Cooperative Subnetworks of Molecularly Similar Interneurons in Mouse Neocortex

Cooperative Subnetworks of Molecularly Similar Interneurons in Mouse Neocortex Karnani Jackson Ayzenshtat Tucciarone Manoocheri Snider Yuste

Tag layer of the DHSS Meta-Tagging Project. The article text is reproduced from the staged copy named in sourceDesc and is not altered; the tags are held separately in standOff so that overlapping and nested spans are preserved.

Cooperative Subnetworks of Molecularly Similar Interneurons in Mouse Neocortex 2016 Neuron 10.1016/j.neuron.2016.02.037 Karnani2016-cooperative-subnetworks-of-molecularly-simil Karnani2016-cooperative-subnetworks-of-molecularly-simil.txt, the project's staged copy f6397a9b05f4a454e22d374fdf9d4352ffb62b4a not-redistributable neuroscience

the paper's own words

the tag layer; a sentence written by the tagger about the paper, always anchored to a quotation but not itself a quotation

0 spans located by searching the stored text for the quote's exact words, allowing any whitespace between them whether members of a molecularly defined interneuron class fire together as a population rather than individually, and what circuit mechanisms would make that happen Quotation withheld because this record is not cleared for redistribution. how interneurons coordinate their activity within a subpopulation, given that the populations inhibit each other and sit in a densely interconnected network Quotation withheld because this record is not cleared for redistribution. four complementary mechanisms that recruit members of the same interneuron class, and a proposal to replace the hierarchical model with reciprocally inhibitory population dyads Quotation withheld because this record is not cleared for redistribution. mechanism Quotation withheld because this record is not cleared for redistribution. within-population co-activity of neocortical interneurons Quotation withheld because this record is not cleared for redistribution. electrophysiology electrophysiology Quotation withheld because this record is not cleared for redistribution. mechanisms Four converging circuit mechanisms recruit members of the same interneuron class: electrical coupling, absent or scarce within-population inhibition, disinhibition via inhibition of the reciprocally inhibitory class, and cholinergic self-excitation of VIPs. duplicate of: mechanisms[4] Quotation withheld because this record is not cleared for redistribution. mechanisms VIPs inhibiting SOMs (and vice versa) removes inhibition from their own population members, so firing in a few cells raises the co-activity of the whole population. Quotation withheld because this record is not cleared for redistribution. mechanisms Acetylcholine produced by VIPs recruits further VIPs that are themselves ACh-sensitive, closing a positive feedback loop. Quotation withheld because this record is not cleared for redistribution. mechanisms Strongly facilitating, slow-kinetics inhibitory synapses between SOMs and VIPs grow during sustained firing, sustaining disinhibition of the partner population over hundreds of milliseconds. Quotation withheld because this record is not cleared for redistribution. four mechanisms recruit members of the same class: electrical coupling, absent or scarce within-population inhibition, disinhibition through the reciprocally inhibitory population, and self-excitation of VIPs Quotation withheld because this record is not cleared for redistribution. and a fifth from outside: the surrounding excitatory network drives each population with its own synaptic dynamics, depressing onto VIPs and facilitating onto SOMs Quotation withheld because this record is not cleared for redistribution. the recruitment of VIPs by exogenous ACh and their mecamylamine-sensitive population-recruitment, combined with their ability to produce ACh, suggest that ambient ACh recruits a positive feedback loop via VIPs Quotation withheld because this record is not cleared for redistribution. separating the excitatory sources of two mutually inhibitory populations is likened to agonist and antagonist muscles, where contracting both at once would waste energy Quotation withheld because this record is not cleared for redistribution. the proposed function is supralinear amplification of one kind of inhibition as demand grows, for instance before excitation overwhelms the circuit Quotation withheld because this record is not cleared for redistribution. The proposed mechanisms are supported mostly by correlation because the cell-type-specific loss-of-function tools needed do not exist. what this limits: every mechanistic claim in the paper: SUITABLE LOSS-OF-FUNCTION EXPERIMENTS TO TEST THE CONTRIBUTION OF THESE MECHANISMS ... DO NOT YET EXIST, so the support is correlational by necessity rather than by choice Quotation withheld because this record is not cleared for redistribution. Broad optogenetic activation also stimulates axon terminals, so the recruitment results may reflect complex network interactions rather than direct drive. what this limits: the recruitment results specifically: THIS BROAD ACTIVATION OF THE LOCAL MICROCIRCUIT INCLUDES AXON TERMINALS EXPRESSING C1V1, THEREFORE THESE RESULTS MAY BE DUE TO MULTIPLE COMPLEX NETWORK INTERACTIONS Quotation withheld because this record is not cleared for redistribution. The marker-defined populations probably contain finer subpopulations that were not resolved here. what this limits: every claim stated at the level of the three marker classes: they are CONVENIENTLY CATEGORIZED BY THE NON-OVERLAPPING MARKERS SOM, PV AND VIP TO UTILIZE TRANSGENIC MICE, and THERE ARE LIKELY FURTHER SUBPO[pulations] - the taxonomy follows the available reagents Quotation withheld because this record is not cleared for redistribution. the mechanisms are implied by correlation rather than shown directly, because the loss-of-function reagent that would be needed does not exist duplicate of: limitations[0] · what this limits: the same scope as limitations[0] - correlation, because the loss-of-function tool does not exist Quotation withheld because this record is not cleared for redistribution. broad optogenetic activation also excites axon terminals, so those results may reflect complex network interactions duplicate of: limitations[1] · what this limits: the same scope as limitations[1] - axon terminals are excited too Quotation withheld because this record is not cleared for redistribution. the three marker classes are convenient rather than fundamental, and the framework is expected to hold on average for populations known to contain further subtypes what this limits: the same scope as limitations[2], with the consequence: the framework is expected to hold on average over classes that are convenient rather than fundamental Quotation withheld because this record is not cleared for redistribution. cell-type labelling was verified in only one of the mouse lines, the rest taken on the strength of prior documentation what this limits: the cell-type identification across lines: labelling was verified in one line, the rest taken on the strength of prior documentation Quotation withheld because this record is not cleared for redistribution. the animals were naive, and other work suggests some of this wiring changes with learning what this limits: the transfer of the wiring result to trained animals: WE USED NAIVE MICE, and other evidence suggests VIPS AND PVS CAN WIRE INTO A MUTUALLY INHIBITORY DYAD DURING LEARNING Quotation withheld because this record is not cleared for redistribution. the compartmentalisation of cooperativity into local teams is offered as a hypothesis for future study what this limits: the team/compartment account specifically: WE HYPOTHESIZE FOR FURTHER STUDIES - offered as a hypothesis, not a result Quotation withheld because this record is not cleared for redistribution. monosynaptic rabies tracing from genetically defined starter cells, used to ask how much local input each class receives Quotation withheld because this record is not cleared for redistribution. sequential stimulation of soma-sized volumes while recording two cells, to ask whether their inputs are shared Quotation withheld because this record is not cleared for redistribution. agonist and antagonist muscle pairs, used as the argument for why two mutually inhibitory populations should not share excitatory drive Quotation withheld because this record is not cleared for redistribution. co-activity of principal cells during ongoing activity, imported as the question to ask of inhibitory cells Quotation withheld because this record is not cleared for redistribution. running the identical protocol on PV cells, where shared presynaptic partners ARE expected eliminates: the protocol's inability to detect shared presynaptic partners as the reason none were found between VIPs and SOMs - a positive control against a null result · outcome (stated by the tagger, no quotation): held Quotation withheld because this record is not cleared for redistribution. comparing baseline firing rate between control and GABA receptor blockade - a manipulation check reported inside a parenthesis eliminates: a shift in baseline firing rate as the explanation of the change in cooperativity: 1.9 +/- 0.2 Hz against 1.9 +/- 0.3 Hz, P = 0.68 · outcome (stated by the tagger, no quotation): held Quotation withheld because this record is not cleared for redistribution. inspecting active and passive electrophysiological properties of VIPs grouped by cooperativity and by their response to GABA blockade - the sentence states the check and not its result, so no outcome is recorded eliminates: that the spread in VIP cooperativity is one population varying, when it may be two subclasses - the check groups cells by cooperativity and by their response to GABA blockade and asks whether the electrophysiology follows · outcome (stated by the tagger, no quotation): absent-not-in-staged-text · would have ruled out, had it been run: distinct VIP subclasses as the explanation of the variation in cooperativity Quotation withheld because this record is not cleared for redistribution. correlation correlation Quotation withheld because this record is not cleared for redistribution. locomotion locomotion Quotation withheld because this record is not cleared for redistribution. THE EXPERIMENT THAT WOULD SETTLE IT IS NAMED, AND THE REAGENT DOES NOT EXIST: suitable loss-of-function tests would need synaptic blockers simultaneously specific to both the presynaptic and the postsynaptic population, so the mechanisms are implied by correlation rather than directly Quotation withheld because this record is not cleared for redistribution. and the single exception is marked as such - one figure directly demonstrates the GABAergic role in cooperative recruitment, and the paper says which one Quotation withheld because this record is not cleared for redistribution. A POSITIVE CONTROL CHOSEN ON THE STRENGTH OF AN EARLIER MEASUREMENT IN THE SAME DATASET: PVs are used to validate the shared-input assay because they had the highest subthreshold cross-correlations here, so they should share presynaptic partners Quotation withheld because this record is not cleared for redistribution. and the null is only interpretable because that control worked - shared inputs were detectable in PV pairs and indistinguishable from zero in VIP-SOM pairs Quotation withheld because this record is not cleared for redistribution. THE MEASURE IS BUILT BY SUBTRACTING THE BASELINE INTO EXISTENCE: laser-EPSCs are the post-stimulus count minus the pre-stimulus count, so the quantity itself is a difference and can legitimately come out negative Quotation withheld because this record is not cleared for redistribution. the correlation analysis is repeated on binarized data specifically to avoid systematic biases in the fluorescence measure it was first computed on Quotation withheld because this record is not cleared for redistribution. THE CATEGORIES ARE DECLARED CONVENIENT RATHER THAN FUNDAMENTAL: the three markers are used because transgenic lines exist for them, and further subpopulations are expected inside each Quotation withheld because this record is not cleared for redistribution. AND THE UNIT OF COMPUTATION IS PROPOSED TO CHANGE: groups rather than individual neurons, and reciprocally inhibitory dyads rather than a sequential hierarchy Quotation withheld because this record is not cleared for redistribution. a distance threshold is set from the data rather than chosen: correlations approach zero beyond 150 micrometres, which is read as the range over which inputs are shared and then used as the analysis cutoff Quotation withheld because this record is not cleared for redistribution. claim_type: mechanism a second reading reaching the same value. Recorded: mechanism — note: a second reading reaching the same value from a different sentence. Not applied - a person decides. Quotation withheld because this record is not cleared for redistribution. method: correlation of calcium signals in labelled cells in awake mice, then paired and quadruple patch clamp, rabies tracing and two-photon input mapping in slices a second reading DIFFERING from the recorded value. Recorded: electrophysiology — note: a second reading, verbatim-verified, disagreeing with the value on the record. Not applied - a person decides. Quotation withheld because this record is not cleared for redistribution. phenomenon: within-population co-activity: members of one interneuron class active together during ongoing activity a second reading DIFFERING from the recorded value. Recorded: within-population co-activity of neocortical interneurons — note: a second reading, verbatim-verified, disagreeing with the value on the record. Not applied - a person decides. Quotation withheld because this record is not cleared for redistribution. contradicts → Pfeffer et al. (2013) on target-selective inhibition by PV interneurons Quotation withheld because this record is not cleared for redistribution. extends → Hofer et al. (2011) on PV interneuron co-activity Quotation withheld because this record is not cleared for redistribution. replicates → Hu et al. (2011) and Ma et al. (2012) on the absence of SOM-to-SOM inhibition Quotation withheld because this record is not cleared for redistribution. contradicts → the prevailing hierarchical disinhibitory model built from activating one interneuron population at a time Quotation withheld because this record is not cleared for redistribution. extends → the demonstration that parvalbumin interneurons are co-active during ongoing activity, generalised here to two more classes Quotation withheld because this record is not cleared for redistribution. qualifies → the account of VIP interneurons as specialists in long-range excitation, by showing comparable local input Quotation withheld because this record is not cleared for redistribution. supports → reports that within-population inhibition is absent among somatostatin cells Quotation withheld because this record is not cleared for redistribution. qualifies → its own generalisation, since the mice were naive and some of this wiring is reported to change with learning Quotation withheld because this record is not cleared for redistribution. VIP —distinct-from→ SOM VIP SOM strength of the relation: 1 · note: two non-overlapping populations, each shown to be active as a population rather than individually Quotation withheld because this record is not cleared for redistribution. locomotion —causes→ correlation locomotion correlation direction on the measure: increases · strength of the relation: 1 · note: which is why the data are split by movement state before anything is compared Quotation withheld because this record is not cleared for redistribution. locomotion —causes→ VIP locomotion VIP strength of the relation: 1 Audit 2026-09-28: DOUBTFUL - read against its span and kept, but the span says VIP and SOM activity is modulated by movement; 'locomotion causes VIP' drops 'activity'. Quotation withheld because this record is not cleared for redistribution. depressing —contrasted-with→ facilitating depressing facilitating strength of the relation: 1 · note: the same excitatory cells drive the two populations with opposite short-term dynamics, so they peak at different times during a train Quotation withheld because this record is not cleared for redistribution. disinhibition —distinct-from→ electrical coupling disinhibition electrical coupling strength of the relation: 1 · note: FOUR MECHANISMS, NOT ONE: electrical coupling, absent or scarce within-population inhibition, disinhibition through the opposing population, and self-excitation Audit 2026-09-28: DOUBTFUL - read against its span and kept, but the span lists coupling and disinhibition among recruitment mechanisms; distinct-from is loose. Quotation withheld because this record is not cleared for redistribution. GABAergic —causes→ cooperativity GABAergic cooperativity strength of the relation: 1 · note: AND THE ONE DIRECT DEMONSTRATION IS MARKED AS THE EXCEPTION, with the figure named Audit 2026-09-28: DOUBTFUL - read against its span and kept, but the span shows GABAergic inhibition's role in cooperative recruitment of SOMs; 'cooperativity' is terse. Quotation withheld because this record is not cleared for redistribution. dyad —contrasted-with→ hierarchy dyad hierarchy strength of the relation: 1 · note: AND THE UNIT OF COMPUTATION CHANGES WITH THE MODEL: groups rather than individual neurons, dyads rather than a sequence Quotation withheld because this record is not cleared for redistribution. RETRACTED (A DECLARED UNKNOWN, AND THE EDGE IS A TRUISM THE SPAN DOES NOT ASSERT) · interneuron —part-of→ population interneuron population RETRACTED: A DECLARED UNKNOWN, AND THE EDGE IS A TRUISM THE SPAN DOES NOT ASSERT. The edge is 'interneuron --part-of--> population'. The span: '…unclear how interneurons coordinate their activity within…' The sentence is about what is NOT known. That interneurons belong to populations is background vocabulary it uses in passing, not a relation it establishes. — strength of the relation: 1 · note: and the puzzle stated plainly - how mutually inhibitory populations manage to operate together inside a densely interconnected network Audit 2026-09-28: DOUBTFUL - read against its span and kept, but interneurons within subpopulations; part-of is loose. Quotation withheld because this record is not cleared for redistribution. paired t-test paired t-test Quotation withheld because this record is not cleared for redistribution. t-test t-test Quotation withheld because this record is not cleared for redistribution. correlation correlation Quotation withheld because this record is not cleared for redistribution. z-test z-test Quotation withheld because this record is not cleared for redistribution. standard error of the mean standard error of the mean Quotation withheld because this record is not cleared for redistribution. cross-correlation cross-correlation Quotation withheld because this record is not cleared for redistribution. surrogate/shuffled data test surrogate/shuffled data test Quotation withheld because this record is not cleared for redistribution. paired T-test paired T-test Quotation withheld because this record is not cleared for redistribution. z-test z-test Quotation withheld because this record is not cleared for redistribution. correlation coefficients correlation coefficients Quotation withheld because this record is not cleared for redistribution. In awake mice, VIP interneurons were more strongly correlated with other VIPs than with non-VIP cells, showing within-population co-activity during ongoing activity. Quotation withheld because this record is not cleared for redistribution. Nearby pyramidal cells contact VIPs and SOMs at similar rates but with opposite short-term dynamics, depressing onto VIPs and facilitating onto SOMs. Quotation withheld because this record is not cleared for redistribution. Two-photon input mapping found essentially no shared presynaptic excitatory cells between a VIP and a SOM, whereas PV pairs shared about a fifth of their inputs. Quotation withheld because this record is not cleared for redistribution. VIP-SOM pairs had significantly lower membrane-potential correlations than all other pairs, supporting innervation by distinct sets of pyramidal cells. Quotation withheld because this record is not cleared for redistribution. Forcing three SOM interneurons to fire significantly increased the spontaneous firing of a fourth, uninduced SOM, demonstrating cooperative within-population recruitment. Quotation withheld because this record is not cleared for redistribution. Blocking GABA-A and GABA-B receptors sharply reduced the SOM cooperative recruitment effect, identifying disinhibition as a main driver. Quotation withheld because this record is not cleared for redistribution. VIP cooperativity was insensitive to GABA blockade but reduced by nicotinic blockade, implicating acetylcholine released by VIPs in recruiting other VIPs. Quotation withheld because this record is not cleared for redistribution. Inhibitory connectivity was lowest within SOMs (0%), within VIPs, and from VIPs to PVs. Quotation withheld because this record is not cleared for redistribution. Contrary to an earlier report, PV interneurons were found to inhibit other interneuron types without selectivity. Quotation withheld because this record is not cleared for redistribution. members of two non-overlapping interneuron populations are active as populations rather than individually Quotation withheld because this record is not cleared for redistribution. and the conclusion is that these populations work as cooperative units, generating an amplifying nonlinearity in circuit output Quotation withheld because this record is not cleared for redistribution. co-activity of principal cells had been studied extensively while that of inhibitory cells had received little attention Quotation withheld because this record is not cleared for redistribution. ongoing activity is argued to be the optimal condition, because it shows circuit flow without thalamic drive Quotation withheld because this record is not cleared for redistribution. THE PUZZLE STATED PLAINLY: how mutually inhibitory populations manage to operate together inside a densely interconnected network Quotation withheld because this record is not cleared for redistribution. both classes were more correlated with their own members than with anything else, in awake mice in a dark room Quotation withheld because this record is not cleared for redistribution. THE ANALYSIS IS REPEATED ON BINARIZED DATA to avoid systematic biases in the fluorescence correlations Quotation withheld because this record is not cleared for redistribution. and the co-activity counts are held against shuffled surrogates as well as against the other cells in the field Quotation withheld because this record is not cleared for redistribution. the data are split by movement state because locomotion modulates both populations Quotation withheld because this record is not cleared for redistribution. and locomotion raised the within-population correlations in both Quotation withheld because this record is not cleared for redistribution. the same excitatory cells drive the two populations with opposite short-term dynamics, depressing onto one and facilitating onto the other Quotation withheld because this record is not cleared for redistribution. which shows up in network activation: one class fires a single spike and falls silent while the other ramps up Quotation withheld because this record is not cleared for redistribution. with a caveat attached in the same paragraph, that broad activation reaches axon terminals too Quotation withheld because this record is not cleared for redistribution. rabies tracing was run in a different cortical area precisely because the published version had been done in the other one Quotation withheld because this record is not cleared for redistribution. VIPs and SOMs received input from nearby cells to a similar extent, most heavily from L2/3. Quotation withheld because this record is not cleared for redistribution. shared spontaneous depolarisations were frequent within a class and rare across the two Quotation withheld because this record is not cleared for redistribution. correlations approached zero beyond 150 um intersomatic distance, suggesting they reflect shared input from nearby cells Quotation withheld because this record is not cleared for redistribution. correlation coefficients were not significantly different between SOM-SOM, VIP-PV and VIP-VIP pairs, while VIP-SOM pairs had significantly lower correlations Quotation withheld because this record is not cleared for redistribution. THE MEASURE IS A DIFFERENCE BY CONSTRUCTION: laser-EPSCs are the post-stimulus count minus the pre-stimulus count Quotation withheld because this record is not cleared for redistribution. with the stimulation first shown to have recruited presynaptic cells at all Quotation withheld because this record is not cleared for redistribution. A POSITIVE CONTROL CHOSEN ON THE STRENGTH OF AN EARLIER MEASUREMENT IN THE SAME DATASET: the class with the highest subthreshold cross-correlations, which should therefore share inputs Quotation withheld because this record is not cleared for redistribution. and it worked - both shared and unique inputs were detected in those pairs Quotation withheld because this record is not cleared for redistribution. while the shared proportion across the two test classes was indistinguishable from zero Quotation withheld because this record is not cleared for redistribution. with the conclusion hedged in the sentence itself - innervated at least in part by non-overlapping excitatory subnetworks Quotation withheld because this record is not cleared for redistribution. AN INHERITED ASSUMPTION WITH ONE SPOT-CHECK: since the high reliability of cell-type specific labelling in these mouse lines has been documented before, labelling was only confirmed in one of them with immunohistochemistry Quotation withheld because this record is not cleared for redistribution. two cortical areas were checked to agree before being pooled, with the separated data kept in a supplement Quotation withheld because this record is not cleared for redistribution. only PVs showed high within-population inhibition; the lowest rates were within SOMs Quotation withheld because this record is not cleared for redistribution. FOUR MECHANISMS, NOT ONE: electrical coupling, absent or scarce within-population inhibition, disinhibition through the opposing population, and self-excitation Quotation withheld because this record is not cleared for redistribution. THE EXPERIMENT THAT WOULD SETTLE IT IS NAMED AND THE REAGENT DOES NOT EXIST: blockers simultaneously specific to both the presynaptic and the postsynaptic population would be needed Quotation withheld because this record is not cleared for redistribution. and the single direct demonstration is marked as the exception, with the figure named Quotation withheld because this record is not cleared for redistribution. ambient acetylcholine is proposed to close a positive feedback loop through cells that both respond to it and produce it Quotation withheld because this record is not cleared for redistribution. THE CATEGORIES ARE DECLARED CONVENIENT RATHER THAN FUNDAMENTAL, with further subpopulations expected inside each and the framework claimed to hold on average Quotation withheld because this record is not cleared for redistribution. and a compartmentalisation hypothesis is offered for future work - nearby members cooperating while a distant group of the same class stays independent Quotation withheld because this record is not cleared for redistribution. the argument for segregated excitatory drive is borrowed from agonist and antagonist muscle pairs Quotation withheld because this record is not cleared for redistribution. with the generalisation offered that non-overlapping input may hold for every reciprocally inhibitory pair Quotation withheld because this record is not cleared for redistribution. and a limit stated: the mice were naive, and this wiring is reported to change with learning Quotation withheld because this record is not cleared for redistribution. AND THE UNIT OF COMPUTATION IS PROPOSED TO CHANGE: groups rather than individual neurons, dyads rather than a sequential hierarchy Quotation withheld because this record is not cleared for redistribution. with the model being replaced traced to a design that activates one population at a time Quotation withheld because this record is not cleared for redistribution. the proposed function is supralinear amplification of one kind of inhibition as demand grows Quotation withheld because this record is not cleared for redistribution. ensembles ensembles Quotation withheld because this record is not cleared for redistribution. laser-EPSCs laser-EPSCs Quotation withheld because this record is not cleared for redistribution. summation ratio summation ratio Quotation withheld because this record is not cleared for redistribution. increased firing increased firing Quotation withheld because this record is not cleared for redistribution. L-EPSC L-EPSC Quotation withheld because this record is not cleared for redistribution. co-activation co-activation Quotation withheld because this record is not cleared for redistribution. dyad dyad Quotation withheld because this record is not cleared for redistribution. positive control positive control Quotation withheld because this record is not cleared for redistribution. GABA receptor blockade GABA receptor blockade Quotation withheld because this record is not cleared for redistribution. spontaneous firing rate spontaneous firing rate Quotation withheld because this record is not cleared for redistribution. baseline firing rate baseline firing rate Quotation withheld because this record is not cleared for redistribution. correlation correlation Quotation withheld because this record is not cleared for redistribution. L-EPSC L-EPSC Quotation withheld because this record is not cleared for redistribution. intersomatic intersomatic Quotation withheld because this record is not cleared for redistribution. locomotion locomotion Quotation withheld because this record is not cleared for redistribution. ~85% — share of neocortical interneurons covered by the three non-overlapping PV, SOM and VIP populations Quotation withheld because this record is not cleared for redistribution. 68%, 8643/12641 cells — proportion of rabies-labelled presynaptic cells with clear pyramidal morphology Quotation withheld because this record is not cleared for redistribution. 68 ± 22 % — increase in spontaneous firing of the fourth SOM when three other SOMs were driven to fire Quotation withheld because this record is not cleared for redistribution. 22.6 ± 5.6 % vs −0.8 ± 4.9 % — estimated proportion of shared presynaptic cells for PV-PV pairs versus VIP-SOM pairs Quotation withheld because this record is not cleared for redistribution. R = 0.07 ± 0.01 to R = 0.13 ± 0.03, P = 0.038 — rise in within-VIP pairwise correlation from stationary to locomotion epochs Quotation withheld because this record is not cleared for redistribution. 86.3% / 94.9% — specificity and sensitivity of the VIP-cre reporter line checked by immunohistochemistry Quotation withheld because this record is not cleared for redistribution. 55 % / 8.8 % / 24 % — share of spontaneous depolarizing activations shared within SOM pairs, across SOM-VIP pairs, and within VIP pairs Quotation withheld because this record is not cleared for redistribution. 16 ± 7 %, P = 0.007 — residual SOM firing-rate increase surviving GABA-A and GABA-B receptor blockade Quotation withheld because this record is not cleared for redistribution. 150 mm — the intersomatic distance beyond which subthreshold correlations vanish, read from the data and then used as the analysis cutoff Quotation withheld because this record is not cleared for redistribution. 86.3% — labelled cells that were immunoreactive in the one line checked, against 94.9 percent the other way Quotation withheld because this record is not cleared for redistribution. 55 % — spontaneous activations shared within a class, against 8.8 percent across the two Quotation withheld because this record is not cleared for redistribution. 26 ± 7 Hz — average firing rate SOMs ramped up to during the light pulse, while VIPs tended to fire a single AP at the beginning followed by quiescence Quotation withheld because this record is not cleared for redistribution.

The full text of this article is not redistributable, so it is not included here. The evidence sentences are withheld with it; the tag layer and its grounding status remain. The article itself is at https://doi.org/10.1016/j.neuron.2016.02.037.