A multimodal gradient architecture for the human pulvinar

2097 

Abstract Submission 

Gianpaolo Basile¹, Salvatore Bertino², Giuseppina Rizzo², Demetrio Milardi¹, Giuseppe Anastasi¹, Alberto Cacciola¹

¹Dpt of Biomedical, Dental Sciences and Morphological and Functional Images, University of Messina, Messina, Italy, ²Dpt of Clinical and Experimental Medicine, University of Messina, Messina, Italy

Gianpaolo Basile, Dr.  
Dpt of Biomedical, Dental Sciences and Morphological and Functional Images, University of Messina
Messina, Italy

Salvatore Bertino, Dr.  
Dpt of Clinical and Experimental Medicine, University of Messina
Messina, Italy

Giuseppina Rizzo, Dr  
Dpt of Clinical and Experimental Medicine, University of Messina
Messina, Italy

Demetrio Milardi, Prof.  
Dpt of Biomedical, Dental Sciences and Morphological and Functional Images, University of Messina
Messina, Italy

Giuseppe Anastasi, Prof.  
Dpt of Biomedical, Dental Sciences and Morphological and Functional Images, University of Messina
Messina, Italy

Alberto Cacciola  
Dpt of Biomedical, Dental Sciences and Morphological and Functional Images, University of Messina
Messina, Italy

Integrative processes are fundamental for our brain to transform input signals from multiple sensory channels into a detailed, hierarchically organized, and meaningful representation of the environment. Pulvinar, the largest thalamic nucleus, is likely to play a leading role in such associative processes by means of a rich set of topographical connections to various cortical areas (1). Anatomical evidence in non-human primates suggests that such connectivity patterns are not limited to spatially discrete units, such as histological sub-nuclei; rather, they are organized in a continuum that mirrors cortico-cortical connectivity, reflecting cortical hierarchies of information transfer, and corresponding to some extent to cytochemical markers (2). Here, we directly address the question of continuous transitions in pulvinar-cortical connectivity profiles and their relationship to cortico-cortical connectivity as well as to neurotransmitter expression, by leveraging state-of-art gradient mapping techniques on multimodal imaging data.

We employed 3T structural, diffusion, and resting-state functional MRI datasets of 210 healthy subjects (males=92, females=118, age range 22-36 years) from the HCP repository (3). Receptor expression data were obtained from a recently published, multicentric, multi-tracer positron emission tomography atlas (4). For each left and right pulvinar voxel, the following measures were assessed: 1) functional connectivity (BOLD signal Pearson's correlation) to 400 cortical regions from preprocessed, denoised rs-fMRI data; 2) structural connectivity to 400 cortical regions from preprocessed diffusion data (CSD signal modeling; probabilistic whole brain tractography; connectivity measure: streamline count); 3) averaged, normalized density values for 30 receptors including serotonergic, dopaminergic, cholinergic, and glutamatergic markers. For each of these voxel-wise features independently, diffusion embedding was employed to infer gradient maps of spatial variability within the pulvinar (distance metric: cosine similarity; alpha=0.5) (5). Gradient-weighted structural and functional connectivity maps were retrieved to explore the cortical connectivity patterns associated with each gradient map. Receptor expression gradients were identified by their top 5 most correlated receptor markers. Cortico-cortical functional and structural connectivity gradients were also obtained from the same data for further comparison.

We identified three functional connectivity gradients explaining ~90% of the variance of our data. Their cortical connectivity profiles were highly correlated to the first three cortico-cortical connectivity gradients. The main gradient, spanning from unimodal to transmodal cortical areas, was organized on the dorso-ventral axis of the pulvinar and showed high correlation to the main receptor expression gradient (left: r=-0.69 ; right: r=-0.78) reflecting changes in the expression of serotonergic and dopaminergic markers, and to the secondary structural connectivity gradient (left: r=0.71; right: r=0.78). The secondary gradient was correlated to the third cortico-cortical gradient, reflecting a transition in connectivity from multiple-demand to paralimbic areas. Finally, the third gradient, associated with the secondary cortico-cortical gradient (visual to sensorimotor region), was organized on the medio-lateral axis; it was found correlated to the main structural connectivity gradient (left: r=-0.58 ; right: r=-0.60), and to the secondary receptor expression gradient (left: r=-0.47; right: r=-0.71)

The human pulvinar hosts multiple representations of cortical connectivity, reflecting cortico-cortical gradient hierarchy. Our results support the hypothesis that cortico-pulvinar connectivity mirrors cortico-cortical connectivity and provide novel insights on the relation between pulvinar connectivity patterns and major neuromodulator systems.

Connectivity (eg. functional, effective, structural)

Task-Independent and Resting-State Analysis

Anatomy and Functional Systems ¹

Subcortical Structures ²

Transmitter Receptors

CHEMOARCHITECTURE

Data analysis

FUNCTIONAL MRI

Modeling

Positron Emission Tomography (PET)

RECEPTORS

Sub-Cortical

Thalamus

Tractography

Other - Gradient