Synthesis: what the mouse-cortex E→E record establishes and where it falls short

Synthesis: what the mouse-cortex E→E record establishes and where it falls short

This section is represented as a source-backed SciDEX wiki artifact. The full source remains in the original computational-review repository.

Source: https://github.com/AllenNeuralDynamics/ComputationalReviewRecurrence/blob/79ce062d54a924ce05953ec90aa9d26044d2b48f/content/16_synthesis.md

Citation anchors captured: 17

Citation contexts

• ^{1Citationpaper:paper-6c94ef0ab8aaInhibition-stabilised recurrent amplification of cortical inputs: supported, with the bracket on inhibition flagged. The mouse-cortex E→E record now contains direct evidence of holographic recruitment of small ensembles producing supra-linear, behaviourally relevant amplification in V1 [Marshel2019], of feature-tuned and topographically structured recurrent gain across higher visual areas at panoramic field of v...content/16_synthesis.md:line 8Open reference} Inhibition-stabilised recurrent amplification of cortical inputs: supported, with the bracket on inhibition flagged. The mouse-cortex E→E record now contains direct evidence of holographic recruitment of small ensembles producing supra-linear, behaviourally relevant amplification in V1 ^{1Citationpaper:paper-6c94ef0ab8aaInhibition-stabilised recurrent amplification of cortical inputs: supported, with the bracket on inhibition flagged. The mouse-cortex E→E record now contains direct evidence of holographic recruitment of small ensembles producing supra-linear, behaviourally relevant amplification in V1 [Marshel2019], of feature-tuned and topographically structured recurrent gain across higher visual areas at panoramic field of v...content/16_synthesis.md:line 8Open reference}, of feature-tuned and topographically structured recurrent gain across higher visual areas at panoramic field of v...

• ^{2Citationpaper:paper-8d923acf5f56Inhibition-stabilised recurrent amplification of cortical inputs: supported, with the bracket on inhibition flagged. The mouse-cortex E→E record now contains direct evidence of holographic recruitment of small ensembles producing supra-linear, behaviourally relevant amplification in V1 [Marshel2019], of feature-tuned and topographically structured recurrent gain across higher visual areas at panoramic field of v...content/16_synthesis.md:line 8Open reference} Inhibition-stabilised recurrent amplification of cortical inputs: supported, with the bracket on inhibition flagged. The mouse-cortex E→E record now contains direct evidence of holographic recruitment of small ensembles producing supra-linear, behaviourally relevant amplification in V1 ^{1Citationpaper:paper-6c94ef0ab8aaInhibition-stabilised recurrent amplification of cortical inputs: supported, with the bracket on inhibition flagged. The mouse-cortex E→E record now contains direct evidence of holographic recruitment of small ensembles producing supra-linear, behaviourally relevant amplification in V1 [Marshel2019], of feature-tuned and topographically structured recurrent gain across higher visual areas at panoramic field of v...content/16_synthesis.md:line 8Open reference}, of feature-tuned and topographically structured recurrent gain across higher visual areas at panoramic field of v...

• ^{3Citationpaper:paper-fb9d6c04dc50Inhibition-stabilised recurrent amplification of cortical inputs: supported, with the bracket on inhibition flagged. The mouse-cortex E→E record now contains direct evidence of holographic recruitment of small ensembles producing supra-linear, behaviourally relevant amplification in V1 [Marshel2019], of feature-tuned and topographically structured recurrent gain across higher visual areas at panoramic field of v...content/16_synthesis.md:line 8Open reference} Inhibition-stabilised recurrent amplification of cortical inputs: supported, with the bracket on inhibition flagged. The mouse-cortex E→E record now contains direct evidence of holographic recruitment of small ensembles producing supra-linear, behaviourally relevant amplification in V1 ^{1Citationpaper:paper-6c94ef0ab8aaInhibition-stabilised recurrent amplification of cortical inputs: supported, with the bracket on inhibition flagged. The mouse-cortex E→E record now contains direct evidence of holographic recruitment of small ensembles producing supra-linear, behaviourally relevant amplification in V1 [Marshel2019], of feature-tuned and topographically structured recurrent gain across higher visual areas at panoramic field of v...content/16_synthesis.md:line 8Open reference}, of feature-tuned and topographically structured recurrent gain across higher visual areas at panoramic field of v...

• ^{4Citationpaper:paper-b1711402e644Bilateral-redundancy and discrete-choice attractor dynamics: supported. Large-scale inactivation shows that contrast-discrimination decisions in mouse depend on a small set of visual and frontal areas, with VIS and MOs cortices carrying the causal load despite widespread dorsal-cortex movement-related activity [ZatkaHaas2021]. Convergence of multi-cortical input onto single claustrum neurons, with selective loss...content/16_synthesis.md:line 10Open reference} Bilateral-redundancy and discrete-choice attractor dynamics: supported. Large-scale inactivation shows that contrast-discrimination decisions in mouse depend on a small set of visual and frontal areas, with VIS and MOs cortices carrying the causal load despite widespread dorsal-cortex movement-related activity ^{4Citationpaper:paper-b1711402e644Bilateral-redundancy and discrete-choice attractor dynamics: supported. Large-scale inactivation shows that contrast-discrimination decisions in mouse depend on a small set of visual and frontal areas, with VIS and MOs cortices carrying the causal load despite widespread dorsal-cortex movement-related activity [ZatkaHaas2021]. Convergence of multi-cortical input onto single claustrum neurons, with selective loss...content/16_synthesis.md:line 10Open reference}. Convergence of multi-cortical input onto single claustrum neurons, with selective loss...

• ^{5Citationpaper:paper-4a2a205c7670Bilateral-redundancy and discrete-choice attractor dynamics: supported. Large-scale inactivation shows that contrast-discrimination decisions in mouse depend on a small set of visual and frontal areas, with VIS and MOs cortices carrying the causal load despite widespread dorsal-cortex movement-related activity [ZatkaHaas2021]. Convergence of multi-cortical input onto single claustrum neurons, with selective loss...content/16_synthesis.md:line 10Open reference} Bilateral-redundancy and discrete-choice attractor dynamics: supported. Large-scale inactivation shows that contrast-discrimination decisions in mouse depend on a small set of visual and frontal areas, with VIS and MOs cortices carrying the causal load despite widespread dorsal-cortex movement-related activity ^{4Citationpaper:paper-b1711402e644Bilateral-redundancy and discrete-choice attractor dynamics: supported. Large-scale inactivation shows that contrast-discrimination decisions in mouse depend on a small set of visual and frontal areas, with VIS and MOs cortices carrying the causal load despite widespread dorsal-cortex movement-related activity [ZatkaHaas2021]. Convergence of multi-cortical input onto single claustrum neurons, with selective loss...content/16_synthesis.md:line 10Open reference}. Convergence of multi-cortical input onto single claustrum neurons, with selective loss...

• ^{1Citationpaper:paper-6c94ef0ab8aaInhibition-stabilised recurrent amplification of cortical inputs: supported, with the bracket on inhibition flagged. The mouse-cortex E→E record now contains direct evidence of holographic recruitment of small ensembles producing supra-linear, behaviourally relevant amplification in V1 [Marshel2019], of feature-tuned and topographically structured recurrent gain across higher visual areas at panoramic field of v...content/16_synthesis.md:line 8Open reference}0 Hopfield-style content-addressable attractors in mouse cortex: unsupported. The symmetry-of-weights and pattern-completion benchmarks demanded by {ref}sec-pattern-completion cannot be assessed against the partial connectomes currently in hand, and the closest comparative anatomy — a macaque ventrolateral PFC excitatory projectome — establishes only the topographic prerequisite, not the symmetry. **Canonical bu...

• ^{1Citationpaper:paper-6c94ef0ab8aaInhibition-stabilised recurrent amplification of cortical inputs: supported, with the bracket on inhibition flagged. The mouse-cortex E→E record now contains direct evidence of holographic recruitment of small ensembles producing supra-linear, behaviourally relevant amplification in V1 [Marshel2019], of feature-tuned and topographically structured recurrent gain across higher visual areas at panoramic field of v...content/16_synthesis.md:line 8Open reference}1 Hopfield-style content-addressable attractors in mouse cortex: unsupported. The symmetry-of-weights and pattern-completion benchmarks demanded by {ref}sec-pattern-completion cannot be assessed against the partial connectomes currently in hand, and the closest comparative anatomy — a macaque ventrolateral PFC excitatory projectome — establishes only the topographic prerequisite, not the symmetry. **Canonical bu...

• ^{1Citationpaper:paper-6c94ef0ab8aaInhibition-stabilised recurrent amplification of cortical inputs: supported, with the bracket on inhibition flagged. The mouse-cortex E→E record now contains direct evidence of holographic recruitment of small ensembles producing supra-linear, behaviourally relevant amplification in V1 [Marshel2019], of feature-tuned and topographically structured recurrent gain across higher visual areas at panoramic field of v...content/16_synthesis.md:line 8Open reference}2 Predictive coding and the manifold / reservoir framing: partial. Allen Brain Observatory analyses show that visually evoked responses are weaker for more spatially predictable image patches, with the effect concentrated in higher-level features rather than pixel statistics ^{1Citationpaper:paper-6c94ef0ab8aaInhibition-stabilised recurrent amplification of cortical inputs: supported, with the bracket on inhibition flagged. The mouse-cortex E→E record now contains direct evidence of holographic recruitment of small ensembles producing supra-linear, behaviourally relevant amplification in V1 [Marshel2019], of feature-tuned and topographically structured recurrent gain across higher visual areas at panoramic field of v...content/16_synthesis.md:line 8Open reference}3, the direction expected from the canonical microcircuit. Normative models map subtractive and divisive prediction-error comput...

• ^{1Citationpaper:paper-6c94ef0ab8aaInhibition-stabilised recurrent amplification of cortical inputs: supported, with the bracket on inhibition flagged. The mouse-cortex E→E record now contains direct evidence of holographic recruitment of small ensembles producing supra-linear, behaviourally relevant amplification in V1 [Marshel2019], of feature-tuned and topographically structured recurrent gain across higher visual areas at panoramic field of v...content/16_synthesis.md:line 8Open reference}4 A millimetre-scale dynamic connectome — co-registered EM wiring with single-trial functional readout across cortical depth — would test symmetry-of-weights and bump-attractor predictions against the wiring that generates them. Current tooling sits at hundreds of simultaneously targeted neurons via spiral scanning, temporal focusing, and three-dimensional computer-generated holography ^{1Citationpaper:paper-6c94ef0ab8aaInhibition-stabilised recurrent amplification of cortical inputs: supported, with the bracket on inhibition flagged. The mouse-cortex E→E record now contains direct evidence of holographic recruitment of small ensembles producing supra-linear, behaviourally relevant amplification in V1 [Marshel2019], of feature-tuned and topographically structured recurrent gain across higher visual areas at panoramic field of v...content/16_synthesis.md:line 8Open reference}5, with axial po...

• ^{1Citationpaper:paper-6c94ef0ab8aaInhibition-stabilised recurrent amplification of cortical inputs: supported, with the bracket on inhibition flagged. The mouse-cortex E→E record now contains direct evidence of holographic recruitment of small ensembles producing supra-linear, behaviourally relevant amplification in V1 [Marshel2019], of feature-tuned and topographically structured recurrent gain across higher visual areas at panoramic field of v...content/16_synthesis.md:line 8Open reference}6 A millimetre-scale dynamic connectome — co-registered EM wiring with single-trial functional readout across cortical depth — would test symmetry-of-weights and bump-attractor predictions against the wiring that generates them. Current tooling sits at hundreds of simultaneously targeted neurons via spiral scanning, temporal focusing, and three-dimensional computer-generated holography ^{1Citationpaper:paper-6c94ef0ab8aaInhibition-stabilised recurrent amplification of cortical inputs: supported, with the bracket on inhibition flagged. The mouse-cortex E→E record now contains direct evidence of holographic recruitment of small ensembles producing supra-linear, behaviourally relevant amplification in V1 [Marshel2019], of feature-tuned and topographically structured recurrent gain across higher visual areas at panoramic field of v...content/16_synthesis.md:line 8Open reference}7, with axial po...

• ^{1Citationpaper:paper-6c94ef0ab8aaInhibition-stabilised recurrent amplification of cortical inputs: supported, with the bracket on inhibition flagged. The mouse-cortex E→E record now contains direct evidence of holographic recruitment of small ensembles producing supra-linear, behaviourally relevant amplification in V1 [Marshel2019], of feature-tuned and topographically structured recurrent gain across higher visual areas at panoramic field of v...content/16_synthesis.md:line 8Open reference}8 A millimetre-scale dynamic connectome — co-registered EM wiring with single-trial functional readout across cortical depth — would test symmetry-of-weights and bump-attractor predictions against the wiring that generates them. Current tooling sits at hundreds of simultaneously targeted neurons via spiral scanning, temporal focusing, and three-dimensional computer-generated holography ^{1Citationpaper:paper-6c94ef0ab8aaInhibition-stabilised recurrent amplification of cortical inputs: supported, with the bracket on inhibition flagged. The mouse-cortex E→E record now contains direct evidence of holographic recruitment of small ensembles producing supra-linear, behaviourally relevant amplification in V1 [Marshel2019], of feature-tuned and topographically structured recurrent gain across higher visual areas at panoramic field of v...content/16_synthesis.md:line 8Open reference}9, with axial po...

• ^{2Citationpaper:paper-8d923acf5f56Inhibition-stabilised recurrent amplification of cortical inputs: supported, with the bracket on inhibition flagged. The mouse-cortex E→E record now contains direct evidence of holographic recruitment of small ensembles producing supra-linear, behaviourally relevant amplification in V1 [Marshel2019], of feature-tuned and topographically structured recurrent gain across higher visual areas at panoramic field of v...content/16_synthesis.md:line 8Open reference}0 A millimetre-scale dynamic connectome — co-registered EM wiring with single-trial functional readout across cortical depth — would test symmetry-of-weights and bump-attractor predictions against the wiring that generates them. Current tooling sits at hundreds of simultaneously targeted neurons via spiral scanning, temporal focusing, and three-dimensional computer-generated holography ^{2Citationpaper:paper-8d923acf5f56Inhibition-stabilised recurrent amplification of cortical inputs: supported, with the bracket on inhibition flagged. The mouse-cortex E→E record now contains direct evidence of holographic recruitment of small ensembles producing supra-linear, behaviourally relevant amplification in V1 [Marshel2019], of feature-tuned and topographically structured recurrent gain across higher visual areas at panoramic field of v...content/16_synthesis.md:line 8Open reference}1, with axial po...

• ^{2Citationpaper:paper-8d923acf5f56Inhibition-stabilised recurrent amplification of cortical inputs: supported, with the bracket on inhibition flagged. The mouse-cortex E→E record now contains direct evidence of holographic recruitment of small ensembles producing supra-linear, behaviourally relevant amplification in V1 [Marshel2019], of feature-tuned and topographically structured recurrent gain across higher visual areas at panoramic field of v...content/16_synthesis.md:line 8Open reference}2 Fast in-vivo plasticity at identified pyramidal–pyramidal synapses is the bottleneck for predictive coding and requires voltage indicators usable in awake mouse cortex on the relevant timescales. JEDI-2P enables per-cell voltage recording for >30 min with random-access microscopy ^{2Citationpaper:paper-8d923acf5f56Inhibition-stabilised recurrent amplification of cortical inputs: supported, with the bracket on inhibition flagged. The mouse-cortex E→E record now contains direct evidence of holographic recruitment of small ensembles producing supra-linear, behaviourally relevant amplification in V1 [Marshel2019], of feature-tuned and topographically structured recurrent gain across higher visual areas at panoramic field of v...content/16_synthesis.md:line 8Open reference}3; time-multiplexed pulsed 2P excitation separates voltage signal from photostimulation crosstalk in all-optical experiments;...

• ^{2Citationpaper:paper-8d923acf5f56Inhibition-stabilised recurrent amplification of cortical inputs: supported, with the bracket on inhibition flagged. The mouse-cortex E→E record now contains direct evidence of holographic recruitment of small ensembles producing supra-linear, behaviourally relevant amplification in V1 [Marshel2019], of feature-tuned and topographically structured recurrent gain across higher visual areas at panoramic field of v...content/16_synthesis.md:line 8Open reference}4 Fast in-vivo plasticity at identified pyramidal–pyramidal synapses is the bottleneck for predictive coding and requires voltage indicators usable in awake mouse cortex on the relevant timescales. JEDI-2P enables per-cell voltage recording for >30 min with random-access microscopy ^{2Citationpaper:paper-8d923acf5f56Inhibition-stabilised recurrent amplification of cortical inputs: supported, with the bracket on inhibition flagged. The mouse-cortex E→E record now contains direct evidence of holographic recruitment of small ensembles producing supra-linear, behaviourally relevant amplification in V1 [Marshel2019], of feature-tuned and topographically structured recurrent gain across higher visual areas at panoramic field of v...content/16_synthesis.md:line 8Open reference}5; time-multiplexed pulsed 2P excitation separates voltage signal from photostimulation crosstalk in all-optical experiments;...

• ^{2Citationpaper:paper-8d923acf5f56Inhibition-stabilised recurrent amplification of cortical inputs: supported, with the bracket on inhibition flagged. The mouse-cortex E→E record now contains direct evidence of holographic recruitment of small ensembles producing supra-linear, behaviourally relevant amplification in V1 [Marshel2019], of feature-tuned and topographically structured recurrent gain across higher visual areas at panoramic field of v...content/16_synthesis.md:line 8Open reference}6 Fast in-vivo plasticity at identified pyramidal–pyramidal synapses is the bottleneck for predictive coding and requires voltage indicators usable in awake mouse cortex on the relevant timescales. JEDI-2P enables per-cell voltage recording for >30 min with random-access microscopy ^{2Citationpaper:paper-8d923acf5f56Inhibition-stabilised recurrent amplification of cortical inputs: supported, with the bracket on inhibition flagged. The mouse-cortex E→E record now contains direct evidence of holographic recruitment of small ensembles producing supra-linear, behaviourally relevant amplification in V1 [Marshel2019], of feature-tuned and topographically structured recurrent gain across higher visual areas at panoramic field of v...content/16_synthesis.md:line 8Open reference}7; time-multiplexed pulsed 2P excitation separates voltage signal from photostimulation crosstalk in all-optical experiments;...

• ^{2Citationpaper:paper-8d923acf5f56Inhibition-stabilised recurrent amplification of cortical inputs: supported, with the bracket on inhibition flagged. The mouse-cortex E→E record now contains direct evidence of holographic recruitment of small ensembles producing supra-linear, behaviourally relevant amplification in V1 [Marshel2019], of feature-tuned and topographically structured recurrent gain across higher visual areas at panoramic field of v...content/16_synthesis.md:line 8Open reference}8 A head-to-head trained-versus-untrained reservoir comparison, informed by the line-attractor dynamics in prefrontal reversal-probability subspaces identified in primate PFC and matched RNNs ^{2Citationpaper:paper-8d923acf5f56Inhibition-stabilised recurrent amplification of cortical inputs: supported, with the bracket on inhibition flagged. The mouse-cortex E→E record now contains direct evidence of holographic recruitment of small ensembles producing supra-linear, behaviourally relevant amplification in V1 [Marshel2019], of feature-tuned and topographically structured recurrent gain across higher visual areas at panoramic field of v...content/16_synthesis.md:line 8Open reference}9, would discriminate trained-recurrence from echo-state accounts of motor-cortex computation. Pan-cortical 2P mesoscopes that image nearly all dorsal and lateral neocortex, dual-plane three-photon imaging with non-u...

• ^{1Citationpaper:paper-6c94ef0ab8aaInhibition-stabilised recurrent amplification of cortical inputs: supported, with the bracket on inhibition flagged. The mouse-cortex E→E record now contains direct evidence of holographic recruitment of small ensembles producing supra-linear, behaviourally relevant amplification in V1 [Marshel2019], of feature-tuned and topographically structured recurrent gain across higher visual areas at panoramic field of v...content/16_synthesis.md:line 8Open reference}0 Inhibitory contribution and neuromodulation. The operating-point and gain-control arguments developed in {ref}sec-amplification-isn are bracketed throughout, even as pathway-specific cortico-cortical inhibition has been shown to substantially modulate primary somatosensory responses in awake mouse cortex via long-range projection silencing ^{1Citationpaper:paper-6c94ef0ab8aaInhibition-stabilised recurrent amplification of cortical inputs: supported, with the bracket on inhibition flagged. The mouse-cortex E→E record now contains direct evidence of holographic recruitment of small ensembles producing supra-linear, behaviourally relevant amplification in V1 [Marshel2019], of feature-tuned and topographically structured recurrent gain across higher visual areas at panoramic field of v...content/16_synthesis.md:line 8Open reference}1, and as recurrent local circuits implicated in surround-...