Relative spike timing spread was 11.5% lower with neighbor light in 94 adult mouse TRN maps under stated corrections

We tested spike timing dispersion in public adult mouse thalamic reticular nucleus (TRN) recordings. We used 477 neighbor tests from 94 maps on 71 dates. We selected valid paired trials without a new baseline coefficient of variation (CV) cutoff or a coupling-call requirement. Under stated light corrections, the equal-map geometric mean ratios were 0.930 for mean latency, 0.823 for timing standard deviation (SD), and 0.885 for CV. The 95% date bootstrap range for CV was 0.833 to 0.945. These were planned secondary summaries. The two primary within-map distance tests used 470 tests from 87 maps. SD and CV ratio changes per extra 25 micrometres were 1.054 and 1.038. Both had Holm-adjusted p=0.112. A planned analysis used separate trial halves for mean change and dispersion. Its pooled CV factor was 0.955 per 0.1 lower training log mean ratio, with range 0.922 to 0.980. The early-to-late direction remained uncertain. Correction methods differed. A model added after these results showed that faster threshold crossing can reduce CV with unchanged input noise. The result does not identify noise suppression, gap conductance, direct anatomical edges, or in vivo function. All 12,800 trial latencies were rebuilt exactly from 640 raw files. Independent numerical checks passed. A clean replay reproduced 43 result files. The package preserves compact observations, source versions and hashes, code, all results, three figures, and repeat steps.

electrical synapses · thalamic reticular nucleus · spike timing · latency variability · connectome · light stimulation · paired trials · noise