Tracing cortical functional connectivity of histological defined thalamic nuclei

Introduction: The thalamus is made up of a number of individual nuclei, which are closely connected to each other and to cortical and subcortical brain regions (1, 2). However, despite its central role cortical and subcortical thalamic connections were mainly experimentally determined in animal models (3). Only the use of diffusion imaging and resting state fMRI provides novel insights into the in vivo human brain anatomy including the thalamus. DTI allows to examine structural connectivity of the thalamus to identify connections to cortical areas (4), and rsfMRI can determine the functional connectivity between the thalamus and cortical areas (5). Here we examined in functional connections using a 3D anatomical atlas thalamic (6) to determine the cortical connectivity of cytoarchitectonically defined thalamic nuclei. Methods: The 3D volumes of 29 thalamic nuclei from the atlas of Morel (7) were transformed into the MNI space and a connectivity analysis was performed by using rs-fMRI data obtained from the Human Connectome Project (HCP) data set (8). Four resting state sessions (Gradient-echo EPI, 1200 scans per session, TR: 720ms, TE: 33.1 ms, FOV: 208x180, 2 mm isotropic, 72 slices, multiband factor: 8) of 40 subjects were chosen from the HCP data set. Data were preprocessed and ICA denoised using the available HCP pipeline (9) and FSL Fix tool. Data were smoothed with 3 mm kernel and partial regressions were computed for all thalamic nuclei to cerebral cortex (10). The fixed effect analysis was performed to calculate group maps and winner-takes-it-all maps on the level of cortical surface. Results: Anatomy: According to anatomy (6) thalamic nuclei are divided in 29 defined nuclei (s. Fig. 1): anterior (AD, AM, AV, LD), lateral (VA, VL, VM, VPI, VPL, VPM); posterior (LGN, Li, LP, MGN, Po, PuA, Pul, PuL, PuM, SG) and medial (CeM, CL, CM, Hb, MD, MV, Pf, Pv, sPf). Connectivity: The fixed effect maps (s. Fig. 2) are depicted in the according color code (threshold 0.002). In both hemispheres, the majority of projections are confined to the occipital, parietal, lateral prefrontal lobes and peri-central region, while the temporal, medial, and inferior frontal lobes are spared. In addition, a number of nuclei project to multiple sites and show a considerable overlap. Winner-takes-it-all: In a second step, each vertex point was assigned to the strongest correlated nuclei maps (s. Fig. 3). This revealed that a number of nuclei are connected to a single cortical region (left: LD, LP, MD, Pf, Po, PuA, Pul; and right: CM, Hb, LD, LGN, Li, Po, PuA, Pul, PuM, VA), while other are connected to 2-10 cortical clusters (left: Left CM, Li, PuL, VPM, VPL, AM, MV nuclei and right PuL, pf, LP, VPM, SG, VPL, VL, AM). In addition, a number of nuclei (left LGN, CL, sPf, AV, VL, AD, VPI and right sPf, AV, VL, AD, VPI) are connect to 17-54 cortical clusters. Hemispheric differences: The sum of cortical areas was determined (s. Fig 4A). The areas size varies among the nuclei (left: 0 – 20246; right: 0 - 10002) and hemispheres. (s. Fig 4B). Nuclei PuA, Pf, VPM, LP, SG, VPL, AM, CL, and AV were right dominant (area 22 – 2733 mm2), while Po, MV, MD, Li, sPf, AD, LGN, VPI and VL nucleus were left dominant (area 99-3249 mm2)