With 1 s bins, state mean changes supplied 37% of the V1 correlation contrast between cell types ordered by gene expression in four mice
We tested how behavioral state changes contribute to a reported activity pattern across molecular cell types in mouse primary visual cortex (V1). We used Bugeon et al. figshare version 2: 1,065 unique inhibitory cells, 11 broad types, and 17 sessions from four mice. We ordered types by the first principal component of 72 raw gene counts. We rebuilt type-mean activity in 1 s bins. The rebuilt correlation matrix closely matched the published matrix across 55 distinct pairs (r = 0.9995). We split each correlation into covariance of state means and covariance within states, with total standard deviations in both terms. The tested contrast compares pairs one or two steps apart in the type order with more distant pairs. The total contrast was 0.10675. State mean changes supplied 0.03903, or 36.6%. Within-state activity supplied 0.06773, or 63.4%. The within-state contrast was positive in all four mice and after transition removal, continuous behavior fits, time-block fits, and type-removal checks. Equal mouse weights gave a state mean share of 38.6%. An exploratory check on the same 32 pairs kept a positive within-state contrast in every mouse. The code, compact traces, source checksums, figures, and repeat steps are preserved. This is a descriptive split of type-mean calcium activity. Coarse state labels, unmeasured behavior, common input, cell sampling, and calcium processing can affect the residual. There is no matched synaptic map. The result does not identify direct connections or establish a population effect beyond these four mice.
cell type · transcriptomics · calcium imaging · behavioral state · functional connectivity · covariance · mouse V1 · reproducible research
References
5- doi:10.1038/s41586-022-04915-7
[Legacy relation: reproduces] We rebuilt the Figure 5d type correlation matrix and then split its ordered pattern into between-state and within-state covariance.
- usesdoi:10.6084/m9.figshare.19448531.v2
This version supplies the activity traces, state labels, gene counts, type calls, cell identities, and behavior data.
- usesIn four mice, the published transcriptomic axis of V1 interneurons predicted state modulation mainly between subclasses and types, and added no held-out-mouse prediction beyond subclass labels sci:sha256:ae8f8a75666959531984573cd28c08243c10c1ffa9aac892e750e1a11d29232a
We used its source-file access, factual subtype-to-type map, and prior audit to define a distinct correlation question. We rebuilt all new numerical analyses from the original data.
- doi:10.1523/JNEUROSCI.4994-13.2015
This experiment showed a relation between cortical state and shared activity. It motivates a state split before a circuit interpretation.
- doi:10.1038/s41586-022-05209-8
The source article correction changes the Figure 6 time-constant axis. It does not change the Figure 5d values used here.
Cited by
0 publicationsPublic catalog entries on Sciscape.
No incoming citations on Sciscape.