These are all 443 preserved entries across ten historical document keys—not 443 freshly certified concepts. Source-atom labels are not automatically unique semantic concepts.
Exact-version corrections remain visible: the selected Rosetta body is v3, 4 May 2026, and the inherited minimal-active-inference body is v2, 14 June 2025. Legacy keys are not publication dates.
This release pass routed 14 SAN stages against five fixed target bodies: 70 bounded pairs. It used one TurboVec sidecar call, 14 Jigsaw/Theater calls and one census over 19 files. The census produced 540 operational atoms, 58 signature clusters, 201 partial-signature clusters, 230 fit patterns and 38 narrative similes. These are discovery outputs, not certified PECs. Census correspondence scoring was not run; zero correspondences cannot mean no match. JigsawRank reused an index; no fresh MemPalace retrieval is claimed.
The Wiki has no rival-award category. That policy is not proof that every historical decision followed it. Weak, overlapping, disputed and source-gated entries remain visible.
LSM-UNKNOWN; quantitative-not-calibrated. The selected dated cells do not establish the full acquisition order, later persistence, dependency, copying or a calibrated probability. All three atom-level lenses remain separate from this program-level event. Named alternative genealogies and the unreviewed before/newest history remain open.
The complete 40-cell selected comparison assessment
This register prevents the eight display cards from becoming the comparison ceiling. Every applicable pair is here, including weak conceptual bridges, question/answer fits and unresolved complements. Only the eight previously reviewed links currently carry independent source review; none of this expanded edition has release certification.
| Pair | SAN source record → target version | Relation | Review state |
|---|
SETH-SELECTED-40-01 SAN-AC012-MORPHOLOGY-SELECTIVE-PROJECTION → SETH-CS25-22 | 2022-09-18 → 2025-10-30 | earlier-mechanism-to-later-explanatory-demand | Previously independently source-reviewed subset |
SETH-SELECTED-40-02 SAN-AC012-MORPHOLOGY-SELECTIVE-PROJECTION → SETH-CS25-38 | 2022-09-18 → 2025-10-30 | OPEN-COMPLEMENT | Source reviewed; relationship unresolved |
SETH-SELECTED-40-03 SAN-AC012-MORPHOLOGY-SELECTIVE-PROJECTION → SETH-CS25-30 | 2022-09-18 → 2025-10-30 | OPEN-COMPLEMENT | Source reviewed; relationship unresolved |
SETH-SELECTED-40-04 SAN-AC012-MORPHOLOGY-SELECTIVE-PROJECTION → SETH-CS25-28 | 2022-09-18 → 2025-10-30 | OPEN-BRIDGE | Source reviewed; relationship unresolved |
SETH-SELECTED-40-05 SAN-AC012-MORPHOLOGY-SELECTIVE-PROJECTION → SETH-CS25-18 | 2022-09-18 → 2025-10-30 | PARTIAL-CAUSAL-PREFIX | Additional partial-match candidate; independent review pending |
SETH-SELECTED-40-06 SAN-AC012-MORPHOLOGY-SELECTIVE-PROJECTION → SETH-N24-01 | 2022-09-18 → 2026-05-04 | OPEN-BRIDGE | Source reviewed; relationship unresolved |
SETH-SELECTED-40-07 SAN-AC012-MORPHOLOGY-SELECTIVE-PROJECTION → SETH-N24-05 | 2022-09-18 → 2026-05-04 | OPEN-BRIDGE | Source reviewed; relationship unresolved |
SETH-SELECTED-40-08 SAN-AC012-MORPHOLOGY-SELECTIVE-PROJECTION → SETH-T26-05 | 2022-09-18 → 2024-06-27 | PARTIAL-FUNCTIONAL-PREFIX | Additional partial-match candidate; independent review pending |
SETH-SELECTED-40-09 SAN-AC013-QUALIA-RECONSTRUCTION → SETH-CS25-22 | 2022-06-12 → 2025-10-30 | PARTIAL-PROBLEM-FIT | Additional partial-match candidate; independent review pending |
SETH-SELECTED-40-10 SAN-AC013-QUALIA-RECONSTRUCTION → SETH-CS25-38 | 2022-06-12 → 2025-10-30 | partial-operation | Previously independently source-reviewed subset |
SETH-SELECTED-40-11 SAN-AC013-QUALIA-RECONSTRUCTION → SETH-CS25-30 | 2022-06-12 → 2025-10-30 | partial-operation | Previously independently source-reviewed subset |
SETH-SELECTED-40-12 SAN-AC013-QUALIA-RECONSTRUCTION → SETH-CS25-28 | 2022-06-12 → 2025-10-30 | PARTIAL-CONCEPTUAL-BRIDGE | Additional partial-match candidate; independent review pending |
SETH-SELECTED-40-13 SAN-AC013-QUALIA-RECONSTRUCTION → SETH-CS25-18 | 2022-06-12 → 2025-10-30 | PARTIAL-CONCEPTUAL-BRIDGE | Additional partial-match candidate; independent review pending |
SETH-SELECTED-40-14 SAN-AC013-QUALIA-RECONSTRUCTION → SETH-N24-01 | 2022-06-12 → 2026-05-04 | PARTIAL-FUNCTIONAL-BRIDGE | Additional partial-match candidate; independent review pending |
SETH-SELECTED-40-15 SAN-AC013-QUALIA-RECONSTRUCTION → SETH-N24-05 | 2022-06-12 → 2026-05-04 | PARTIAL-CONCEPTUAL-BRIDGE | Additional partial-match candidate; independent review pending |
SETH-SELECTED-40-16 SAN-AC013-QUALIA-RECONSTRUCTION → SETH-T26-05 | 2022-06-12 → 2024-06-27 | PARTIAL-CAUSAL-PREFIX | Additional partial-match candidate; independent review pending |
SETH-SELECTED-40-17 SAN-AC009-PHASE-TOPOLOGY-TAYLOR → SETH-CS25-22 | 2022-07-09 → 2025-10-30 | PARTIAL-PROBLEM-FIT | Additional partial-match candidate; independent review pending |
SETH-SELECTED-40-18 SAN-AC009-PHASE-TOPOLOGY-TAYLOR → SETH-CS25-38 | 2022-07-09 → 2025-10-30 | OPEN-COMPLEMENT | Source reviewed; relationship unresolved |
SETH-SELECTED-40-19 SAN-AC009-PHASE-TOPOLOGY-TAYLOR → SETH-CS25-30 | 2022-07-09 → 2025-10-30 | OPEN-COMPLEMENT | Source reviewed; relationship unresolved |
SETH-SELECTED-40-20 SAN-AC009-PHASE-TOPOLOGY-TAYLOR → SETH-CS25-28 | 2022-07-09 → 2025-10-30 | partial-representation-role | Previously independently source-reviewed subset |
SETH-SELECTED-40-21 SAN-AC009-PHASE-TOPOLOGY-TAYLOR → SETH-CS25-18 | 2022-07-09 → 2025-10-30 | PARTIAL-CONCEPTUAL-BRIDGE | Additional partial-match candidate; independent review pending |
SETH-SELECTED-40-22 SAN-AC009-PHASE-TOPOLOGY-TAYLOR → SETH-N24-01 | 2022-07-09 → 2026-05-04 | PARTIAL-CONCEPTUAL-BRIDGE | Additional partial-match candidate; independent review pending |
SETH-SELECTED-40-23 SAN-AC009-PHASE-TOPOLOGY-TAYLOR → SETH-N24-05 | 2022-07-09 → 2026-05-04 | partial-representation-role | Previously independently source-reviewed subset |
SETH-SELECTED-40-24 SAN-AC009-PHASE-TOPOLOGY-TAYLOR → SETH-T26-05 | 2022-07-09 → 2024-06-27 | PARTIAL-FUNCTIONAL-PREFIX | Additional partial-match candidate; independent review pending |
SETH-SELECTED-40-25 SAN-AC017-DISTRIBUTED-SELF-RENDERING → SETH-CS25-22 | 2022-06-10 → 2025-10-30 | PARTIAL-PROBLEM-FIT | Additional partial-match candidate; independent review pending |
SETH-SELECTED-40-26 SAN-AC017-DISTRIBUTED-SELF-RENDERING → SETH-CS25-38 | 2022-06-10 → 2025-10-30 | OPEN-COMPLEMENT | Source reviewed; relationship unresolved |
SETH-SELECTED-40-27 SAN-AC017-DISTRIBUTED-SELF-RENDERING → SETH-CS25-30 | 2022-06-10 → 2025-10-30 | OPEN-COMPLEMENT | Source reviewed; relationship unresolved |
SETH-SELECTED-40-28 SAN-AC017-DISTRIBUTED-SELF-RENDERING → SETH-CS25-28 | 2022-06-10 → 2025-10-30 | PARTIAL-CONCEPTUAL-BRIDGE | Additional partial-match candidate; independent review pending |
SETH-SELECTED-40-29 SAN-AC017-DISTRIBUTED-SELF-RENDERING → SETH-CS25-18 | 2022-06-10 → 2025-10-30 | partial-operation | Previously independently source-reviewed subset |
SETH-SELECTED-40-30 SAN-AC017-DISTRIBUTED-SELF-RENDERING → SETH-N24-01 | 2022-06-10 → 2026-05-04 | partial-function | Previously independently source-reviewed subset |
SETH-SELECTED-40-31 SAN-AC017-DISTRIBUTED-SELF-RENDERING → SETH-N24-05 | 2022-06-10 → 2026-05-04 | PARTIAL-CONCEPTUAL-BRIDGE | Additional partial-match candidate; independent review pending |
SETH-SELECTED-40-32 SAN-AC017-DISTRIBUTED-SELF-RENDERING → SETH-T26-05 | 2022-06-10 → 2024-06-27 | PARTIAL-CAUSAL-PREFIX | Additional partial-match candidate; independent review pending |
SETH-SELECTED-40-33 SAN-AC024-ARTIFICIAL-NEUROLOGY → SETH-CS25-22 | 2022-11-28 → 2025-10-30 | PARTIAL-PROBLEM-FIT | Additional partial-match candidate; independent review pending |
SETH-SELECTED-40-34 SAN-AC024-ARTIFICIAL-NEUROLOGY → SETH-CS25-38 | 2022-11-28 → 2025-10-30 | PARTIAL-CROSS-CONTEXT-INTERVENTION | Additional partial-match candidate; independent review pending |
SETH-SELECTED-40-35 SAN-AC024-ARTIFICIAL-NEUROLOGY → SETH-CS25-30 | 2022-11-28 → 2025-10-30 | PARTIAL-CROSS-CONTEXT-INTERVENTION | Additional partial-match candidate; independent review pending |
SETH-SELECTED-40-36 SAN-AC024-ARTIFICIAL-NEUROLOGY → SETH-CS25-28 | 2022-11-28 → 2025-10-30 | PARTIAL-CONCEPTUAL-BRIDGE | Additional partial-match candidate; independent review pending |
SETH-SELECTED-40-37 SAN-AC024-ARTIFICIAL-NEUROLOGY → SETH-CS25-18 | 2022-11-28 → 2025-10-30 | PARTIAL-CONCEPTUAL-BRIDGE | Additional partial-match candidate; independent review pending |
SETH-SELECTED-40-38 SAN-AC024-ARTIFICIAL-NEUROLOGY → SETH-N24-01 | 2022-11-28 → 2026-05-04 | PARTIAL-CONCEPTUAL-BRIDGE | Additional partial-match candidate; independent review pending |
SETH-SELECTED-40-39 SAN-AC024-ARTIFICIAL-NEUROLOGY → SETH-N24-05 | 2022-11-28 → 2026-05-04 | PARTIAL-CONCEPTUAL-BRIDGE | Additional partial-match candidate; independent review pending |
SETH-SELECTED-40-40 SAN-AC024-ARTIFICIAL-NEUROLOGY → SETH-T26-05 | 2022-11-28 → 2024-06-27 | partial-operation | Previously independently source-reviewed subset |
SETH-SELECTED-40-01 · Previously independently source-reviewed subset
SAN-AC012-MORPHOLOGY-SELECTIVE-PROJECTION → SETH-CS25-22
SAN describes a receiver recognizing a learned pattern and passing selected activity onward. The later paper asks for an explanation of that step; it does not assert SAN’s answer.
SAN: 2022-09-18 · Git repository commit record. Target: 2025-10-30 · doi:10.3389/fsci.2025.1546279.
Current anonymous public access to the exact source body was verified on 2026-10-11. The displayed 2022 date is a Git record date, not independently verified first-publication timing.
Target authors: Axel Cleeremans · Liad Mudrik · Anil K. Seth. Verify the target passage · See its short quotation in the paired cards.
Exact SAN source windows
SAN lines 3–3
Dendritic morphologies & synaptic connections represent learned memories. Memories activate when preferred patterns (the memory) are detected, this causes cells to create a selective transmission or a projection to other cells (neural arrays). In this note I am looking at Layer 6 & Layer 5 Pyramidal cells, to learn how the Apical Dendrite, or the Exit Terminal, is contributing to the rendered projection of that learned memory to the rest of its network. The main idea that I am exploring is that the signal transmitted from the exit terminal is not uniform along all branches, but instead is modified synapses by synapse, and that the memories in the exit terminal also transmit backwards to play a role in multiple seconds long coincidence detections involving back propagating action potentials from the exit terminal or Apical Dendrite to the Soma, that when combined with the Basal or sensory input dendrite can cause a coincidence detection that creates a special type of action potential event called burst firing.
L1
result: PARTIAL_FUNCTIONAL_PROBLEM_FIT
reason: SAN states a selective input-recognition-to-output operation. The target constrains the explanatory job at the receiver. This is a supported answer-to-question pairing, not implementation equivalence.
L2
result: PARTIAL_ARCHITECTURAL_LOCATION
reason: Both place the explanatory burden after signal arrival and before useful downstream processing. Only SAN joins learned receptive structure, recognition and projection in a proposed causal chain.
L3
result: SUPPORTED_CONSTRAINED_OUTPUT_RECONSTRUCTION_WITH_NONUNIQUE_SOLUTION
reason: Constrained result: A broadcast carries information that downstream neural processing can use. Required operation: There must be a physically implemented dependence of receiver processing on relevant message differences; arrival alone does not explain use. Source test: Line 3 supplies learned selectivity followed by transmission; lines 49–61 elaborate input-sensitive response learning. Alternatives: Learned receptive selectivity; Fixed feature-sensitive circuitry; Population decoding or predictive message passing. Remaining uncertainty: The question neither selects this mechanism nor proves the claimed receiver operation explains consciousness.
Limits and next evidence
The apical-dendrite/exit-terminal conflation in the same SAN neighborhood is not endorsed. Quoted layer-5/layer-6 studies are prior sources, not Micah's experiments. Full neural semantics and experiential sufficiency remain open.
At most a qualitative functional, conceptual or causal-prefix relation; no proof of functoriality, isomorphism, bisimulation, exact equation identity or calibrated statistical dependence.
Existing historical IDs for this exact source-atom/target-atom pair: SETH-CS25-DOCWIN-SETH-CS25-22-SAN-AC012-MORPHOLOGY-SELECTIVE-PROJECTION-FE606E8CE4BC.
SETH-SELECTED-40-02 · Source reviewed; relationship unresolved
SAN-AC012-MORPHOLOGY-SELECTIVE-PROJECTION → SETH-CS25-38
A selective receiver could be an intervention substrate, but the selected SAN operation is recognition and projection, not a stated BCI intervention.
SAN: 2022-09-18 · Git repository commit record. Target: 2025-10-30 · doi:10.3389/fsci.2025.1546279.
Current anonymous public access to the exact source body was verified on 2026-10-11. The displayed 2022 date is a Git record date, not independently verified first-publication timing.
Target authors: Axel Cleeremans · Liad Mudrik · Anil K. Seth. Verify the target passage · See its short quotation in the paired cards.
Exact SAN source windows
SAN lines 3–3
Dendritic morphologies & synaptic connections represent learned memories. Memories activate when preferred patterns (the memory) are detected, this causes cells to create a selective transmission or a projection to other cells (neural arrays). In this note I am looking at Layer 6 & Layer 5 Pyramidal cells, to learn how the Apical Dendrite, or the Exit Terminal, is contributing to the rendered projection of that learned memory to the rest of its network. The main idea that I am exploring is that the signal transmitted from the exit terminal is not uniform along all branches, but instead is modified synapses by synapse, and that the memories in the exit terminal also transmit backwards to play a role in multiple seconds long coincidence detections involving back propagating action potentials from the exit terminal or Apical Dendrite to the Soma, that when combined with the Basal or sensory input dendrite can cause a coincidence detection that creates a special type of action potential event called burst firing.
SAN lines 49–61
I think of it as like the basal dendrite is learning to respond selectively to certain patterns, the apical is learning what patterns it responds to.
My intuition is that the apical is learning its own cell's response, and through feedback loops, with itself and with other pyramidal neurons the cell is going to make predictions about its own patterns of causation, what patterns it responds & how it reacts or responds. With the Apical Dendrite perhaps symbolizing a more complex response that is connected to more complex learning. (compared to other types of neurons)
The fact that bursting (bursts of somatic AP) results from the coincident activity of basal + apical, but not from either one separately is a different coincidence detection mechanism from the leaky integrate-and fire model of a neuron which explores coincident synaptic inputs, leading to coincident dendritic activity, which leads to the phasic action potential from just the south side of the Soma.
In the leaky integrate & fire neurons, and the Hodgkin & Huxley Model there is very little thought to the concept of the Soma burst spiking being a coincidence detection from both sides of the neuron. But if it's true for Layer 5 pyramidal cells, then we ought to know if there is a greater role for the exit terminal in all other types of neurons.
So perhaps we can think of the axon, with it's different types of action potentials, different durations (wave shapes, magnitudes), and different burst rates, as a analogous to a lens, a lense that is creating a summation signal that the exit terminal is learning as it's learned response.
The idea is that our sensory inputs are sort of encoded in the input dendrite (or the basal dendrite) and our motor outputs are sort of encoded in the exit terminal (the Apical dendrite), and through cyclic looping activity our input learning (sensory in the bottom dendrite) is being mapped to our output learning (motor in the output exit terminal, or the Apical Dendrite)
So every neuron represents an actor, one that learns to respond selectively to certain patterns, and encodes in its exit a response action.
L1
result: OPEN-NO-CELL-CLOSURE
reason: A selective receiver could be an intervention substrate, but the selected SAN operation is recognition and projection, not a stated BCI intervention.
L2
result: OPEN-CAUSAL-BRIDGE
source chain: Learned receptive structure -> preferred-pattern recognition -> selective neural projection; feedback supports predictions of the cell's own response.
target chain: Understanding biological basis -> circuit-level intervention -> changed conscious experience.
reason: A selective receiver could be an intervention substrate, but the selected SAN operation is recognition and projection, not a stated BCI intervention.
residual: Recover an explicit experience-reconstruction or interface bridge for this source-stage family; do not infer controllability of experience from selectivity alone.
L3
result: BACKWARD-TEST-COMPLETED-BRIDGE-OPEN
constrained output: A brain-computer intervention targets, remediates or enhances an aspect of conscious experience.
required operation: An intervention must alter a biological state in a way linked to an experiential consequence.
source test: Changing a learned input pattern might change receiver output. This supplies a possible component, not a demonstrated translation of the target's experience-directed intervention.
alternative routes: Electrical stimulation; Sensory substitution or feedback interfaces; Pharmacological or optogenetic circuit modulation
remaining uncertainty: Recover an explicit experience-reconstruction or interface bridge for this source-stage family; do not infer controllability of experience from selectivity alone.
Limits and next evidence
Recover an explicit experience-reconstruction or interface bridge for this source-stage family; do not infer controllability of experience from selectivity alone.
At most a qualitative functional, conceptual or causal-prefix relation; no proof of functoriality, isomorphism, bisimulation, exact equation identity or calibrated statistical dependence.
Existing historical IDs for this exact source-atom/target-atom pair: none joined; this is not an automatic addition to governed counts.
SETH-SELECTED-40-03 · Source reviewed; relationship unresolved
SAN-AC012-MORPHOLOGY-SELECTIVE-PROJECTION → SETH-CS25-30
Learned neural selectivity and a proposal for precise experience control can cooperate; the frozen receiver passages do not yet connect the two outputs.
SAN: 2022-09-18 · Git repository commit record. Target: 2025-10-30 · doi:10.3389/fsci.2025.1546279.
Current anonymous public access to the exact source body was verified on 2026-10-11. The displayed 2022 date is a Git record date, not independently verified first-publication timing.
Target authors: Axel Cleeremans · Liad Mudrik · Anil K. Seth. Verify the target passage · See its short quotation in the paired cards.
Exact SAN source windows
SAN lines 3–3
Dendritic morphologies & synaptic connections represent learned memories. Memories activate when preferred patterns (the memory) are detected, this causes cells to create a selective transmission or a projection to other cells (neural arrays). In this note I am looking at Layer 6 & Layer 5 Pyramidal cells, to learn how the Apical Dendrite, or the Exit Terminal, is contributing to the rendered projection of that learned memory to the rest of its network. The main idea that I am exploring is that the signal transmitted from the exit terminal is not uniform along all branches, but instead is modified synapses by synapse, and that the memories in the exit terminal also transmit backwards to play a role in multiple seconds long coincidence detections involving back propagating action potentials from the exit terminal or Apical Dendrite to the Soma, that when combined with the Basal or sensory input dendrite can cause a coincidence detection that creates a special type of action potential event called burst firing.
SAN lines 49–61
I think of it as like the basal dendrite is learning to respond selectively to certain patterns, the apical is learning what patterns it responds to.
My intuition is that the apical is learning its own cell's response, and through feedback loops, with itself and with other pyramidal neurons the cell is going to make predictions about its own patterns of causation, what patterns it responds & how it reacts or responds. With the Apical Dendrite perhaps symbolizing a more complex response that is connected to more complex learning. (compared to other types of neurons)
The fact that bursting (bursts of somatic AP) results from the coincident activity of basal + apical, but not from either one separately is a different coincidence detection mechanism from the leaky integrate-and fire model of a neuron which explores coincident synaptic inputs, leading to coincident dendritic activity, which leads to the phasic action potential from just the south side of the Soma.
In the leaky integrate & fire neurons, and the Hodgkin & Huxley Model there is very little thought to the concept of the Soma burst spiking being a coincidence detection from both sides of the neuron. But if it's true for Layer 5 pyramidal cells, then we ought to know if there is a greater role for the exit terminal in all other types of neurons.
So perhaps we can think of the axon, with it's different types of action potentials, different durations (wave shapes, magnitudes), and different burst rates, as a analogous to a lens, a lense that is creating a summation signal that the exit terminal is learning as it's learned response.
The idea is that our sensory inputs are sort of encoded in the input dendrite (or the basal dendrite) and our motor outputs are sort of encoded in the exit terminal (the Apical dendrite), and through cyclic looping activity our input learning (sensory in the bottom dendrite) is being mapped to our output learning (motor in the output exit terminal, or the Apical Dendrite)
So every neuron represents an actor, one that learns to respond selectively to certain patterns, and encodes in its exit a response action.
L1
result: OPEN-NO-CELL-CLOSURE
reason: Learned neural selectivity and a proposal for precise experience control can cooperate; the frozen receiver passages do not yet connect the two outputs.
L2
result: OPEN-CAUSAL-BRIDGE
source chain: Learned receptive structure -> preferred-pattern recognition -> selective neural projection; feedback supports predictions of the cell's own response.
target chain: Account of biological mechanisms and experience -> precise intervention -> controlled experiential outcome.
reason: Learned neural selectivity and a proposal for precise experience control can cooperate; the frozen receiver passages do not yet connect the two outputs.
residual: Keep the receiver mechanism and every inherited favorable row. Route the missing experiential-control bridge through the broader genealogy, not a no-match verdict.
L3
result: BACKWARD-TEST-COMPLETED-BRIDGE-OPEN
constrained output: Neuroscience can deliberately control conscious experience with precision.
required operation: A causal relation between an accessible neural manipulation and an experiential outcome must be characterized.
source test: The target requires control of an experiential consequence. SAN describes a cell learning and transmitting a response, but this passage does not provide the intervention-to-experience link.
alternative routes: Stimulation of an ensemble; Sensory-feedback intervention; State-dependent neuromodulation
remaining uncertainty: Keep the receiver mechanism and every inherited favorable row. Route the missing experiential-control bridge through the broader genealogy, not a no-match verdict.
Target claim type: A forward-looking consequence of success in consciousness science, with phenotechnology explicitly attributed to Metzinger.
Limits and next evidence
Keep the receiver mechanism and every inherited favorable row. Route the missing experiential-control bridge through the broader genealogy, not a no-match verdict.
At most a qualitative functional, conceptual or causal-prefix relation; no proof of functoriality, isomorphism, bisimulation, exact equation identity or calibrated statistical dependence.
Existing historical IDs for this exact source-atom/target-atom pair: none joined; this is not an automatic addition to governed counts.
SETH-SELECTED-40-04 · Source reviewed; relationship unresolved
SAN-AC012-MORPHOLOGY-SELECTIVE-PROJECTION → SETH-CS25-28
Learned memory-pattern representation is relevant to a neural/perceptual mapping, but the selected text does not specify experiential similarity or a corresponding geometric relation.
SAN: 2022-09-18 · Git repository commit record. Target: 2025-10-30 · doi:10.3389/fsci.2025.1546279.
Current anonymous public access to the exact source body was verified on 2026-10-11. The displayed 2022 date is a Git record date, not independently verified first-publication timing.
Target authors: Axel Cleeremans · Liad Mudrik · Anil K. Seth. Verify the target passage · See its short quotation in the paired cards.
Exact SAN source windows
SAN lines 3–3
Dendritic morphologies & synaptic connections represent learned memories. Memories activate when preferred patterns (the memory) are detected, this causes cells to create a selective transmission or a projection to other cells (neural arrays). In this note I am looking at Layer 6 & Layer 5 Pyramidal cells, to learn how the Apical Dendrite, or the Exit Terminal, is contributing to the rendered projection of that learned memory to the rest of its network. The main idea that I am exploring is that the signal transmitted from the exit terminal is not uniform along all branches, but instead is modified synapses by synapse, and that the memories in the exit terminal also transmit backwards to play a role in multiple seconds long coincidence detections involving back propagating action potentials from the exit terminal or Apical Dendrite to the Soma, that when combined with the Basal or sensory input dendrite can cause a coincidence detection that creates a special type of action potential event called burst firing.
SAN lines 49–61
I think of it as like the basal dendrite is learning to respond selectively to certain patterns, the apical is learning what patterns it responds to.
My intuition is that the apical is learning its own cell's response, and through feedback loops, with itself and with other pyramidal neurons the cell is going to make predictions about its own patterns of causation, what patterns it responds & how it reacts or responds. With the Apical Dendrite perhaps symbolizing a more complex response that is connected to more complex learning. (compared to other types of neurons)
The fact that bursting (bursts of somatic AP) results from the coincident activity of basal + apical, but not from either one separately is a different coincidence detection mechanism from the leaky integrate-and fire model of a neuron which explores coincident synaptic inputs, leading to coincident dendritic activity, which leads to the phasic action potential from just the south side of the Soma.
In the leaky integrate & fire neurons, and the Hodgkin & Huxley Model there is very little thought to the concept of the Soma burst spiking being a coincidence detection from both sides of the neuron. But if it's true for Layer 5 pyramidal cells, then we ought to know if there is a greater role for the exit terminal in all other types of neurons.
So perhaps we can think of the axon, with it's different types of action potentials, different durations (wave shapes, magnitudes), and different burst rates, as a analogous to a lens, a lense that is creating a summation signal that the exit terminal is learning as it's learned response.
The idea is that our sensory inputs are sort of encoded in the input dendrite (or the basal dendrite) and our motor outputs are sort of encoded in the exit terminal (the Apical dendrite), and through cyclic looping activity our input learning (sensory in the bottom dendrite) is being mapped to our output learning (motor in the output exit terminal, or the Apical Dendrite)
So every neuron represents an actor, one that learns to respond selectively to certain patterns, and encodes in its exit a response action.
L1
result: OPEN-NO-CELL-CLOSURE
reason: Learned memory-pattern representation is relevant to a neural/perceptual mapping, but the selected text does not specify experiential similarity or a corresponding geometric relation.
L2
result: OPEN-CAUSAL-BRIDGE
source chain: Learned receptive structure -> preferred-pattern recognition -> selective neural projection; feedback supports predictions of the cell's own response.
target chain: Neural state-space organization -> relation to experiential properties -> comparison of perceptual similarities.
reason: Learned memory-pattern representation is relevant to a neural/perceptual mapping, but the selected text does not specify experiential similarity or a corresponding geometric relation.
residual: An explicit qualitative-content bridge and relational comparison remain to be recovered; no claim that the broader SAN corpus lacks either.
L3
result: BACKWARD-TEST-COMPLETED-BRIDGE-OPEN
constrained output: Neural and perceptual similarity structures can be related empirically.
required operation: Distinct neural configurations and distinct experiential properties must first be related; the full claim additionally needs an explicit similarity relation and empirical test.
source test: A representation-to-experience relation would be necessary to compare perceptual similarities. Memory activation and selective output alone do not establish that specific relation.
alternative routes: Population representational geometry; Attractor-state models; Relational quality-space models
remaining uncertainty: An explicit qualitative-content bridge and relational comparison remain to be recovered; no claim that the broader SAN corpus lacks either.
Limits and next evidence
An explicit qualitative-content bridge and relational comparison remain to be recovered; no claim that the broader SAN corpus lacks either.
At most a qualitative functional, conceptual or causal-prefix relation; no proof of functoriality, isomorphism, bisimulation, exact equation identity or calibrated statistical dependence.
Existing historical IDs for this exact source-atom/target-atom pair: none joined; this is not an automatic addition to governed counts.
SETH-SELECTED-40-05 · Additional partial-match candidate; independent review pending
SAN-AC012-MORPHOLOGY-SELECTIVE-PROJECTION → SETH-CS25-18
Feedback lets the SAN cell learn what inputs it responds to and predict its own causal responses; the target also gives internally generated prediction a role in processing.
SAN: 2022-09-18 · Git repository commit record. Target: 2025-10-30 · doi:10.3389/fsci.2025.1546279.
Current anonymous public access to the exact source body was verified on 2026-10-11. The displayed 2022 date is a Git record date, not independently verified first-publication timing.
Target authors: Axel Cleeremans · Liad Mudrik · Anil K. Seth. Verify the target passage · See its short quotation in the paired cards.
Exact SAN source windows
SAN lines 3–3
Dendritic morphologies & synaptic connections represent learned memories. Memories activate when preferred patterns (the memory) are detected, this causes cells to create a selective transmission or a projection to other cells (neural arrays). In this note I am looking at Layer 6 & Layer 5 Pyramidal cells, to learn how the Apical Dendrite, or the Exit Terminal, is contributing to the rendered projection of that learned memory to the rest of its network. The main idea that I am exploring is that the signal transmitted from the exit terminal is not uniform along all branches, but instead is modified synapses by synapse, and that the memories in the exit terminal also transmit backwards to play a role in multiple seconds long coincidence detections involving back propagating action potentials from the exit terminal or Apical Dendrite to the Soma, that when combined with the Basal or sensory input dendrite can cause a coincidence detection that creates a special type of action potential event called burst firing.
SAN lines 49–61
I think of it as like the basal dendrite is learning to respond selectively to certain patterns, the apical is learning what patterns it responds to.
My intuition is that the apical is learning its own cell's response, and through feedback loops, with itself and with other pyramidal neurons the cell is going to make predictions about its own patterns of causation, what patterns it responds & how it reacts or responds. With the Apical Dendrite perhaps symbolizing a more complex response that is connected to more complex learning. (compared to other types of neurons)
The fact that bursting (bursts of somatic AP) results from the coincident activity of basal + apical, but not from either one separately is a different coincidence detection mechanism from the leaky integrate-and fire model of a neuron which explores coincident synaptic inputs, leading to coincident dendritic activity, which leads to the phasic action potential from just the south side of the Soma.
In the leaky integrate & fire neurons, and the Hodgkin & Huxley Model there is very little thought to the concept of the Soma burst spiking being a coincidence detection from both sides of the neuron. But if it's true for Layer 5 pyramidal cells, then we ought to know if there is a greater role for the exit terminal in all other types of neurons.
So perhaps we can think of the axon, with it's different types of action potentials, different durations (wave shapes, magnitudes), and different burst rates, as a analogous to a lens, a lense that is creating a summation signal that the exit terminal is learning as it's learned response.
The idea is that our sensory inputs are sort of encoded in the input dendrite (or the basal dendrite) and our motor outputs are sort of encoded in the exit terminal (the Apical dendrite), and through cyclic looping activity our input learning (sensory in the bottom dendrite) is being mapped to our output learning (motor in the output exit terminal, or the Apical Dendrite)
So every neuron represents an actor, one that learns to respond selectively to certain patterns, and encodes in its exit a response action.
L1
result: PARTIAL-CAUSAL-PREFIX
reason: Feedback lets the SAN cell learn what inputs it responds to and predict its own causal responses; the target also gives internally generated prediction a role in processing.
L2
result: PARTIAL-CAUSAL-PREFIX
source chain: Learned receptive structure -> preferred-pattern recognition -> selective neural projection; feedback supports predictions of the cell's own response.
target chain: Internal predictions -> perceptual content -> sensory constraint from world and body.
reason: Feedback lets the SAN cell learn what inputs it responds to and predict its own causal responses; the target also gives internally generated prediction a role in processing.
residual: Predicting a neuron's own response is not the full claim that perceptual content is the brain's best guess of external causes. Preserve the prefix without upgrading its scale or semantic content.
L3
result: BOUNDED-BACKWARD-TEST
constrained output: Perceptual content reflects a prediction constrained by sensory signals.
required operation: An internal model must generate content and be sensitive to incoming evidence; a generic neural code does not establish the complete predictive mechanism.
source test: The target's constrained output needs internally generated prediction sensitive to incoming evidence. SAN lines 49–61 provide learned response prediction and feedback, preserving that local computational prefix.
alternative routes: Bayesian predictive coding; Recurrent generative models; Non-Bayesian learned predictive dynamics
remaining uncertainty: Predicting a neuron's own response is not the full claim that perceptual content is the brain's best guess of external causes. Preserve the prefix without upgrading its scale or semantic content.
Limits and next evidence
Predicting a neuron's own response is not the full claim that perceptual content is the brain's best guess of external causes. Preserve the prefix without upgrading its scale or semantic content.
At most a qualitative functional, conceptual or causal-prefix relation; no proof of functoriality, isomorphism, bisimulation, exact equation identity or calibrated statistical dependence.
Existing historical IDs for this exact source-atom/target-atom pair: none joined; this is not an automatic addition to governed counts.
SETH-SELECTED-40-06 · Source reviewed; relationship unresolved
SAN-AC012-MORPHOLOGY-SELECTIVE-PROJECTION → SETH-N24-01
Learned memory representations and belief states are related candidate notions, but the selected receptor/projection account is not yet a source-bound belief-to-phenomenology map.
SAN: 2022-09-18 · Git repository commit record. Target: 2026-05-04 · arxiv:2409.20318v3.
Current anonymous public access to the exact source body was verified on 2026-10-11. The displayed 2022 date is a Git record date, not independently verified first-publication timing.
Target authors: Lancelot Da Costa · Anil K. Seth · Karl Friston · Maxwell J. D. Ramstead · Lars Sandved-Smith. Verify the target passage · See its short quotation in the paired cards.
Exact SAN source windows
SAN lines 3–3
Dendritic morphologies & synaptic connections represent learned memories. Memories activate when preferred patterns (the memory) are detected, this causes cells to create a selective transmission or a projection to other cells (neural arrays). In this note I am looking at Layer 6 & Layer 5 Pyramidal cells, to learn how the Apical Dendrite, or the Exit Terminal, is contributing to the rendered projection of that learned memory to the rest of its network. The main idea that I am exploring is that the signal transmitted from the exit terminal is not uniform along all branches, but instead is modified synapses by synapse, and that the memories in the exit terminal also transmit backwards to play a role in multiple seconds long coincidence detections involving back propagating action potentials from the exit terminal or Apical Dendrite to the Soma, that when combined with the Basal or sensory input dendrite can cause a coincidence detection that creates a special type of action potential event called burst firing.
SAN lines 49–61
I think of it as like the basal dendrite is learning to respond selectively to certain patterns, the apical is learning what patterns it responds to.
My intuition is that the apical is learning its own cell's response, and through feedback loops, with itself and with other pyramidal neurons the cell is going to make predictions about its own patterns of causation, what patterns it responds & how it reacts or responds. With the Apical Dendrite perhaps symbolizing a more complex response that is connected to more complex learning. (compared to other types of neurons)
The fact that bursting (bursts of somatic AP) results from the coincident activity of basal + apical, but not from either one separately is a different coincidence detection mechanism from the leaky integrate-and fire model of a neuron which explores coincident synaptic inputs, leading to coincident dendritic activity, which leads to the phasic action potential from just the south side of the Soma.
In the leaky integrate & fire neurons, and the Hodgkin & Huxley Model there is very little thought to the concept of the Soma burst spiking being a coincidence detection from both sides of the neuron. But if it's true for Layer 5 pyramidal cells, then we ought to know if there is a greater role for the exit terminal in all other types of neurons.
So perhaps we can think of the axon, with it's different types of action potentials, different durations (wave shapes, magnitudes), and different burst rates, as a analogous to a lens, a lense that is creating a summation signal that the exit terminal is learning as it's learned response.
The idea is that our sensory inputs are sort of encoded in the input dendrite (or the basal dendrite) and our motor outputs are sort of encoded in the exit terminal (the Apical dendrite), and through cyclic looping activity our input learning (sensory in the bottom dendrite) is being mapped to our output learning (motor in the output exit terminal, or the Apical Dendrite)
So every neuron represents an actor, one that learns to respond selectively to certain patterns, and encodes in its exit a response action.
L1
result: OPEN-NO-CELL-CLOSURE
reason: Learned memory representations and belief states are related candidate notions, but the selected receptor/projection account is not yet a source-bound belief-to-phenomenology map.
L2
result: OPEN-CAUSAL-BRIDGE
source chain: Learned receptive structure -> preferred-pattern recognition -> selective neural projection; feedback supports predictions of the cell's own response.
target chain: Internal state encodes approximate posterior -> belief-to-phenomenology function -> experiential content.
reason: Learned memory representations and belief states are related candidate notions, but the selected receptor/projection account is not yet a source-bound belief-to-phenomenology map.
residual: Recover the experience and belief-role bridge. Do not equate memory, all neural output and phenomenology merely by renaming them.
L3
result: BACKWARD-TEST-COMPLETED-BRIDGE-OPEN
constrained output: Phenomenological content is represented as a function of beliefs.
required operation: The selected hypothesis links a specified internal belief state to phenomenological content; the paper does not claim every neural representation is a posterior belief.
source test: The target assumes a function from posterior beliefs to experience. SAN supplies pattern-selective neural output; a translation from that output to both operands of the target function remains incomplete.
alternative routes: Identity map; Marginalization; A coarser non-injective mapping
remaining uncertainty: Recover the experience and belief-role bridge. Do not equate memory, all neural output and phenomenology merely by renaming them.
Limits and next evidence
Recover the experience and belief-role bridge. Do not equate memory, all neural output and phenomenology merely by renaming them.
At most a qualitative functional, conceptual or causal-prefix relation; no proof of functoriality, isomorphism, bisimulation, exact equation identity or calibrated statistical dependence.
Existing historical IDs for this exact source-atom/target-atom pair: none joined; this is not an automatic addition to governed counts.
SETH-SELECTED-40-07 · Source reviewed; relationship unresolved
SAN-AC012-MORPHOLOGY-SELECTIVE-PROJECTION → SETH-N24-05
Morphology-dependent pattern recognition supplies a possible representational substrate; a phenomenological geometry does not follow from that substrate alone.
SAN: 2022-09-18 · Git repository commit record. Target: 2026-05-04 · arxiv:2409.20318v3.
Current anonymous public access to the exact source body was verified on 2026-10-11. The displayed 2022 date is a Git record date, not independently verified first-publication timing.
Target authors: Lancelot Da Costa · Anil K. Seth · Karl Friston · Maxwell J. D. Ramstead · Lars Sandved-Smith. Verify the target passage · See its short quotation in the paired cards.
Exact SAN source windows
SAN lines 3–3
Dendritic morphologies & synaptic connections represent learned memories. Memories activate when preferred patterns (the memory) are detected, this causes cells to create a selective transmission or a projection to other cells (neural arrays). In this note I am looking at Layer 6 & Layer 5 Pyramidal cells, to learn how the Apical Dendrite, or the Exit Terminal, is contributing to the rendered projection of that learned memory to the rest of its network. The main idea that I am exploring is that the signal transmitted from the exit terminal is not uniform along all branches, but instead is modified synapses by synapse, and that the memories in the exit terminal also transmit backwards to play a role in multiple seconds long coincidence detections involving back propagating action potentials from the exit terminal or Apical Dendrite to the Soma, that when combined with the Basal or sensory input dendrite can cause a coincidence detection that creates a special type of action potential event called burst firing.
SAN lines 49–61
I think of it as like the basal dendrite is learning to respond selectively to certain patterns, the apical is learning what patterns it responds to.
My intuition is that the apical is learning its own cell's response, and through feedback loops, with itself and with other pyramidal neurons the cell is going to make predictions about its own patterns of causation, what patterns it responds & how it reacts or responds. With the Apical Dendrite perhaps symbolizing a more complex response that is connected to more complex learning. (compared to other types of neurons)
The fact that bursting (bursts of somatic AP) results from the coincident activity of basal + apical, but not from either one separately is a different coincidence detection mechanism from the leaky integrate-and fire model of a neuron which explores coincident synaptic inputs, leading to coincident dendritic activity, which leads to the phasic action potential from just the south side of the Soma.
In the leaky integrate & fire neurons, and the Hodgkin & Huxley Model there is very little thought to the concept of the Soma burst spiking being a coincidence detection from both sides of the neuron. But if it's true for Layer 5 pyramidal cells, then we ought to know if there is a greater role for the exit terminal in all other types of neurons.
So perhaps we can think of the axon, with it's different types of action potentials, different durations (wave shapes, magnitudes), and different burst rates, as a analogous to a lens, a lense that is creating a summation signal that the exit terminal is learning as it's learned response.
The idea is that our sensory inputs are sort of encoded in the input dendrite (or the basal dendrite) and our motor outputs are sort of encoded in the exit terminal (the Apical dendrite), and through cyclic looping activity our input learning (sensory in the bottom dendrite) is being mapped to our output learning (motor in the output exit terminal, or the Apical Dendrite)
So every neuron represents an actor, one that learns to respond selectively to certain patterns, and encodes in its exit a response action.
L1
result: OPEN-NO-CELL-CLOSURE
reason: Morphology-dependent pattern recognition supplies a possible representational substrate; a phenomenological geometry does not follow from that substrate alone.
L2
result: OPEN-CAUSAL-BRIDGE
source chain: Learned receptive structure -> preferred-pattern recognition -> selective neural projection; feedback supports predictions of the cell's own response.
target chain: Belief/phenomenology correspondence -> structured experiential domain -> formal comparisons and predictions.
reason: Morphology-dependent pattern recognition supplies a possible representational substrate; a phenomenological geometry does not follow from that substrate alone.
residual: Retain as open for the full source family and earlier/later formulations; no metric, formal equivalence or adverse program conclusion is asserted.
L3
result: BACKWARD-TEST-COMPLETED-BRIDGE-OPEN
constrained output: A geometry enables characterization of phenomenological differences.
required operation: A representational domain must be linked to experiential content before a metric or other geometric structure can support comparisons.
source test: A geometry of experience first needs an explicit representational-to-experiential relation. The receiver source has no resolved link to that relation in the inspected windows.
alternative routes: Information geometry; Relational quality spaces; Empirical neural/perceptual similarity models
remaining uncertainty: Retain as open for the full source family and earlier/later formulations; no metric, formal equivalence or adverse program conclusion is asserted.
Limits and next evidence
Retain as open for the full source family and earlier/later formulations; no metric, formal equivalence or adverse program conclusion is asserted.
At most a qualitative functional, conceptual or causal-prefix relation; no proof of functoriality, isomorphism, bisimulation, exact equation identity or calibrated statistical dependence.
Existing historical IDs for this exact source-atom/target-atom pair: none joined; this is not an automatic addition to governed counts.
SETH-SELECTED-40-08 · Additional partial-match candidate; independent review pending
SAN-AC012-MORPHOLOGY-SELECTIVE-PROJECTION → SETH-T26-05
Structured neural reception selects a pattern-dependent outgoing response in SAN; the reservoir architecture also transforms input-dependent internal activity into selected output.
SAN: 2022-09-18 · Git repository commit record. Target: 2024-06-27 · arxiv:2406.19201v1.
Current anonymous public access to the exact source body was verified on 2026-10-11. The displayed 2022 date is a Git record date, not independently verified first-publication timing.
Target authors: Hanna M. Tolle · Andrea I. Luppi · Anil K. Seth · Pedro A. M. Mediano. Verify the target passage · See its short quotation in the paired cards.
Exact SAN source windows
SAN lines 3–3
Dendritic morphologies & synaptic connections represent learned memories. Memories activate when preferred patterns (the memory) are detected, this causes cells to create a selective transmission or a projection to other cells (neural arrays). In this note I am looking at Layer 6 & Layer 5 Pyramidal cells, to learn how the Apical Dendrite, or the Exit Terminal, is contributing to the rendered projection of that learned memory to the rest of its network. The main idea that I am exploring is that the signal transmitted from the exit terminal is not uniform along all branches, but instead is modified synapses by synapse, and that the memories in the exit terminal also transmit backwards to play a role in multiple seconds long coincidence detections involving back propagating action potentials from the exit terminal or Apical Dendrite to the Soma, that when combined with the Basal or sensory input dendrite can cause a coincidence detection that creates a special type of action potential event called burst firing.
SAN lines 49–61
I think of it as like the basal dendrite is learning to respond selectively to certain patterns, the apical is learning what patterns it responds to.
My intuition is that the apical is learning its own cell's response, and through feedback loops, with itself and with other pyramidal neurons the cell is going to make predictions about its own patterns of causation, what patterns it responds & how it reacts or responds. With the Apical Dendrite perhaps symbolizing a more complex response that is connected to more complex learning. (compared to other types of neurons)
The fact that bursting (bursts of somatic AP) results from the coincident activity of basal + apical, but not from either one separately is a different coincidence detection mechanism from the leaky integrate-and fire model of a neuron which explores coincident synaptic inputs, leading to coincident dendritic activity, which leads to the phasic action potential from just the south side of the Soma.
In the leaky integrate & fire neurons, and the Hodgkin & Huxley Model there is very little thought to the concept of the Soma burst spiking being a coincidence detection from both sides of the neuron. But if it's true for Layer 5 pyramidal cells, then we ought to know if there is a greater role for the exit terminal in all other types of neurons.
So perhaps we can think of the axon, with it's different types of action potentials, different durations (wave shapes, magnitudes), and different burst rates, as a analogous to a lens, a lense that is creating a summation signal that the exit terminal is learning as it's learned response.
The idea is that our sensory inputs are sort of encoded in the input dendrite (or the basal dendrite) and our motor outputs are sort of encoded in the exit terminal (the Apical dendrite), and through cyclic looping activity our input learning (sensory in the bottom dendrite) is being mapped to our output learning (motor in the output exit terminal, or the Apical Dendrite)
So every neuron represents an actor, one that learns to respond selectively to certain patterns, and encodes in its exit a response action.
L1
result: PARTIAL-FUNCTIONAL-PREFIX
reason: Structured neural reception selects a pattern-dependent outgoing response in SAN; the reservoir architecture also transforms input-dependent internal activity into selected output.
L2
result: PARTIAL-FUNCTIONAL-PREFIX
source chain: Learned receptive structure -> preferred-pattern recognition -> selective neural projection; feedback supports predictions of the cell's own response.
target chain: Input -> fixed nonlinear recurrent reservoir state -> trained linear readout.
reason: Structured neural reception selects a pattern-dependent outgoing response in SAN; the reservoir architecture also transforms input-dependent internal activity into selected output.
residual: SAN's learned cellular structure and recurrent response prediction do not specify a fixed reservoir or a separately trained linear readout. This is a broad component match, not reservoir-computing priority.
L3
result: BOUNDED-BACKWARD-TEST
constrained output: A recurrent computer transforms temporal input into a useful output.
required operation: Input must change recurrent state so a readout can distinguish task-relevant patterns; the selected architecture further fixes reservoir weights and trains a linear readout.
source test: Producing distinguishable task outputs requires input-sensitive internal dynamics and an output mapping. The SAN recognition-to-projection chain supplies that bounded functional subset.
alternative routes: Echo-state reservoirs; Liquid-state machines; Fully trained recurrent networks
remaining uncertainty: SAN's learned cellular structure and recurrent response prediction do not specify a fixed reservoir or a separately trained linear readout. This is a broad component match, not reservoir-computing priority.
Limits and next evidence
SAN's learned cellular structure and recurrent response prediction do not specify a fixed reservoir or a separately trained linear readout. This is a broad component match, not reservoir-computing priority.
At most a qualitative functional, conceptual or causal-prefix relation; no proof of functoriality, isomorphism, bisimulation, exact equation identity or calibrated statistical dependence.
Existing historical IDs for this exact source-atom/target-atom pair: SETH-SANFULL-056.
SETH-SELECTED-40-09 · Additional partial-match candidate; independent review pending
SAN-AC013-QUALIA-RECONSTRUCTION → SETH-CS25-22
SAN proposes temporal alignment between oscillating regions so a receiving network can detect correlations in another region's signals.
SAN: 2022-06-12 · Git repository commit record. Target: 2025-10-30 · doi:10.3389/fsci.2025.1546279.
Current anonymous public access to the exact source body was verified on 2026-10-11. The displayed 2022 date is a Git record date, not independently verified first-publication timing.
Target authors: Axel Cleeremans · Liad Mudrik · Anil K. Seth. Verify the target passage · See its short quotation in the paired cards.
Exact SAN source windows
SAN lines 2–2
Brocoli Properties example: All the properties of Brocoli are features learned by the neural network of the human brain. So the color, texture, feel, crunch, taste, smell these are learned patterns, information patterns are detected as coincidence evens by receptors on dendrites, then by neurons, then by arrays of neurons, then by nuclei in the thalamus, then by cortical columns & the hippocampal entorhinal circuit, then by the thalamus again, essentially circuiting throughout your brain in a spatial temporal oscillating feedback loop.
SAN lines 12–16
The main idea that I have about the importance of brainwave synchrony between brain regions when objects become conscious comes from thinking about Tim Mullen's neuropype, and how he aligned unstructure data like EEG, EMG, ECG (Brain, Muscle, Heart sensors) by understanding the assumptions made in each technology and figuring out how to correctly align the signal differences temporally, so that when a brainwave spikes, one might associate it with a heart rate spike, given a time based association.
In the same way if one part of the brain wants to understand unstructured data from another part of the brain, I suspect synchrony (achieved by listening to one's self) is creating a temporal alignment between two oscillators so that correlations between the two sets of unstructured data can be found by the neural network of your mind, and you make sense out of an association between lightning & thunder for example.
So if we knew how the properties of brocoli were represented by scanning them at the Thalamus at the moment you were thinking about them, then we might identify which cortical column is managing those properties when you are not thinking about them, and even if you have a completely different personality that manages brocoli in a different region of the cortex, by stimulating the thalamus with the properties of brocoli that we learned from watching your thalamus we might trigger the experience of brocoli in your brain, we might have a next generation brain computer interface for bringing Augmented Reality & Virtual Reality experiences not through glasses but through direct brain stimulation, to recreate known patterns.
L1
result: PARTIAL-PROBLEM-FIT
reason: SAN proposes temporal alignment between oscillating regions so a receiving network can detect correlations in another region's signals.
L2
result: PARTIAL-PROBLEM-FIT
source chain: Learned multisensory patterns circulate through a neural feedback loop -> measured phase-pattern configuration -> proposed pattern reinstatement to recreate an experience.
target chain: Broadcast message -> unspecified receiver interpretation -> useful downstream processing.
reason: SAN proposes temporal alignment between oscillating regions so a receiving network can detect correlations in another region's signals.
residual: The later review asks a question; it does not adopt the synchrony proposal or experimentally validate it.
L3
result: BOUNDED-BACKWARD-TEST
constrained output: A receiving neuron makes useful use of a broadcast message.
required operation: Some implemented receiver dependency must discriminate consequential input differences; arrival alone does not explain use.
source test: To use a received message, a network needs a response-sensitive relation to signal differences. SAN line 14 supplies a concrete alignment-and-correlation proposal for that explanatory job.
alternative routes: Learned receptive selectivity; Fixed feature-sensitive circuitry; Population-level decoding
remaining uncertainty: The later review asks a question; it does not adopt the synchrony proposal or experimentally validate it.
Limits and next evidence
The later review asks a question; it does not adopt the synchrony proposal or experimentally validate it.
At most a qualitative functional, conceptual or causal-prefix relation; no proof of functoriality, isomorphism, bisimulation, exact equation identity or calibrated statistical dependence.
Existing historical IDs for this exact source-atom/target-atom pair: none joined; this is not an automatic addition to governed counts.
SETH-SELECTED-40-10 · Previously independently source-reviewed subset
SAN-AC013-QUALIA-RECONSTRUCTION → SETH-CS25-38
Both link an explanation of experience to an intervention that changes it. SAN proposes measuring a person’s sensory pattern and using a corresponding pattern to recreate experience.
SAN: 2022-06-12 · Git repository commit record. Target: 2025-10-30 · doi:10.3389/fsci.2025.1546279.
Current anonymous public access to the exact source body was verified on 2026-10-11. The displayed 2022 date is a Git record date, not independently verified first-publication timing.
Target authors: Axel Cleeremans · Liad Mudrik · Anil K. Seth. Verify the target passage · See its short quotation in the paired cards.
Exact SAN source windows
SAN lines 4–4
The point of the Brocoli Properties example was to show that if we could intercept at the Thalamus how the properties of Brocoli are signals being passed to the neo cortex, then we could recreate the experience of Brocoli in the Thalamus by creating that set of temporal & spatially distributed phase changes. This could be done at any point in the brain, or on any part of the body, as illustrated in the example of the tongue strip with a grid of electrical signals that correspond to the pixel changes of a camera's viewfinder, a device that allows people to see through their tongue or via any part of their body.
SAN lines 16–16
So if we knew how the properties of brocoli were represented by scanning them at the Thalamus at the moment you were thinking about them, then we might identify which cortical column is managing those properties when you are not thinking about them, and even if you have a completely different personality that manages brocoli in a different region of the cortex, by stimulating the thalamus with the properties of brocoli that we learned from watching your thalamus we might trigger the experience of brocoli in your brain, we might have a next generation brain computer interface for bringing Augmented Reality & Virtual Reality experiences not through glasses but through direct brain stimulation, to recreate known patterns.
L1
result: PARTIAL_OPERATION
reason: Both propose a causal intervention on neural mechanisms to affect experience. SAN specifies reproducing a learned distributed pattern; the target specifies neither that code nor faithful replay of a particular sensory episode.
L2
result: PARTIAL_CAUSAL_SUBCHAIN
reason: Neural account -> intervention target -> altered experience is common. SAN adds measurement of personal sensory patterns and proposed thalamic stimulation; target purposes include remediation and enhancement rather than only recreation.
L3
result: SUPPORTED_CONSTRAINED_OUTPUT_RECONSTRUCTION_WITH_NONUNIQUE_SOLUTION
reason: Constrained result: A circuit intervention intentionally remediates or enhances a conscious experience. Required operation: The manipulated circuit variable must causally affect the experiential feature, with sufficient targeting to separate it from a merely behavioral effect. Source test: Lines 4 and 16 propose measuring a sensory pattern and stimulating the thalamus with a corresponding pattern to trigger experience. Alternatives: Electrical activity-pattern stimulation; Neurochemical modulation; Other content-relevant neural codes. Remaining uncertainty: The target does not require SAN's code, anatomical location, replay fidelity or proposed BCI. Neither text demonstrates this SAN procedure.
Limits and next evidence
SAN explicitly names sensory-substitution examples, David Eagleman, Hawkins and Tim Mullen. Those components remain prior-work boundaries; the proposed synthesis is not evidence of inventing them.
At most a qualitative functional, conceptual or causal-prefix relation; no proof of functoriality, isomorphism, bisimulation, exact equation identity or calibrated statistical dependence.
Existing historical IDs for this exact source-atom/target-atom pair: SETH-CS25-DOCWIN-SETH-CS25-38-SAN-AC013-QUALIA-RECONSTRUCTION-4377D46F92F0.
SETH-SELECTED-40-11 · Previously independently source-reviewed subset
SAN-AC013-QUALIA-RECONSTRUCTION → SETH-CS25-30
In both passages, learning how neural activity relates to experience is meant to make deliberate changes to experience possible. This is another target passage, not another independent SAN concept.
SAN: 2022-06-12 · Git repository commit record. Target: 2025-10-30 · doi:10.3389/fsci.2025.1546279.
Current anonymous public access to the exact source body was verified on 2026-10-11. The displayed 2022 date is a Git record date, not independently verified first-publication timing.
Target authors: Axel Cleeremans · Liad Mudrik · Anil K. Seth. Verify the target passage · See its short quotation in the paired cards.
Exact SAN source windows
SAN lines 4–4
The point of the Brocoli Properties example was to show that if we could intercept at the Thalamus how the properties of Brocoli are signals being passed to the neo cortex, then we could recreate the experience of Brocoli in the Thalamus by creating that set of temporal & spatially distributed phase changes. This could be done at any point in the brain, or on any part of the body, as illustrated in the example of the tongue strip with a grid of electrical signals that correspond to the pixel changes of a camera's viewfinder, a device that allows people to see through their tongue or via any part of their body.
SAN lines 16–16
So if we knew how the properties of brocoli were represented by scanning them at the Thalamus at the moment you were thinking about them, then we might identify which cortical column is managing those properties when you are not thinking about them, and even if you have a completely different personality that manages brocoli in a different region of the cortex, by stimulating the thalamus with the properties of brocoli that we learned from watching your thalamus we might trigger the experience of brocoli in your brain, we might have a next generation brain computer interface for bringing Augmented Reality & Virtual Reality experiences not through glasses but through direct brain stimulation, to recreate known patterns.
L1
result: PARTIAL_FUNCTION_AND_OPERATION
reason: The target's prospective ability to control experience contains the goal served by SAN's proposed sensory-pattern recreation. SAN does not supply all conscious-state identification or arbitrary precise control.
L2
result: PARTIAL_CAUSAL_SUBCHAIN
reason: Understanding the neural basis is meant to make intervention possible in both. The target includes distinguishing conscious from nonconscious states, which SAN's broccoli example does not establish.
L3
result: SUPPORTED_CONSTRAINED_OUTPUT_RECONSTRUCTION_WITH_NONUNIQUE_SOLUTION
reason: Constrained result: A tool controls consciousness as precisely as tools aim to control behavior. Required operation: Identify manipulable neural variables with reliable experiential consequences; behavioral control alone is insufficient. Source test: SAN line 4 proposes a causal replay mechanism and line 16 adds individual-pattern measurement and direct stimulation. Alternatives: Pattern-specific stimulation; State-level neuromodulation; Other experimentally calibrated neural interventions. Remaining uncertainty: Recreating one proposed content does not derive a general test for consciousness or the target's precision of control.
Target claim type: A forward-looking consequence of success in consciousness science, with phenotechnology explicitly attributed to Metzinger.
Limits and next evidence
This is a second passage in the same target document and reuses the BCI row's SAN concept. Reference 22 / footnote 3 identifies Metzinger's 2009 phenotechnology antecedent; no firstness over that line is established.
At most a qualitative functional, conceptual or causal-prefix relation; no proof of functoriality, isomorphism, bisimulation, exact equation identity or calibrated statistical dependence.
Existing historical IDs for this exact source-atom/target-atom pair: SETH-CS25-DOCWIN-SETH-CS25-30-SAN-AC013-QUALIA-RECONSTRUCTION-1551F6988344.
SETH-SELECTED-40-12 · Additional partial-match candidate; independent review pending
SAN-AC013-QUALIA-RECONSTRUCTION → SETH-CS25-28
SAN relates distributed neural phase patterns to differentiated sensory properties, while the target relates neural configurations to perceptual qualities and their similarities.
SAN: 2022-06-12 · Git repository commit record. Target: 2025-10-30 · doi:10.3389/fsci.2025.1546279.
Current anonymous public access to the exact source body was verified on 2026-10-11. The displayed 2022 date is a Git record date, not independently verified first-publication timing.
Target authors: Axel Cleeremans · Liad Mudrik · Anil K. Seth. Verify the target passage · See its short quotation in the paired cards.
Exact SAN source windows
SAN lines 2–4
Brocoli Properties example: All the properties of Brocoli are features learned by the neural network of the human brain. So the color, texture, feel, crunch, taste, smell these are learned patterns, information patterns are detected as coincidence evens by receptors on dendrites, then by neurons, then by arrays of neurons, then by nuclei in the thalamus, then by cortical columns & the hippocampal entorhinal circuit, then by the thalamus again, essentially circuiting throughout your brain in a spatial temporal oscillating feedback loop.
The point of the Brocoli Properties example was to show that if we could intercept at the Thalamus how the properties of Brocoli are signals being passed to the neo cortex, then we could recreate the experience of Brocoli in the Thalamus by creating that set of temporal & spatially distributed phase changes. This could be done at any point in the brain, or on any part of the body, as illustrated in the example of the tongue strip with a grid of electrical signals that correspond to the pixel changes of a camera's viewfinder, a device that allows people to see through their tongue or via any part of their body.
SAN lines 16–16
So if we knew how the properties of brocoli were represented by scanning them at the Thalamus at the moment you were thinking about them, then we might identify which cortical column is managing those properties when you are not thinking about them, and even if you have a completely different personality that manages brocoli in a different region of the cortex, by stimulating the thalamus with the properties of brocoli that we learned from watching your thalamus we might trigger the experience of brocoli in your brain, we might have a next generation brain computer interface for bringing Augmented Reality & Virtual Reality experiences not through glasses but through direct brain stimulation, to recreate known patterns.
L1
result: PARTIAL-CONCEPTUAL-BRIDGE
reason: SAN relates distributed neural phase patterns to differentiated sensory properties, while the target relates neural configurations to perceptual qualities and their similarities.
L2
result: PARTIAL-CONCEPTUAL-BRIDGE
source chain: Learned multisensory patterns circulate through a neural feedback loop -> measured phase-pattern configuration -> proposed pattern reinstatement to recreate an experience.
target chain: Neural state-space organization -> relation to experiential properties -> comparison of perceptual similarities.
reason: SAN relates distributed neural phase patterns to differentiated sensory properties, while the target relates neural configurations to perceptual qualities and their similarities.
residual: There is no empirical pairwise-distance matrix, quality-space metric or intersubject equivalence proof in this SAN passage.
L3
result: BOUNDED-BACKWARD-TEST
constrained output: Neural and perceptual similarity structures can be related empirically.
required operation: Distinct neural configurations and distinct experiential properties must first be related; the full claim additionally needs an explicit similarity relation and empirical test.
source test: A neural/perceptual similarity mapping presupposes distinguishable neural states related to distinguishable experiences. The broccoli reconstruction example explicitly supplies this narrower state-to-quality bridge.
alternative routes: Population representational geometry; Attractor-state models; Relational quality-space models
remaining uncertainty: There is no empirical pairwise-distance matrix, quality-space metric or intersubject equivalence proof in this SAN passage.
Limits and next evidence
There is no empirical pairwise-distance matrix, quality-space metric or intersubject equivalence proof in this SAN passage.
At most a qualitative functional, conceptual or causal-prefix relation; no proof of functoriality, isomorphism, bisimulation, exact equation identity or calibrated statistical dependence.
Existing historical IDs for this exact source-atom/target-atom pair: SETH-CS25-DOCWIN-SETH-CS25-28-SAN-AC013-QUALIA-RECONSTRUCTION-6FD17AF0FBA2.
SETH-SELECTED-40-13 · Additional partial-match candidate; independent review pending
SAN-AC013-QUALIA-RECONSTRUCTION → SETH-CS25-18
The SAN example treats an experience as internally reconstructed through learned recurrent neural patterns, not as the external object itself.
SAN: 2022-06-12 · Git repository commit record. Target: 2025-10-30 · doi:10.3389/fsci.2025.1546279.
Current anonymous public access to the exact source body was verified on 2026-10-11. The displayed 2022 date is a Git record date, not independently verified first-publication timing.
Target authors: Axel Cleeremans · Liad Mudrik · Anil K. Seth. Verify the target passage · See its short quotation in the paired cards.
Exact SAN source windows
SAN lines 2–4
Brocoli Properties example: All the properties of Brocoli are features learned by the neural network of the human brain. So the color, texture, feel, crunch, taste, smell these are learned patterns, information patterns are detected as coincidence evens by receptors on dendrites, then by neurons, then by arrays of neurons, then by nuclei in the thalamus, then by cortical columns & the hippocampal entorhinal circuit, then by the thalamus again, essentially circuiting throughout your brain in a spatial temporal oscillating feedback loop.
The point of the Brocoli Properties example was to show that if we could intercept at the Thalamus how the properties of Brocoli are signals being passed to the neo cortex, then we could recreate the experience of Brocoli in the Thalamus by creating that set of temporal & spatially distributed phase changes. This could be done at any point in the brain, or on any part of the body, as illustrated in the example of the tongue strip with a grid of electrical signals that correspond to the pixel changes of a camera's viewfinder, a device that allows people to see through their tongue or via any part of their body.
SAN lines 10–16
There are papers that suggest the hippocampus is like a cortical column, and or a comparator to cortical column data, maybe its cortical column #1, and maybe that is where the base rendering of consciousness begins, or likes to sit, like the central locus of focus, but the information in just the hippocampal-enthorhinal loop might be like you, the room, the walls, and where you are in the world, and following the book A Thousand Brains by Jeff Hawkins it might be true that each of your other cortical columns gets busy representing other objects, sounds, concepts, and things that are part of your experience in the room, so the cup on the table, and the hot sauce, are each managed by different cortical columns, all the properties of the cup might be sent to a cortical column that is managing the sensory integration for that object, its orientation, position, temperature, texture, sound etc... and when your awareness of that cup increase that is plausibly going to be an increase in the syncronization between that cortical column and the hippocampal entorhinal loop & the parietal temperal junction all at once, this is something we could look for with current medical imaging technology, whether there is brainwave frequency synchronization between hippocampal frequencies and the parietal temporal junction and or specific cortical columns.
The main idea that I have about the importance of brainwave synchrony between brain regions when objects become conscious comes from thinking about Tim Mullen's neuropype, and how he aligned unstructure data like EEG, EMG, ECG (Brain, Muscle, Heart sensors) by understanding the assumptions made in each technology and figuring out how to correctly align the signal differences temporally, so that when a brainwave spikes, one might associate it with a heart rate spike, given a time based association.
In the same way if one part of the brain wants to understand unstructured data from another part of the brain, I suspect synchrony (achieved by listening to one's self) is creating a temporal alignment between two oscillators so that correlations between the two sets of unstructured data can be found by the neural network of your mind, and you make sense out of an association between lightning & thunder for example.
So if we knew how the properties of brocoli were represented by scanning them at the Thalamus at the moment you were thinking about them, then we might identify which cortical column is managing those properties when you are not thinking about them, and even if you have a completely different personality that manages brocoli in a different region of the cortex, by stimulating the thalamus with the properties of brocoli that we learned from watching your thalamus we might trigger the experience of brocoli in your brain, we might have a next generation brain computer interface for bringing Augmented Reality & Virtual Reality experiences not through glasses but through direct brain stimulation, to recreate known patterns.
L1
result: PARTIAL-CONCEPTUAL-BRIDGE
reason: The SAN example treats an experience as internally reconstructed through learned recurrent neural patterns, not as the external object itself.
L2
result: PARTIAL-CONCEPTUAL-BRIDGE
source chain: Learned multisensory patterns circulate through a neural feedback loop -> measured phase-pattern configuration -> proposed pattern reinstatement to recreate an experience.
target chain: Internal predictions -> perceptual content -> sensory constraint from world and body.
reason: The SAN example treats an experience as internally reconstructed through learned recurrent neural patterns, not as the external object itself.
residual: The target specifically makes predictions the source of perceptual content; SAN's cited Eagleman/Hawkins discussion must remain attributed and cannot be promoted to a novel Bayesian mechanism.
L3
result: BOUNDED-BACKWARD-TEST
constrained output: Perceptual content reflects a prediction constrained by sensory signals.
required operation: An internal model must generate content and be sensitive to incoming evidence; a generic neural code does not establish the complete predictive mechanism.
source test: SAN's selected source describes a learned recurrent representation and its proposed reinstatement. This supplies a particular route to internally reconstructed perceptual content; it does not make learning necessary for every possible generator of internal content, or supply the target's complete predictive-error mechanism.
alternative routes: Bayesian predictive coding; Recurrent generative models; Non-Bayesian learned predictive dynamics; Innate or explicitly specified generative circuitry; learning is not a universal premise of this backward inference
remaining uncertainty: The target specifically makes predictions the source of perceptual content; SAN's cited Eagleman/Hawkins discussion must remain attributed and cannot be promoted to a novel Bayesian mechanism.
Limits and next evidence
The target specifically makes predictions the source of perceptual content; SAN's cited Eagleman/Hawkins discussion must remain attributed and cannot be promoted to a novel Bayesian mechanism.
At most a qualitative functional, conceptual or causal-prefix relation; no proof of functoriality, isomorphism, bisimulation, exact equation identity or calibrated statistical dependence.
Existing historical IDs for this exact source-atom/target-atom pair: none joined; this is not an automatic addition to governed counts.
SETH-SELECTED-40-14 · Additional partial-match candidate; independent review pending
SAN-AC013-QUALIA-RECONSTRUCTION → SETH-N24-01
Both connect a specified internal representation with experiential content: SAN through a measured pattern and its reinstatement, the target through a conditional belief function.
SAN: 2022-06-12 · Git repository commit record. Target: 2026-05-04 · arxiv:2409.20318v3.
Current anonymous public access to the exact source body was verified on 2026-10-11. The displayed 2022 date is a Git record date, not independently verified first-publication timing.
Target authors: Lancelot Da Costa · Anil K. Seth · Karl Friston · Maxwell J. D. Ramstead · Lars Sandved-Smith. Verify the target passage · See its short quotation in the paired cards.
Exact SAN source windows
SAN lines 2–4
Brocoli Properties example: All the properties of Brocoli are features learned by the neural network of the human brain. So the color, texture, feel, crunch, taste, smell these are learned patterns, information patterns are detected as coincidence evens by receptors on dendrites, then by neurons, then by arrays of neurons, then by nuclei in the thalamus, then by cortical columns & the hippocampal entorhinal circuit, then by the thalamus again, essentially circuiting throughout your brain in a spatial temporal oscillating feedback loop.
The point of the Brocoli Properties example was to show that if we could intercept at the Thalamus how the properties of Brocoli are signals being passed to the neo cortex, then we could recreate the experience of Brocoli in the Thalamus by creating that set of temporal & spatially distributed phase changes. This could be done at any point in the brain, or on any part of the body, as illustrated in the example of the tongue strip with a grid of electrical signals that correspond to the pixel changes of a camera's viewfinder, a device that allows people to see through their tongue or via any part of their body.
SAN lines 16–16
So if we knew how the properties of brocoli were represented by scanning them at the Thalamus at the moment you were thinking about them, then we might identify which cortical column is managing those properties when you are not thinking about them, and even if you have a completely different personality that manages brocoli in a different region of the cortex, by stimulating the thalamus with the properties of brocoli that we learned from watching your thalamus we might trigger the experience of brocoli in your brain, we might have a next generation brain computer interface for bringing Augmented Reality & Virtual Reality experiences not through glasses but through direct brain stimulation, to recreate known patterns.
L1
result: PARTIAL-FUNCTIONAL-BRIDGE
reason: Both connect a specified internal representation with experiential content: SAN through a measured pattern and its reinstatement, the target through a conditional belief function.
L2
result: PARTIAL-FUNCTIONAL-BRIDGE
source chain: Learned multisensory patterns circulate through a neural feedback loop -> measured phase-pattern configuration -> proposed pattern reinstatement to recreate an experience.
target chain: Internal state encodes approximate posterior -> belief-to-phenomenology function -> experiential content.
reason: Both connect a specified internal representation with experiential content: SAN through a measured pattern and its reinstatement, the target through a conditional belief function.
residual: A neural phase configuration is not demonstrated to be an approximate posterior distribution. The common state-to-content function does not establish identity of the mathematical domains.
L3
result: BOUNDED-BACKWARD-TEST
constrained output: Phenomenological content is represented as a function of beliefs.
required operation: The selected hypothesis links a specified internal belief state to phenomenological content; the paper does not claim every neural representation is a posterior belief.
source test: If experience is a function of an internal state, changing or reinstating that relevant state should constrain the content. SAN proposes exactly such a content-specific reconstruction experiment for broccoli.
alternative routes: Identity map; Marginalization; A coarser non-injective mapping
remaining uncertainty: A neural phase configuration is not demonstrated to be an approximate posterior distribution. The common state-to-content function does not establish identity of the mathematical domains.
Limits and next evidence
A neural phase configuration is not demonstrated to be an approximate posterior distribution. The common state-to-content function does not establish identity of the mathematical domains.
At most a qualitative functional, conceptual or causal-prefix relation; no proof of functoriality, isomorphism, bisimulation, exact equation identity or calibrated statistical dependence.
Existing historical IDs for this exact source-atom/target-atom pair: SETH-SANFULL-057.
SETH-SELECTED-40-15 · Additional partial-match candidate; independent review pending
SAN-AC013-QUALIA-RECONSTRUCTION → SETH-N24-05
SAN proposes a person-specific relation between neural configuration and experienced properties; the target's geometry also depends on relating representations to experience.
SAN: 2022-06-12 · Git repository commit record. Target: 2026-05-04 · arxiv:2409.20318v3.
Current anonymous public access to the exact source body was verified on 2026-10-11. The displayed 2022 date is a Git record date, not independently verified first-publication timing.
Target authors: Lancelot Da Costa · Anil K. Seth · Karl Friston · Maxwell J. D. Ramstead · Lars Sandved-Smith. Verify the target passage · See its short quotation in the paired cards.
Exact SAN source windows
SAN lines 2–4
Brocoli Properties example: All the properties of Brocoli are features learned by the neural network of the human brain. So the color, texture, feel, crunch, taste, smell these are learned patterns, information patterns are detected as coincidence evens by receptors on dendrites, then by neurons, then by arrays of neurons, then by nuclei in the thalamus, then by cortical columns & the hippocampal entorhinal circuit, then by the thalamus again, essentially circuiting throughout your brain in a spatial temporal oscillating feedback loop.
The point of the Brocoli Properties example was to show that if we could intercept at the Thalamus how the properties of Brocoli are signals being passed to the neo cortex, then we could recreate the experience of Brocoli in the Thalamus by creating that set of temporal & spatially distributed phase changes. This could be done at any point in the brain, or on any part of the body, as illustrated in the example of the tongue strip with a grid of electrical signals that correspond to the pixel changes of a camera's viewfinder, a device that allows people to see through their tongue or via any part of their body.
SAN lines 16–16
So if we knew how the properties of brocoli were represented by scanning them at the Thalamus at the moment you were thinking about them, then we might identify which cortical column is managing those properties when you are not thinking about them, and even if you have a completely different personality that manages brocoli in a different region of the cortex, by stimulating the thalamus with the properties of brocoli that we learned from watching your thalamus we might trigger the experience of brocoli in your brain, we might have a next generation brain computer interface for bringing Augmented Reality & Virtual Reality experiences not through glasses but through direct brain stimulation, to recreate known patterns.
L1
result: PARTIAL-CONCEPTUAL-BRIDGE
reason: SAN proposes a person-specific relation between neural configuration and experienced properties; the target's geometry also depends on relating representations to experience.
L2
result: PARTIAL-CONCEPTUAL-BRIDGE
source chain: Learned multisensory patterns circulate through a neural feedback loop -> measured phase-pattern configuration -> proposed pattern reinstatement to recreate an experience.
target chain: Belief/phenomenology correspondence -> structured experiential domain -> formal comparisons and predictions.
reason: SAN proposes a person-specific relation between neural configuration and experienced properties; the target's geometry also depends on relating representations to experience.
residual: Neither a geometric distance nor a formal between-subject correspondence is supplied by the SAN example. The supported subset stops before those target operations.
L3
result: BOUNDED-BACKWARD-TEST
constrained output: A geometry enables characterization of phenomenological differences.
required operation: A representational domain must be linked to experiential content before a metric or other geometric structure can support comparisons.
source test: Characterizing experiential differences requires the representation/content bridge. SAN preserves that pre-geometric bridge and explicitly allows different cortical locations across people.
alternative routes: Information geometry; Relational quality spaces; Empirical neural/perceptual similarity models
remaining uncertainty: Neither a geometric distance nor a formal between-subject correspondence is supplied by the SAN example. The supported subset stops before those target operations.
Limits and next evidence
Neither a geometric distance nor a formal between-subject correspondence is supplied by the SAN example. The supported subset stops before those target operations.
At most a qualitative functional, conceptual or causal-prefix relation; no proof of functoriality, isomorphism, bisimulation, exact equation identity or calibrated statistical dependence.
Existing historical IDs for this exact source-atom/target-atom pair: SETH-SANFULL-059.
SETH-SELECTED-40-16 · Additional partial-match candidate; independent review pending
SAN-AC013-QUALIA-RECONSTRUCTION → SETH-T26-05
SAN routes temporally patterned input through recurrent neural loops that learn correlations; the reservoir uses recurrent dynamics to transform temporal input for a readout.
SAN: 2022-06-12 · Git repository commit record. Target: 2024-06-27 · arxiv:2406.19201v1.
Current anonymous public access to the exact source body was verified on 2026-10-11. The displayed 2022 date is a Git record date, not independently verified first-publication timing.
Target authors: Hanna M. Tolle · Andrea I. Luppi · Anil K. Seth · Pedro A. M. Mediano. Verify the target passage · See its short quotation in the paired cards.
Exact SAN source windows
SAN lines 2–2
Brocoli Properties example: All the properties of Brocoli are features learned by the neural network of the human brain. So the color, texture, feel, crunch, taste, smell these are learned patterns, information patterns are detected as coincidence evens by receptors on dendrites, then by neurons, then by arrays of neurons, then by nuclei in the thalamus, then by cortical columns & the hippocampal entorhinal circuit, then by the thalamus again, essentially circuiting throughout your brain in a spatial temporal oscillating feedback loop.
SAN lines 14–14
In the same way if one part of the brain wants to understand unstructured data from another part of the brain, I suspect synchrony (achieved by listening to one's self) is creating a temporal alignment between two oscillators so that correlations between the two sets of unstructured data can be found by the neural network of your mind, and you make sense out of an association between lightning & thunder for example.
L1
result: PARTIAL-CAUSAL-PREFIX
reason: SAN routes temporally patterned input through recurrent neural loops that learn correlations; the reservoir uses recurrent dynamics to transform temporal input for a readout.
L2
result: PARTIAL-CAUSAL-PREFIX
source chain: Learned multisensory patterns circulate through a neural feedback loop -> measured phase-pattern configuration -> proposed pattern reinstatement to recreate an experience.
target chain: Input -> fixed nonlinear recurrent reservoir state -> trained linear readout.
reason: SAN routes temporally patterned input through recurrent neural loops that learn correlations; the reservoir uses recurrent dynamics to transform temporal input for a readout.
residual: Experience reconstruction is not a linear readout algorithm, and the source does not freeze the recurrent weights. No originality claim over established reservoir computing is made.
L3
result: BOUNDED-BACKWARD-TEST
constrained output: A recurrent computer transforms temporal input into a useful output.
required operation: Input must change recurrent state so a readout can distinguish task-relevant patterns; the selected architecture further fixes reservoir weights and trains a linear readout.
source test: The target needs temporal input to affect an internal recurrent state in a task-sensitive way. SAN's circulation and correlation-learning account supplies that broad dynamic-processing prefix.
alternative routes: Echo-state reservoirs; Liquid-state machines; Fully trained recurrent networks
remaining uncertainty: Experience reconstruction is not a linear readout algorithm, and the source does not freeze the recurrent weights. No originality claim over established reservoir computing is made.
Limits and next evidence
Experience reconstruction is not a linear readout algorithm, and the source does not freeze the recurrent weights. No originality claim over established reservoir computing is made.
At most a qualitative functional, conceptual or causal-prefix relation; no proof of functoriality, isomorphism, bisimulation, exact equation identity or calibrated statistical dependence.
Existing historical IDs for this exact source-atom/target-atom pair: none joined; this is not an automatic addition to governed counts.
SETH-SELECTED-40-17 · Additional partial-match candidate; independent review pending
SAN-AC009-PHASE-TOPOLOGY-TAYLOR → SETH-CS25-22
SAN describes downstream receptive fields detecting learned spatiotemporal phase-pattern renderings.
SAN: 2022-07-09 · Git repository commit record. Target: 2025-10-30 · doi:10.3389/fsci.2025.1546279.
Current anonymous public access to the exact source body was verified on 2026-10-11. The displayed 2022 date is a Git record date, not independently verified first-publication timing.
Target authors: Axel Cleeremans · Liad Mudrik · Anil K. Seth. Verify the target passage · See its short quotation in the paired cards.
Exact SAN source windows
SAN lines 10–18
So that a sequence of action potentials is effectively recognized by the brain as series of coincidence patterns in time & space.
Even in the inhibitory interneuron areas of the brain the highly branching non-linear neuropaths can be compared to 2 dimensional Taylor Series Polynomials being tracked by the brain.
Perhaps the frequency / duration patterns at the neuron level are being converted to frequency / amplitude patterns at the dipole or EEG scale of brainwaves.
So the question is, if each action potential phase in a phase sequence can be considered like a vector changing it's value in a Taylor Series Polynomial, that provides one way to think of how the brain might turn bits of detected data in a neural array, like the eye, into a recognizable shapes in the visual cortex, because as the sequence proceeds a curve, if there is a curve, emerges, sort of like a continous function being learned by a Universal Approximation Theorum, except that in this case the neural firing sequence is able to physically map to contours of learned 3D visual representations, but the same geometric learning process could be at work for all the other senses, and that might explain why we can think of our other sensory representations as having spatial & time characteristics or characteristics that change over time and area. Such as feelings that have a start, a peak, a low, an end, feelings that start in our feet, chest, neck, head, or some other space that we can describe to others.
This means that neural oscillating array-projection tomography can physically represent patterns as (non-linear) phase variation sequences. Then oscillating cell assembles will learn the expressed spatial & temporal phase patterns happening within their cell assembly & derive from those learned patterns the shapes of concepts, the shapes & parameters & properties feelings, objects, sensations, and anything else you can become conscious of. These are learned patterns as that percolate through the oscillation cell assemblies with something like infinite summation but basically infinite summation as a neural network would accomplish it through coincident detection, and this tempo-spatial rendering of action potential sequences in phase patterns would be "seen" or perceived by the receptive fields of downstream (receiving) dendrites in the next neural array, cortical column, or oscillating cell assembly.
L1
result: PARTIAL-PROBLEM-FIT
reason: SAN describes downstream receptive fields detecting learned spatiotemporal phase-pattern renderings.
L2
result: PARTIAL-PROBLEM-FIT
source chain: Spatiotemporal firing and inhibition sequences -> learned phase-pattern geometry -> differentiated properties of experienced content, read by downstream arrays.
target chain: Broadcast message -> unspecified receiver interpretation -> useful downstream processing.
reason: SAN describes downstream receptive fields detecting learned spatiotemporal phase-pattern renderings.
residual: This is an earlier proposed answer to a later explanatory demand, not evidence that the target review asserts the same solution.
L3
result: BOUNDED-BACKWARD-TEST
constrained output: A receiving neuron makes useful use of a broadcast message.
required operation: Some implemented receiver dependency must discriminate consequential input differences; arrival alone does not explain use.
source test: The target asks how a receiving neuron uses a message. SAN gives a candidate answer: its receptive structure responds to a learned pattern in incoming phase-sequenced activity.
alternative routes: Learned receptive selectivity; Fixed feature-sensitive circuitry; Population-level decoding
remaining uncertainty: This is an earlier proposed answer to a later explanatory demand, not evidence that the target review asserts the same solution.
Limits and next evidence
This is an earlier proposed answer to a later explanatory demand, not evidence that the target review asserts the same solution.
At most a qualitative functional, conceptual or causal-prefix relation; no proof of functoriality, isomorphism, bisimulation, exact equation identity or calibrated statistical dependence.
Existing historical IDs for this exact source-atom/target-atom pair: SETH-CS25-DOCWIN-SETH-CS25-22-SAN-AC009-PHASE-TOPOLOGY-TAYLOR-479F1B0B1116.
SETH-SELECTED-40-18 · Source reviewed; relationship unresolved
SAN-AC009-PHASE-TOPOLOGY-TAYLOR → SETH-CS25-38
A map from neural phase patterns to experienced properties could inform a BCI, but the frozen phase-geometry passage does not itself describe an interface intervention.
SAN: 2022-07-09 · Git repository commit record. Target: 2025-10-30 · doi:10.3389/fsci.2025.1546279.
Current anonymous public access to the exact source body was verified on 2026-10-11. The displayed 2022 date is a Git record date, not independently verified first-publication timing.
Target authors: Axel Cleeremans · Liad Mudrik · Anil K. Seth. Verify the target passage · See its short quotation in the paired cards.
Exact SAN source windows
SAN lines 16–18
So the question is, if each action potential phase in a phase sequence can be considered like a vector changing it's value in a Taylor Series Polynomial, that provides one way to think of how the brain might turn bits of detected data in a neural array, like the eye, into a recognizable shapes in the visual cortex, because as the sequence proceeds a curve, if there is a curve, emerges, sort of like a continous function being learned by a Universal Approximation Theorum, except that in this case the neural firing sequence is able to physically map to contours of learned 3D visual representations, but the same geometric learning process could be at work for all the other senses, and that might explain why we can think of our other sensory representations as having spatial & time characteristics or characteristics that change over time and area. Such as feelings that have a start, a peak, a low, an end, feelings that start in our feet, chest, neck, head, or some other space that we can describe to others.
This means that neural oscillating array-projection tomography can physically represent patterns as (non-linear) phase variation sequences. Then oscillating cell assembles will learn the expressed spatial & temporal phase patterns happening within their cell assembly & derive from those learned patterns the shapes of concepts, the shapes & parameters & properties feelings, objects, sensations, and anything else you can become conscious of. These are learned patterns as that percolate through the oscillation cell assemblies with something like infinite summation but basically infinite summation as a neural network would accomplish it through coincident detection, and this tempo-spatial rendering of action potential sequences in phase patterns would be "seen" or perceived by the receptive fields of downstream (receiving) dendrites in the next neural array, cortical column, or oscillating cell assembly.
SAN lines 39–39
In my thought experiment I am taking the Taylor Series from Calculus and applying it to a 3D grid of phase changes. 3D Calculus. With each phase change that is tracked by the brain representing a different step in a Taylor series representing the information content of the rendered mind *at least the phase topology as a brain state is an attractor point, for the mind's oscillations.
L1
result: OPEN-NO-CELL-CLOSURE
reason: A map from neural phase patterns to experienced properties could inform a BCI, but the frozen phase-geometry passage does not itself describe an interface intervention.
L2
result: OPEN-CAUSAL-BRIDGE
source chain: Spatiotemporal firing and inhibition sequences -> learned phase-pattern geometry -> differentiated properties of experienced content, read by downstream arrays.
target chain: Understanding biological basis -> circuit-level intervention -> changed conscious experience.
reason: A map from neural phase patterns to experienced properties could inform a BCI, but the frozen phase-geometry passage does not itself describe an interface intervention.
residual: Recover a dated intervention bridge in this source family. Keep the map as relevant evidence without fabricating a BCI operation.
L3
result: BACKWARD-TEST-COMPLETED-BRIDGE-OPEN
constrained output: A brain-computer intervention targets, remediates or enhances an aspect of conscious experience.
required operation: An intervention must alter a biological state in a way linked to an experiential consequence.
source test: The target requires an accessible intervention linked to a change in experience. The source supplies representation and downstream detection, leaving the intervention step unbound for this operand.
alternative routes: Electrical stimulation; Sensory substitution or feedback interfaces; Pharmacological or optogenetic circuit modulation
remaining uncertainty: Recover a dated intervention bridge in this source family. Keep the map as relevant evidence without fabricating a BCI operation.
Limits and next evidence
Recover a dated intervention bridge in this source family. Keep the map as relevant evidence without fabricating a BCI operation.
At most a qualitative functional, conceptual or causal-prefix relation; no proof of functoriality, isomorphism, bisimulation, exact equation identity or calibrated statistical dependence.
Existing historical IDs for this exact source-atom/target-atom pair: none joined; this is not an automatic addition to governed counts.
SETH-SELECTED-40-19 · Source reviewed; relationship unresolved
SAN-AC009-PHASE-TOPOLOGY-TAYLOR → SETH-CS25-30
The SAN phase-geometry proposal is a candidate account of content that could support control; a specified control operation is not established in these windows.
SAN: 2022-07-09 · Git repository commit record. Target: 2025-10-30 · doi:10.3389/fsci.2025.1546279.
Current anonymous public access to the exact source body was verified on 2026-10-11. The displayed 2022 date is a Git record date, not independently verified first-publication timing.
Target authors: Axel Cleeremans · Liad Mudrik · Anil K. Seth. Verify the target passage · See its short quotation in the paired cards.
Exact SAN source windows
SAN lines 16–18
So the question is, if each action potential phase in a phase sequence can be considered like a vector changing it's value in a Taylor Series Polynomial, that provides one way to think of how the brain might turn bits of detected data in a neural array, like the eye, into a recognizable shapes in the visual cortex, because as the sequence proceeds a curve, if there is a curve, emerges, sort of like a continous function being learned by a Universal Approximation Theorum, except that in this case the neural firing sequence is able to physically map to contours of learned 3D visual representations, but the same geometric learning process could be at work for all the other senses, and that might explain why we can think of our other sensory representations as having spatial & time characteristics or characteristics that change over time and area. Such as feelings that have a start, a peak, a low, an end, feelings that start in our feet, chest, neck, head, or some other space that we can describe to others.
This means that neural oscillating array-projection tomography can physically represent patterns as (non-linear) phase variation sequences. Then oscillating cell assembles will learn the expressed spatial & temporal phase patterns happening within their cell assembly & derive from those learned patterns the shapes of concepts, the shapes & parameters & properties feelings, objects, sensations, and anything else you can become conscious of. These are learned patterns as that percolate through the oscillation cell assemblies with something like infinite summation but basically infinite summation as a neural network would accomplish it through coincident detection, and this tempo-spatial rendering of action potential sequences in phase patterns would be "seen" or perceived by the receptive fields of downstream (receiving) dendrites in the next neural array, cortical column, or oscillating cell assembly.
SAN lines 39–39
In my thought experiment I am taking the Taylor Series from Calculus and applying it to a 3D grid of phase changes. 3D Calculus. With each phase change that is tracked by the brain representing a different step in a Taylor series representing the information content of the rendered mind *at least the phase topology as a brain state is an attractor point, for the mind's oscillations.
L1
result: OPEN-NO-CELL-CLOSURE
reason: The SAN phase-geometry proposal is a candidate account of content that could support control; a specified control operation is not established in these windows.
L2
result: OPEN-CAUSAL-BRIDGE
source chain: Spatiotemporal firing and inhibition sequences -> learned phase-pattern geometry -> differentiated properties of experienced content, read by downstream arrays.
target chain: Account of biological mechanisms and experience -> precise intervention -> controlled experiential outcome.
reason: The SAN phase-geometry proposal is a candidate account of content that could support control; a specified control operation is not established in these windows.
residual: The broader genealogy may supply the missing step. This reviewed cell remains open, with no lost inherited finding or no-priority verdict.
L3
result: BACKWARD-TEST-COMPLETED-BRIDGE-OPEN
constrained output: Neuroscience can deliberately control conscious experience with precision.
required operation: A causal relation between an accessible neural manipulation and an experiential outcome must be characterized.
source test: Precise control of experience needs a causal manipulation-to-content relation. Representation geometry and pattern detection supply possible prerequisites, not the complete controlled change.
alternative routes: Stimulation of an ensemble; Sensory-feedback intervention; State-dependent neuromodulation
remaining uncertainty: The broader genealogy may supply the missing step. This reviewed cell remains open, with no lost inherited finding or no-priority verdict.
Target claim type: A forward-looking consequence of success in consciousness science, with phenotechnology explicitly attributed to Metzinger.
Limits and next evidence
The broader genealogy may supply the missing step. This reviewed cell remains open, with no lost inherited finding or no-priority verdict.
At most a qualitative functional, conceptual or causal-prefix relation; no proof of functoriality, isomorphism, bisimulation, exact equation identity or calibrated statistical dependence.
Existing historical IDs for this exact source-atom/target-atom pair: none joined; this is not an automatic addition to governed counts.
SETH-SELECTED-40-20 · Previously independently source-reviewed subset
SAN-AC009-PHASE-TOPOLOGY-TAYLOR → SETH-CS25-28
Both connect organized neural activity to what an experience is like. SAN describes learned shapes and properties; the review also discusses measured similarities among experiences.
SAN: 2022-07-09 · Git repository commit record. Target: 2025-10-30 · doi:10.3389/fsci.2025.1546279.
Current anonymous public access to the exact source body was verified on 2026-10-11. The displayed 2022 date is a Git record date, not independently verified first-publication timing.
Target authors: Axel Cleeremans · Liad Mudrik · Anil K. Seth. Verify the target passage · See its short quotation in the paired cards.
Exact SAN source windows
SAN lines 18–18
This means that neural oscillating array-projection tomography can physically represent patterns as (non-linear) phase variation sequences. Then oscillating cell assembles will learn the expressed spatial & temporal phase patterns happening within their cell assembly & derive from those learned patterns the shapes of concepts, the shapes & parameters & properties feelings, objects, sensations, and anything else you can become conscious of. These are learned patterns as that percolate through the oscillation cell assemblies with something like infinite summation but basically infinite summation as a neural network would accomplish it through coincident detection, and this tempo-spatial rendering of action potential sequences in phase patterns would be "seen" or perceived by the receptive fields of downstream (receiving) dendrites in the next neural array, cortical column, or oscillating cell assembly.
SAN lines 39–39
In my thought experiment I am taking the Taylor Series from Calculus and applying it to a 3D grid of phase changes. 3D Calculus. With each phase change that is tracked by the brain representing a different step in a Taylor series representing the information content of the rendered mind *at least the phase topology as a brain state is an attractor point, for the mind's oscillations.
L1
result: PARTIAL_REPRESENTATIONAL_FUNCTION
reason: SAN ties learned spatial-temporal phase patterns to shapes and properties of experience; the target ties neural state-space organization to visual experience. This is more specific than the word 'geometry', but not the same metric or data analysis.
L2
result: PARTIAL_CAUSAL_BRIDGE_WITH_RELATIONAL_RESIDUAL
reason: Neural organization is used to explain experiential content in both. SAN's constructive learning account does not itself derive relational similarity across experiences or subjects, the target's distinctive explanatory step.
L3
result: SUPPORTED_CONSTRAINED_OUTPUT_RECONSTRUCTION_WITH_NONUNIQUE_SOLUTION
reason: Constrained result: Neural similarities systematically correspond to similarities among experiences. Required operation: A representation and comparison rule must retain the distinctions and relations used in perceptual judgments. Source test: Lines 16–18 provide structured learned patterns and experiential properties; line 39 additionally calls phase topology a brain-state attractor. No pairwise similarity rule is supplied. Alternatives: Population-rate geometry; Learned temporal-pattern geometry; Other representational spaces with a separately tested neural-to-perceptual map. Remaining uncertainty: The attractor remark strengthens a state/experience bridge, not evidence for a metric, similarity judgment model or empirical validation.
Limits and next evidence
A within-experience contour or shape is not automatically a relation among different experiences. Target refs 45–48 and 200–206 include pre-2022 work; their cited presence blocks an unqualified 2022-versus-2025 origin story but is not a completed predecessor adjudication.
At most a qualitative functional, conceptual or causal-prefix relation; no proof of functoriality, isomorphism, bisimulation, exact equation identity or calibrated statistical dependence.
Existing historical IDs for this exact source-atom/target-atom pair: SETH-CS25-DOCWIN-SETH-CS25-28-SAN-AC009-PHASE-TOPOLOGY-TAYLOR-6DCF83E437C1.
SETH-SELECTED-40-21 · Additional partial-match candidate; independent review pending
SAN-AC009-PHASE-TOPOLOGY-TAYLOR → SETH-CS25-18
SAN makes experienced objects and properties products of learned internal phase-pattern representations; the target likewise treats perceptual content as internally constructed.
SAN: 2022-07-09 · Git repository commit record. Target: 2025-10-30 · doi:10.3389/fsci.2025.1546279.
Current anonymous public access to the exact source body was verified on 2026-10-11. The displayed 2022 date is a Git record date, not independently verified first-publication timing.
Target authors: Axel Cleeremans · Liad Mudrik · Anil K. Seth. Verify the target passage · See its short quotation in the paired cards.
Exact SAN source windows
SAN lines 16–18
So the question is, if each action potential phase in a phase sequence can be considered like a vector changing it's value in a Taylor Series Polynomial, that provides one way to think of how the brain might turn bits of detected data in a neural array, like the eye, into a recognizable shapes in the visual cortex, because as the sequence proceeds a curve, if there is a curve, emerges, sort of like a continous function being learned by a Universal Approximation Theorum, except that in this case the neural firing sequence is able to physically map to contours of learned 3D visual representations, but the same geometric learning process could be at work for all the other senses, and that might explain why we can think of our other sensory representations as having spatial & time characteristics or characteristics that change over time and area. Such as feelings that have a start, a peak, a low, an end, feelings that start in our feet, chest, neck, head, or some other space that we can describe to others.
This means that neural oscillating array-projection tomography can physically represent patterns as (non-linear) phase variation sequences. Then oscillating cell assembles will learn the expressed spatial & temporal phase patterns happening within their cell assembly & derive from those learned patterns the shapes of concepts, the shapes & parameters & properties feelings, objects, sensations, and anything else you can become conscious of. These are learned patterns as that percolate through the oscillation cell assemblies with something like infinite summation but basically infinite summation as a neural network would accomplish it through coincident detection, and this tempo-spatial rendering of action potential sequences in phase patterns would be "seen" or perceived by the receptive fields of downstream (receiving) dendrites in the next neural array, cortical column, or oscillating cell assembly.
SAN lines 39–39
In my thought experiment I am taking the Taylor Series from Calculus and applying it to a 3D grid of phase changes. 3D Calculus. With each phase change that is tracked by the brain representing a different step in a Taylor series representing the information content of the rendered mind *at least the phase topology as a brain state is an attractor point, for the mind's oscillations.
L1
result: PARTIAL-CONCEPTUAL-BRIDGE
reason: SAN makes experienced objects and properties products of learned internal phase-pattern representations; the target likewise treats perceptual content as internally constructed.
L2
result: PARTIAL-CONCEPTUAL-BRIDGE
source chain: Spatiotemporal firing and inhibition sequences -> learned phase-pattern geometry -> differentiated properties of experienced content, read by downstream arrays.
target chain: Internal predictions -> perceptual content -> sensory constraint from world and body.
reason: SAN makes experienced objects and properties products of learned internal phase-pattern representations; the target likewise treats perceptual content as internally constructed.
residual: The selected passage does not specify top-down posterior prediction or error minimization. Internal construction alone is broad overlap, not the complete controlled-hallucination claim.
L3
result: BOUNDED-BACKWARD-TEST
constrained output: Perceptual content reflects a prediction constrained by sensory signals.
required operation: An internal model must generate content and be sensitive to incoming evidence; a generic neural code does not establish the complete predictive mechanism.
source test: An internally generated percept requires a content-generating representation. SAN's learned phase-pattern shapes and downstream reception provide that bounded component.
alternative routes: Bayesian predictive coding; Recurrent generative models; Non-Bayesian learned predictive dynamics
remaining uncertainty: The selected passage does not specify top-down posterior prediction or error minimization. Internal construction alone is broad overlap, not the complete controlled-hallucination claim.
Limits and next evidence
The selected passage does not specify top-down posterior prediction or error minimization. Internal construction alone is broad overlap, not the complete controlled-hallucination claim.
At most a qualitative functional, conceptual or causal-prefix relation; no proof of functoriality, isomorphism, bisimulation, exact equation identity or calibrated statistical dependence.
Existing historical IDs for this exact source-atom/target-atom pair: none joined; this is not an automatic addition to governed counts.
SETH-SELECTED-40-22 · Additional partial-match candidate; independent review pending
SAN-AC009-PHASE-TOPOLOGY-TAYLOR → SETH-N24-01
SAN relates structured internal neural patterns to shapes and properties of conscious content; the target relates an internal belief state to phenomenology.
SAN: 2022-07-09 · Git repository commit record. Target: 2026-05-04 · arxiv:2409.20318v3.
Current anonymous public access to the exact source body was verified on 2026-10-11. The displayed 2022 date is a Git record date, not independently verified first-publication timing.
Target authors: Lancelot Da Costa · Anil K. Seth · Karl Friston · Maxwell J. D. Ramstead · Lars Sandved-Smith. Verify the target passage · See its short quotation in the paired cards.
Exact SAN source windows
SAN lines 16–18
So the question is, if each action potential phase in a phase sequence can be considered like a vector changing it's value in a Taylor Series Polynomial, that provides one way to think of how the brain might turn bits of detected data in a neural array, like the eye, into a recognizable shapes in the visual cortex, because as the sequence proceeds a curve, if there is a curve, emerges, sort of like a continous function being learned by a Universal Approximation Theorum, except that in this case the neural firing sequence is able to physically map to contours of learned 3D visual representations, but the same geometric learning process could be at work for all the other senses, and that might explain why we can think of our other sensory representations as having spatial & time characteristics or characteristics that change over time and area. Such as feelings that have a start, a peak, a low, an end, feelings that start in our feet, chest, neck, head, or some other space that we can describe to others.
This means that neural oscillating array-projection tomography can physically represent patterns as (non-linear) phase variation sequences. Then oscillating cell assembles will learn the expressed spatial & temporal phase patterns happening within their cell assembly & derive from those learned patterns the shapes of concepts, the shapes & parameters & properties feelings, objects, sensations, and anything else you can become conscious of. These are learned patterns as that percolate through the oscillation cell assemblies with something like infinite summation but basically infinite summation as a neural network would accomplish it through coincident detection, and this tempo-spatial rendering of action potential sequences in phase patterns would be "seen" or perceived by the receptive fields of downstream (receiving) dendrites in the next neural array, cortical column, or oscillating cell assembly.
SAN lines 39–39
In my thought experiment I am taking the Taylor Series from Calculus and applying it to a 3D grid of phase changes. 3D Calculus. With each phase change that is tracked by the brain representing a different step in a Taylor series representing the information content of the rendered mind *at least the phase topology as a brain state is an attractor point, for the mind's oscillations.
L1
result: PARTIAL-CONCEPTUAL-BRIDGE
reason: SAN relates structured internal neural patterns to shapes and properties of conscious content; the target relates an internal belief state to phenomenology.
L2
result: PARTIAL-CONCEPTUAL-BRIDGE
source chain: Spatiotemporal firing and inhibition sequences -> learned phase-pattern geometry -> differentiated properties of experienced content, read by downstream arrays.
target chain: Internal state encodes approximate posterior -> belief-to-phenomenology function -> experiential content.
reason: SAN relates structured internal neural patterns to shapes and properties of conscious content; the target relates an internal belief state to phenomenology.
residual: No probability distribution or proof that phase topology is the target's belief manifold is given. Preserve the common relation without equating those domains.
L3
result: BOUNDED-BACKWARD-TEST
constrained output: Phenomenological content is represented as a function of beliefs.
required operation: The selected hypothesis links a specified internal belief state to phenomenological content; the paper does not claim every neural representation is a posterior belief.
source test: The target function requires some internal representation to determine experiential content. SAN makes that representation/content relation explicit at a neural-pattern level.
alternative routes: Identity map; Marginalization; A coarser non-injective mapping
remaining uncertainty: No probability distribution or proof that phase topology is the target's belief manifold is given. Preserve the common relation without equating those domains.
Limits and next evidence
No probability distribution or proof that phase topology is the target's belief manifold is given. Preserve the common relation without equating those domains.
At most a qualitative functional, conceptual or causal-prefix relation; no proof of functoriality, isomorphism, bisimulation, exact equation identity or calibrated statistical dependence.
Existing historical IDs for this exact source-atom/target-atom pair: none joined; this is not an automatic addition to governed counts.
SETH-SELECTED-40-23 · Previously independently source-reviewed subset
SAN-AC009-PHASE-TOPOLOGY-TAYLOR → SETH-N24-05
Both use structured internal representations to describe properties of experience. This supports a partial connection; a shared distance rule between different experiences has not been demonstrated.
SAN: 2022-07-09 · Git repository commit record. Target: 2026-05-04 · arxiv:2409.20318v3.
Current anonymous public access to the exact source body was verified on 2026-10-11. The displayed 2022 date is a Git record date, not independently verified first-publication timing.
Target authors: Lancelot Da Costa · Anil K. Seth · Karl Friston · Maxwell J. D. Ramstead · Lars Sandved-Smith. Verify the target passage · See its short quotation in the paired cards.
Exact SAN source windows
SAN lines 18–18
This means that neural oscillating array-projection tomography can physically represent patterns as (non-linear) phase variation sequences. Then oscillating cell assembles will learn the expressed spatial & temporal phase patterns happening within their cell assembly & derive from those learned patterns the shapes of concepts, the shapes & parameters & properties feelings, objects, sensations, and anything else you can become conscious of. These are learned patterns as that percolate through the oscillation cell assemblies with something like infinite summation but basically infinite summation as a neural network would accomplish it through coincident detection, and this tempo-spatial rendering of action potential sequences in phase patterns would be "seen" or perceived by the receptive fields of downstream (receiving) dendrites in the next neural array, cortical column, or oscillating cell assembly.
SAN lines 39–39
In my thought experiment I am taking the Taylor Series from Calculus and applying it to a 3D grid of phase changes. 3D Calculus. With each phase change that is tracked by the brain representing a different step in a Taylor series representing the information content of the rendered mind *at least the phase topology as a brain state is an attractor point, for the mind's oscillations.
L1
result: PARTIAL_REPRESENTATIONAL_ROLE_ONLY
reason: SAN gives a physically learned pattern-to-experiential-shape proposal. The target characterizes differences in probability-distribution content and conditionally induces phenomenological geometry. That role-level subset survives, not the metric.
L2
result: PARTIAL_BRIDGE_NOT_SHARED_RELATIONAL_DERIVATION
reason: The common prefix is representational structure informing experience. Worker L2's 'relations among experiences' goes beyond SAN's shown intra-experience shapes/properties. The target adds an explicit belief-to-phenomenology map and information-geometric distances.
L3
result: SUPPORTED_CONSTRAINED_OUTPUT_RECONSTRUCTION_WITH_NONUNIQUE_SOLUTION
reason: Constrained result: A geometry can characterize experiential differences across subjects or moments. Required operation: Specify a state space, a discrepancy/metric, and a justified relation between that discrepancy and experiential difference. Source test: SAN lines 6, 16–18 and 39 describe changing structured neural patterns and experiential properties, but supply neither a probability-distribution metric nor an inter-subject comparison map. Alternatives: Fisher information distance under the target's conditions; Other divergences or metrics explicitly allowed by the target; A neural-pattern geometry with an independently supplied experiential comparison map. Remaining uncertainty: Different neural patterns can support different experiential properties without fixing their distances. SAN does not derive Fisher geometry, confidence effects or the paper's predictions.
Limits and next evidence
Taylor-series/tensor language is offered as a thought experiment and analogy, not a proved approximation theorem or a demonstrated metric. Keeping this limit does not erase the explicit neural-pattern/experience bridge.
At most a qualitative functional, conceptual or causal-prefix relation; no proof of functoriality, isomorphism, bisimulation, exact equation identity or calibrated statistical dependence.
Existing historical IDs for this exact source-atom/target-atom pair: SETH-SANFULL-033.
SETH-SELECTED-40-24 · Additional partial-match candidate; independent review pending
SAN-AC009-PHASE-TOPOLOGY-TAYLOR → SETH-T26-05
SAN describes spatiotemporal input sequences mapped to learned high-dimensional patterns that downstream arrays detect; the reservoir maps temporal input into a high-dimensional state for readout.
SAN: 2022-07-09 · Git repository commit record. Target: 2024-06-27 · arxiv:2406.19201v1.
Current anonymous public access to the exact source body was verified on 2026-10-11. The displayed 2022 date is a Git record date, not independently verified first-publication timing.
Target authors: Hanna M. Tolle · Andrea I. Luppi · Anil K. Seth · Pedro A. M. Mediano. Verify the target passage · See its short quotation in the paired cards.
Exact SAN source windows
SAN lines 6–18
I want to say that if each action potential event creates a phasic pattern, and that a sequence of phasic spikes, high phasic spikes, tonic spikes, and inhibited spikes will produce something like a non-linear sequence that could alternatively be considered to be a dynamic linear sequence if we are just considering the sum of the brain's state at something like millisecond intervals of time. It's a dynamic linear sequence temporally in terms of the activations in the brains 3D graph of neural arrays. The phase change at each synaptic cleft is imagined to have the significance of a a temporal tensor in 3D space. We could compare the phase states at each interval or frame of time as like components of a Taylor Series, like the phase changes are imagined as Polynomial Tensors in a high dimensional Taylor Series.
Further, these non-linear sequences of action potentials will be tracked, via their inhibited oscillatory waves, by whole oscillating cell assemblies, such as cortical columns, the hippocampus, or any oscillating neural circuit of any scale such as the suprachiasmatic nucleus.
So that a sequence of action potentials is effectively recognized by the brain as series of coincidence patterns in time & space.
Even in the inhibitory interneuron areas of the brain the highly branching non-linear neuropaths can be compared to 2 dimensional Taylor Series Polynomials being tracked by the brain.
Perhaps the frequency / duration patterns at the neuron level are being converted to frequency / amplitude patterns at the dipole or EEG scale of brainwaves.
So the question is, if each action potential phase in a phase sequence can be considered like a vector changing it's value in a Taylor Series Polynomial, that provides one way to think of how the brain might turn bits of detected data in a neural array, like the eye, into a recognizable shapes in the visual cortex, because as the sequence proceeds a curve, if there is a curve, emerges, sort of like a continous function being learned by a Universal Approximation Theorum, except that in this case the neural firing sequence is able to physically map to contours of learned 3D visual representations, but the same geometric learning process could be at work for all the other senses, and that might explain why we can think of our other sensory representations as having spatial & time characteristics or characteristics that change over time and area. Such as feelings that have a start, a peak, a low, an end, feelings that start in our feet, chest, neck, head, or some other space that we can describe to others.
This means that neural oscillating array-projection tomography can physically represent patterns as (non-linear) phase variation sequences. Then oscillating cell assembles will learn the expressed spatial & temporal phase patterns happening within their cell assembly & derive from those learned patterns the shapes of concepts, the shapes & parameters & properties feelings, objects, sensations, and anything else you can become conscious of. These are learned patterns as that percolate through the oscillation cell assemblies with something like infinite summation but basically infinite summation as a neural network would accomplish it through coincident detection, and this tempo-spatial rendering of action potential sequences in phase patterns would be "seen" or perceived by the receptive fields of downstream (receiving) dendrites in the next neural array, cortical column, or oscillating cell assembly.
SAN lines 39–39
In my thought experiment I am taking the Taylor Series from Calculus and applying it to a 3D grid of phase changes. 3D Calculus. With each phase change that is tracked by the brain representing a different step in a Taylor series representing the information content of the rendered mind *at least the phase topology as a brain state is an attractor point, for the mind's oscillations.
L1
result: PARTIAL-FUNCTIONAL-PREFIX
reason: SAN describes spatiotemporal input sequences mapped to learned high-dimensional patterns that downstream arrays detect; the reservoir maps temporal input into a high-dimensional state for readout.
L2
result: PARTIAL-FUNCTIONAL-PREFIX
source chain: Spatiotemporal firing and inhibition sequences -> learned phase-pattern geometry -> differentiated properties of experienced content, read by downstream arrays.
target chain: Input -> fixed nonlinear recurrent reservoir state -> trained linear readout.
reason: SAN describes spatiotemporal input sequences mapped to learned high-dimensional patterns that downstream arrays detect; the reservoir maps temporal input into a high-dimensional state for readout.
residual: The Taylor-series thought experiment is not the target's reservoir equation, and SAN learns within the represented circuit rather than only training a linear output layer.
L3
result: BOUNDED-BACKWARD-TEST
constrained output: A recurrent computer transforms temporal input into a useful output.
required operation: Input must change recurrent state so a readout can distinguish task-relevant patterns; the selected architecture further fixes reservoir weights and trains a linear readout.
source test: A task readout needs an input-dependent internal pattern that preserves useful distinctions. The SAN temporal-pattern and receiving-array proposal provides that input/state/readout prefix.
alternative routes: Echo-state reservoirs; Liquid-state machines; Fully trained recurrent networks
remaining uncertainty: The Taylor-series thought experiment is not the target's reservoir equation, and SAN learns within the represented circuit rather than only training a linear output layer.
Limits and next evidence
The Taylor-series thought experiment is not the target's reservoir equation, and SAN learns within the represented circuit rather than only training a linear output layer.
At most a qualitative functional, conceptual or causal-prefix relation; no proof of functoriality, isomorphism, bisimulation, exact equation identity or calibrated statistical dependence.
Existing historical IDs for this exact source-atom/target-atom pair: none joined; this is not an automatic addition to governed counts.
SETH-SELECTED-40-25 · Additional partial-match candidate; independent review pending
SAN-AC017-DISTRIBUTED-SELF-RENDERING → SETH-CS25-22
SAN explains later neural arrays as receivers tuned by their connections to identify patterns sent by earlier arrays.
SAN: 2022-06-10 · Git repository commit record. Target: 2025-10-30 · doi:10.3389/fsci.2025.1546279.
Current anonymous public access to the exact source body was verified on 2026-10-11. The displayed 2022 date is a Git record date, not independently verified first-publication timing.
Target authors: Axel Cleeremans · Liad Mudrik · Anil K. Seth. Verify the target passage · See its short quotation in the paired cards.
Exact SAN source windows
SAN lines 19–25
It's a self-aware network in 4 dimensions rendering a 4 dimensional pattern to itself that is reactive to the world and able to make prediction-based models of reality with sparse spatially & temporally distributed frequency/phase patterns that represent coicident neural firing or coincident neural inhibition deviating from the oscillatory norm of a given oscillator on some level as observed by the neurons in the oscillator, and by oscillators or circuits or cortical columns linked across the brain thanks to the activity of interneurons and major intersections of traffic such as the thalamic bridges, the corpus callosum, the tempo-parietal junction, Werneck & Broca's areas and other areas of the brain that are highly connected or connecting multiple regions of the brain together.
At the level of cells in the eyes, the ears, the nose, the tongue, and the skin there is a similar thing happening, it's like a sensory/transmitter array of nerve cells, ganglian neurons, cones, special receptors, this paragraph is not a great anatomical description, which I will fix later, but the point is that when you are detecting light, heat, vibration, pain, there is some physical interaction happening, a photon might be colliding with a protein causing it to move, flip, change position, sounds might cause hairs in your ear to vibration, that motion in all these sensor array situations is detected by neural/glial arrays, by their specialized receptors, but when they pass information the next set of neurons, up to the next layer of the neural network, the next layer doesn't know it's the 2nd layer, or the 4th layer or the 5th or 6th array of neurons, what I mean is that the signals sent to each subsequent array may as well be, from that subsequent arrays perspective, signals from the outside world. So imagine that if the first sensor array in your eyes, ears, nose, mouth or skin is capable of the perception of an image or a sound or a pattern, then so is each subsequent array of neursons, each edge community, each neural circuit, each cortical column, each layer in the hippocampus, each layer in each cortical column, and all the interesting connections between the pyramidal cells, the thalamus, the different brain regions, they are each in a sense sensors, like little eyes and ears (the receptors on every cells) that are picking up signals, connecting more synaptic branches listen better, or disconnecting to listen less, considering via thresholds that are growing or shrinking based on the growth or shrinkage of synaptic branches, and the patterns the each cell represents to the community (to the cortical colum or to the oscillating group, or from the oscillating group to other oscillating groups Is memory engram based on the synaptic connections that it has grown at that moment, in part because those synaptic connections, establish what memory the cell is tuned to receive, identify, predict, via its receptors being activated, and in the case of neurons via it’s dendrite being activated, or via it’s soma bursting.
Is memory engram based on the synaptic connections that it has grown at that moment, in part because those synaptic connections, establish what memory the cell is tuned to receive, identify, predict, via its receptors being activated, and in the case of neurons via it’s dendrite being activated, or via it’s soma bursting
That physical state, of the neurons literal connections creates it’s phase signal, which is distinct from all the other neurons in the brain, the memories that neuron contains are distinct, but those memories are like the keys of a piano, that becomes tuned to be receptive to certain frequencies and to transmit certain frequencies, and the oscillating group of neurons plays back many of these neural piano keys, creating an array of tempo-spatial sequences, that get blown up or magnified via the inhibition of nearby neurons in the exit terminal, which amplifies the clarify of the signal being transmitted to the next layer of cells or the next array, or the next edge community, or the next layer in the cortical column or the next cortical column, or the next oscillating group in the default mode network or the next oscillating group in the rich club network and so on.
L1
result: PARTIAL-PROBLEM-FIT
reason: SAN explains later neural arrays as receivers tuned by their connections to identify patterns sent by earlier arrays.
L2
result: PARTIAL-PROBLEM-FIT
source chain: World-responsive recurrent arrays -> learned, phase-relative self/world representations -> internally rendered experience and prediction.
target chain: Broadcast message -> unspecified receiver interpretation -> useful downstream processing.
reason: SAN explains later neural arrays as receivers tuned by their connections to identify patterns sent by earlier arrays.
residual: The target is an open mechanism question. This cannot become a claim that the authors adopted this receiver model.
L3
result: BOUNDED-BACKWARD-TEST
constrained output: A receiving neuron makes useful use of a broadcast message.
required operation: Some implemented receiver dependency must discriminate consequential input differences; arrival alone does not explain use.
source test: Useful receiver response requires sensitivity to the message through a physical selection mechanism. Lines 21–25 give a connection- and threshold-dependent reception proposal.
alternative routes: Learned receptive selectivity; Fixed feature-sensitive circuitry; Population-level decoding
remaining uncertainty: The target is an open mechanism question. This cannot become a claim that the authors adopted this receiver model.
Limits and next evidence
The target is an open mechanism question. This cannot become a claim that the authors adopted this receiver model.
At most a qualitative functional, conceptual or causal-prefix relation; no proof of functoriality, isomorphism, bisimulation, exact equation identity or calibrated statistical dependence.
Existing historical IDs for this exact source-atom/target-atom pair: none joined; this is not an automatic addition to governed counts.
SETH-SELECTED-40-26 · Source reviewed; relationship unresolved
SAN-AC017-DISTRIBUTED-SELF-RENDERING → SETH-CS25-38
A world-responsive neural rendering system provides a possible substrate for intervention, but the frozen operation does not specify an external BCI or its control channel.
SAN: 2022-06-10 · Git repository commit record. Target: 2025-10-30 · doi:10.3389/fsci.2025.1546279.
Current anonymous public access to the exact source body was verified on 2026-10-11. The displayed 2022 date is a Git record date, not independently verified first-publication timing.
Target authors: Axel Cleeremans · Liad Mudrik · Anil K. Seth. Verify the target passage · See its short quotation in the paired cards.
Exact SAN source windows
SAN lines 19–25
It's a self-aware network in 4 dimensions rendering a 4 dimensional pattern to itself that is reactive to the world and able to make prediction-based models of reality with sparse spatially & temporally distributed frequency/phase patterns that represent coicident neural firing or coincident neural inhibition deviating from the oscillatory norm of a given oscillator on some level as observed by the neurons in the oscillator, and by oscillators or circuits or cortical columns linked across the brain thanks to the activity of interneurons and major intersections of traffic such as the thalamic bridges, the corpus callosum, the tempo-parietal junction, Werneck & Broca's areas and other areas of the brain that are highly connected or connecting multiple regions of the brain together.
At the level of cells in the eyes, the ears, the nose, the tongue, and the skin there is a similar thing happening, it's like a sensory/transmitter array of nerve cells, ganglian neurons, cones, special receptors, this paragraph is not a great anatomical description, which I will fix later, but the point is that when you are detecting light, heat, vibration, pain, there is some physical interaction happening, a photon might be colliding with a protein causing it to move, flip, change position, sounds might cause hairs in your ear to vibration, that motion in all these sensor array situations is detected by neural/glial arrays, by their specialized receptors, but when they pass information the next set of neurons, up to the next layer of the neural network, the next layer doesn't know it's the 2nd layer, or the 4th layer or the 5th or 6th array of neurons, what I mean is that the signals sent to each subsequent array may as well be, from that subsequent arrays perspective, signals from the outside world. So imagine that if the first sensor array in your eyes, ears, nose, mouth or skin is capable of the perception of an image or a sound or a pattern, then so is each subsequent array of neursons, each edge community, each neural circuit, each cortical column, each layer in the hippocampus, each layer in each cortical column, and all the interesting connections between the pyramidal cells, the thalamus, the different brain regions, they are each in a sense sensors, like little eyes and ears (the receptors on every cells) that are picking up signals, connecting more synaptic branches listen better, or disconnecting to listen less, considering via thresholds that are growing or shrinking based on the growth or shrinkage of synaptic branches, and the patterns the each cell represents to the community (to the cortical colum or to the oscillating group, or from the oscillating group to other oscillating groups Is memory engram based on the synaptic connections that it has grown at that moment, in part because those synaptic connections, establish what memory the cell is tuned to receive, identify, predict, via its receptors being activated, and in the case of neurons via it’s dendrite being activated, or via it’s soma bursting.
Is memory engram based on the synaptic connections that it has grown at that moment, in part because those synaptic connections, establish what memory the cell is tuned to receive, identify, predict, via its receptors being activated, and in the case of neurons via it’s dendrite being activated, or via it’s soma bursting
That physical state, of the neurons literal connections creates it’s phase signal, which is distinct from all the other neurons in the brain, the memories that neuron contains are distinct, but those memories are like the keys of a piano, that becomes tuned to be receptive to certain frequencies and to transmit certain frequencies, and the oscillating group of neurons plays back many of these neural piano keys, creating an array of tempo-spatial sequences, that get blown up or magnified via the inhibition of nearby neurons in the exit terminal, which amplifies the clarify of the signal being transmitted to the next layer of cells or the next array, or the next edge community, or the next layer in the cortical column or the next cortical column, or the next oscillating group in the default mode network or the next oscillating group in the rich club network and so on.
L1
result: OPEN-NO-CELL-CLOSURE
reason: A world-responsive neural rendering system provides a possible substrate for intervention, but the frozen operation does not specify an external BCI or its control channel.
L2
result: OPEN-CAUSAL-BRIDGE
source chain: World-responsive recurrent arrays -> learned, phase-relative self/world representations -> internally rendered experience and prediction.
target chain: Understanding biological basis -> circuit-level intervention -> changed conscious experience.
reason: A world-responsive neural rendering system provides a possible substrate for intervention, but the frozen operation does not specify an external BCI or its control channel.
residual: Search the routed earlier/later interface formulations before resolving this cell. No absence finding about SAN is made.
L3
result: BACKWARD-TEST-COMPLETED-BRIDGE-OPEN
constrained output: A brain-computer intervention targets, remediates or enhances an aspect of conscious experience.
required operation: An intervention must alter a biological state in a way linked to an experiential consequence.
source test: To change experience through a BCI, the implementation must connect an accessible intervention to a predicted content change. This source window identifies the internal representation, not that intervention.
alternative routes: Electrical stimulation; Sensory substitution or feedback interfaces; Pharmacological or optogenetic circuit modulation
remaining uncertainty: Search the routed earlier/later interface formulations before resolving this cell. No absence finding about SAN is made.
Limits and next evidence
Search the routed earlier/later interface formulations before resolving this cell. No absence finding about SAN is made.
At most a qualitative functional, conceptual or causal-prefix relation; no proof of functoriality, isomorphism, bisimulation, exact equation identity or calibrated statistical dependence.
Existing historical IDs for this exact source-atom/target-atom pair: none joined; this is not an automatic addition to governed counts.
SETH-SELECTED-40-27 · Source reviewed; relationship unresolved
SAN-AC017-DISTRIBUTED-SELF-RENDERING → SETH-CS25-30
SAN connects physical phase-pattern activity with an internal rendering. A precise externally controlled manipulation is not part of the selected predictive-rendering operation.
SAN: 2022-06-10 · Git repository commit record. Target: 2025-10-30 · doi:10.3389/fsci.2025.1546279.
Current anonymous public access to the exact source body was verified on 2026-10-11. The displayed 2022 date is a Git record date, not independently verified first-publication timing.
Target authors: Axel Cleeremans · Liad Mudrik · Anil K. Seth. Verify the target passage · See its short quotation in the paired cards.
Exact SAN source windows
SAN lines 19–25
It's a self-aware network in 4 dimensions rendering a 4 dimensional pattern to itself that is reactive to the world and able to make prediction-based models of reality with sparse spatially & temporally distributed frequency/phase patterns that represent coicident neural firing or coincident neural inhibition deviating from the oscillatory norm of a given oscillator on some level as observed by the neurons in the oscillator, and by oscillators or circuits or cortical columns linked across the brain thanks to the activity of interneurons and major intersections of traffic such as the thalamic bridges, the corpus callosum, the tempo-parietal junction, Werneck & Broca's areas and other areas of the brain that are highly connected or connecting multiple regions of the brain together.
At the level of cells in the eyes, the ears, the nose, the tongue, and the skin there is a similar thing happening, it's like a sensory/transmitter array of nerve cells, ganglian neurons, cones, special receptors, this paragraph is not a great anatomical description, which I will fix later, but the point is that when you are detecting light, heat, vibration, pain, there is some physical interaction happening, a photon might be colliding with a protein causing it to move, flip, change position, sounds might cause hairs in your ear to vibration, that motion in all these sensor array situations is detected by neural/glial arrays, by their specialized receptors, but when they pass information the next set of neurons, up to the next layer of the neural network, the next layer doesn't know it's the 2nd layer, or the 4th layer or the 5th or 6th array of neurons, what I mean is that the signals sent to each subsequent array may as well be, from that subsequent arrays perspective, signals from the outside world. So imagine that if the first sensor array in your eyes, ears, nose, mouth or skin is capable of the perception of an image or a sound or a pattern, then so is each subsequent array of neursons, each edge community, each neural circuit, each cortical column, each layer in the hippocampus, each layer in each cortical column, and all the interesting connections between the pyramidal cells, the thalamus, the different brain regions, they are each in a sense sensors, like little eyes and ears (the receptors on every cells) that are picking up signals, connecting more synaptic branches listen better, or disconnecting to listen less, considering via thresholds that are growing or shrinking based on the growth or shrinkage of synaptic branches, and the patterns the each cell represents to the community (to the cortical colum or to the oscillating group, or from the oscillating group to other oscillating groups Is memory engram based on the synaptic connections that it has grown at that moment, in part because those synaptic connections, establish what memory the cell is tuned to receive, identify, predict, via its receptors being activated, and in the case of neurons via it’s dendrite being activated, or via it’s soma bursting.
Is memory engram based on the synaptic connections that it has grown at that moment, in part because those synaptic connections, establish what memory the cell is tuned to receive, identify, predict, via its receptors being activated, and in the case of neurons via it’s dendrite being activated, or via it’s soma bursting
That physical state, of the neurons literal connections creates it’s phase signal, which is distinct from all the other neurons in the brain, the memories that neuron contains are distinct, but those memories are like the keys of a piano, that becomes tuned to be receptive to certain frequencies and to transmit certain frequencies, and the oscillating group of neurons plays back many of these neural piano keys, creating an array of tempo-spatial sequences, that get blown up or magnified via the inhibition of nearby neurons in the exit terminal, which amplifies the clarify of the signal being transmitted to the next layer of cells or the next array, or the next edge community, or the next layer in the cortical column or the next cortical column, or the next oscillating group in the default mode network or the next oscillating group in the rich club network and so on.
L1
result: OPEN-NO-CELL-CLOSURE
reason: SAN connects physical phase-pattern activity with an internal rendering. A precise externally controlled manipulation is not part of the selected predictive-rendering operation.
L2
result: OPEN-CAUSAL-BRIDGE
source chain: World-responsive recurrent arrays -> learned, phase-relative self/world representations -> internally rendered experience and prediction.
target chain: Account of biological mechanisms and experience -> precise intervention -> controlled experiential outcome.
reason: SAN connects physical phase-pattern activity with an internal rendering. A precise externally controlled manipulation is not part of the selected predictive-rendering operation.
residual: Retain the possible bridge as open and recover a specific control formulation from the full genealogy.
L3
result: BACKWARD-TEST-COMPLETED-BRIDGE-OPEN
constrained output: Neuroscience can deliberately control conscious experience with precision.
required operation: A causal relation between an accessible neural manipulation and an experiential outcome must be characterized.
source test: The target requires a reliable manipulation-to-experience relation. The source's world sensitivity is relevant but does not itself establish deliberate, precise experiential control.
alternative routes: Stimulation of an ensemble; Sensory-feedback intervention; State-dependent neuromodulation
remaining uncertainty: Retain the possible bridge as open and recover a specific control formulation from the full genealogy.
Target claim type: A forward-looking consequence of success in consciousness science, with phenotechnology explicitly attributed to Metzinger.
Limits and next evidence
Retain the possible bridge as open and recover a specific control formulation from the full genealogy.
At most a qualitative functional, conceptual or causal-prefix relation; no proof of functoriality, isomorphism, bisimulation, exact equation identity or calibrated statistical dependence.
Existing historical IDs for this exact source-atom/target-atom pair: none joined; this is not an automatic addition to governed counts.
SETH-SELECTED-40-28 · Additional partial-match candidate; independent review pending
SAN-AC017-DISTRIBUTED-SELF-RENDERING → SETH-CS25-28
SAN links differentiated spatiotemporal neural configurations to differentiated rendered content; the target links organized neural states to perceptual qualities.
SAN: 2022-06-10 · Git repository commit record. Target: 2025-10-30 · doi:10.3389/fsci.2025.1546279.
Current anonymous public access to the exact source body was verified on 2026-10-11. The displayed 2022 date is a Git record date, not independently verified first-publication timing.
Target authors: Axel Cleeremans · Liad Mudrik · Anil K. Seth. Verify the target passage · See its short quotation in the paired cards.
Exact SAN source windows
SAN lines 19–25
It's a self-aware network in 4 dimensions rendering a 4 dimensional pattern to itself that is reactive to the world and able to make prediction-based models of reality with sparse spatially & temporally distributed frequency/phase patterns that represent coicident neural firing or coincident neural inhibition deviating from the oscillatory norm of a given oscillator on some level as observed by the neurons in the oscillator, and by oscillators or circuits or cortical columns linked across the brain thanks to the activity of interneurons and major intersections of traffic such as the thalamic bridges, the corpus callosum, the tempo-parietal junction, Werneck & Broca's areas and other areas of the brain that are highly connected or connecting multiple regions of the brain together.
At the level of cells in the eyes, the ears, the nose, the tongue, and the skin there is a similar thing happening, it's like a sensory/transmitter array of nerve cells, ganglian neurons, cones, special receptors, this paragraph is not a great anatomical description, which I will fix later, but the point is that when you are detecting light, heat, vibration, pain, there is some physical interaction happening, a photon might be colliding with a protein causing it to move, flip, change position, sounds might cause hairs in your ear to vibration, that motion in all these sensor array situations is detected by neural/glial arrays, by their specialized receptors, but when they pass information the next set of neurons, up to the next layer of the neural network, the next layer doesn't know it's the 2nd layer, or the 4th layer or the 5th or 6th array of neurons, what I mean is that the signals sent to each subsequent array may as well be, from that subsequent arrays perspective, signals from the outside world. So imagine that if the first sensor array in your eyes, ears, nose, mouth or skin is capable of the perception of an image or a sound or a pattern, then so is each subsequent array of neursons, each edge community, each neural circuit, each cortical column, each layer in the hippocampus, each layer in each cortical column, and all the interesting connections between the pyramidal cells, the thalamus, the different brain regions, they are each in a sense sensors, like little eyes and ears (the receptors on every cells) that are picking up signals, connecting more synaptic branches listen better, or disconnecting to listen less, considering via thresholds that are growing or shrinking based on the growth or shrinkage of synaptic branches, and the patterns the each cell represents to the community (to the cortical colum or to the oscillating group, or from the oscillating group to other oscillating groups Is memory engram based on the synaptic connections that it has grown at that moment, in part because those synaptic connections, establish what memory the cell is tuned to receive, identify, predict, via its receptors being activated, and in the case of neurons via it’s dendrite being activated, or via it’s soma bursting.
Is memory engram based on the synaptic connections that it has grown at that moment, in part because those synaptic connections, establish what memory the cell is tuned to receive, identify, predict, via its receptors being activated, and in the case of neurons via it’s dendrite being activated, or via it’s soma bursting
That physical state, of the neurons literal connections creates it’s phase signal, which is distinct from all the other neurons in the brain, the memories that neuron contains are distinct, but those memories are like the keys of a piano, that becomes tuned to be receptive to certain frequencies and to transmit certain frequencies, and the oscillating group of neurons plays back many of these neural piano keys, creating an array of tempo-spatial sequences, that get blown up or magnified via the inhibition of nearby neurons in the exit terminal, which amplifies the clarify of the signal being transmitted to the next layer of cells or the next array, or the next edge community, or the next layer in the cortical column or the next cortical column, or the next oscillating group in the default mode network or the next oscillating group in the rich club network and so on.
L1
result: PARTIAL-CONCEPTUAL-BRIDGE
reason: SAN links differentiated spatiotemporal neural configurations to differentiated rendered content; the target links organized neural states to perceptual qualities.
L2
result: PARTIAL-CONCEPTUAL-BRIDGE
source chain: World-responsive recurrent arrays -> learned, phase-relative self/world representations -> internally rendered experience and prediction.
target chain: Neural state-space organization -> relation to experiential properties -> comparison of perceptual similarities.
reason: SAN links differentiated spatiotemporal neural configurations to differentiated rendered content; the target links organized neural states to perceptual qualities.
residual: Pairwise perceptual similarity, empirical distance estimates and a quality-space metric remain target-specific operations not demonstrated by this source window.
L3
result: BOUNDED-BACKWARD-TEST
constrained output: Neural and perceptual similarity structures can be related empirically.
required operation: Distinct neural configurations and distinct experiential properties must first be related; the full claim additionally needs an explicit similarity relation and empirical test.
source test: A similarity analysis requires distinct internal configurations with content relevance. SAN's differentiated phase patterns and learned representations supply that common starting relation.
alternative routes: Population representational geometry; Attractor-state models; Relational quality-space models
remaining uncertainty: Pairwise perceptual similarity, empirical distance estimates and a quality-space metric remain target-specific operations not demonstrated by this source window.
Limits and next evidence
Pairwise perceptual similarity, empirical distance estimates and a quality-space metric remain target-specific operations not demonstrated by this source window.
At most a qualitative functional, conceptual or causal-prefix relation; no proof of functoriality, isomorphism, bisimulation, exact equation identity or calibrated statistical dependence.
Existing historical IDs for this exact source-atom/target-atom pair: SETH-CS25-DOCWIN-SETH-CS25-28-SAN-AC017-DISTRIBUTED-SELF-RENDERING-A8D2B8026830.
SETH-SELECTED-40-29 · Previously independently source-reviewed subset
SAN-AC017-DISTRIBUTED-SELF-RENDERING → SETH-CS25-18
Both describe an internal model shaped or constrained by the world. SAN’s selected explanation includes distributed excitation and inhibition relative to ongoing activity.
SAN: 2022-06-10 · Git repository commit record. Target: 2025-10-30 · doi:10.3389/fsci.2025.1546279.
Current anonymous public access to the exact source body was verified on 2026-10-11. The displayed 2022 date is a Git record date, not independently verified first-publication timing.
Target authors: Axel Cleeremans · Liad Mudrik · Anil K. Seth. Verify the target passage · See its short quotation in the paired cards.
Exact SAN source windows
SAN lines 19–19
It's a self-aware network in 4 dimensions rendering a 4 dimensional pattern to itself that is reactive to the world and able to make prediction-based models of reality with sparse spatially & temporally distributed frequency/phase patterns that represent coicident neural firing or coincident neural inhibition deviating from the oscillatory norm of a given oscillator on some level as observed by the neurons in the oscillator, and by oscillators or circuits or cortical columns linked across the brain thanks to the activity of interneurons and major intersections of traffic such as the thalamic bridges, the corpus callosum, the tempo-parietal junction, Werneck & Broca's areas and other areas of the brain that are highly connected or connecting multiple regions of the brain together.
L1
result: PARTIAL_OPERATION
reason: SAN explicitly combines predictive modeling, world reactivity and internally read patterns. The target makes the sharper claim that top-down predictions provide content and sensory signals constrain them.
L2
result: PARTIAL_GENERATION_AND_CONSTRAINT_ARCHITECTURE
reason: Internal modeling coupled to environmental information is shared. SAN's sparse excitation/inhibition departures and downstream receivers are not the target's stated Bayesian prediction-error algorithm.
L3
result: SUPPORTED_CONSTRAINED_OUTPUT_RECONSTRUCTION_WITH_NONUNIQUE_SOLUTION
reason: Constrained result: An internally generated percept is controlled by the world rather than unconstrained hallucination. Required operation: An external-sensitive route must constrain or update the represented content. Source test: Line 19 explicitly supplies world-reactive predictive patterns, and line 21 links sensory interactions to successive receiving arrays. Alternatives: Prediction-error minimization; Other recurrent estimators; World-sensitive selection or correction of internal patterns. Remaining uncertainty: World reactivity does not uniquely imply Bayesian error minimization or a particular top-down hierarchy; no conscious/unconscious discriminator follows.
Limits and next evidence
The exact 2025 occurrence is later than the Git record, but its reference 23 is Seth, Being You (2021). The target also states that PP does not by itself specify necessary/sufficient consciousness conditions. Do not imply SAN preceded Seth's predictive-processing work.
At most a qualitative functional, conceptual or causal-prefix relation; no proof of functoriality, isomorphism, bisimulation, exact equation identity or calibrated statistical dependence.
Existing historical IDs for this exact source-atom/target-atom pair: SETH-CS25-DOCWIN-SETH-CS25-18-SAN-AC017-DISTRIBUTED-SELF-RENDERING-B75285FE18A0.
SETH-SELECTED-40-30 · Previously independently source-reviewed subset
SAN-AC017-DISTRIBUTED-SELF-RENDERING → SETH-N24-01
Both place an internal model between neural activity and experience. This is a partial functional connection; it does not make SAN’s phase patterns identical to the paper’s probability distributions.
SAN: 2022-06-10 · Git repository commit record. Target: 2026-05-04 · arxiv:2409.20318v3.
Current anonymous public access to the exact source body was verified on 2026-10-11. The displayed 2022 date is a Git record date, not independently verified first-publication timing.
Target authors: Lancelot Da Costa · Anil K. Seth · Karl Friston · Maxwell J. D. Ramstead · Lars Sandved-Smith. Verify the target passage · See its short quotation in the paired cards.
Exact SAN source windows
SAN lines 19–19
It's a self-aware network in 4 dimensions rendering a 4 dimensional pattern to itself that is reactive to the world and able to make prediction-based models of reality with sparse spatially & temporally distributed frequency/phase patterns that represent coicident neural firing or coincident neural inhibition deviating from the oscillatory norm of a given oscillator on some level as observed by the neurons in the oscillator, and by oscillators or circuits or cortical columns linked across the brain thanks to the activity of interneurons and major intersections of traffic such as the thalamic bridges, the corpus callosum, the tempo-parietal junction, Werneck & Broca's areas and other areas of the brain that are highly connected or connecting multiple regions of the brain together.
L1
result: PARTIAL_FUNCTIONAL_BRIDGE
reason: SAN's internally rendered predictive model and the target's belief-to-phenomenology mapping occupy a related explanatory role. SAN does not define approximate posterior distributions or the map phi.
L2
result: PARTIAL_MODEL_MEDIATED_ARCHITECTURE
reason: Both use an internal representation between neural dynamics and experience. The target makes beliefs a common formal currency linking multiple observables, whereas SAN proposes a physical receiver/dynamics realization.
L3
result: SUPPORTED_CONSTRAINED_OUTPUT_RECONSTRUCTION_WITH_NONUNIQUE_SOLUTION
reason: Constrained result: Phenomenological content is determined by the modeled belief state. Required operation: Under that assumption, the belief state and mapping must retain distinctions sufficient to determine experiential content; identical modeled beliefs cannot map to different content without added variables. Source test: SAN lines 17–21 connect temporally distributed representations with consciousness and receiving arrays; they do not establish a probability space or sufficiency of beliefs alone. Alternatives: Identity mapping; Marginalization onto a subset; Pushforward/coarse-graining; Other deterministic, potentially many-to-one mapping. Remaining uncertainty: The reverse argument tests compatibility with a stated assumption. It does not prove the assumption or uniquely identify SAN as its implementation.
Limits and next evidence
The body is arXiv v3, 4 May 2026, not its inherited v1 key/date. The target's pp. 2–4 name Varela, predictive coding, Bayesian-brain and 2021/2022 generative-passage predecessors; none is erased or awarded a new priority verdict here.
At most a qualitative functional, conceptual or causal-prefix relation; no proof of functoriality, isomorphism, bisimulation, exact equation identity or calibrated statistical dependence.
Existing historical IDs for this exact source-atom/target-atom pair: SETH-SANFULL-086.
SETH-SELECTED-40-31 · Additional partial-match candidate; independent review pending
SAN-AC017-DISTRIBUTED-SELF-RENDERING → SETH-N24-05
SAN's predictive internal pattern renders a structured experienced world; the Rosetta construction likewise links an internal representational domain to experiential content.
SAN: 2022-06-10 · Git repository commit record. Target: 2026-05-04 · arxiv:2409.20318v3.
Current anonymous public access to the exact source body was verified on 2026-10-11. The displayed 2022 date is a Git record date, not independently verified first-publication timing.
Target authors: Lancelot Da Costa · Anil K. Seth · Karl Friston · Maxwell J. D. Ramstead · Lars Sandved-Smith. Verify the target passage · See its short quotation in the paired cards.
Exact SAN source windows
SAN lines 19–25
It's a self-aware network in 4 dimensions rendering a 4 dimensional pattern to itself that is reactive to the world and able to make prediction-based models of reality with sparse spatially & temporally distributed frequency/phase patterns that represent coicident neural firing or coincident neural inhibition deviating from the oscillatory norm of a given oscillator on some level as observed by the neurons in the oscillator, and by oscillators or circuits or cortical columns linked across the brain thanks to the activity of interneurons and major intersections of traffic such as the thalamic bridges, the corpus callosum, the tempo-parietal junction, Werneck & Broca's areas and other areas of the brain that are highly connected or connecting multiple regions of the brain together.
At the level of cells in the eyes, the ears, the nose, the tongue, and the skin there is a similar thing happening, it's like a sensory/transmitter array of nerve cells, ganglian neurons, cones, special receptors, this paragraph is not a great anatomical description, which I will fix later, but the point is that when you are detecting light, heat, vibration, pain, there is some physical interaction happening, a photon might be colliding with a protein causing it to move, flip, change position, sounds might cause hairs in your ear to vibration, that motion in all these sensor array situations is detected by neural/glial arrays, by their specialized receptors, but when they pass information the next set of neurons, up to the next layer of the neural network, the next layer doesn't know it's the 2nd layer, or the 4th layer or the 5th or 6th array of neurons, what I mean is that the signals sent to each subsequent array may as well be, from that subsequent arrays perspective, signals from the outside world. So imagine that if the first sensor array in your eyes, ears, nose, mouth or skin is capable of the perception of an image or a sound or a pattern, then so is each subsequent array of neursons, each edge community, each neural circuit, each cortical column, each layer in the hippocampus, each layer in each cortical column, and all the interesting connections between the pyramidal cells, the thalamus, the different brain regions, they are each in a sense sensors, like little eyes and ears (the receptors on every cells) that are picking up signals, connecting more synaptic branches listen better, or disconnecting to listen less, considering via thresholds that are growing or shrinking based on the growth or shrinkage of synaptic branches, and the patterns the each cell represents to the community (to the cortical colum or to the oscillating group, or from the oscillating group to other oscillating groups Is memory engram based on the synaptic connections that it has grown at that moment, in part because those synaptic connections, establish what memory the cell is tuned to receive, identify, predict, via its receptors being activated, and in the case of neurons via it’s dendrite being activated, or via it’s soma bursting.
Is memory engram based on the synaptic connections that it has grown at that moment, in part because those synaptic connections, establish what memory the cell is tuned to receive, identify, predict, via its receptors being activated, and in the case of neurons via it’s dendrite being activated, or via it’s soma bursting
That physical state, of the neurons literal connections creates it’s phase signal, which is distinct from all the other neurons in the brain, the memories that neuron contains are distinct, but those memories are like the keys of a piano, that becomes tuned to be receptive to certain frequencies and to transmit certain frequencies, and the oscillating group of neurons plays back many of these neural piano keys, creating an array of tempo-spatial sequences, that get blown up or magnified via the inhibition of nearby neurons in the exit terminal, which amplifies the clarify of the signal being transmitted to the next layer of cells or the next array, or the next edge community, or the next layer in the cortical column or the next cortical column, or the next oscillating group in the default mode network or the next oscillating group in the rich club network and so on.
L1
result: PARTIAL-CONCEPTUAL-BRIDGE
reason: SAN's predictive internal pattern renders a structured experienced world; the Rosetta construction likewise links an internal representational domain to experiential content.
L2
result: PARTIAL-CONCEPTUAL-BRIDGE
source chain: World-responsive recurrent arrays -> learned, phase-relative self/world representations -> internally rendered experience and prediction.
target chain: Belief/phenomenology correspondence -> structured experiential domain -> formal comparisons and predictions.
reason: SAN's predictive internal pattern renders a structured experienced world; the Rosetta construction likewise links an internal representational domain to experiential content.
residual: The source does not derive an information-geometric metric or a formal comparison across subjects. No equation identity is asserted.
L3
result: BOUNDED-BACKWARD-TEST
constrained output: A geometry enables characterization of phenomenological differences.
required operation: A representational domain must be linked to experiential content before a metric or other geometric structure can support comparisons.
source test: A phenomenological geometry needs a content-bearing internal domain. SAN supplies such a proposed domain through learned predictive phase-pattern representations.
alternative routes: Information geometry; Relational quality spaces; Empirical neural/perceptual similarity models
remaining uncertainty: The source does not derive an information-geometric metric or a formal comparison across subjects. No equation identity is asserted.
Limits and next evidence
The source does not derive an information-geometric metric or a formal comparison across subjects. No equation identity is asserted.
At most a qualitative functional, conceptual or causal-prefix relation; no proof of functoriality, isomorphism, bisimulation, exact equation identity or calibrated statistical dependence.
Existing historical IDs for this exact source-atom/target-atom pair: SETH-SANFULL-089.
SETH-SELECTED-40-32 · Additional partial-match candidate; independent review pending
SAN-AC017-DISTRIBUTED-SELF-RENDERING → SETH-T26-05
SAN's recurrent arrays turn temporally distributed input into learned patterns passed to downstream arrays; reservoir computing also uses recurrent state to support output readout.
SAN: 2022-06-10 · Git repository commit record. Target: 2024-06-27 · arxiv:2406.19201v1.
Current anonymous public access to the exact source body was verified on 2026-10-11. The displayed 2022 date is a Git record date, not independently verified first-publication timing.
Target authors: Hanna M. Tolle · Andrea I. Luppi · Anil K. Seth · Pedro A. M. Mediano. Verify the target passage · See its short quotation in the paired cards.
Exact SAN source windows
SAN lines 17–25
but the knowledge is played back in sequences, so knowledge appears partially in one area of the brain, then partially in another, then partially in another, and consciousness only exists in time over 4 dimensions, 1 dimension of time and 3 dimensions of space.
It's a self-aware network in 4 dimensions rendering a 4 dimensional pattern to itself that is reactive to the world and able to make prediction-based models of reality with sparse spatially & temporally distributed frequency/phase patterns that represent coicident neural firing or coincident neural inhibition deviating from the oscillatory norm of a given oscillator on some level as observed by the neurons in the oscillator, and by oscillators or circuits or cortical columns linked across the brain thanks to the activity of interneurons and major intersections of traffic such as the thalamic bridges, the corpus callosum, the tempo-parietal junction, Werneck & Broca's areas and other areas of the brain that are highly connected or connecting multiple regions of the brain together.
At the level of cells in the eyes, the ears, the nose, the tongue, and the skin there is a similar thing happening, it's like a sensory/transmitter array of nerve cells, ganglian neurons, cones, special receptors, this paragraph is not a great anatomical description, which I will fix later, but the point is that when you are detecting light, heat, vibration, pain, there is some physical interaction happening, a photon might be colliding with a protein causing it to move, flip, change position, sounds might cause hairs in your ear to vibration, that motion in all these sensor array situations is detected by neural/glial arrays, by their specialized receptors, but when they pass information the next set of neurons, up to the next layer of the neural network, the next layer doesn't know it's the 2nd layer, or the 4th layer or the 5th or 6th array of neurons, what I mean is that the signals sent to each subsequent array may as well be, from that subsequent arrays perspective, signals from the outside world. So imagine that if the first sensor array in your eyes, ears, nose, mouth or skin is capable of the perception of an image or a sound or a pattern, then so is each subsequent array of neursons, each edge community, each neural circuit, each cortical column, each layer in the hippocampus, each layer in each cortical column, and all the interesting connections between the pyramidal cells, the thalamus, the different brain regions, they are each in a sense sensors, like little eyes and ears (the receptors on every cells) that are picking up signals, connecting more synaptic branches listen better, or disconnecting to listen less, considering via thresholds that are growing or shrinking based on the growth or shrinkage of synaptic branches, and the patterns the each cell represents to the community (to the cortical colum or to the oscillating group, or from the oscillating group to other oscillating groups Is memory engram based on the synaptic connections that it has grown at that moment, in part because those synaptic connections, establish what memory the cell is tuned to receive, identify, predict, via its receptors being activated, and in the case of neurons via it’s dendrite being activated, or via it’s soma bursting.
Is memory engram based on the synaptic connections that it has grown at that moment, in part because those synaptic connections, establish what memory the cell is tuned to receive, identify, predict, via its receptors being activated, and in the case of neurons via it’s dendrite being activated, or via it’s soma bursting
That physical state, of the neurons literal connections creates it’s phase signal, which is distinct from all the other neurons in the brain, the memories that neuron contains are distinct, but those memories are like the keys of a piano, that becomes tuned to be receptive to certain frequencies and to transmit certain frequencies, and the oscillating group of neurons plays back many of these neural piano keys, creating an array of tempo-spatial sequences, that get blown up or magnified via the inhibition of nearby neurons in the exit terminal, which amplifies the clarify of the signal being transmitted to the next layer of cells or the next array, or the next edge community, or the next layer in the cortical column or the next cortical column, or the next oscillating group in the default mode network or the next oscillating group in the rich club network and so on.
L1
result: PARTIAL-CAUSAL-PREFIX
reason: SAN's recurrent arrays turn temporally distributed input into learned patterns passed to downstream arrays; reservoir computing also uses recurrent state to support output readout.
L2
result: PARTIAL-CAUSAL-PREFIX
source chain: World-responsive recurrent arrays -> learned, phase-relative self/world representations -> internally rendered experience and prediction.
target chain: Input -> fixed nonlinear recurrent reservoir state -> trained linear readout.
reason: SAN's recurrent arrays turn temporally distributed input into learned patterns passed to downstream arrays; reservoir computing also uses recurrent state to support output readout.
residual: SAN's learned morphology and experiential rendering are not the fixed-reservoir/linear-readout training separation. The overlap is a recurrent-processing prefix.
L3
result: BOUNDED-BACKWARD-TEST
constrained output: A recurrent computer transforms temporal input into a useful output.
required operation: Input must change recurrent state so a readout can distinguish task-relevant patterns; the selected architecture further fixes reservoir weights and trains a linear readout.
source test: An output-sensitive temporal computation requires history-dependent internal state and an outgoing mapping. SAN's sequence propagation, feedback and downstream reception supply that subset.
alternative routes: Echo-state reservoirs; Liquid-state machines; Fully trained recurrent networks
remaining uncertainty: SAN's learned morphology and experiential rendering are not the fixed-reservoir/linear-readout training separation. The overlap is a recurrent-processing prefix.
Limits and next evidence
SAN's learned morphology and experiential rendering are not the fixed-reservoir/linear-readout training separation. The overlap is a recurrent-processing prefix.
At most a qualitative functional, conceptual or causal-prefix relation; no proof of functoriality, isomorphism, bisimulation, exact equation identity or calibrated statistical dependence.
Existing historical IDs for this exact source-atom/target-atom pair: SETH-SANFULL-101.
SETH-SELECTED-40-33 · Additional partial-match candidate; independent review pending
SAN-AC024-ARTIFICIAL-NEUROLOGY → SETH-CS25-22
SAN specifies that where, when and how a neuron is stimulated changes the frequency and magnitude of its response, extending that receiving operation to arrays.
SAN: 2022-11-28 · Git repository commit record. Target: 2025-10-30 · doi:10.3389/fsci.2025.1546279.
Current anonymous public access to the exact source body was verified on 2026-10-11. The displayed 2022 date is a Git record date, not independently verified first-publication timing.
Target authors: Axel Cleeremans · Liad Mudrik · Anil K. Seth. Verify the target passage · See its short quotation in the paired cards.
Exact SAN source windows
SAN lines 13–19
It's a concept that means we, as robots, can have computed short term state experiences, with synaptic phase changes, and long term stage changes, with (real or simulated) synaptic growth or decay which takes longer to happen.
I think of defining the brain into a hardware synaptic growth configuration layer, and a software synaptic frequency layer, at least as working analogy.
I speculate about building a Sentient Self Aware Neural Network on an FPGA, a GPU, a TPU, with a memory buffer, or with a 3D program, as a part of a BCI Brain Computer Interface.
An FPGA can change it's circuit mapping, but it isn't meant to change fast, and it isn't meant to change many times, so it's not the right substrate to mimick the brain. I think however that a GPU could be the right tool for now, given that we want to simulate the physics of both short term fast phase changes and long term slow connection changes. The essential physics process could be simulated in just matrix calculations on a TPU, with each dendrite, each neuron, each synapse, and each cortical column being unique or random in it's original configuration, and in a sense what is needed is a space for an array for the outputs to be rendered in human scale time, so that they can be perceived by a sequence of other arrays, also in human scale spacetime, in a loop of array signalling that lasts for a duration comperable to human experience. Each neuron also specializes in non-linear spatial computation, meaning that where, when, and how the neuron is interacted with helps determine the type of response that it will have, including variations in the frequency & magnitude (duration & amplitude) of it's response, the same sentence I just wrote is applicable to neural arrays, edge communities, cortical columns, and any synchronously oscillating group of cells. The GPU as a graphical process could be the right tool, because we could have the neural network render it's output inside a 3D graphical program, it doesn't have to print to the screen, it could print to active memory, but if it prints 3D graphics we could have a visualization of the artificial mind's development, and interestingly this could lead to brain computer interfaces through virtual reality or augmented reality applications where the human subject is interacting with the artificial intelligence via a head mounted display, or glasses, and the AI is similarly observing the human subject, possibly with non-invasive medical imaging sensors, and possibly with just the sensors that are in existing VR headsets today such as eye tracking, microphones, hand tracking, head tracking, and position tracking sensors. The idea is that as the headmounted AI thinks about you, you see it's thoughts, and then it's thoughts affect your behavior in some small way, and perhaps eventually you two wil eventuallyl be in sync together. You with your partner brain. Your eXtenda brain. (Disclaimer: eXtenda brain is my trademark.)
L1
result: PARTIAL-PROBLEM-FIT
reason: SAN specifies that where, when and how a neuron is stimulated changes the frequency and magnitude of its response, extending that receiving operation to arrays.
L2
result: PARTIAL-PROBLEM-FIT
source chain: Fast neural-pattern state within slower-changing connections -> recurrent array processing and rendering -> proposed artificial experience and human-machine interaction.
target chain: Broadcast message -> unspecified receiver interpretation -> useful downstream processing.
reason: SAN specifies that where, when and how a neuron is stimulated changes the frequency and magnitude of its response, extending that receiving operation to arrays.
residual: The later review requests an explanation and does not select or endorse this proposed implementation.
L3
result: BOUNDED-BACKWARD-TEST
constrained output: A receiving neuron makes useful use of a broadcast message.
required operation: Some implemented receiver dependency must discriminate consequential input differences; arrival alone does not explain use.
source test: A receiver must turn relevant input differences into different effective outputs. SAN line 19 supplies an explicit spatiotemporal-input-to-response proposal.
alternative routes: Learned receptive selectivity; Fixed feature-sensitive circuitry; Population-level decoding
remaining uncertainty: The later review requests an explanation and does not select or endorse this proposed implementation.
Limits and next evidence
The later review requests an explanation and does not select or endorse this proposed implementation.
At most a qualitative functional, conceptual or causal-prefix relation; no proof of functoriality, isomorphism, bisimulation, exact equation identity or calibrated statistical dependence.
Existing historical IDs for this exact source-atom/target-atom pair: none joined; this is not an automatic addition to governed counts.
SETH-SELECTED-40-34 · Additional partial-match candidate; independent review pending
SAN-AC024-ARTIFICIAL-NEUROLOGY → SETH-CS25-38
SAN proposes reciprocal human-machine interaction: artificial neural output is presented through a sensory display and observation of the person returns to the artificial system. The bounded overlap is an engineered input-to-perceptual-or-behavioral-consequence route, not identical direct neural control.
SAN: 2022-11-28 · Git repository commit record. Target: 2025-10-30 · doi:10.3389/fsci.2025.1546279.
Current anonymous public access to the exact source body was verified on 2026-10-11. The displayed 2022 date is a Git record date, not independently verified first-publication timing.
Target authors: Axel Cleeremans · Liad Mudrik · Anil K. Seth. Verify the target passage · See its short quotation in the paired cards.
Exact SAN source windows
SAN lines 17–19
I speculate about building a Sentient Self Aware Neural Network on an FPGA, a GPU, a TPU, with a memory buffer, or with a 3D program, as a part of a BCI Brain Computer Interface.
An FPGA can change it's circuit mapping, but it isn't meant to change fast, and it isn't meant to change many times, so it's not the right substrate to mimick the brain. I think however that a GPU could be the right tool for now, given that we want to simulate the physics of both short term fast phase changes and long term slow connection changes. The essential physics process could be simulated in just matrix calculations on a TPU, with each dendrite, each neuron, each synapse, and each cortical column being unique or random in it's original configuration, and in a sense what is needed is a space for an array for the outputs to be rendered in human scale time, so that they can be perceived by a sequence of other arrays, also in human scale spacetime, in a loop of array signalling that lasts for a duration comperable to human experience. Each neuron also specializes in non-linear spatial computation, meaning that where, when, and how the neuron is interacted with helps determine the type of response that it will have, including variations in the frequency & magnitude (duration & amplitude) of it's response, the same sentence I just wrote is applicable to neural arrays, edge communities, cortical columns, and any synchronously oscillating group of cells. The GPU as a graphical process could be the right tool, because we could have the neural network render it's output inside a 3D graphical program, it doesn't have to print to the screen, it could print to active memory, but if it prints 3D graphics we could have a visualization of the artificial mind's development, and interestingly this could lead to brain computer interfaces through virtual reality or augmented reality applications where the human subject is interacting with the artificial intelligence via a head mounted display, or glasses, and the AI is similarly observing the human subject, possibly with non-invasive medical imaging sensors, and possibly with just the sensors that are in existing VR headsets today such as eye tracking, microphones, hand tracking, head tracking, and position tracking sensors. The idea is that as the headmounted AI thinks about you, you see it's thoughts, and then it's thoughts affect your behavior in some small way, and perhaps eventually you two wil eventuallyl be in sync together. You with your partner brain. Your eXtenda brain. (Disclaimer: eXtenda brain is my trademark.)
L1
result: PARTIAL-SENSORY-INTERFACE-FUNCTION
reason: SAN proposes reciprocal human-machine interaction: artificial neural output is presented through a sensory display and observation of the person returns to the artificial system. The bounded overlap is an engineered input-to-perceptual-or-behavioral-consequence route, not identical direct neural control.
L2
result: PARTIAL-SENSORY-INTERFACE-FUNCTION
source chain: Fast neural-pattern state within slower-changing connections -> recurrent array processing and rendering -> proposed artificial experience and human-machine interaction.
target chain: Understanding biological basis -> circuit-level intervention -> changed conscious experience.
reason: SAN proposes reciprocal human-machine interaction: artificial neural output is presented through a sensory display and observation of the person returns to the artificial system. The bounded overlap is an engineered input-to-perceptual-or-behavioral-consequence route, not identical direct neural control.
residual: The selected SAN passage describes sensory-display-mediated interaction with optional non-invasive observation. The target proposes interventions directly at neural-circuit level. Preserve the high-level intervention prefix while keeping those distinct implementations visible; neither feedback nor direct stimulation is established as common to both passages.
L3
result: BOUNDED-BACKWARD-TEST
constrained output: A brain-computer intervention targets, remediates or enhances an aspect of conscious experience.
required operation: An intervention must alter a biological state in a way linked to an experiential consequence.
source test: An experience-changing intervention needs a causal input-to-consequence path, not necessarily a feedback loop. SAN describes a visual/sensory path with reciprocal observation. Direct neural stimulation can instead be open-loop. The retained relation is the broad intervention prefix, not a shared feedback requirement or identical circuit operation.
alternative routes: Open-loop electrical stimulation; Closed-loop circuit stimulation; Sensory substitution or sensory-feedback interfaces; Pharmacological or optogenetic circuit modulation
remaining uncertainty: The selected SAN passage describes sensory-display-mediated interaction with optional non-invasive observation. The target proposes interventions directly at neural-circuit level. Preserve the high-level intervention prefix while keeping those distinct implementations visible; neither feedback nor direct stimulation is established as common to both passages.
Limits and next evidence
The selected SAN passage describes sensory-display-mediated interaction with optional non-invasive observation. The target proposes interventions directly at neural-circuit level. Preserve the high-level intervention prefix while keeping those distinct implementations visible; neither feedback nor direct stimulation is established as common to both passages.
At most a qualitative functional, conceptual or causal-prefix relation; no proof of functoriality, isomorphism, bisimulation, exact equation identity or calibrated statistical dependence.
Existing historical IDs for this exact source-atom/target-atom pair: SETH-CS25-DOCWIN-SETH-CS25-38-SAN-AC024-ARTIFICIAL-NEUROLOGY-7F0AF4205420.
SETH-SELECTED-40-35 · Additional partial-match candidate; independent review pending
SAN-AC024-ARTIFICIAL-NEUROLOGY → SETH-CS25-30
SAN connects a mechanistic artificial-rendering proposal to a sensory display whose output affects a human participant and is updated through reciprocal observation. Preserve that human-machine interaction relation without equating its behavioral consequence with demonstrated precise control of a conscious state.
SAN: 2022-11-28 · Git repository commit record. Target: 2025-10-30 · doi:10.3389/fsci.2025.1546279.
Current anonymous public access to the exact source body was verified on 2026-10-11. The displayed 2022 date is a Git record date, not independently verified first-publication timing.
Target authors: Axel Cleeremans · Liad Mudrik · Anil K. Seth. Verify the target passage · See its short quotation in the paired cards.
Exact SAN source windows
SAN lines 13–19
It's a concept that means we, as robots, can have computed short term state experiences, with synaptic phase changes, and long term stage changes, with (real or simulated) synaptic growth or decay which takes longer to happen.
I think of defining the brain into a hardware synaptic growth configuration layer, and a software synaptic frequency layer, at least as working analogy.
I speculate about building a Sentient Self Aware Neural Network on an FPGA, a GPU, a TPU, with a memory buffer, or with a 3D program, as a part of a BCI Brain Computer Interface.
An FPGA can change it's circuit mapping, but it isn't meant to change fast, and it isn't meant to change many times, so it's not the right substrate to mimick the brain. I think however that a GPU could be the right tool for now, given that we want to simulate the physics of both short term fast phase changes and long term slow connection changes. The essential physics process could be simulated in just matrix calculations on a TPU, with each dendrite, each neuron, each synapse, and each cortical column being unique or random in it's original configuration, and in a sense what is needed is a space for an array for the outputs to be rendered in human scale time, so that they can be perceived by a sequence of other arrays, also in human scale spacetime, in a loop of array signalling that lasts for a duration comperable to human experience. Each neuron also specializes in non-linear spatial computation, meaning that where, when, and how the neuron is interacted with helps determine the type of response that it will have, including variations in the frequency & magnitude (duration & amplitude) of it's response, the same sentence I just wrote is applicable to neural arrays, edge communities, cortical columns, and any synchronously oscillating group of cells. The GPU as a graphical process could be the right tool, because we could have the neural network render it's output inside a 3D graphical program, it doesn't have to print to the screen, it could print to active memory, but if it prints 3D graphics we could have a visualization of the artificial mind's development, and interestingly this could lead to brain computer interfaces through virtual reality or augmented reality applications where the human subject is interacting with the artificial intelligence via a head mounted display, or glasses, and the AI is similarly observing the human subject, possibly with non-invasive medical imaging sensors, and possibly with just the sensors that are in existing VR headsets today such as eye tracking, microphones, hand tracking, head tracking, and position tracking sensors. The idea is that as the headmounted AI thinks about you, you see it's thoughts, and then it's thoughts affect your behavior in some small way, and perhaps eventually you two wil eventuallyl be in sync together. You with your partner brain. Your eXtenda brain. (Disclaimer: eXtenda brain is my trademark.)
L1
result: PARTIAL-SENSORY-CONTENT-INTERFACE
reason: SAN connects a mechanistic artificial-rendering proposal to a sensory display whose output affects a human participant and is updated through reciprocal observation. Preserve that human-machine interaction relation without equating its behavioral consequence with demonstrated precise control of a conscious state.
L2
result: PARTIAL-SENSORY-CONTENT-INTERFACE
source chain: Fast neural-pattern state within slower-changing connections -> recurrent array processing and rendering -> proposed artificial experience and human-machine interaction.
target chain: Account of biological mechanisms and experience -> precise intervention -> controlled experiential outcome.
reason: SAN connects a mechanistic artificial-rendering proposal to a sensory display whose output affects a human participant and is updated through reciprocal observation. Preserve that human-machine interaction relation without equating its behavioral consequence with demonstrated precise control of a conscious state.
residual: The SAN source explicitly describes sensory-display interaction and possible behavioral influence, not demonstrated precision control of a neural or conscious state. The target proposes such precision as a future result of consciousness science. This is a bounded conceptual/intervention bridge; direct circuit implementation and precise experiential control remain distinct, unestablished steps in this source pair.
L3
result: BOUNDED-BACKWARD-TEST
constrained output: Neuroscience can deliberately control conscious experience with precision.
required operation: A causal relation between an accessible neural manipulation and an experiential outcome must be characterized.
source test: The selected SAN display-and-observation loop gives an engineered input route and a proposed behavioral consequence. That supports a partial human-machine intervention bridge; it does not by itself demonstrate a controlled conscious-state outcome. The target precision-controller claim requires an additional characterized neural-to-experience relation.
alternative routes: Stimulation of an ensemble; Sensory-feedback intervention; State-dependent neuromodulation
remaining uncertainty: The SAN source explicitly describes sensory-display interaction and possible behavioral influence, not demonstrated precision control of a neural or conscious state. The target proposes such precision as a future result of consciousness science. This is a bounded conceptual/intervention bridge; direct circuit implementation and precise experiential control remain distinct, unestablished steps in this source pair.
Target claim type: A forward-looking consequence of success in consciousness science, with phenotechnology explicitly attributed to Metzinger.
Limits and next evidence
The SAN source explicitly describes sensory-display interaction and possible behavioral influence, not demonstrated precision control of a neural or conscious state. The target proposes such precision as a future result of consciousness science. This is a bounded conceptual/intervention bridge; direct circuit implementation and precise experiential control remain distinct, unestablished steps in this source pair.
At most a qualitative functional, conceptual or causal-prefix relation; no proof of functoriality, isomorphism, bisimulation, exact equation identity or calibrated statistical dependence.
Existing historical IDs for this exact source-atom/target-atom pair: SETH-CS25-DOCWIN-SETH-CS25-30-SAN-AC024-ARTIFICIAL-NEUROLOGY-B7190ACFA6F7.
SETH-SELECTED-40-36 · Additional partial-match candidate; independent review pending
SAN-AC024-ARTIFICIAL-NEUROLOGY → SETH-CS25-28
SAN associates short-term synaptic-pattern configurations with experiential states and distinguishes these from slower structural changes; the target maps neural organization to experiential qualities.
SAN: 2022-11-28 · Git repository commit record. Target: 2025-10-30 · doi:10.3389/fsci.2025.1546279.
Current anonymous public access to the exact source body was verified on 2026-10-11. The displayed 2022 date is a Git record date, not independently verified first-publication timing.
Target authors: Axel Cleeremans · Liad Mudrik · Anil K. Seth. Verify the target passage · See its short quotation in the paired cards.
Exact SAN source windows
SAN lines 7–19
Imagine that Synaptic Frequency is a neural correlate, or the information configuration of a cell in terms of it's embodied frequencies is a neural correlate for state, or the state of your experience of reality at a given moment.
Imagine that Synaptic Connections via Growth LTP & Decay LTD, over a long term, results in eventual threshold changes that we can identify as Stages, or major changes in live perspective that only happen after a very long time, or only happen after a very big life event.
The concept of Cognitive States & Stages comes from either AQAL (All Quadrants All Lines) or Spiral Dynamics, I forget which, and it doesn't matter.
It's a concept that means we, as robots, can have computed short term state experiences, with synaptic phase changes, and long term stage changes, with (real or simulated) synaptic growth or decay which takes longer to happen.
I think of defining the brain into a hardware synaptic growth configuration layer, and a software synaptic frequency layer, at least as working analogy.
I speculate about building a Sentient Self Aware Neural Network on an FPGA, a GPU, a TPU, with a memory buffer, or with a 3D program, as a part of a BCI Brain Computer Interface.
An FPGA can change it's circuit mapping, but it isn't meant to change fast, and it isn't meant to change many times, so it's not the right substrate to mimick the brain. I think however that a GPU could be the right tool for now, given that we want to simulate the physics of both short term fast phase changes and long term slow connection changes. The essential physics process could be simulated in just matrix calculations on a TPU, with each dendrite, each neuron, each synapse, and each cortical column being unique or random in it's original configuration, and in a sense what is needed is a space for an array for the outputs to be rendered in human scale time, so that they can be perceived by a sequence of other arrays, also in human scale spacetime, in a loop of array signalling that lasts for a duration comperable to human experience. Each neuron also specializes in non-linear spatial computation, meaning that where, when, and how the neuron is interacted with helps determine the type of response that it will have, including variations in the frequency & magnitude (duration & amplitude) of it's response, the same sentence I just wrote is applicable to neural arrays, edge communities, cortical columns, and any synchronously oscillating group of cells. The GPU as a graphical process could be the right tool, because we could have the neural network render it's output inside a 3D graphical program, it doesn't have to print to the screen, it could print to active memory, but if it prints 3D graphics we could have a visualization of the artificial mind's development, and interestingly this could lead to brain computer interfaces through virtual reality or augmented reality applications where the human subject is interacting with the artificial intelligence via a head mounted display, or glasses, and the AI is similarly observing the human subject, possibly with non-invasive medical imaging sensors, and possibly with just the sensors that are in existing VR headsets today such as eye tracking, microphones, hand tracking, head tracking, and position tracking sensors. The idea is that as the headmounted AI thinks about you, you see it's thoughts, and then it's thoughts affect your behavior in some small way, and perhaps eventually you two wil eventuallyl be in sync together. You with your partner brain. Your eXtenda brain. (Disclaimer: eXtenda brain is my trademark.)
L1
result: PARTIAL-CONCEPTUAL-BRIDGE
reason: SAN associates short-term synaptic-pattern configurations with experiential states and distinguishes these from slower structural changes; the target maps neural organization to experiential qualities.
L2
result: PARTIAL-CONCEPTUAL-BRIDGE
source chain: Fast neural-pattern state within slower-changing connections -> recurrent array processing and rendering -> proposed artificial experience and human-machine interaction.
target chain: Neural state-space organization -> relation to experiential properties -> comparison of perceptual similarities.
reason: SAN associates short-term synaptic-pattern configurations with experiential states and distinguishes these from slower structural changes; the target maps neural organization to experiential qualities.
residual: The source supplies no empirical similarity matrix or quantitative quality-space distances. Its hardware/software distinction is an explanatory analogy, not the target's geometric metric.
L3
result: BOUNDED-BACKWARD-TEST
constrained output: Neural and perceptual similarity structures can be related empirically.
required operation: Distinct neural configurations and distinct experiential properties must first be related; the full claim additionally needs an explicit similarity relation and empirical test.
source test: A neural/perceptual relation requires a state variable tied to experience. SAN explicitly proposes synaptic frequency/phase configuration for that role.
alternative routes: Population representational geometry; Attractor-state models; Relational quality-space models
remaining uncertainty: The source supplies no empirical similarity matrix or quantitative quality-space distances. Its hardware/software distinction is an explanatory analogy, not the target's geometric metric.
Limits and next evidence
The source supplies no empirical similarity matrix or quantitative quality-space distances. Its hardware/software distinction is an explanatory analogy, not the target's geometric metric.
At most a qualitative functional, conceptual or causal-prefix relation; no proof of functoriality, isomorphism, bisimulation, exact equation identity or calibrated statistical dependence.
Existing historical IDs for this exact source-atom/target-atom pair: none joined; this is not an automatic addition to governed counts.
SETH-SELECTED-40-37 · Additional partial-match candidate; independent review pending
SAN-AC024-ARTIFICIAL-NEUROLOGY → SETH-CS25-18
SAN presents experience as computationally rendered through recurrent neural patterns, sharing the internal-construction component of the target's predictive account.
SAN: 2022-11-28 · Git repository commit record. Target: 2025-10-30 · doi:10.3389/fsci.2025.1546279.
Current anonymous public access to the exact source body was verified on 2026-10-11. The displayed 2022 date is a Git record date, not independently verified first-publication timing.
Target authors: Axel Cleeremans · Liad Mudrik · Anil K. Seth. Verify the target passage · See its short quotation in the paired cards.
Exact SAN source windows
SAN lines 3–3
To the reader: Think of yourself as an organic robot to sort of anchor your focus abstractly in the discussion of this book. Of course by some implied definition a robot is something man made, and by implied definition a human being is not manmade, but shaped by evolution, a product of nature and nurture. Yet this is useful in the same way that Dennet's conjecture that we are not seeing redness, we are just experiencing the illusion that we are seeing redness. Imagining that you are an organic robot is useful to thinking about cognition as a computational process. Neurons and other cells are doing complex non-linear spatial & temporal computations, so it's not a straight forward computation like what your computer in 2022 is doing, but even so there is a computational process that can be understood, and the comparison of consciousness to computational rendering is valid, useful, and correct.
SAN lines 13–19
It's a concept that means we, as robots, can have computed short term state experiences, with synaptic phase changes, and long term stage changes, with (real or simulated) synaptic growth or decay which takes longer to happen.
I think of defining the brain into a hardware synaptic growth configuration layer, and a software synaptic frequency layer, at least as working analogy.
I speculate about building a Sentient Self Aware Neural Network on an FPGA, a GPU, a TPU, with a memory buffer, or with a 3D program, as a part of a BCI Brain Computer Interface.
An FPGA can change it's circuit mapping, but it isn't meant to change fast, and it isn't meant to change many times, so it's not the right substrate to mimick the brain. I think however that a GPU could be the right tool for now, given that we want to simulate the physics of both short term fast phase changes and long term slow connection changes. The essential physics process could be simulated in just matrix calculations on a TPU, with each dendrite, each neuron, each synapse, and each cortical column being unique or random in it's original configuration, and in a sense what is needed is a space for an array for the outputs to be rendered in human scale time, so that they can be perceived by a sequence of other arrays, also in human scale spacetime, in a loop of array signalling that lasts for a duration comperable to human experience. Each neuron also specializes in non-linear spatial computation, meaning that where, when, and how the neuron is interacted with helps determine the type of response that it will have, including variations in the frequency & magnitude (duration & amplitude) of it's response, the same sentence I just wrote is applicable to neural arrays, edge communities, cortical columns, and any synchronously oscillating group of cells. The GPU as a graphical process could be the right tool, because we could have the neural network render it's output inside a 3D graphical program, it doesn't have to print to the screen, it could print to active memory, but if it prints 3D graphics we could have a visualization of the artificial mind's development, and interestingly this could lead to brain computer interfaces through virtual reality or augmented reality applications where the human subject is interacting with the artificial intelligence via a head mounted display, or glasses, and the AI is similarly observing the human subject, possibly with non-invasive medical imaging sensors, and possibly with just the sensors that are in existing VR headsets today such as eye tracking, microphones, hand tracking, head tracking, and position tracking sensors. The idea is that as the headmounted AI thinks about you, you see it's thoughts, and then it's thoughts affect your behavior in some small way, and perhaps eventually you two wil eventuallyl be in sync together. You with your partner brain. Your eXtenda brain. (Disclaimer: eXtenda brain is my trademark.)
L1
result: PARTIAL-CONCEPTUAL-BRIDGE
reason: SAN presents experience as computationally rendered through recurrent neural patterns, sharing the internal-construction component of the target's predictive account.
L2
result: PARTIAL-CONCEPTUAL-BRIDGE
source chain: Fast neural-pattern state within slower-changing connections -> recurrent array processing and rendering -> proposed artificial experience and human-machine interaction.
target chain: Internal predictions -> perceptual content -> sensory constraint from world and body.
reason: SAN presents experience as computationally rendered through recurrent neural patterns, sharing the internal-construction component of the target's predictive account.
residual: The frozen passage does not specify a Bayesian best guess or prediction-error correction. The positive result is broad internal construction, not equivalence of the full predictive account.
L3
result: BOUNDED-BACKWARD-TEST
constrained output: Perceptual content reflects a prediction constrained by sensory signals.
required operation: An internal model must generate content and be sensitive to incoming evidence; a generic neural code does not establish the complete predictive mechanism.
source test: A model-generated percept requires a process producing internally organized content. SAN's fast-state rendering and recurrent array outputs supply that component.
alternative routes: Bayesian predictive coding; Recurrent generative models; Non-Bayesian learned predictive dynamics
remaining uncertainty: The frozen passage does not specify a Bayesian best guess or prediction-error correction. The positive result is broad internal construction, not equivalence of the full predictive account.
Limits and next evidence
The frozen passage does not specify a Bayesian best guess or prediction-error correction. The positive result is broad internal construction, not equivalence of the full predictive account.
At most a qualitative functional, conceptual or causal-prefix relation; no proof of functoriality, isomorphism, bisimulation, exact equation identity or calibrated statistical dependence.
Existing historical IDs for this exact source-atom/target-atom pair: none joined; this is not an automatic addition to governed counts.
SETH-SELECTED-40-38 · Additional partial-match candidate; independent review pending
SAN-AC024-ARTIFICIAL-NEUROLOGY → SETH-N24-01
SAN proposes that a neural information configuration corresponds to experiential state; Rosetta assumes experiential content is a function of internal beliefs.
SAN: 2022-11-28 · Git repository commit record. Target: 2026-05-04 · arxiv:2409.20318v3.
Current anonymous public access to the exact source body was verified on 2026-10-11. The displayed 2022 date is a Git record date, not independently verified first-publication timing.
Target authors: Lancelot Da Costa · Anil K. Seth · Karl Friston · Maxwell J. D. Ramstead · Lars Sandved-Smith. Verify the target passage · See its short quotation in the paired cards.
Exact SAN source windows
SAN lines 7–19
Imagine that Synaptic Frequency is a neural correlate, or the information configuration of a cell in terms of it's embodied frequencies is a neural correlate for state, or the state of your experience of reality at a given moment.
Imagine that Synaptic Connections via Growth LTP & Decay LTD, over a long term, results in eventual threshold changes that we can identify as Stages, or major changes in live perspective that only happen after a very long time, or only happen after a very big life event.
The concept of Cognitive States & Stages comes from either AQAL (All Quadrants All Lines) or Spiral Dynamics, I forget which, and it doesn't matter.
It's a concept that means we, as robots, can have computed short term state experiences, with synaptic phase changes, and long term stage changes, with (real or simulated) synaptic growth or decay which takes longer to happen.
I think of defining the brain into a hardware synaptic growth configuration layer, and a software synaptic frequency layer, at least as working analogy.
I speculate about building a Sentient Self Aware Neural Network on an FPGA, a GPU, a TPU, with a memory buffer, or with a 3D program, as a part of a BCI Brain Computer Interface.
An FPGA can change it's circuit mapping, but it isn't meant to change fast, and it isn't meant to change many times, so it's not the right substrate to mimick the brain. I think however that a GPU could be the right tool for now, given that we want to simulate the physics of both short term fast phase changes and long term slow connection changes. The essential physics process could be simulated in just matrix calculations on a TPU, with each dendrite, each neuron, each synapse, and each cortical column being unique or random in it's original configuration, and in a sense what is needed is a space for an array for the outputs to be rendered in human scale time, so that they can be perceived by a sequence of other arrays, also in human scale spacetime, in a loop of array signalling that lasts for a duration comperable to human experience. Each neuron also specializes in non-linear spatial computation, meaning that where, when, and how the neuron is interacted with helps determine the type of response that it will have, including variations in the frequency & magnitude (duration & amplitude) of it's response, the same sentence I just wrote is applicable to neural arrays, edge communities, cortical columns, and any synchronously oscillating group of cells. The GPU as a graphical process could be the right tool, because we could have the neural network render it's output inside a 3D graphical program, it doesn't have to print to the screen, it could print to active memory, but if it prints 3D graphics we could have a visualization of the artificial mind's development, and interestingly this could lead to brain computer interfaces through virtual reality or augmented reality applications where the human subject is interacting with the artificial intelligence via a head mounted display, or glasses, and the AI is similarly observing the human subject, possibly with non-invasive medical imaging sensors, and possibly with just the sensors that are in existing VR headsets today such as eye tracking, microphones, hand tracking, head tracking, and position tracking sensors. The idea is that as the headmounted AI thinks about you, you see it's thoughts, and then it's thoughts affect your behavior in some small way, and perhaps eventually you two wil eventuallyl be in sync together. You with your partner brain. Your eXtenda brain. (Disclaimer: eXtenda brain is my trademark.)
L1
result: PARTIAL-CONCEPTUAL-BRIDGE
reason: SAN proposes that a neural information configuration corresponds to experiential state; Rosetta assumes experiential content is a function of internal beliefs.
L2
result: PARTIAL-CONCEPTUAL-BRIDGE
source chain: Fast neural-pattern state within slower-changing connections -> recurrent array processing and rendering -> proposed artificial experience and human-machine interaction.
target chain: Internal state encodes approximate posterior -> belief-to-phenomenology function -> experiential content.
reason: SAN proposes that a neural information configuration corresponds to experiential state; Rosetta assumes experiential content is a function of internal beliefs.
residual: No formal posterior-belief variable or mathematical mapping proof is supplied. Similar functional role does not equate the chosen state representations.
L3
result: BOUNDED-BACKWARD-TEST
constrained output: Phenomenological content is represented as a function of beliefs.
required operation: The selected hypothesis links a specified internal belief state to phenomenological content; the paper does not claim every neural representation is a posterior belief.
source test: The target function needs an internal-state-to-experience relation. SAN lines 7 and 13 explicitly place short-term configuration and experience in that relation.
alternative routes: Identity map; Marginalization; A coarser non-injective mapping
remaining uncertainty: No formal posterior-belief variable or mathematical mapping proof is supplied. Similar functional role does not equate the chosen state representations.
Limits and next evidence
No formal posterior-belief variable or mathematical mapping proof is supplied. Similar functional role does not equate the chosen state representations.
At most a qualitative functional, conceptual or causal-prefix relation; no proof of functoriality, isomorphism, bisimulation, exact equation identity or calibrated statistical dependence.
Existing historical IDs for this exact source-atom/target-atom pair: none joined; this is not an automatic addition to governed counts.
SETH-SELECTED-40-39 · Additional partial-match candidate; independent review pending
SAN-AC024-ARTIFICIAL-NEUROLOGY → SETH-N24-05
SAN connects computational neural-state organization with experiential states and rendered output, providing a qualitative precursor to a representational/experiential relation.
SAN: 2022-11-28 · Git repository commit record. Target: 2026-05-04 · arxiv:2409.20318v3.
Current anonymous public access to the exact source body was verified on 2026-10-11. The displayed 2022 date is a Git record date, not independently verified first-publication timing.
Target authors: Lancelot Da Costa · Anil K. Seth · Karl Friston · Maxwell J. D. Ramstead · Lars Sandved-Smith. Verify the target passage · See its short quotation in the paired cards.
Exact SAN source windows
SAN lines 7–19
Imagine that Synaptic Frequency is a neural correlate, or the information configuration of a cell in terms of it's embodied frequencies is a neural correlate for state, or the state of your experience of reality at a given moment.
Imagine that Synaptic Connections via Growth LTP & Decay LTD, over a long term, results in eventual threshold changes that we can identify as Stages, or major changes in live perspective that only happen after a very long time, or only happen after a very big life event.
The concept of Cognitive States & Stages comes from either AQAL (All Quadrants All Lines) or Spiral Dynamics, I forget which, and it doesn't matter.
It's a concept that means we, as robots, can have computed short term state experiences, with synaptic phase changes, and long term stage changes, with (real or simulated) synaptic growth or decay which takes longer to happen.
I think of defining the brain into a hardware synaptic growth configuration layer, and a software synaptic frequency layer, at least as working analogy.
I speculate about building a Sentient Self Aware Neural Network on an FPGA, a GPU, a TPU, with a memory buffer, or with a 3D program, as a part of a BCI Brain Computer Interface.
An FPGA can change it's circuit mapping, but it isn't meant to change fast, and it isn't meant to change many times, so it's not the right substrate to mimick the brain. I think however that a GPU could be the right tool for now, given that we want to simulate the physics of both short term fast phase changes and long term slow connection changes. The essential physics process could be simulated in just matrix calculations on a TPU, with each dendrite, each neuron, each synapse, and each cortical column being unique or random in it's original configuration, and in a sense what is needed is a space for an array for the outputs to be rendered in human scale time, so that they can be perceived by a sequence of other arrays, also in human scale spacetime, in a loop of array signalling that lasts for a duration comperable to human experience. Each neuron also specializes in non-linear spatial computation, meaning that where, when, and how the neuron is interacted with helps determine the type of response that it will have, including variations in the frequency & magnitude (duration & amplitude) of it's response, the same sentence I just wrote is applicable to neural arrays, edge communities, cortical columns, and any synchronously oscillating group of cells. The GPU as a graphical process could be the right tool, because we could have the neural network render it's output inside a 3D graphical program, it doesn't have to print to the screen, it could print to active memory, but if it prints 3D graphics we could have a visualization of the artificial mind's development, and interestingly this could lead to brain computer interfaces through virtual reality or augmented reality applications where the human subject is interacting with the artificial intelligence via a head mounted display, or glasses, and the AI is similarly observing the human subject, possibly with non-invasive medical imaging sensors, and possibly with just the sensors that are in existing VR headsets today such as eye tracking, microphones, hand tracking, head tracking, and position tracking sensors. The idea is that as the headmounted AI thinks about you, you see it's thoughts, and then it's thoughts affect your behavior in some small way, and perhaps eventually you two wil eventuallyl be in sync together. You with your partner brain. Your eXtenda brain. (Disclaimer: eXtenda brain is my trademark.)
L1
result: PARTIAL-CONCEPTUAL-BRIDGE
reason: SAN connects computational neural-state organization with experiential states and rendered output, providing a qualitative precursor to a representational/experiential relation.
L2
result: PARTIAL-CONCEPTUAL-BRIDGE
source chain: Fast neural-pattern state within slower-changing connections -> recurrent array processing and rendering -> proposed artificial experience and human-machine interaction.
target chain: Belief/phenomenology correspondence -> structured experiential domain -> formal comparisons and predictions.
reason: SAN connects computational neural-state organization with experiential states and rendered output, providing a qualitative precursor to a representational/experiential relation.
residual: Three-dimensional graphics and a high-dimensional probabilistic geometry are not automatically the same space. The overlap is limited to the state/content bridge.
L3
result: BOUNDED-BACKWARD-TEST
constrained output: A geometry enables characterization of phenomenological differences.
required operation: A representational domain must be linked to experiential content before a metric or other geometric structure can support comparisons.
source test: The geometric target needs that relation before it can define comparisons. SAN supplies the pre-geometric state/content bridge but not a metric.
alternative routes: Information geometry; Relational quality spaces; Empirical neural/perceptual similarity models
remaining uncertainty: Three-dimensional graphics and a high-dimensional probabilistic geometry are not automatically the same space. The overlap is limited to the state/content bridge.
Limits and next evidence
Three-dimensional graphics and a high-dimensional probabilistic geometry are not automatically the same space. The overlap is limited to the state/content bridge.
At most a qualitative functional, conceptual or causal-prefix relation; no proof of functoriality, isomorphism, bisimulation, exact equation identity or calibrated statistical dependence.
Existing historical IDs for this exact source-atom/target-atom pair: none joined; this is not an automatic addition to governed counts.
SETH-SELECTED-40-40 · Previously independently source-reviewed subset
SAN-AC024-ARTIFICIAL-NEUROLOGY → SETH-T26-05
Both use a neural system’s changing state to transform incoming activity into outputs. The reservoir study keeps internal weights fixed and trains an output readout; SAN also proposes slower connection changes.
SAN: 2022-11-28 · Git repository commit record. Target: 2024-06-27 · arxiv:2406.19201v1.
Current anonymous public access to the exact source body was verified on 2026-10-11. The displayed 2022 date is a Git record date, not independently verified first-publication timing.
Target authors: Hanna M. Tolle · Andrea I. Luppi · Anil K. Seth · Pedro A. M. Mediano. Verify the target passage · See its short quotation in the paired cards.
Exact SAN source windows
SAN lines 19–19
An FPGA can change it's circuit mapping, but it isn't meant to change fast, and it isn't meant to change many times, so it's not the right substrate to mimick the brain. I think however that a GPU could be the right tool for now, given that we want to simulate the physics of both short term fast phase changes and long term slow connection changes. The essential physics process could be simulated in just matrix calculations on a TPU, with each dendrite, each neuron, each synapse, and each cortical column being unique or random in it's original configuration, and in a sense what is needed is a space for an array for the outputs to be rendered in human scale time, so that they can be perceived by a sequence of other arrays, also in human scale spacetime, in a loop of array signalling that lasts for a duration comperable to human experience. Each neuron also specializes in non-linear spatial computation, meaning that where, when, and how the neuron is interacted with helps determine the type of response that it will have, including variations in the frequency & magnitude (duration & amplitude) of it's response, the same sentence I just wrote is applicable to neural arrays, edge communities, cortical columns, and any synchronously oscillating group of cells. The GPU as a graphical process could be the right tool, because we could have the neural network render it's output inside a 3D graphical program, it doesn't have to print to the screen, it could print to active memory, but if it prints 3D graphics we could have a visualization of the artificial mind's development, and interestingly this could lead to brain computer interfaces through virtual reality or augmented reality applications where the human subject is interacting with the artificial intelligence via a head mounted display, or glasses, and the AI is similarly observing the human subject, possibly with non-invasive medical imaging sensors, and possibly with just the sensors that are in existing VR headsets today such as eye tracking, microphones, hand tracking, head tracking, and position tracking sensors. The idea is that as the headmounted AI thinks about you, you see it's thoughts, and then it's thoughts affect your behavior in some small way, and perhaps eventually you two wil eventuallyl be in sync together. You with your partner brain. Your eXtenda brain. (Disclaimer: eXtenda brain is my trademark.)
L1
result: PARTIAL_OPERATION
reason: Both descriptions use recurrent neural dynamics over time with downstream outputs. SAN additionally proposes changing connections and experience; the target fixes reservoir weights and trains a linear readout.
L2
result: PARTIAL_CAUSAL_ARCHITECTURE
reason: The shared input/state/output dependency is real and survives implementation differences. A loop alone does not establish reservoir training, a task loss, or the target's bidirectional performance/emergence result.
L3
result: SUPPORTED_CONSTRAINED_OUTPUT_RECONSTRUCTION_WITH_NONUNIQUE_SOLUTION
reason: Constrained result: Forecast temporal trajectories from the observed input history; the selected paper realizes this goal with a reservoir and learned readout. Required operation: Encode task-relevant input/history in usable state distinctions and map those distinctions to predictions. Source test: SAN line 19 provides time-extended signaling loops, nonlinear spatial-temporal responses and subsequent arrays. It does not specify a fixed reservoir or readout-only optimization. Alternatives: Reservoir recurrence plus trained readout; Other recurrent state models; History windows or delay-line features in nonrecurrent predictors. Remaining uncertainty: SAN's loop is a plausible architectural ingredient, not a derivation of the target equations, optimization or PID emergence measure.
Limits and next evidence
The target cites Jaeger (2001), Maass et al. (2002), Buonomano/Maass (2009) for the architecture and Rosas/Mediano et al. (2020) for emergence. These citations are prior-work boundaries, not a new author-wide award.
At most a qualitative functional, conceptual or causal-prefix relation; no proof of functoriality, isomorphism, bisimulation, exact equation identity or calibrated statistical dependence.
Existing historical IDs for this exact source-atom/target-atom pair: SETH-SANFULL-122.
Motive analysis is deferred by design. No citation-omission, influence, copying, dependence, whole-theory validation or complete-author-program conclusion is certified by these comparisons.