High-resolution subregion parcellation of subthalamic nucleus based on voxel-level connectivity
-
1
Korea Institute of Science and Technology, Center for Functional Connectomics, Republic of Korea
-
2
University of Science and Technology (UST), Neuroscience Program, Republic of Korea
Mapping detailed connectivity and delineating brain regions by anatomical and functional profiles are crucial for understanding brain function. With the advent of novel imaging techniques and data analysis algorithms, connectivity-based approaches have shown to further divide previously defined topographical region into distinct subregions based on long-range connectivity patterns in cortex and other brain areas. We aim to map the connectivity-based subdivision of the subthalamic nucleus (STN), an area known to play a critical role in information processing within basal ganglia circuitry. Recently, STN has become the primary target of deep brain stimulation (DBS) in treatments of various neurological and psychiatric disorders, making it essential to understand detailed functional subregions within and topological patterns of inputs to STN. Still, partly because of its small size and previous difficulties in precisely delineating its structure, the functional and anatomical connectivity patterns of the STN circuitry require more study.
Previous works, based on functional activation pattern similarity of regions from functional Magnetic Resonance Imaging (fMRI) and tractography using Diffusion Tensor Imaging (DTI) to find functional/structural connectivity patterns, were often limited to large cortical-region-driven approaches instead of full voxel-by-voxel analysis between brain regions. Moreover, because of the low resolution offered by these techniques, identified subdivisions were often relatively coarsely partitioned, making these techniques inapplicable to small brain regions. On the other hand, fluorescent tracer injection studies can provide more precise cellular-level, anatomically accurate connectivity information that can be used for fine-grained parcellations of even very small brain areas. Also, it can easily be extended to study cell-type specific connection patterns, providing additional valuable information that can be used as criteria for more complex functional parcellation.
Despite the ability to segment arbitrary regions into finer subdivisions, full voxel-by-voxel approaches are often hindered by inter-subject variability and lack of measurement data. Since the number of unknowns (number of voxels in region of interest) greatly exceed the number of available experiments, such an approach is generally considered intractable without accepting the hypothesis that the data can be well represented in low-rank subspace or by the use of appropriate constraints that not only efficiently narrow down the solution space but that also capture intrinsic biological properties of the data.
Motivated by this, we focus on two challenging key tasks: (1) using fluorescent protein-expressing viral tracers for constrained full voxel-by-voxel structural connectivity analysis in external globus pallidus (GPe)-STN and cortico-subthalamic circuits; (2) using nonparametric clustering of connectivity patterns in a 3-D spatial map for fine-level subregion parcellation of both GPe and STN.
We apply a dictionary learning based Blind Source Separation (BSS) algorithm to find the sparse representation of axonal projection data with dictionary atoms that effectively capture biological properties. We then estimate a voxel-level connectivity matrix using multivariate sparse regression techniques with constraints that promote smooth and spatially coherent/localized projection patterns. Finally, we propose a non-parametric probabilistic connectivity-based clustering method that automatically estimates the number of subgroups and also imposes spatial contingencies within subregions.
Keywords:
connectomics,
parcellation,
Brain Mapping,
tracer injections,
connectivity,
axonal projection,
STN-DBS,
Subthalamic Nucleus
Conference:
Neuroinformatics 2016, Reading, United Kingdom, 3 Sep - 4 Sep, 2016.
Presentation Type:
Poster
Topic:
General neuroinformatics
Citation:
Jeon
H,
Feng
L,
Lee
H,
Oh
W and
Kim
J
(2016). High-resolution subregion parcellation of subthalamic nucleus based on voxel-level connectivity.
Front. Neuroinform.
Conference Abstract:
Neuroinformatics 2016.
doi: 10.3389/conf.fninf.2016.20.00054
Copyright:
The abstracts in this collection have not been subject to any Frontiers peer review or checks, and are not endorsed by Frontiers.
They are made available through the Frontiers publishing platform as a service to conference organizers and presenters.
The copyright in the individual abstracts is owned by the author of each abstract or his/her employer unless otherwise stated.
Each abstract, as well as the collection of abstracts, are published under a Creative Commons CC-BY 4.0 (attribution) licence (https://creativecommons.org/licenses/by/4.0/) and may thus be reproduced, translated, adapted and be the subject of derivative works provided the authors and Frontiers are attributed.
For Frontiers’ terms and conditions please see https://www.frontiersin.org/legal/terms-and-conditions.
Received:
30 Apr 2016;
Published Online:
18 Jul 2016.
*
Correspondence:
Mr. Hyungju Jeon, Korea Institute of Science and Technology, Center for Functional Connectomics, Seoul, 136-791, Republic of Korea, hyungju.jeon@gmail.com