Neural Hardware: Full Domain Analysis

Status: Full 12-step methodology analysis. Neural hardware as the physical realization layer of the cognitive information substrate — the biological computational tissue that cognition runs on. Domain kind (R1): Realization-layer domain. Neural hardware is to cognition what digital computing + physical hardware is to entity system, and what biochemistry + chemistry is to biology. It's the physical structure that realizes the information substrate. Builds on: exploration-cognitive-chain-graph-structure.md (graph structure), exploration-ssa-overlay-three-substrates.md (SSA pattern confirmation), entity_domain_analysis/exploration-cognitive-hardware-parallel.md (initial neural hardware sketch), v1_revision/v1_biology_domain_analysis/bio_v2/exploration-cognitive-information-system-full-analysis.md (cognitive substrate analysis) Parallel to: entity_domain_analysis/analysis-physical-hardware.md (digital hardware domain), entity_domain_analysis/analysis-digital-computing.md (digital computing domain), v1_revision/v1_biology_domain_analysis/bio_v1/biology-domain.md (biology substrate — chemistry realization)

Step 1 — Information Gathering

1.1 What is being analyzed

The physical structure that the cognitive information substrate runs on. Not the cognitive primitives themselves (that's the substrate: {Rp,Ct,As,Sq,Sy,Ev}), not what minds DO (that's cognitive architecture: {Kw,Sk,Dc,Pl,Co,Jd,Cr,Si,Id}), but the BIOLOGICAL COMPUTATIONAL TISSUE that makes cognition physically possible.

This is the same realization question as:

• What physical structure does the entity system run on? → Digital computing on physical hardware
• What physical structure does biology run on? → Biochemistry on chemistry

For cognition: What physical structure does the cognitive substrate run on? → Neural hardware.

Neural hardware is BIOLOGICAL — it's living tissue, produced by organism development, maintained by ongoing biological processes, operated by electrochemistry. This biological character distinguishes it from digital hardware (inert after fabrication) and places it in the same domain family as biology's own substrate chemistry (alive, active, metabolically maintained).

1.2 Relevant literature

• Cellular neuroscience: Kandel, Schwartz & Jessell "Principles of Neural Science" — neuron doctrine, synaptic transmission, ion channel biophysics
• Computational neuroscience: Dayan & Abbott "Theoretical Neuroscience" — neural coding, population dynamics, network models
• Neuroanatomy: Brodmann areas, cortical layers, subcortical nuclei, connectomics (Human Connectome Project)
• Developmental neuroscience: Sanes, Reh & Harris "Development of the Nervous System" — neurogenesis, axon guidance, synaptogenesis, critical periods
• Glial biology: Barres, Fields — astrocyte function, myelination, neurovascular coupling, glial computation
• Comparative neuroscience: Striedter "Principles of Brain Evolution" — nervous system diversity, convergent solutions
• Evolutionary neuroscience: Arendt et al. on neuron type evolution, Moroz on independent origins of nervous systems
• Systems neuroscience: Buzsáki "Rhythms of the Brain" — oscillatory dynamics, temporal coding
• Synaptic plasticity: Hebb, Bliss & Lømo (LTP), Malenka & Bear (LTD), Kandel (molecular mechanisms)

1.3 The key biological constraint

Neural hardware is alive. Every component is a living cell or the product of living cells. This means:

• Every primitive has a biological maintenance requirement (not just an energy requirement)
• The hardware changes during operation (plasticity — learning modifies the hardware)
• The hardware has a developmental trajectory (it's grown, not fabricated)
• The hardware degrades with age (neurodegeneration, synaptic loss)
• The hardware has immune surveillance (microglia, neuroinflammation)

These are not analogies — they are structural features of the domain that any analysis must capture.

Step 2 — Landscape Analysis

2.1 Nervous systems across the tree of life

The landscape of neural hardware spans ~600 million years of evolution, from the simplest nerve nets to the human brain. The key methodological insight: what is INVARIANT across all nervous systems is likely to be a primitive. What varies is a partial level.

2.2 What is invariant across ALL nervous systems

Looking at what EVERY nervous system has, from Hydra to human:

1. Excitable cells — neurons (or neuron-like cells) that generate and propagate electrical signals. Universal across all nervous systems. Even Hydra has them.

2. Connections between cells — synapses (chemical or electrical/gap junctions) that transfer information between neurons. Universal. C. elegans has both types.

3. Energy metabolism — all neural tissue requires metabolic support. Neurons are among the most metabolically demanding cells. Universal.

4. Sensory transduction — converting physical stimuli (light, pressure, chemicals) into neural signals. Universal — even the simplest nerve nets respond to stimuli.

5. Some form of temporal dynamics — rhythmic or patterned activity. Oscillations in vertebrates, central pattern generators in invertebrates, rhythmic swimming in lamprey. Some form of temporal organization appears universal above nerve net level.

6. Some form of spatial organization — from diffuse nerve nets to ganglia to layered cortex. The DEGREE of organization varies enormously, but some spatial structure is always present.

2.3 What varies across nervous systems

2.4 Independent origins

Nervous systems appear to have evolved independently at least 2-3 times:

• Cnidarian nerve nets — different molecular toolkit from bilaterian nervous systems
• Ctenophore nervous systems — Moroz (2014) argues independent origin, different neurotransmitter system (no serotonin, different glutamate receptors)
• Bilaterian nervous systems — the main lineage (insects, mollusks, vertebrates)

If nervous systems evolved independently and converged on similar structures, those shared structures are likely reflecting deep constraints — genuine primitives, not historical accidents.

2.5 Convergent evolution within bilaterians

Even within bilaterians, complex brains evolved independently multiple times:

• Vertebrate neocortex — layered sheet architecture
• Insect mushroom bodies — parallel fiber architecture (similar computational function, different anatomy)
• Cephalopod vertical lobe — another parallel fiber architecture, independently evolved
• Corvid pallium — nuclear (not layered) but functionally equivalent to mammalian neocortex

These convergences suggest that certain computational architectures are ATTRACTORS in neural design space — regardless of the specific anatomical implementation.

Step 3 — Primitive Extraction

3.1 Candidates from the landscape

From the invariants and the convergences, six candidate primitives:

Candidate 1: Excitable Cell (Ec)

The fundamental processing element — a cell that can receive input, integrate it, and generate an output signal (action potential or graded potential).

• Structural minimality: Without excitable cells, no neural computation — remove neurons and there is no nervous system. ✓
• Compositional productivity: Ec combines with connections (signal transfer between cells), with temporal dynamics (rhythmic patterns of excitation), with metabolism (energy to maintain excitability). ✓
• Empirical recurrence: Universal across all nervous systems, including independently evolved ones. Hydra neurons, C. elegans neurons, human cortical pyramidal cells all share: ion channel-based excitability, membrane potential dynamics, signal generation. ✓

Why not "Neuron (Nr)"? "Neuron" implies a specific cell type from the bilaterian lineage. Ctenophore excitable cells may not be homologous to bilaterian neurons. "Excitable cell" captures the STRUCTURAL ROLE — the processing element — regardless of evolutionary origin. But for convenience, we'll use Nr as the abbreviation while understanding it means "the excitable processing element" generally.

Candidate 2: Synapse (Sy)

The connection between processing elements — the site of information transfer from one cell to another. Includes chemical synapses (neurotransmitter release) and electrical synapses (gap junctions).

• Structural minimality: Without synapses, neurons are isolated — no signal transfer, no network, no computation. ✓
• Compositional productivity: Sy combines with Nr (excitable cell + connection = circuit), with plasticity (modifiable connections = learning substrate), with metabolism (synaptic maintenance, vesicle recycling). ✓
• Empirical recurrence: Universal. C. elegans has both chemical and electrical. Hydra has chemical synapses (though simpler). Vertebrates have elaborate chemical synapses with dozens of neurotransmitter types. ✓

Key property: Synapses are PLASTIC — they change strength with use (LTP, LTD, homeostatic plasticity). This is not a separate primitive but a LEVEL of Sy (fixed connections → plastic connections). Plasticity means the hardware CHANGES during operation — the distinguishing feature of neural vs digital hardware.

Candidate 3: Organization (Og)

The spatial/structural arrangement of neurons and synapses into functional units — from nerve nets through ganglia to cortical columns to brain regions.

• Structural minimality: Without organization, neurons and synapses form a random network with no functional architecture. Organization determines WHAT the network computes. ✓
• Compositional productivity: Og combines with Nr (organized neurons = circuits), with Sy (organized connections = pathways), with temporal dynamics (organized oscillations = functional rhythms). ✓
• Empirical recurrence: Universal at some level. Even Hydra's nerve net has preferential pathways. Ganglia organize neurons into functional clusters. Cortical columns organize neurons into computational units. ✓

Why "Organization" not "Circuit"? "Circuit" implies a specific level of organization (the mesoscale). But organization spans from the subcellular (dendritic branching pattern) through the mesoscale (columns, nuclei) to the macroscale (lobes, hemispheres). It's the same primitive at different levels — how processing elements are ARRANGED in space.

Testing against the previous candidate "Circuit (Cr)": The exploration proposed Cr (circuit) — columns, layers, nuclei, ganglia. But there's also subcellular organization (dendritic arbor shape affects computation — a neuron with a wide arbor integrates differently from a narrow one) and macroscale organization (hemispheric specialization, brain region connectivity). These all seem like the SAME concern at different scales: how is the network SPATIALLY ORGANIZED?

If we call it Organization (Og), the partial levels are:

• Og0: no organization (random connectivity)
• Og1: nerve net (distributed, no preferential pathways)
• Og2: ganglionic (clustered, specialized groups)
• Og3: layered (cortical sheets, laminar structure)
• Og4: columnar/modular (functional units within layers)
• Full Og: hierarchical + lateralized (brain regions, hemispheric specialization, long-range connectivity)

Candidate 4: Oscillation (Ol)

Rhythmic patterned activity — the temporal organization of neural hardware operation. Includes neural oscillations (theta, alpha, beta, gamma), central pattern generators, and sleep-wake cycles.

• Structural minimality: Without oscillation/temporal structure, neural activity is random in time — no coordination, no timing, no phase relationships. ✓
• Compositional productivity: Ol combines with Nr (rhythmic firing = temporal codes), with Sy (oscillatory coupling binds distributed representations), with Og (oscillatory patterns organize across brain regions). ✓
• Empirical recurrence: Present in some form in all nervous systems above the simplest nerve nets. Central pattern generators in lamprey. Gamma oscillations in insects. The full spectrum (theta/alpha/beta/gamma) in mammals. ✓

Testing independence from Organization: Could temporal organization be a level of spatial organization? No — you can have well-organized spatial structure with irregular temporal dynamics (some lesion states), and you can have strong oscillatory organization with minimal spatial structure (central pattern generators in simple circuits). Spatial and temporal organization are independent axes.

Testing independence from Excitable Cell: Could oscillation be a property of excitable cells rather than a separate primitive? Individual neurons can oscillate (pacemaker neurons). But network oscillations emerge from POPULATIONS of neurons with specific connectivity — they're not a single-cell property. And oscillatory dynamics have their own partial levels (simple rhythms → multiple frequency bands → cross-frequency coupling → phase-amplitude coordination) that don't map onto cell-level properties. Keep as independent.

Candidate 5: Metabolism (Mb)

The biological maintenance and energy supply system — everything that keeps neural tissue alive and functional. Includes energy delivery (glucose, oxygen via blood supply), waste removal (glymphatic system), protein turnover (synaptic protein synthesis), immune surveillance (microglia), and structural maintenance (astrocyte support, myelination).

• Structural minimality: Without metabolism, neurons die within minutes. No neural hardware operates without metabolic support. ✓
• Compositional productivity: Mb combines with Nr (energy for firing, protein for structure), with Sy (vesicle recycling, receptor turnover, plasticity support), with Ol (metabolic state modulates oscillatory regime — low glucose disrupts oscillations), with Og (neurovascular coupling routes energy to active regions). ✓
• Empirical recurrence: Universal. Every nervous system requires metabolic support. From C. elegans (glial-like sheath cells) to human (astrocytes, oligodendrocytes, microglia, pericytes, endothelial cells). ✓

Why this is MORE than "power supply": In digital hardware, Power (Pw) is simple — constant voltage, stable current. In neural hardware, metabolism:

• Supports computation: astrocytes release gliotransmitters that modulate synaptic transmission
• Enables learning: synaptic plasticity requires new protein synthesis (biological manufacture during operation)
• Routes energy dynamically: neurovascular coupling sends blood to active circuits (not static power)
• Clears waste: glymphatic system (primarily during sleep) removes metabolic byproducts that impair computation
• Maintains structure: oligodendrocytes maintain myelin sheaths, astrocytes maintain ion homeostasis
• Provides immune defense: microglia surveil for damage and infection, prune excess synapses during development

Metabolism IS the biological character of neural hardware. It's what makes this domain biological rather than purely physical.

Candidate 6: Transduction (Td)

The interface between the physical world and neural signals — converting physical stimuli into neural activity (sensory transduction) and neural activity into physical effects (motor output).

• Structural minimality: Without transduction, the nervous system is disconnected from the physical world — no input, no output, no function. ✓
• Compositional productivity: Td combines with Nr (transduction triggers excitation), with Og (sensory maps organize transduction), with Mb (transduction requires specialized metabolic support — photoreceptor metabolism is among the highest in the body). ✓
• Empirical recurrence: Universal. Every nervous system interfaces with the physical world. Hydra's touch-sensitive cells. C. elegans' chemoreceptors. Vertebrate photoreceptors, mechanoreceptors, chemoreceptors. ✓

Why not "Receptor (Rc)"? "Receptor" implies only the sensory side. Neural hardware also interfaces with the physical world through MOTOR output — neuromuscular junctions, glandular secretion. Transduction is bidirectional: world → neural (sensory) and neural → world (motor). The STRUCTURAL ROLE is the same: converting between physical and neural signal domains.

Testing against a broader framing: Is transduction just a special case of synapse? A neuromuscular junction IS a synapse — between a motor neuron and a muscle fiber. A sensory receptor IS a specialized neuron. Could transduction be absorbed into Sy (for the junction) and Nr (for the specialized cell)?

Counter-argument: transduction involves a MEDIUM CHANGE — from non-neural (photons, pressure, chemicals) to neural (action potentials), or from neural to non-neural (muscle contraction, glandular secretion). Synapses transfer within the neural medium. Transduction crosses medium boundaries. This medium-crossing function is structurally distinct.

Keep as independent. Transduction is the neural hardware's PORT — its interface to the physical world.

3.2 Six primitives confirmed

Changes from the initial sketch {Nr, Sy, Cr?, Ol, Mb, Rc}:

• Cr (Circuit) → Og (Organization): broadened to capture all scales of spatial arrangement, not just mesoscale
• Rc (Receptor) → Td (Transduction): broadened to capture bidirectional physical ↔ neural interface, not just sensory input

The fused storage/processing property of Sy (synapses are BOTH connections AND memory — through plasticity) is preserved. This is the structural feature that distinguishes neural hardware from digital hardware, where Ic (interconnect) and St (storage) are separate primitives.

Step 3b — Partial Levels

Excitable Cell (Nr)

Phase transition: Nr2→Nr3. Complex integration. Below: neurons are simple switches (fire or don't fire based on input sum). Above: neurons are COMPUTATIONAL ELEMENTS — dendritic branches compute separately, different input zones interact, the cell's morphology shapes its computation. This is where the single neuron becomes more than a point processor. The transition corresponds to the emergence of complex nervous systems in protostomes (insects, mollusks).

Synapse (Sy)

Phase transition: Sy2���Sy3. Plasticity. Below: connections are fixed or minimally adjustable (C. elegans has ~90% stereotyped wiring). Above: connections change strength based on activity — the hardware LEARNS. This is the transition from fixed circuit to adaptive circuit. It's the neural equivalent of the difference between ROM and RAM, except the neural "RAM" is non-volatile (learned changes persist without power).

Phase transition: Sy4→Full Sy. Tripartite integration. Below: the synapse is a two-party affair (presynaptic neuron → postsynaptic neuron). Above: the synapse is a THREE-PARTY system (neuron → astrocyte → neuron), where astrocytes actively modulate transmission, release gliotransmitters, and participate in plasticity. This is where the biological character of neural hardware becomes irreducible — the synapse isn't just an electrical junction; it's a living biological micro-environment.

Organization (Og)

Og levels describe COMPUTATIONAL ORGANIZATION CAPABILITY — the functional capacity for structured processing — not specific anatomical implementation. Different anatomies (laminar cortex, nuclear pallium, distributed ganglia) can achieve the same functional level through different physical arrangements. This is the same principle as digital hardware, where Sw (switch) describes functional capability regardless of whether it's CMOS, FinFET, or GAA transistor.

Phase transition: Og2→Og3. Centralized integration. Below: processing is distributed across modules — each handles local concerns. Above: a CENTRAL structure integrates information from multiple sources and coordinates global behavior. This is the transition from distributed to centralized processing, corresponding to the emergence of brains in bilaterians.

Phase transition: Og4→Full Og. Asymmetric specialization. Below: the brain's hemispheres are functionally symmetric. Above: the hemispheres SPECIALIZE differently — in humans, left for language/sequential, right for spatial/holistic. This asymmetry maximizes utilization by assigning different specializations rather than maintaining redundant copies.

Convergent implementations at Og4: Mammalian neocortex (6-layer laminar sheet) and corvid pallium (dense nuclear clusters) are DIFFERENT ANATOMIES achieving the same Og4 computational function — deep hierarchical feature extraction with functional modularity. The convergence confirms that Og4 describes a computational ATTRACTOR, not a specific anatomical plan. Corvids pack ~1.5 billion neurons into dense nuclear arrangements, achieving neuron-per-function ratios comparable to mammals with much larger brains.

Note on octopus: The octopus nervous system (~500M neurons, 2/3 in arms) presents a distributed-with-integration architecture. Central brain provides Og3 (centralized integration) while arm ganglia add distributed local autonomy. This is a hybrid — Og3 centrally with Og2 extensions. The methodology captures this as Og3 overall with a distinctive distributed character, rather than requiring a new level.

Oscillation (Ol)

Phase transition: Ol2→Ol3. Cross-regional coordination. Below: oscillations serve local functions (motor rhythms, local circuit timing). Above: oscillations BIND distant brain regions into functional assemblies — a gamma burst in visual cortex phase-locked to hippocampal theta during memory encoding. This is where oscillation becomes a COORDINATION mechanism for distributed processing, not just a local timing device.

Metabolism (Mb)

Phase transition: Mb3→Mb4. Neurovascular coupling. Below: blood supply is relatively uniform — the whole brain gets similar energy delivery. Above: the vascular system is COMPUTATIONALLY COUPLED — active brain regions get more blood flow within seconds. This makes metabolism an active participant in neural computation, not just background infrastructure. The BOLD signal in fMRI IS this coupling.

The biological character emerges at Mb3+. Below Mb3, metabolic support is relatively simple and could be modeled as just "power supply." At Mb3 and above, the metabolic system becomes an ACTIVE computational participant — astrocytes modulate synapses, microglia prune connections, myelination changes signal speed, neurovascular coupling routes energy. This is where neural hardware becomes irreducibly biological.

Transduction (Td)

Phase transition: Td3→Td4. Mapped transduction. Below: sensory information arrives as unstructured neural signals. Above: sensory information is TOPOGRAPHICALLY MAPPED — adjacent receptors project to adjacent neurons, preserving spatial structure. This is where the nervous system begins to build an internal model of external spatial structure through the transduction layer itself.

Step 4 — Dependencies

4.1 Primitive-presence dependencies

Nr → (nothing; foundation — excitable cells are the base element)
Sy → Nr (synapses connect excitable cells — no cells, nothing to connect)
Og → Nr + Sy (organization requires cells AND connections to organize)
Ol → Nr + Sy (oscillation requires cells that fire AND connections that transmit)
Mb → Nr (metabolism supports excitable cells — but Nr minimally functions without Mb at Mb1)
Td → Nr (transduction is a specialized function of excitable cells / at cell interfaces)

DAG:

Nr (hub)
  ├── Sy → Og
  ├── Ol (requires Nr + Sy)
  ├── Mb
  └── Td

Nr is the hub — everything depends on excitable cells. Sy is the secondary hub — Og and Ol both require connections.

4.2 Partial-level dependencies

More interesting than presence dependencies:

• Og3 requires Sy2+ — centralized brains need directed chemical synapses, not just gap junctions. You can't build a hierarchical brain on bidirectional electrical coupling alone.
• Og4 requires Nr4+ — hierarchical modular organization requires specialized cell types (whether laminar pyramidal cells or nuclear projection neurons). Deep hierarchical processing needs functionally distinct cell populations.
• Ol3 requires Og3+ — cross-regional oscillatory coupling requires distinct brain regions to couple between. No regions, no cross-regional coupling.
• Ol4 requires Sy3+ — cross-frequency coupling involves plasticity-dependent mechanisms (STDP interacts with oscillatory phase). Fixed synapses can't support cross-frequency modulation.
• Mb4 requires Og3+ — neurovascular coupling requires a vascularized centralized brain. Ganglionic nervous systems don't have the blood-brain barrier or astrocytic networks needed.
• Full Sy requires Mb3+ — tripartite synapses require astrocytes, which require vascular metabolic support. The most sophisticated synapse type IS biological.
• Full Td requires Og4+ — topographic sensory maps and fine motor control require hierarchical modular organization.

4.3 The dependency structure tells a build-up story

Minimal nervous system: Nr1 + Sy1 + Mb1 + Td1 + Og1 + Ol0
  = Hydra-like nerve net. Excitable cells with gap junctions, basic metabolism, simple touch sensing, no organization, no temporal structure.

Ganglionic: Nr2 + Sy2 + Mb2 + Td2 + Og2 + Ol1
  = C. elegans / Drosophila. Differentiated cell types, chemical synapses, dedicated glia, multiple senses, ganglia, simple rhythms.

Vertebrate brain: Nr3-4 + Sy3 + Mb3 + Td3-4 + Og3-4 + Ol3
  = Rat / cat. Complex neurons, plastic synapses, vascular support, sensory maps, cortical layers, coordinated oscillations.

Human brain: Full Nr + Full Sy + Full Mb + Full Td + Full Og + Full Ol
  = Maximum observed elaboration on all axes.

Step 5 — Pair Enumeration

C(6,2) = 15 pairs.

Step 6 — Load Classification

Heavy pairs

Nr-Sy (Excitable Cell - Synapse): HEAVY.

• The foundational pair of all neuroscience. The neuron doctrine + synaptic transmission = the basis of neural communication.
• Most of electrophysiology, synaptic pharmacology, and connectomics lives here.
• Failure modes: excitotoxicity, synaptic depletion, channelopathies.
• Emergent: neural circuits, signal propagation, basic learning.

Sy-Og (Synapse - Organization): HEAVY.

• Connectivity architecture. The connectome. Wiring diagrams.
• Developmental neuroscience (axon guidance, synaptogenesis, pruning) = how Sy+Og are built.
• Failure modes: miswiring, connectivity disorders (autism spectrum involves altered Sy-Og).
• Emergent: circuit motifs (feedforward inhibition, recurrent excitation, lateral inhibition), hierarchical processing.

Sy-Mb (Synapse - Metabolism): HEAVY.

• THE biological pair. Where neural hardware is irreducibly biological.
• Tripartite synapse (neuron-astrocyte-neuron). Astrocyte modulation. Synaptic protein synthesis.
• Failure modes: neurodegeneration (synapse loss from metabolic failure), neuroinflammation.
• Emergent: activity-dependent maintenance, sleep-dependent consolidation, learning as biological process.

Og-Ol (Organization - Oscillation): HEAVY.

• Functional connectivity. Oscillatory binding. Default mode network.
• Which regions oscillate together determines what the brain is currently computing.
• Failure modes: epilepsy (pathological Og-Ol — organization drives runaway oscillation), schizophrenia (disrupted Og-Ol coupling).
• Emergent: brain states (attention, rest, sleep stages), consciousness correlates.

Medium pairs

Nr-Og: MEDIUM. Cell types constrain organization (pyramidal cells in layers 2/3/5, interneurons distributed) but the relationship is mediated through Sy.

Nr-Ol: MEDIUM. Individual neurons participate in oscillations but the oscillatory dynamics are a network-level phenomenon.

Nr-Mb: MEDIUM. Every neuron requires metabolic support, but the relationship is largely "support" rather than "interaction."

Sy-Ol: MEDIUM. STDP depends on oscillatory phase, but this is a specific mechanism rather than a broad structural relationship.

Og-Mb: MEDIUM. Neurovascular coupling is significant but primarily enables rather than shapes computation.

Og-Td: MEDIUM. Sensory maps are important but are a specific application of organization, not a general relationship.

Light pairs

Nr-Td: LIGHT. Sensory neurons are specialized Nr, but the transduction function is largely independent of general Nr properties.

Sy-Td: LIGHT. Sensorimotor synapses exist but are not structurally distinct from other synapses.

Ol-Mb: LIGHT. Sleep-wake affects oscillations via metabolic state, but indirect.

Ol-Td: LIGHT. Oscillatory sampling of sensory input exists but is a specific mechanism.

Mb-Td: LIGHT. Sensory transduction has high metabolic demand but the relationship is primarily energetic.

Summary

4 heavy of 15 = 26.7%.

Step 7 — Coherent Sub-lattice

7.1 Dependencies

Nr → nothing (hub)
Sy → Nr
Og → Nr, Sy
Ol → Nr, Sy
Mb → Nr
Td → Nr

7.2 Counting valid subsets

Total: 2^6 = 64.

Invalid subsets — those violating a dependency:

• Sy without Nr: subsets containing Sy but not Nr. Sy fixed in, Nr fixed out, 4 others vary: 2^4 = 16
• Og without Nr: contains Og, not Nr. But Og also requires Sy. So Og without Nr is already invalid via Og→Nr. Counted separately: Og in, Nr out, 4 others vary = 16. But some overlap with Sy-without-Nr. Og in + Sy in + Nr out = 2^3 = 8. These are already counted. Og in + Sy out + Nr out = invalid for two reasons (Og→Nr and Og→Sy). 2^3 = 8 new.
• Og without Sy: Og in, Sy out. Nr can be in or out. 2^3 × 2 = ... Let me use inclusion-exclusion.

Let me enumerate dependency violations:

Required implications:

• Sy → Nr
• Og → Nr AND Og → Sy
• Ol → Nr AND Ol → Sy
• Mb → Nr
• Td → Nr

Valid means: ALL of the following hold:

1. If Sy ∈ S then Nr ∈ S
2. If Og ∈ S then Nr ∈ S AND Sy ∈ S
3. If Ol ∈ S then Nr ∈ S AND Sy ∈ S
4. If Mb ∈ S then Nr ∈ S
5. If Td ∈ S then Nr ∈ S

Equivalently: valid if Nr ∈ S whenever any of {Sy, Og, Ol, Mb, Td} ∈ S, AND Sy ∈ S whenever {Og, Ol} ∈ S.

Case 1: Nr ∉ S. Then none of {Sy, Og, Ol, Mb, Td} can be in S. Only valid subset: {}. Count: 1.

Case 2: Nr �� S. Now {Sy, Og, Ol, Mb, Td} can be present subject to: Sy must be in if Og or Ol is in.

Sub-case 2a: Sy ∉ S. Then Og and Ol must be out. Mb and Td are free. Count: 2^2 = 4. (Subsets of {Mb, Td}, all with Nr present, Sy/Og/Ol absent.)

Sub-case 2b: Sy ∈ S. Then Og, Ol, Mb, Td are all free. Count: 2^4 = 16. (Subsets of {Og, Ol, Mb, Td}, all with Nr + Sy present.)

Total valid: 1 + 4 + 16 = 21 of 64.

Filter: 21/64 = 32.8%.

7.3 Comparison

Neural hardware at 32.8% is close to digital hardware at 34.4%. Both hardware layers are relatively LOOSE — lots of valid partial configurations. This makes sense: hardware is the most physically grounded layer, and physical constraints create many viable configurations (simple nerve nets through complex brains, simple circuits through complex chips).

Both hardware domains are much looser than their respective information substrates (biology 12.5%, entity system 14%, cognitive substrate 26.6%). Information substrates are more tightly constrained than the hardware they run on.

Step 8 — Hasse Diagram Walks

8.1 The canonical build-up (evolutionary path)

{} → {Nr} → {Nr,Sy} → {Nr,Sy,Mb} → {Nr,Sy,Mb,Td} → {Nr,Sy,Mb,Td,Og} → {Nr,Sy,Mb,Td,Og,Ol}

Step 0→1: Excitable cells emerge. Cells develop ion channels and electrical excitability. No connections yet — isolated excitable cells.

Step 1→2: Synapses form. Excitable cells develop connections — gap junctions first, then chemical synapses. Now there's a NETWORK.

Step 2→3: Metabolic support develops. Dedicated support cells (glia) emerge to maintain the neural network. The network becomes biologically sustainable.

Step 3→4: Transduction specializes. Some cells at the body surface specialize for detecting physical stimuli (light, pressure, chemicals) and some for motor output. The network connects to the physical world.

Step 4→5: Organization emerges. The network develops spatial structure — ganglia, regions, layers. From diffuse to organized.

Step 5→6: Oscillatory dynamics develop. The organized network develops temporal patterning — rhythmic activity that coordinates distributed processing.

8.2 The evolutionary record match

This build-up matches the actual evolutionary record remarkably well:

The model PREDICTS the evolutionary order. This is the same finding as in the entity domain analysis (the Hasse walk predicts the historical development order for digital computing).

8.3 Alternative walk: the clinical degradation path

Neural hardware can also be walked DOWNWARD — what degrades in neurological disease:

{Nr,Sy,Mb,Td,Og,Ol} → lose Ol first (anesthesia: oscillations disrupted, consciousness lost)
  → lose Og (brain damage: regional organization disrupted)
    → lose Td (sensory loss, motor paralysis)
      → lose Mb (metabolic failure: ischemia, cell death begins)
        → lose Sy (synaptic failure: connections lost)
          → lose Nr (cell death: neurons die)
            → {}

The degradation path is roughly the REVERSE of the build-up. The most recently evolved capabilities (oscillatory coordination, organizational complexity) are the most fragile. The most ancient (basic excitability, basic connections) are the most robust. This matches the neurological principle of "evolution in reverse" — neurodegeneration often recapitulates evolution backwards.

Step 9 — Load-Bearing Compositions

9.1 Core triad identification

From the 4 heavy pairs: Nr-Sy, Sy-Og, Sy-Mb, Og-Ol.

Candidate core triad: {Nr, Sy, Og} — Cell + Connection + Organization.

• Nr-Sy: HEAVY ✓
• Sy-Og: HEAVY ✓
• Nr-Og: Medium (not heavy)

Nr-Og is medium, not heavy. This means {Nr,Sy,Og} is a load-bearing triangle but not technically a core triad (requires all three pairs heavy).

Candidate: {Sy, Og, Ol} — Connection + Organization + Oscillation.

• Sy-Og: HEAVY ✓
• Og-Ol: HEAVY ✓
• Sy-Ol: Medium

Same issue — Sy-Ol is medium.

Candidate: {Nr, Sy, Mb} — Cell + Connection + Metabolism.

• Nr-Sy: HEAVY ✓
• Sy-Mb: HEAVY ✓
• Nr-Mb: Medium

Again, one medium pair.

9.2 Assessment

No clean core triad with all three pairs heavy. The heavy pairs form a PATH rather than a triangle:

Nr-Sy — Sy-Og — Og-Ol
            |
         Sy-Mb

Sy (Synapse) is the HUB of the heavy pairs — appearing in 3 of 4 heavy pairs. This makes Sy the most load-bearing primitive, consistent with neuroscience's emphasis on synaptic function as the central concern.

The load-bearing spine is {Nr, Sy, Og, Ol} — Cell + Connection + Organization + Oscillation — with Mb as the biological support that makes it all possible.

9.3 What this means

The absence of a clean core triad (with all 3 pairs heavy) but the presence of a heavy spine (Nr-Sy-Og-Ol + Sy-Mb) suggests that neural hardware is LINEARLY organized rather than triangularly organized. The structural content flows along a spine from individual cells through connections through spatial organization to temporal dynamics, with biological metabolism supporting the whole chain.

This is structurally different from biology {G,T,R core triad} and entity system {E,I,T core triad}, where the core triad forms a tight triangle. Neural hardware's spine structure reflects its nature as a REALIZATION LAYER — it provides the physical platform, and the tight triangular structure appears in the information substrate it supports (cognitive substrate {Rp,Ct,Sy core triad}).

9.4 Activation mapping and reconciliation

Activation mapping (data/domains/neural-hardware.v1.json, all discriminating; one-home-per-construct): the three JSON-flagged phase transitions each carry a single-driver partial_level.emergent — Nr4 (dendritic computation; §3b Nr PT), Sy2 (synaptic plasticity — the structural divergence from digital hardware, §3b Sy PT + §10.2 + §11.1), Og4 (hierarchical/laminar architecture, a convergent attractor; §3b Og PT + §2.5). The §10.2 Og2→Og3 centralization → behavioral-flexibility prediction maps to an Og.emergent_phases band [3,3] (non-flagged within-primitive regime; only Og4 is a JSON phase_transition → band, not a fabricated PT flag). The §9.1 load-bearing triangle {Nr,Sy,Og} carries a composition.emergent (presence τ — the file's existing emergent_property, the foundational neural-computation unit; §9.1 notes it is not a clean core triad since Nr-Og is medium, but it is the analyst's named computational-substrate unit). A new {Sy,Og,Ol} triad composition is added to host §10.2's explicit "triad for cognitive computation" (Step-10-requires-construct add — §10.2 predicts the emergent but the §9 compositions discussion, which finds no clean core triad, listed no such triad; conjunction τ Sy:2,Og:4,Ol:3, where §10.2's "Sy3-plastic" maps to JSON-Sy2-flagged under the v1-compressed Sy levels). This triad also folds in §10.2's standalone Ol2→Ol3 cross-regional-binding prediction (one-home: in the v1-compressed Ol levels, Ol is load-bearing precisely as part of the cognitive-computation triad — documented here, not separately declared, per the chemistry-Supramolecular / organism-arch-Cognition fold-in precedent). A full-6 {Nr,Sy,Og,Ol,Mb,Td} higher composition is added per the per-domain template (presence τ — §8.1 build-up endpoint + §10.1 "Full → symbolic cognition" row; no explicit analyst "at all Full"). Non-over-flag: the heavy pair Sy-Mb ("THE biological pair") and all medium/light pairs get NO standalone emergent — §9.2/§9.3 explicitly establish there is no clean core triad and the heavy pairs form a spine, not flagged triangles; §10.2 gives Sy-Mb no emergent prediction; Mb's domain-wide load-bearing biological-support role is captured in the full-6 composition (parallel to the chemistry firm-triangles non-over-flag).

Reconciliation (structure-fix): the JSON dependencies array was missing the Ol ⇒ Sy presence edge that this analysis models consistently in §3.1 Candidate 4, §4.1, §4.3, §7.1 and §7.2 (network oscillation is a population phenomenon requiring synaptic connectivity, not a single-cell property). Added BFS over the corrected presence-dependency set confirms the coherent sub-lattice = 21/64 = 32.8%, matching this analysis's §7.2 derivation and the already-correct stored filter_stringency (the JSON's value was right; only its dependency model was incomplete — JSON-as-was BFS'd to 25/64). filter_stringency is unverified descriptive metadata not enforced by validate/check_bounds, so the omission had silently persisted; per the standing filter-discipline rule every file's presence-dependency set is independently BFS-recomputed during the per-file loop.

Step 10 — Emergent Property Predictions

10.1 Properties at specific partial-level regimes

10.2 Phase-transition-driven predictions

Sy2→Sy3 (plasticity onset): Prediction: Systems below Sy3 can learn only through developmental wiring changes (slow, generational). Systems at Sy3+ can learn within a single lifetime (fast, individual). This is testable by comparing learning timescales across organisms with different plasticity levels.

Og2→Og3 (centralization): Prediction: Centralized brains enable behavioral FLEXIBILITY that ganglionic systems can't match — novel responses to novel situations rather than fixed behavioral routines. This is testable by comparing behavioral repertoire complexity in ganglionic vs centralized nervous systems.

Ol2→Ol3 (cross-regional binding): Prediction: Cross-regional oscillatory coupling enables UNIFIED PERCEPTION — binding distributed features (color, shape, motion) into coherent objects. Below Ol3, perception is modular (separate feature channels). This is testable and IS tested — oscillatory binding hypothesis (Singer, Gray).

Sy3 + Og4 + Ol3 (the triad for cognitive computation): Prediction: ALL THREE are required for abstract representation and flexible cognition. Plastic synapses (Sy3) + laminar organization (Og4) + cross-regional oscillations (Ol3) together enable the cortical computation that supports cognitive substrate primitives. This predicts that animals with all three (mammals, possibly corvids with functional equivalent) are capable of abstract thought, while those missing any one are limited to concrete, stimulus-bound behavior.

Step 11 — Structural Pattern Observations

11.1 Parallel to digital hardware

Key structural difference: Digital hardware separates Interconnect (Ic) from Storage (St). Neural hardware FUSES them in the Synapse (Sy) — connections ARE memory (through plasticity). This fusion eliminates the von Neumann bottleneck (no need to move data between separate storage and processing) but makes the hardware's computational function inseparable from its physical state.

11.2 The heavy pair pattern

Neural hardware's heavy pairs form a spine: Nr-Sy — Sy-Og — Og-Ol + Sy-Mb.

Digital hardware's heavy pairs form a more triangular structure. The difference: neural hardware is a GROWTH structure (added incrementally through evolution and development) while digital hardware is a DESIGN structure (planned as a coherent whole). Growth produces spines; design produces triangles.

11.3 The biological character as a domain-level feature

Mb (Metabolism) appears in only 1 heavy pair (Sy-Mb) but is LOAD-BEARING for the entire domain — remove it and everything dies within minutes. Its heavy pairing with Sy specifically (not with Nr or Og) reflects the fact that the SYNAPSE is where biological maintenance is most computationally relevant: synaptic plasticity requires protein synthesis, neurotransmitter recycling, astrocyte modulation.

This means: the biological character of neural hardware is concentrated at the SYNAPSE. Neurons fire through electrochemistry (relatively physics-like). Oscillations are electromagnetic (physics-like). But synapses LEARN through BIOLOGY — protein synthesis, receptor trafficking, structural remodeling. The synapse is where physics meets biology in neural hardware.

Step 12 — Literature Alignment and Cross-Domain Mapping

12.1 Literature alignment

The primitive set {Nr, Sy, Og, Ol, Mb, Td} aligns with major subfields of neuroscience:

Each primitive maps to a recognized subfield. No primitive spans multiple subfields confusingly. No major subfield is left unmapped.

12.2 Cross-domain mapping to cognitive substrate

The bridge from neural hardware to cognitive substrate:

The mapping is many-to-many but structured. Nr and Og primarily support Rp. Sy supports Ct and As. Ol supports Sq. Td provides input to Rp and output from Sy. Mb enables everything.

12.3 Cross-domain mapping to digital hardware

Summary

Primitives

Key structural features

• Filter: 32.8% (21/64) — relatively loose, comparable to digital hardware (34.4%)
• Hub: Nr (excitable cell) — everything depends on it
• Load-bearing spine: Nr-Sy-Og-Ol (with Sy-Mb as biological anchor)
• No clean core triad — heavy pairs form a spine rather than a triangle. Sy is the most connected heavy node (3 of 4 heavy pairs).
• Build-up matches evolution — the Hasse walk predicts the actual evolutionary order of nervous system elaboration
• Degradation matches clinical neurology — the reverse walk matches neurological disease progression

Distinguishing features from digital hardware

1. Sy fuses connection + storage. No von Neumann bottleneck but no hardware/software separation.
2. Mb is biologically active. Not just power supply — active participant in computation through astrocytes, myelination, neurovascular coupling.
3. The hardware LEARNS. Synaptic plasticity means the hardware changes during operation. Digital hardware is (mostly) fixed.
4. Organization is grown, not designed. Neural architecture develops through biological processes, not engineering. Partial levels track evolutionary elaboration.
5. Oscillations are multiple and imprecise. Not a single precise clock but a landscape of interacting rhythms at different frequencies.

Referenced by the model

Cited as a source by 1 model record (browse the model census):

• neural-hardware — domain cognition/sc1