in green). A connector hub will exhibit high PC.

Figure 2.

Cortical functional networks and thalamic atlases. A, Cortical functional networks and thalamic parcellation derived from functional connectivity analyses between the thalamus and each cortical network using rs-fMRI data. Network abbreviations (based on its most predominant anatomical location) are as follows: mO, medial occipital; SM, somato-motor; T, temporal; and latO, lateral occipital. B, Structural connectivity-based segmentation of the thalamus using the Oxford-FSL atlas. Each thalamic subdivision was labeled based on the cortical region it is most structurally connected with. C, Histology-based thalamic parcellation using the Morel atlas. Abbreviations for thalamic nuclei are as follows: MGN, medial geniculate nucleus; PuI, inferior pulvinar nucleus; PuL, lateral pulvinar nucleus; PuA, anterior pulvinar nucleus; Po, posterior nucleus; VM, ventral medial.

Figure 3.

Nodal properties of the thalamus and cortical ROIs. A, Kernel density plot of WMD values for thalamic voxels and cortical ROIs. Thalamic voxels were categorized into two categories of thalamic nuclei (Sherman and Guillery, 2013; Sherman, 2016). B, Kernel density plot of PC values for thalamic voxels and cortical ROIs. Thalamic voxels were categorized into two categories of thalamic nuclei. First-order thalamic nuclei included AN, LGN, MGN, VL, and VP. Higher-order thalamic nuclei included IL, MD, LP, posterior nucleus (Po), pulvinar, VA, and ventral medial (VM). All graph metrics were averaged across network densities from 0.01 to 0.15.

Figure 4.

Spatial distribution of network metrics. A, WMD and PC values of cortical ROIs. B, WMD values of thalamic voxels. C, PC values of thalamic voxels. D, Location of voxels with strong connector (dark navy blue), provincial (dark green), or connector plus provincial hub properties (gold) in the thalamus. In D, only thalamic voxels that exhibited PC and/or WMD values of >90% of all cortical ROIs are displayed. All graph metrics averaged across network densities ranged from 0.01 to 0.15.

Figure 5.

Graph metrics of each thalamic subdivision. A–F, Box plots summarizing WMD values (A–C) and PC values (D–F) for all thalamic voxels for each thalamic subdivision in each thalamic atlas. The horizontal blue dashed line represents WMD or PC values of cortical provincial or connector hubs (arbitrarily defined as >90% of all cortical ROIs). The horizontal gold bar in each individual box plot represents the graph metrics calculated on the level of thalamic subdivision. Abbreviations for the Oxford-FSL atlas are as follows: M, motor; O, occipital; PFC, prefrontal; PL, parietal; pM, premotor; S, somatosensory; T, temporal. Box plot percentiles (5th and 95th for outer whiskers; 25th and 75th for box edges) calculated across voxels for each thalamic subdivision. All graph metrics were averaged across network densities (see Materials and Methods).

Figure 6.

Distributive pattern of thalamocortical connectivity for thalamic nuclei. Cortical functional networks were most strongly connected with the following thalamic nuclei: AN, LGN, VL, VA, ventral medial (VM), IL, LP, MD, and PuM. Thalamic nuclei (labeled in yellow) are displayed on axial MR images. The bar graphs represent the distribution of connectivity strength between thalamic nuclei and each of the nine cortical functional networks. The dashed line represents the expected proportion of total connectivity if connections were equally distributed across networks.

Figure 7.

Cognitive flexibility score and cognitive components. A, Kernel density plot and box plots of cognitive flexibility scores for thalamic voxels and cortical ROIs. Thalamic voxels were categorized into two categories of thalamic nuclei. Box plot percentiles (5th and 95th percentiles for outer whiskers; 25th and 75th for box edges) were calculated across voxels for each type of thalamic nucleus or across cortical ROIs for each type of cortical ROI. B, Spatial distribution of brain activity engaged by each cognitive component recruited by the thalamic nucleus VL.

Figure 8.

The effect of a focal thalamic lesion on cortical network organization. A, MRI scans of thalamic lesions (marked in red) in four patients (S1–S4). The lesion size for each patient in each thalamic subdivision (left, based on Morel atlas; right, based on the functional parcellation atlas) is summarized in bar graphs. B, Individual patient's normalized modularity score across network densities (x-axis) for the whole cerebral cortex. C, Individual patient's normalized connectivity strength for between-network and within-network connectivity.