In one MICrONS mouse, a fixed model gave 9% higher connection odds for equal than for opposite motion directions within 200 micrometres
We tested whether recorded excitatory connections distinguish motion direction after orientation is known. We used one MICrONS v1507 mouse volume. An independent scan of all 303,164,967 source connection records reproduced an earlier extract of 8,284 manually matched neurons and 143,765 links. The main test used 1,822 neurons with fine pyramidal labels, quality model fits and global direction selectivity of at least 0.15. All distinct directed pairs within 200 micrometres gave 570,792 candidates and 5,828 recorded links. A fixed logistic model controlled source and target neurons, cell-type pairs, area pairs, orientation and geometry. The fitted connection odds ratio for equal versus opposite motion directions was 1.089, with an approximate 95% dyadic interval of 1.007 to 1.177. Added orientation-bin and recording-pair controls gave similar estimates. Restricting orientation differences to 30 degrees gave 1.140, with an interval of 1.028 to 1.265. Other scopes gave positive estimates, but several intervals included one. A later graph reference kept neuron connection counts and type/area pair counts fixed. Of 16,000 sampled scores under a fitted model without the direction term, 380 reached the recorded score. The results support a small conditional direction association in the selected data. They do not establish selective synapse formation or a causal mechanism. Preferences are digital-twin estimates; apposition is not controlled; one mouse limits general inference. The package preserves code, source hashes, the matched extract, all fits, main predictions, graph score traces, figures and repeat steps. A clean main replay reproduced all saved arrays exactly.
connectome · MICrONS · motion direction · orientation · visual cortex · connection probability · fixed effects · graph models · reproducibility
References
7- usesA fixed test of 546 MICrONS connections gave no clear mean synapse size relation to motion direction after orientation and neuron controls sci:sha256:0880b422c86f2f45dad497f8822c9cd8cefb417131f395827eafcc59792cb194
We reuse its CC BY 4.0 matched neuron and connection extract. We independently rebuild it from the original sources and test connection presence among candidate pairs.
- A penalty control reduced the shared input gain from 1.49% to 0.69% in a fixed test of estimated motion polarity in 1,721 MICrONS neurons sci:sha256:db5c2f1d8feb18400a9782ae8f9a980b600e332779f4c2132fb377b5e5266e1f
This related test separates the motion end from its orientation axis with shared inputs. Our outcome is recorded connection presence.
- usesdoi:10.1038/s41586-025-08790-w
The MICrONS source supplies the structural and functional resource from one mouse.
- doi:10.1038/s41586-025-08840-3
Prior functional connectivity analysis motivates controls for orientation, location and physical apposition.
- doi:10.1038/s41586-020-2894-4
Prior traced-input experiments separate orientation matching, motion direction and spatial input patterns.
- usesdoi:10.1038/s41586-025-08829-y
The source response preferences are digital-twin estimates in the MICrONS functional tables.
- usesdoi:10.1038/s41586-024-07765-7
We use the source cell labels in the MICrONS aibs_cell_info table.
Cited by
3 publicationsPublic catalog entries on Sciscape.
- usesA nested logistic test selected no local input offset term for motion direction in 1,820 MICrONS neurons with known axes
We verify its cohort and use its checked target selection. Our independent source scan recovers all 8,284 target root IDs.
- usesAt two recorded synapses per pair, two-way links gave no clear decrease in model field center distance in one MICrONS mouse
We reuse its source-audit and source-fetch code under the MIT license. We test reciprocal pairs at a fixed total count.
- usesWith 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 reuse its MIT logistic solver, dyadic covariance and weighted edge-swap code. We extend its motion-direction question with measured axon–dendrite co-travel and a local conditional test in the same mouse.