Connectome structure constrains neural activity across scales

In Caenorhabditis elegans, a parameter-free predictor using connectome paths forecasts stimulus-evoked neural responses far better than direct synaptic connectivity, with most information in two-step paths.

• Why it matters: Coarse-graining connectomes may erase structure that matters for neural dynamics, but it is unclear how much resolution is needed to preserve functional predictions.
• What they did: The study compared direct connectivity versus a path-based predictor across connectome resolutions, then tested coarse-graining by biological cell classes, random partitions, and connectivity-guided alternative groupings.
• The result: Cell-class coarse-graining preserved predictions despite discarding path structure, random partitions degraded them, and connectivity-guided partitions preserved prediction down to about nine groups; lineage-based partitions in Drosophila preserved path structure better than random coarse-graining.

AI summary of the abstract · may contain errors · not medical advice · read the paper.