Receptor-signed chemical wiring predicted a small part of head-neuron odour and salt responses in 21 and 7 immobilized C. elegans after removal of a gap-junction-organized mode

Question: in whole-brain recordings with identified neurons, how much of the response of non-sensory head neurons to odour and salt does the wiring diagram predict, when the measured sensory responses are the input? Data: NeuroPAL recordings of Yemini et al. (2021), 21 OH16230 and 7 OH15500 immobilized worms, 10 s pulses of 2-butanone, 2,3-pentanedione and NaCl. Maps: Cook 2019 and two Witvliet 2021 adult connectomes, Fenyves 2020 receptor-based synapse signs, the Randi 2023 optogenetic atlas and a peptide map. Results: (1) With a sham-calibrated 1% threshold, 10 of 33 sensory units but only 5 of 71 non-sensory units responded reliably (AIZ, AVH, RIA, RIR, RME). (2) The first principal component of spontaneous activity (reversal against forward neurons) followed gap junctions: a unit's loading correlated with its gap partners' mean loading at r = 0.45, 0.64 and 0.65 in the three connectomes (rewired networks about -0.03, p < 0.002), also with soma-distance and cell-category controls; chemical synapses gave 0.27, -0.10 and -0.12. (3) After removal of this mode component, one-step chemical wiring with receptor-based signs predicted the remaining responses in all six tests (3 connectomes x 2 strains): gain over shuffled sensory identities 0.09-0.13 in Spearman rho, p = 0.008-0.056. Shuffled sign labels removed the prediction (p = 0.003; p < 0.003). Unsigned wiring, the gap junction map and the functional atlas did not predict the residual responses. (4) Mean responses correlated with the mode loadings (rho -0.55 at pentanedione onset), but sham windows gave -0.66 to +0.36, and per-worm tests found no robust stimulus-locked mode shift. (5) The random left/right position of AWC-ON and AWC-OFF gave no detectable lateral effect in AWC partners (p = 0.16). We also report data pitfalls: a start transient, cyclic stimulus orders that confound carry-over with identity, a crosstalk rule that removes 17% of traces, and AWC labels whose odour responses are reverse to Wes and Bargmann 2001. Main limitation: the gains are small (rho 0.12-0.17), the signs are expression-based predictions for 342 of 979 sensory edges, the worms were paralyzed, and the connectomes come from other animals.

C. elegans · NeuroPAL · whole-brain calcium imaging · chemosensation · connectome · gap junctions · synapse sign · global brain dynamics · structure-function · functional atlas · Witvliet · Cook 2019 · AWC asymmetry · negative control calibration