After identity, pre-stimulus signal and distance controls, nine unsigned wiring terms gave no clear target selection gain in a ridge test of 1,225 stimuli from 103 C. elegans worms

We tested whether wiring helps select the recorded downstream neuron with the largest measured fluorescence change after a stimulus. The test used 1,225 stimuli and 52,140 target measurements from 103 C. elegans worms. All test worms were absent from fitting and tuning. The reference used target identity, pre-stimulus signal values and soma distance. A ridge rule added nine unsigned wiring features. Mean selected response above the event mean rose from 0.553 to 0.596 in transformed units. The paired gain was 0.042. Its central 95% range was -0.005 to +0.099 when predictions stayed fixed. In 199 whole-worm resamples with model refits and fixed selected penalties, the range was -0.038 to +0.079. One worm with one usable stimulus supplied 46% of the net paired gain. Thus the main test gave no clear added choice value. The outcome is asinh of green fractional change divided by 0.1. Selection was among usable recorded targets and did not establish a causal effect or the true response order. The package provides a benchmark for target selection, source code, exact inputs, model states, predictions, figures, source versions and checksums, and steps to repeat the study. It also preserves the planned sensitivity tests and independent source and solver checks.

connectome · target selection · Caenorhabditis elegans · optogenetics · calcium imaging · animal validation · structure and function