Sensory recurrence and convergent integration in the C. elegans connectome
- 1Concordia University
Abstract
The synaptic wiring of Caenorhabditis elegans has a global feed-forward architecture, from sensory to motor neurons. Yet, whole-brain optogenetic mapping shows that signals do not flow predominantly forward, and sensory neurons respond to more neurons than they drive. Where the wiring contains pathways leading back to their starting point is unclear. Here, we locate this recurrence using the feedback coefficient, derived from the Kubo-Martin-Schwinger framework for directed networks, which quantifies the proportion of walks reaching a neuron that originate from itself. The connectome has about three times more recurrence than strength-matched random networks, with the largest excess in sensory neurons, whereas random wiring places it on highly connected neurons. Along the thermotaxis pathway, this produces a recurrence gradient from thermosensory neurons to integrating interneurons. The ordering replicates in five reconstructions, including two independent adults, persists when reciprocal contacts are preserved or gap junctions removed, but vanishes with only short paths or loop counts. Convergence shows the reverse pattern, with every neuron receiving walks from fewer sources than in random wiring, especially sensory neurons. The wiring extends from a sensory periphery folding back on itself to interneurons and motor neurons integrating many sources, a candidate structural substrate for non-feed-forward signalling.
Related
- The C. elegans Connectome Toolbox: an open, extensible framework for multimodal connectomic data integration and analysis
- What an atlas-fitted connectome can and cannot do: evolutionary search over interneuron stimulation in a whole-body C. elegans model
- Clonally Convergent Tph/Tfh-like CD4+ T Cells are Shared Across Recovery and Recurrence in Experimental Periodontitis
- Null-model treatment of the sensory-motor boundary changes an evolutionary connectome comparison
- A C. elegans model for functional analysis of conserved ADPKD variants in cilia, extracellular vesicles, and sensory signaling
BibTeX
@misc{moutuou2026sensory,
title = {{Sensory recurrence and convergent integration in the C. elegans connectome}},
author = {E. Moutuou and H. Benali},
howpublished = {bioRxiv},
year = {2026},
month = oct,
doi = {10.64898/2026.10.29.762790},
url = {https://www.biorxiv.org/content/10.64898/2026.10.29.762790v1},
note = {CC BY},
}
RIS
TY - UNPB
TI - Sensory recurrence and convergent integration in the C. elegans connectome
AU - E. Moutuou
AU - H. Benali
PY - 2026
DA - 2026/10/29
PB - bioRxiv
DO - 10.64898/2026.10.29.762790
UR - https://www.biorxiv.org/content/10.64898/2026.10.29.762790v1
AB - The synaptic wiring of Caenorhabditis elegans has a global feed-forward architecture, from sensory to motor neurons. Yet, whole-brain optogenetic mapping shows that signals do not flow predominantly forward, and sensory neurons respond to more neurons than they drive. Where the wiring contains pathways leading back to their starting point is unclear. Here, we locate this recurrence using the feedback coefficient, derived from the Kubo-Martin-Schwinger framework for directed networks, which quantifies the proportion of walks reaching a neuron that originate from itself. The connectome has about three times more recurrence than strength-matched random networks, with the largest excess in sensory neurons, whereas random wiring places it on highly connected neurons. Along the thermotaxis pathway, this produces a recurrence gradient from thermosensory neurons to integrating interneurons. The ordering replicates in five reconstructions, including two independent adults, persists when reciprocal contacts are preserved or gap junctions removed, but vanishes with only short paths or loop counts. Convergence shows the reverse pattern, with every neuron receiving walks from fewer sources than in random wiring, especially sensory neurons. The wiring extends from a sensory periphery folding back on itself to interneurons and motor neurons integrating many sources, a candidate structural substrate for non-feed-forward signalling.
ER -