With co-travel controls and gDSI ≥ 0.10, a fixed model gave 27% higher connection odds for equal than opposite estimated motion directions in one MICrONS mouse

We tested whether cortical connections distinguish motion direction after orientation and physical opportunity are known. We retained all 295,371 released axon–dendrite proximity pairs from 148 source axons in one MICrONS mouse. Motion fields matched all 12,894 functional records exactly by cell and recording. The main test required global direction selectivity of at least 0.10 and fixed model quality cuts. It had 45,314 candidate pairs and 1,323 recorded connections. A logistic model controlled cell send and receive rates, orientation, position, model field distance and measured co-travel length. Its fitted odds ratio for equal versus opposite directions was 1.270, with an approximate 95% interval of 1.064–1.517. Removing co-travel gave 1.264 on the same fitted support. All 67 source deletion estimates were 1.225–1.313. A later graph reference fixed connection counts and allowed pairs. Of 16,000 stored scores, 130 reached the observed direction score. A later local conditional test removed cell intercepts directly. Its ratio was 1.235, with an approximate interval of 0.955–1.598. Thus, measured physical opportunity did not remove the main fitted direction association. The local conditional test remained uncertain. The data come from one animal, and motion preferences are model estimates. These results do not establish a causal wiring rule or a difference between animals. The package preserves the full opportunity extract, provenance, code, model states and figures.

connectome · MICrONS · motion direction · axon-dendrite proximity · functional specificity · conditional likelihood