Where neurons sit in the spinal cord may matter as much as what they are:
Our new paper is out in Nature Communications: Spatial and network principles behind neural generation of locomotion.
Decades of genetics have given us a rich catalogue of spinal cell types, but the mapping from genetic identity to motor function has stayed stubbornly loose — most functions turn out to be distributed across several classes. We propose that the missing variable is space.
Open access here:
https://lnkd.in/dk5J_tfM
We built a model of the mouse lumbo-sacral cord in which cell positions come from single-nucleus RNA sequencing and spatial transcriptomics, and synapses are sampled probabilistically from cell-type-specific projection patterns. No functional modules are imposed, and no parameters are optimized.
What emerges is an asymmetric “Mexican hat” projectome — local excitation, longer-range inhibition, skewed along the rostro-caudal axis. That single motif produces locomotor rhythm, rotational population dynamics, left–right alternation, and traveling waves of activity down the cord. Descending glutamatergic and serotonergic input reshapes the projectome’s symmetry, and the model transitions from walking to trotting to bounding. Silencing V1 or V0 in silico reproduces the published in vivo phenotypes, including the counterintuitive slowdown after V1 loss.
The practical upshot: a spinal model with explicit physical structure and transcriptomic identity is a natural starting point for thinking about spinal cord injury — which projections are lost, and which would need to be restored.
Congratulations to Salif Komi and all co-authors, and to everyone who helped.
Thank you to our funding:
The Lundbeckfonden / Lundbeck Foundation (no. R366-2021-233, R.W.B.), Novonordisk Foundation (no. NNF23OC0082192, R.W.B.), The Swiss National Science Foundation (no. SNF Postdoc.Mobility P500PB_206824, S.K.), and EIC Pathfinder Open Project 101130161- Project Move2Treat – EU HORIZON EIC 2024 PATHFINDER OPEN