中国神经再生研究(英文版) ›› 2023, Vol. 18 ›› Issue (12): 2661-2662.doi: 10.4103/1673-5374.373708

• 观点:脊髓损伤修复保护与再生 • 上一篇    下一篇

利用光学透明、光片显微镜和基于人工智能的图像分析推进脊髓损伤研究

  

  • 出版日期:2023-12-15 发布日期:2023-06-14

Advancing spinal cord injury research with optical clearing, light sheet microscopy, and artificial intelligence-based image analysis

Qiang Li, Alfredo Sandoval Jr*, Bo Chen*#br#   

  1. Department of Neurobiology, University of Texas Medical Branch, Galveston, TX, USA
  • Online:2023-12-15 Published:2023-06-14
  • Contact: Alfredo Sandoval Jr, BS, alfsando@UTMB.EDU; Bo Chen, PhD,Bochen1@utmb.edu.
  • Supported by:


摘要: https://orcid.org/0000-0002-7230-8325 (Alfredo Sandoval Jr) 
https://orcid.org/0000-0003-3814-2509 (Bo Chen)

Abstract: From the days of Ramon y Cajal’s first sketches, neuroscientists have recognized the importance of visualizing the complex architecture of the central nervous system. In the past century, we have come to appreciate how the rich structural and functional complementarity of axons and cell types in the spinal cord make it uniquely suited for information transfer between the periphery and the brain. However, appreciating these relationships has been limited by the low-throughput histology techniques in common usage. For example, axon projections span many spinal segments, traversing centimeters to meters (depending on the species), with many spinal nuclei also occupying more than one spinal level. Spinal cord injuries also disrupt many levels of the spinal cord, with secondary inflammatory responses reaching far beyond the epicenter and surrounding penumbra after injury. These injuries affect the primary injury site, the penumbra, and descending and ascending white matter tracts that pass through the injury site, ultimately impacting communication between the periphery, the spinal cord, and the supraspinal structures. Therefore, holistically examining spinal structures or injuries that span multiple segments with two-dimensional (2D) sections is extremely challenging. At best, 2D sections can give a snapshot of axons and neighboring cells, however, they provide no information about axon trajectory and only offer variable representations of cells in any given cross-section. Additionally, these approaches are labor-intensive and can only produce short reconstructions of projection axons in sparsely labeled samples due to the optical limitations associated with imaging thick tissue sections. These drawbacks may severely limit the novelty, scope, and impact of studies, as useful data is left uncollected due to the biased nature and technical limitations of traditional 2D histology and imaging techniques. In light of these disadvantages and the difficulty in acquiring samples from complex SCI experiments, tools that can effectively resolve structures across multiple scales, from the tissue to the subcellular level, while preserving spatial relationships, are highly needed.