Back

Neural fiber orientations across spinal cord gray matter

Soerensen, S. F.; Kaur, J.; Berg, R.

2026-08-12 neuroscience
10.64898/2026.08.06.743226 bioRxiv
Show abstract

Understanding connectivity within the nervous system is central to understanding its function, and the network architecture of many brain regions has been mapped in fine detail. The sub-tract organization of neuronal projections within the human cord, particularly in gray matter is essentially unknown. Since full connectomic reconstruction is currently intractable at this scale, a tractable first step is to statistically characterize neurite orientations across spinal regions. Here we combine a [~]400 hour 9.4T ex vivo high-angular-resolution diffusion-weighted MRI spanning all segments (C1-S5) of a post-mortem human spinal cord with structure-tensor analysis of SMI-32 neurofilament fluorescence microscopy to map neurite orientation across the cord. We find that projection geometry varies systematically along the rostro-caudal axis. Cervical, lumbar, and sacral gray matter exhibit low anisotropy, high orientation dispersion, and predominantly transverse fiber populations (>91% of orientations), consistent with dense segment-local circuitry. Thoracic gray matter is categorically distinct from the rest of the cord with elevated axial diffusivity, white-matter-like anisotropy, and a fiber orientation distribution in which 60% of orientations run within a polar angle of 85-95{degrees} i.e., along the the cords rostro-caudal axis. Micron-resolution microscopy independently supports this longitudinal architecture. Together, these findings reveal a regional dissociation in spinal gray-matter wiring and suggest a coupling between projection geometry and the distinct processing demands of each segment.

Matching journals

The top 5 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.