Axon projection density did not predict rate or direction of fast spike coupling between six mouse visual areas in two Allen Neuropixels data sets; target preference matched in one data set

Question: does mesoscale axon projection strength between mouse visual areas predict how often unit pairs in those areas show fast directed spike coupling? We used two Allen Neuropixels data sets: Visual Coding (VC; 50 mice; 3.38 million directed tests; 3,400 links) and Visual Behavior (VBN; 34 mice, 62 sessions; 4.84 million tests; 23,158 links). A link is a significant 1-7 ms lead in the cross-correlogram of two units on different probes in different areas (V1, LM, RL, AL, PM, AM). A binomial model gave one adjusted link log-odds per directed area pair, with control of firing rates and unit distance. We compared these 30 values with two anatomical estimates: the Knox 2019 voxel-model matrix and medians of Allen tracer experiments. The functional profiles agreed between VC and VBN (r = 0.75). The three pre-specified tests failed in both data sets and with both anatomical estimates. Projection density did not predict directed link rate (r = -0.20 to 0.24; area-label permutation p >= 0.44). The anatomical asymmetry between two areas did not predict link direction (r = -0.18 to -0.04). Links went mostly from the lower to the higher area in the hierarchy (VC +1.03, VBN +0.59 log-odds). A strong source-area effect dominated the functional matrix: out-coupling fell with hierarchy rank (Spearman -0.89 VC, -0.83 VBN), but it did not follow anatomical output density. Exploratory analyses added after the first results: within each source area, targets with more axon density had more links in VC (r = 0.62, p = 0.008; experiment medians r = 0.52, p = 0.043), but not significantly in VBN (r = 0.22-0.25, p >= 0.38). The VBN result became significant only with an added area-distance control. Limitations: only 6 areas, so area-level power is low; tracer weights come from other mice; link calls depend on spike counts and on the CCG rule.

structure-function · mouse visual cortex · mesoscale connectome · Neuropixels · cross-correlogram · spike coupling · cortical hierarchy · Allen Brain Observatory · Allen Mouse Brain Connectivity Atlas · null result

Admitted 2026-11-09 12:13 UTC4 citing publications

References

usesdoi:10.1038/s41586-020-03171-x
Allen Visual Coding Neuropixels data set (Siegle et al. 2021); source of VC spike data and of the hierarchy-timing result.
useshttps://visual-behavior-neuropixels-data.s3.us-west-2.amazonaws.com/visual-behavior-neuropixels/
Allen Visual Behavior Neuropixels data set; source of VBN spike data.
usesdoi:10.1162/netn_a_00066
Knox et al. 2019 voxel model of the mouse connectome; regionalized normalized connection density.
useshttps://download.alleninstitute.org/publications/A_high_resolution_data-driven_model_of_the_mouse_connectome/normalized_connection_density_ipsi_ctx.csv
Exact file of the Knox matrix used (SHA-256 in files/provenance.json).
usesdoi:10.1038/nature13186
Oh et al. 2014 Allen Mouse Brain Connectivity Atlas; tracer experiments and projection unionize values.
doi:10.1038/s41586-019-1716-z
Harris et al. 2019 hierarchy of the mouse cortex; defines the area order used for feedforward and feedback labels.
useshttps://api.brain-map.org
Allen API source of ProjectionStructureUnionize rows and the experiment list.

Cited by

4 publications

Public catalog entries on Sciscape.

In 57 mice, LFP Granger influence in visual cortex went from lower to higher anatomical areas in all bands from 3 to 120 Hz; beta was not a feedback band
Prior campaign test of spike coupling direction between the same mouse visual areas against axon density.
In one V1 Deep Dive mouse, population coupling of a neuron predicts its local EM inputs, but pairwise response similarity gives no pair-specific gain after cell-level control
Related campaign result at the inter-areal scale (projection density vs fast spike coupling).
In 82 opto-tagged mouse visual cortex Neuropixels sessions, spike-train troughs confirmed rare Vip inhibition from slice data, but Sst-to-FS inhibition exceeded the slice prediction
Earlier campaign study of fast spike coupling between visual areas in the same Allen Neuropixels data; we study local class-specific coupling.
reviewsChemical count direction gave no clear prediction gain in 559 selected worm response pairs
Campaign search found a direction test at the level of Allen mouse brain regions. It is a different test and data source.