From ab47b0621aaf05250b44696a2f075cf237d903bd Mon Sep 17 00:00:00 2001 From: ocrampal Date: Wed, 10 Jun 2026 11:06:12 +0200 Subject: [PATCH] Create 2026-06-06-tripartite_synapse_v4.md --- .../2026-06-06-tripartite_synapse_v4.md | 793 ++++++++++++++++++ 1 file changed, 793 insertions(+) create mode 100644 elements/neuron/appunti/2026-06-06-tripartite_synapse_v4.md diff --git a/elements/neuron/appunti/2026-06-06-tripartite_synapse_v4.md b/elements/neuron/appunti/2026-06-06-tripartite_synapse_v4.md new file mode 100644 index 0000000..433c584 --- /dev/null +++ b/elements/neuron/appunti/2026-06-06-tripartite_synapse_v4.md @@ -0,0 +1,793 @@ +--- +include_toc: true +--- + +# Neuron, budget, traces and behaviours + +--- + +## Conventions: The Language of the System + +The system operates across two scopes and multiple contexts. The **DAY scope** is the organism's active period — all fast and intermediate behaviors run here, structural variables are read but never permanently written, and every action leaves a trace that may or may not survive to the NIGHT. The **NIGHT scope** is when structural commits execute, budgets replenish, and all DAY traces are either converted into permanent architecture or cleared. Within each scope, each compartment operates in a specific **context** — AP or NOT_AP for the presynapse, bAP or NOT_bAP for the postsynapse, and CONTINUOUS for the astrocyte and dendritic branch — that determines which behaviors are permitted to run. + +Every variable in the system is either a **state variable**, a **trace**, or a **budget**. State variables describe the current condition of a compartment. Traces are modifications deposited by a behavior that persist beyond the moment and bias future behaviors — they are the system's memory between contexts. Budgets are shared resource pools that constrain how much of any behavior can actually execute, and they ensure that nothing in the system is free. + +--- + +## Shared Resource Pools: The Economy Before Any Behavior Runs + +Before any spike fires or any receptor moves, the system's capacity for change is already determined by the state of its shared resource pools. These pools exist at four nested territorial levels, and they are the reason that what happens at one synapse has consequences for all its neighbors. + +The axonal arbor maintains a shared pool of vesicle scaffold proteins — RIM, Munc13, and VGCC subunits — that every bouton on that axon must draw from when it wants to expand its active zone. It also maintains a mitochondrial ATP capacity that sets the ceiling on how many release events can occur before the energy supply is exhausted and boutons begin going silent. A set of tagged boutons tracks which sites are currently in competition for proteins drifting along the axon. + +The dendritic branch maintains its own receptor reserve — a local endosomal pool of AMPA receptors held near the spines that can be rapidly inserted without waiting for the soma — alongside actin machinery for spine enlargement, a protein flux arriving from the soma, a local pool of stored mRNAs that branch ribosomes can translate immediately, and its own mitochondrial ATP budget. The set of tagged spines on the branch tracks which sites are competing for these local resources. + +The astrocyte territory — which in humans wraps hundreds of thousands of synapses — maintains pools of serine racemase enzyme for D-serine synthesis, glutamate transporter proteins for clearance, extracellular matrix proteins for structural sealing, and a finite number of perisynaptic process extensions that it can selectively allocate to different synapses. All of these draw from a single ATP budget that is itself ultimately capped by the glucose supply from the nearest blood capillary — a hard vascular ceiling that cannot be exceeded regardless of demand. + +The soma maintains its own pools of synthesis capacity, receptor production rate, organelles for shipping to branches, and mRNA transcription rate. These are the upstream sources that replenish the branch-level pools, and their rate is set by the CREB transcription state left over from DAY activity — meaning the soma's production capacity during NIGHT is determined by how the DAY went. + +--- + +## Budget Allocation: Constraint Before Commitment + +When any compartment wants to draw structural resources, it does not simply take them. It calls a budget allocation function that checks what is available, applies a priority weighting based on whether that site has planted a synaptic tag, grants only what can be afforded, and decrements the pool accordingly. If the grant is smaller than the request — because the pool is partially depleted by competing synapses — the deficit is queued for the next NIGHT cycle and heterosynaptic depression is triggered in untagged neighbors. The neighbors lose resources not because any signal told them to weaken, but because the pool was exhausted before reaching them. + +The astrocyte allocation has an additional layer: all three of its resource draws — D-serine, ECM proteins, and process extensions — are individually capped by their specific pools, and then the combined ATP cost of all three is checked against the total ATP budget. If the ATP cost would exceed the budget, all three grants are scaled down proportionally. This means a heavily loaded astrocyte serving many simultaneously active synapses delivers partial support to each rather than full support to a few. + +Budget replenishment runs continuously but peaks during the NIGHT scope. The soma's protein synthesis rate, driven by CREB activation accumulated during DAY, determines how fast the dendritic branch pools refill. The astrocyte's pools recover at rates set by enzyme synthesis and ECM production machinery, with process extensions recovering the slowest — on a timescale of hours — which is why a fully deployed astrocyte cannot immediately reallocate its walls to a new synapse. The vascular glucose ceiling is the one parameter in the entire system that cannot be increased by any molecular mechanism — it is set by blood flow and capillary density, and it sets the absolute limit on how much simultaneous structural change the system can support. + +--- + +## DAY Scope — Presynaptic Bouton + +### Context: AP + +When an action potential arrives, the first check is not biological but economic: does this bouton's local mitochondrial capacity have enough ATP to afford the release? If not, the bouton goes silent regardless of the electrical input. This is the energy gate that prevents a depleted axon from continuing to operate at full output — it is not a regulatory decision but a resource constraint. + +If the budget permits, calcium floods in through voltage-gated channels. The key quantity is not the peak calcium but the residual: the calcium left over from previous spikes that has not yet decayed. This residual is the presynaptic memory of recent activity — a trace that accumulates with repeated firing and fades with silence on a timescale of roughly 100 milliseconds. The current release probability is biased by this residual, meaning a bouton that has been firing recently is primed to release more than one that has been silent. Vesicles are released probabilistically from the readily-releasable pool, glutamate fills the cleft, and the pool shrinks by exactly the number released. The mitochondrial budget decrements. + +If enough glutamate escapes the cleft to reach the low-affinity mGluR2/3 receptors on the presynaptic membrane, an autoinhibitory brake engages. The Gi-coupled cascade suppresses adenylyl cyclase, reducing cAMP and directly dampening voltage-gated calcium channel opening. This is a negative feedback loop entirely within the presynaptic compartment: excess output detected, output probability reduced, without any postsynaptic input. The RRP is then partially refilled by the astrocyte's lactate supply, but only to the extent that lactate is available — another resource gate. + +### Context: NOT_AP + +Between spikes, the presynapse is not idle. The residual calcium trace continues to decay passively — if spikes do not keep arriving, the facilitation advantage erodes. If recent firing history has been consistently above 20 Hz, the reserve pool is mobilized toward the readily-releasable pool, making more vesicles available for the next burst — but only if the reserve is non-empty, meaning this mobilization is itself budgeted by whatever was stored during prior quiescence. If firing history has been sparse, release probability drifts downward. + +The more important event in the NOT_AP context is the neuromodulatory broadcast. When dopamine or norepinephrine arrives above threshold, PKA activity rises and a synaptic tag is planted — the bouton is added to the set of axon-tagged boutons competing for drifting structural proteins. This tag is a trace of the neuromodulatory context that will persist into the NIGHT scope: if a structural commit runs before the tag expires, this bouton gets priority access to the axonal protein pool. + +--- + +## DAY Scope — Dendritic Spine + +### Context: NOT_bAP + +The NOT_bAP context is where the Hebbian anticipation window opens. Glutamate arriving from the presynapse activates AMPA receptors, depolarizing the spine membrane. If this depolarization is sufficient — and crucially if D-serine supplied by the astrocyte is present as a co-agonist — the NMDA receptor's magnesium block is ejected and calcium begins to enter. The amplitude and speed of this calcium rise are recorded as traces: amplitude encodes whether the event was strong enough for LTP, speed distinguishes LTP-driving fast rises from LTD-driving slow ones. + +The synaptic tag is planted here, in this context, before the back-propagating AP arrives to confirm coincidence. The trigger for tag planting is the calcium rise combined with NMDA opening — the system is tagging the moment it detects an incoming signal strong enough to potentially be worth encoding, before it knows whether the postsynaptic cell actually fired. This is the Hebbian anticipation: the forward glutamate signal is the candidate event, and the tag is the molecular record of that candidacy. The spine is simultaneously added to the dendritic branch's pool of tagged spines, entering competition for local protein resources. If the neuromodulator context gate has already been set by PKA, the AMPA insertion threshold is lowered, making the eventual structural commit easier. + +### Context: bAP + +When the back-propagating action potential arrives from the soma down the dendritic tree, it depolarizes the spine further. This is the confirmation signal. If a tag was planted just before the bAP — meaning the forward glutamate signal and the retrograde AP signal coincide in time — the calcium rise is amplified beyond the LTP threshold, confirming genuine Hebbian coincidence. The bAP did not cause the calcium rise; it amplified one that was already in progress, retroactively validating the tag that was planted in anticipation. + +If no tag was present — if the bAP arrives at a spine that received no glutamate signal — the membrane depolarizes transiently but no calcium amplification occurs and no trace is deposited. The spine passes through the bAP context without consequence. This asymmetry is the core of spike-timing-dependent plasticity: the order matters, and the tag mechanism enforces the order by requiring the forward signal to precede the backward one by enough time to plant the tag before the bAP confirms it. At the end of the bAP context, the sodium-potassium pump resets the membrane, drawing on the astrocyte's lactate supply, and the soma ATP budget decrements to pay for the propagation cost of the bAP itself. + +--- + +## DAY Scope — Dendritic Branch + +### Context: CONTINUOUS + +The dendritic branch runs continuously as an integrator and distributor, operating in both directions simultaneously. Upward, it integrates the membrane potentials of all spines on its length into a branch voltage that propagates toward the soma and contributes to whether a somatic action potential fires. It also propagates the back-propagating AP downward from the soma toward the spines, but its geometry matters: bAP amplitude decays with distance from the soma, meaning spines at the distal end of a long branch receive a weaker confirmation signal than proximal spines. This geometry is itself a slow structural trace — branches that have grown longer or lost mitochondrial density will propagate bAPs less faithfully. + +The branch accumulates its own calcium signal driven by the bAP and by spillover from active spines, but this calcium decays more slowly than spine calcium — it integrates across multiple spines and persists for hundreds of milliseconds. When enough spines on the same branch are co-active, a branch-level tag is planted. This is the branch's version of the synaptic tag: a marker that this branch as a whole has been meaningfully recruited, which triggers local BDNF release — a trace that signals branch-level survival and growth pressure to the soma. + +When a branch tag is set and the local mRNA pool is non-empty, the branch ribosomes begin translating stored mRNAs into structural proteins locally, without waiting for the soma. This local translation provides a fast protein supply that can support early structural changes within minutes rather than hours. The translation itself draws from the branch mitochondrial budget, so a branch that is energetically depleted cannot translate even if its mRNA pool is full. The acetylcholine broadcast modulates the global LTP threshold continuously in this context — during periods of high attentional drive, the threshold lowers across all branches simultaneously. + +--- + +## DAY Scope — Soma + +### Context: AP + +The soma fires an action potential when integrated dendritic input crosses the firing threshold — itself a variable that is modulated by the neuromodulatory state. When it fires, calcium enters the nucleus — a trace that accumulates with firing rate and decays over seconds. This nuclear calcium is the signal that gates CREB phosphorylation in the soma, and CREB phosphorylation is the trace that determines how much protein synthesis runs during the subsequent NIGHT scope. But CREB phosphorylation requires both nuclear calcium above threshold AND PKA activity above threshold — it is a coincidence gate at the somatic level, just as NMDA opening is a coincidence gate at the spine level. Both the activity pattern (nuclear calcium) and the neuromodulatory context (PKA) must align for the gene expression program to activate. When they do, Arc mRNA and BDNF production begin — both traces that will be shipped to active dendritic branches during the NOT_AP context. + +### Context: NOT_AP + +Between somatic spikes, the soma integrates dendritic inputs and manages the distribution of resources to branches. It ships Arc mRNA preferentially to tagged branches, whose tags were planted during the DAY based on spine-level calcium events. It queues organelle deliveries to branches ranked by their branch calcium levels — the most active branches receive mitochondria first. This shipping process is itself budgeted: the organelle pool is finite, and a soma that has been heavily depleting its pool during a period of high activity will have fewer organelles to ship until NIGHT replenishment runs. + +The neuromodulatory broadcast in the NOT_AP context sets the PKA gate simultaneously across all compartments: GluA1-Ser845 priming lowers the AMPA insertion threshold at all tagged spines, DARPP-32 phosphorylation silences the LTD phosphatase globally, and CREB activation opens the gene expression gate. None of these actions write any structural variable — they are traces that bias what the NIGHT scope will commit. + +--- + +## DAY Scope — Astrocyte + +### Context: CONTINUOUS + +The astrocyte operates without discrete contexts — it monitors its entire territory continuously. Glutamate clearance runs at all times at a rate set by the density of transporter proteins currently deployed, drawing from the ATP budget with every cycle. When spillover crosses the low-affinity mGluR5 threshold — signaling that the cleft is genuinely saturated rather than merely active — the astrocyte's internal calcium rises proportionally to the spillover level. This local calcium rise drives D-serine secretion, but the amount secreted is capped by the current serine racemase enzyme capacity, which may be partially depleted if many neighboring synapses are simultaneously demanding co-agonist support. A busy astrocyte territory delivers partial D-serine to many synapses rather than full D-serine to a few — the support is rationed automatically by the enzyme pool. + +The same spillover simultaneously activates the presynaptic mGluR2/3 brake through a completely different receptor arm, reducing vesicle release probability at the source of the overflow. This push-pull is not coordinated by any signal — it is a structural consequence of the two receptor types sitting on different cells with different G-protein couplings, activated by the same diffusing ligand at the same moment. + +The astrocyte checks continuously whether its local calcium has exceeded the global overload threshold. If it has, a soma-wide calcium wave fires and the shockwave lockdown executes — a circuit-breaker that mass-internalizes postsynaptic AMPA receptors, hyperpolarizes the membrane, and clusters presynaptic calcium channels for maximum signal preservation, all while rapidly depleting the astrocyte's ATP budget. The lactate pipeline runs in parallel at all times, converting glucose to lactate up to the vascular ceiling and distributing it in fractions to the presynapse, postsynapse, and dendritic branch — the fractionation determining which compartment gets refueled first under high demand. + +--- + +## Special Case: Shockwave Lockdown + +The shockwave lockdown can trigger in either scope. It bypasses all budget gates — it is an emergency override that executes regardless of resource availability — but it pays a large ATP cost that leaves the astrocyte depleted and slower to support normal plasticity events for some time afterward. AMPA receptors are mass-internalized and returned to the dendritic reserve pool rather than destroyed, meaning they are recoverable when the emergency passes. + +--- + +## NIGHT Scope — All Compartments + +The NIGHT scope opens with budget replenishment. Soma protein synthesis rate, which peaked because CREB was activated during DAY, now drives the highest protein production of the cycle. Branch-level pools refill from this production — receptor reserves, actin machinery, mRNA pools, and mitochondrial capacity all recover at rates proportional to how much CREB-driven expression ran. The astrocyte's pools replenish on their own schedules, with process extensions recovering the most slowly. + +The soma then processes its shipping queue: organelles deferred during DAY are delivered to the branches that ranked highest by calcium activity. Once the queues are resolved, the structural commit function runs for every synapse that carries a tag. + +### PRE — Structural Commit + +The presynapse draws from the axonal protein pool to expand its active zone. What it receives is proportional to how much remains in the pool after competing tagged boutons have drawn their shares. The expansion commits permanently: the active zone grows, more docking slots are installed, calcium channels cluster more tightly beneath the zone, and baseline release probability rises. These are slow traces that will read by every AP context in every subsequent DAY scope. If the pool was insufficient for the full requested expansion, the deficit is queued for the next NIGHT and the partial expansion is committed immediately. In the LTD branch, the process runs in reverse and the proteins dismantled from the active zone are returned to the shared pool, where they become available to other boutons. + +### POST — Structural Commit + +The postsynapse draws from the branch receptor reserve and actin machinery. CaMKII, activated by the calcium trace deposited during DAY, anchors new AMPA receptors into the membrane surface — the number anchored determined by how many receptors the budget granted. The spine head physically enlarges in proportion to the actin machinery available. Both changes are slow traces that will be read by every NOT_bAP context in every subsequent DAY. Deficits are queued. In the LTD branch, phosphatase PP1 drives receptor internalization, and the internalized receptors are returned to the branch reserve pool — again conserving the total receptor count across the segment rather than destroying it. + +### DEND — Structural Commit + +The dendritic branch receives the organelles shipped from the soma, permanently increasing its local mitochondrial density — a slow trace that will improve bAP propagation fidelity and local translation capacity in future DAY scopes. Arc mRNA stored in the branch pool is translated into structural proteins that expand the local protein flux, making the branch more self-sufficient for future plasticity demands. If BDNF-TrkB signaling crossed its survival threshold during DAY — meaning the branch was genuinely and collectively active — mitochondrial density is further reinforced, stabilizing the branch architecture. If it did not cross the threshold, mitochondrial density declines slightly, making the branch progressively less capable of supporting future plasticity. This is the branch-level version of the use-it-or-lose-it logic: not a binary decision but a continuous drift in one direction or the other depending on whether BDNF traces accumulated during DAY. + +### SOMA — Structural Commit + +The soma's NIGHT work is primarily generative rather than structural in the local sense. CREB-driven transcription produces new proteins, receptors, and mRNAs at the highest rate of the cycle, replenishing all downstream pools. Organelle biogenesis runs — new mitochondria are produced and added to the organelle pool for shipping in future NIGHT cycles. If the overall somatic firing rate during DAY exceeded the homeostatic ceiling — meaning the entire neuron was over-recruited — a global downscaling factor is applied to all synapses simultaneously, reducing AMPA counts and release probabilities network-wide. This homeostatic correction is not targeted to any specific synapse; it is a blanket adjustment that restores the neuron's mean excitability without erasing the relative differences between strong and weak synapses. When all queues are empty and all tags are cleared, CREB phosphorylation and CREB activation are reset to false, closing the gene expression program until the next DAY's activity reactivates it. + +### ASTRO — Structural Commit + +The astrocyte's NIGHT work centers on remodeling its physical relationship to each synapse it wraps. For synapses that were validated for LTP during DAY, the astrocyte draws from its ECM protein pool to secrete Glypicans and Thrombospondins, sealing and structurally reinforcing the synaptic environment. It draws from its process extension budget to retract its walls inward — physically tightening the wrap around the synapse. This tightening is itself a slow trace with two consequences: it reduces the rate at which glutamate diffuses away from the cleft, and it raises the tonic D-serine level available to the NMDA receptor. Both changes make the synapse more sensitive to future events, making LTP self-reinforcing. For synapses committed to LTD, the reverse occurs: MMP enzymes dissolve the ECM, D-serine supply is cut to zero, and the astrocyte's process walls extend outward, loosening the diffusion geometry and making future signals less concentrated and less likely to trigger NMDA opening. The freed process extensions are returned to the pool and become available for reallocation to LTP synapses. + +--- + +## Key Asymmetries and the Closure Summary + +The perisynaptic distance variable is the one slow trace that amplifies rather than merely reflects the outcome. Because it controls both glutamate diffusion geometry and tonic D-serine availability simultaneously, a synapse that has been potentiated becomes progressively easier to potentiate further, and one that has been depressed becomes progressively harder to rescue — not because of any new signal but because the physical environment was remodeled to favor the existing direction. + +The synaptic tag is the bridge that closes the loop between DAY and NIGHT. It is planted in the NOT_bAP context of DAY when a calcium event meets the tagging threshold, and it is consumed in the NIGHT scope when structural resources are available to fulfill the commit. If the neuromodulatory save signal never arrived during DAY — if dopamine or norepinephrine never validated the event as worth storing — the tag expires at the end of NIGHT without triggering a commit, and the DAY's traces are cleared without structural consequence. + +The budget conservation law closes the entire system: LTP draws from shared pools and passively impoverishes untagged neighbors; LTD returns resources to shared pools and passively enriches them. The total resources in the system do not grow through activity — they are redistributed. The only way to increase total capacity is through NIGHT-scope CREB-driven synthesis and organelle biogenesis, and even that is bounded by the vascular glucose ceiling that the astrocyte cannot exceed regardless of demand. Every behavior in the system therefore occurs within a context that permits it, consumes from a budget that constrains it, and leaves a trace that shifts the probability of the next behavior — and the entire architecture is designed so that no single event, however strong, can permanently commit the system without the conjunction of the right context, the right energy state, and the right neuromodulatory validation. + +--- +--- + +# Tripartite Synapse — Pseudocode v4 +# Key change from v3: every trace split into fast_trace (DAY only) and tag (DAY→NIGHT bridge) + +--- + +## Conventions + +``` +SCOPE = { DAY, NIGHT } +CONTEXT = { AP, NOT_AP, bAP, NOT_bAP, CONTINUOUS } + +Variable types: + FIXED = imposed externally — does not change during simulation + VAR = changes dynamically + FAST_TRACE = accumulates and decays within DAY scope only + lifetime: ms to minutes + role: biases the next behavior within the same context + cleared: automatically by decay, never persists to NIGHT + + TAG = set during DAY, persists into NIGHT scope + lifetime: hours — survives the full DAY scope + role: gates structural commit in NIGHT + cleared: explicitly at end of NIGHT after commit or expiry + requires: explicit SET and CLEAR operations + + BUDGET = single energy variable per compartment + all costs deducted here regardless of molecular source + + STRUCTURE = slow architectural variable + READ during DAY, WRITTEN only during NIGHT +``` + +--- + +## Fixed Parameters + +``` +FIXED input_freq // driving spike frequency +FIXED dopamine_level // reward/save signal from VTA +FIXED NE_level // arousal signal from locus coeruleus +FIXED ACh_level // attention signal from basal forebrain +FIXED vascular_glucose_supply // hard energy ceiling +FIXED branch_geometry // dendritic topology — affects bAP decay +FIXED Ca_HIGH // LTP-triggering Ca²⁺ threshold +FIXED Ca_LOW // LTD-triggering Ca²⁺ threshold +FIXED Ca_TAG_threshold // Ca²⁺ level sufficient to set a tag +FIXED spillover_threshold // cleft saturation threshold for mGluR activation +FIXED homeostatic_ceiling // max firing rate before global downscale +FIXED disuse_threshold // silence duration before passive LTD +FIXED recycling_fraction // fraction of ECM proteins recovered after MMP cleavage +FIXED tag_expiry_threshold // minimum tag strength to survive to NIGHT commit +``` + +--- + +## Budgets — One Per Compartment + +``` +VAR pre_budget // bouton ATP + // costs: VGCC opening, vesicle fusion, VATPase refill, AZ maintenance + // (simplifies: axon_mitochondria_capacity + vesicle_protein_pool) + +VAR post_budget // spine ATP + // costs: NMDA current, NaK reset, AMPA trafficking, actin remodeling + // (simplifies: spine_actin_machinery + receptor_reserve) + +VAR dend_budget // branch ATP + // costs: bAP propagation, local translation, branch Ca²⁺ handling + // (simplifies: dend_mitochondria_capacity + local_protein_flux) + +VAR soma_budget // somatic ATP + // costs: AP generation, CREB transcription, protein synthesis, shipping + // (simplifies: soma_ATP + synthesis_rate + organelle_pool) + +VAR astro_budget // astrocyte ATP + // costs: glutamate clearance, D-serine synthesis, ECM secretion, + // process motility, lactate production + // (simplifies: astro_ATP + racemase_cap + ECM_pool + process_extensions) + +VAR astro_lactate // fuel exported to pre, post, dend budgets + // hard cap: vascular_glucose_supply (FIXED) +``` + +--- + +## Traces — Split Into FAST_TRACE and TAG Per Compartment + +``` +// ── PRE ─────────────────────────────────────────────────────────────── + +FAST_TRACE pre_fast_trace // residual Ca²⁺ in bouton + // SET: += spike_Ca_influx() on each AP + // DECAY: *= exp(-Δt / 100ms) continuously + // ROLE: biases vesicle_release_prob in next AP context + // CLEAR: decays automatically — never reaches NIGHT + // (covers: pre_Ca_residual, short-term facilitation) + +TAG pre_tag // commitment candidacy of this bouton + // SET: when dopamine_level > D1_threshold in NOT_AP context + // ROLE: grants priority access to axonal protein pool in NIGHT + // marks bouton as validated for structural expansion + // CLEAR: explicitly at end of NIGHT after commit or expiry + // (covers: axon_tagged_boutons, PKA-driven bouton marking) + +// ── POST ────────────────────────────────────────────────────────────── + +FAST_TRACE post_fast_trace // Ca²⁺ dynamics in spine + // SET: += NMDA_Ca_influx × rise_speed() in NOT_bAP context + // SET: += bAP_Ca_boost() if pre_tag confirmed in bAP context + // DECAY: *= exp(-Δt / τ_Ca) — faster for LTP, slower for LTD + // ROLE: encodes LTP vs LTD instruction via amplitude + speed + // gates whether post_tag is set + // CLEAR: decays automatically — never reaches NIGHT + // (covers: post_Ca_amplitude, post_Ca_rise_speed, CaMKII vs PP1) + +TAG post_tag // commitment candidacy of this spine + // SET: when post_fast_trace > Ca_TAG_threshold in NOT_bAP + // AND confirmed by bAP in bAP context + // ROLE: grants priority access to dend receptor reserve in NIGHT + // gates structural commit of AMPA insertion + spine growth + // CLEAR: explicitly at end of NIGHT after commit or expiry + // (covers: dend_tagged_spines, synaptic tag-and-capture) + +// ── DEND ────────────────────────────────────────────────────────────── + +FAST_TRACE dend_fast_trace // branch Ca²⁺ integration + // SET: += bAP_Ca_influx() + spine_spillover_Ca() + // DECAY: *= exp(-Δt / 300ms) — slower than spine Ca²⁺ + // ROLE: integrates co-active spine count + // gates whether dend_tag is set + // CLEAR: decays automatically — never reaches NIGHT + // (covers: branch_Ca, local NMDA spillover integration) + +TAG dend_tag // commitment candidacy of this branch + // SET: when co_active_spines > branch_tag_threshold + // AND dend_fast_trace > dend_tag_threshold + // ROLE: activates local mRNA translation during DAY + // gates branch structural expansion in NIGHT + // triggers BDNF release to soma + // CLEAR: explicitly at end of NIGHT after commit or expiry + // (covers: branch_tag, BDNF-TrkB branch survival signal) + +// ── SOMA ────────────────────────────────────────────────────────────── + +FAST_TRACE soma_fast_trace // nuclear Ca²⁺ accumulation + // SET: += nuclear_Ca_influx() on each somatic AP + // DECAY: *= exp(-Δt / τ_nuclear) — seconds timescale + // ROLE: gates CREB phosphorylation when combined with PKA + // gates soma_tag when threshold crossed + // CLEAR: decays automatically — never reaches NIGHT + // (covers: nuclear_Ca, AP-driven Ca²⁺ accumulation) + +TAG soma_tag // gene expression mandate + // SET: when soma_fast_trace > CREB_threshold + // AND dopamine_level > D1_threshold (PKA coincidence) + // ROLE: drives protein synthesis rate in NIGHT scope + // determines how much structural protein is available + // for all compartment commits + // CLEAR: explicitly at end of NIGHT after synthesis executed + // (covers: CREB_phospho, Arc_mRNA, BDNF_production mandate) + +// ── ASTRO ───────────────────────────────────────────────────────────── + +FAST_TRACE astro_fast_trace // local Ca²⁺ in perisynaptic process + // SET: += mGluR5_Ca_influx() when glutamate > spillover_threshold + // DECAY: *= exp(-Δt / τ_astro_Ca) — seconds timescale + // ROLE: drives D-serine release proportional to magnitude + // gates astro_tag when overload threshold crossed + // CLEAR: decays automatically — never reaches NIGHT + // (covers: astro_Ca_local, IP3-driven Ca²⁺) + +TAG astro_tag // territory-level activity record + // SET: when astro_fast_trace > astro_tag_threshold + // integrates overflow events across territory + // ROLE: confirms overflow_sensed condition in NIGHT commit + // gates ECM sealing and process remodeling in NIGHT + // CLEAR: explicitly at end of NIGHT after structural commit + // (covers: mGluR5_activation record, overflow validation) +``` + +--- + +## Structural Variables — Written Only in NIGHT + +``` +VAR pre_structure // active zone capacity + // (covers: active_zone_size + VGCC_clustering + RRP_pool_capacity) + +VAR post_structure // spine sensitivity + // (covers: AMPA_count + spine_volume) + +VAR dend_structure // branch capacity + // (covers: mitochondria_density + local_protein_flux_ceiling) + +VAR soma_structure // somatic output capacity + // (covers: protein_synthesis_rate + organelle_pool) + +VAR astro_structure // synaptic environment geometry + // (covers: perisynaptic_distance + ECM_integrity + // + D_serine_tonic_level + glutamate_clearance_rate) +``` + +--- +--- + +# SCOPE: DAY +# FAST_TRACEs written and decay. TAGs set but not cleared. STRUCTUREs read only. + +--- + +## PRE | CONTEXT: AP + +``` +scope DAY | context AP: + + if pre_budget < release_cost: + suppress(release) + exit context + + // Fast trace: deposit and decay + pre_fast_trace += spike_Ca_influx(input_freq) + pre_fast_trace *= decay(τ = 100ms) // FAST_TRACE — will not reach NIGHT + + // Release gated by fast trace (residual Ca²⁺ from recent APs) + release_prob = pre_structure.base_prob × facilitation(pre_fast_trace) + released = binomial(RRP_pool, release_prob) + glutamate = released × quantal_content + pre_budget -= release_cost + + // Overflow: autoinhibitory brake + astro notification + if glutamate > spillover_threshold: + release_prob *= mGluR_brake_factor // Gi arm — no budget cost on PRE + + // Refuel + pre_budget += astro_lactate × pre_fraction +``` + +## PRE | CONTEXT: NOT_AP + +``` +scope DAY | context NOT_AP: + + // Fast trace decays passively + pre_fast_trace *= decay(τ = 100ms) + + // Intermediate tuning from fast trace history + if pre_fast_trace > high_threshold: + mobilize(reserve → RRP) // pool-limited + elif pre_fast_trace < low_threshold: + release_prob *= depression_factor + + // Tag: set when neuromodulator context validates this bouton + if dopamine_level > D1_threshold or NE_level > β_threshold: + pre_tag = SET // TAG — persists to NIGHT + // pre_fast_trace NOT modified + // two variables, two lifetimes +``` + +--- + +## POST | CONTEXT: NOT_bAP + +``` +scope DAY | context NOT_bAP: + + // AMPA current — gated by post_structure (READ) + Vm += glutamate × post_structure.sensitivity + post_budget -= AMPA_cost + + // NMDA gate: depolarization + D-serine coincidence + if Vm > Mg_eject_threshold and astro_D_serine > D_serine_threshold: + Ca_influx = NMDA_influx(glutamate) + post_fast_trace += Ca_influx × rise_speed(Ca_influx) // FAST_TRACE — will not reach NIGHT + post_budget -= NMDA_cost + + // Tag: set when fast trace crosses tagging threshold + // This is the Hebbian anticipation window — before bAP confirms + if post_fast_trace > Ca_TAG_threshold: + post_tag = SET // TAG — persists to NIGHT + // post_fast_trace NOT modified + // tag is a separate variable, separate lifetime +``` + +## POST | CONTEXT: bAP + +``` +scope DAY | context bAP: + + // bAP arrives — confirms or rejects the tag set in NOT_bAP + Vm += bAP_depolarization(dend_structure.geometry) + + if post_tag == SET and post_fast_trace > Ca_TAG_threshold: + post_fast_trace += bAP_Ca_boost() // FAST_TRACE amplified — confirms coincidence + // post_tag remains SET — bAP confirms but does not set the tag + // else: bAP without prior tag — no amplification, no tag set + + post_budget -= bAP_reset_cost + dend_budget -= bAP_prop_cost +``` + +--- + +## DEND | CONTEXT: CONTINUOUS + +``` +scope DAY | context CONTINUOUS: + + // Integrate and propagate + branch_Vm = integrate(POST.Vm, all_spines) + bAP_local = propagate_bAP(SOMA.AP_fired, dend_structure.geometry) + + // Fast trace: branch Ca²⁺ + dend_fast_trace += bAP_Ca_influx(bAP_local) + dend_fast_trace *= decay(τ = 300ms) // FAST_TRACE — will not reach NIGHT + + // Tag: set when branch co-activity crosses threshold + if co_active_spines(branch) > branch_tag_threshold: + dend_tag = SET // TAG — persists to NIGHT + BDNF_signal → soma // signals branch recruitment to soma + + // Local translation: only runs if tag set and budget allows + // Uses dend_fast_trace magnitude to scale translation rate + if dend_tag == SET and dend_budget > translation_cost: + local_proteins = translate(mRNA_pool, dend_fast_trace) + dend_structure.protein_flux += local_proteins // immediate local supply + dend_budget -= translation_cost + + dend_budget -= branch_maintenance_cost + dend_budget += astro_lactate × dend_fraction +``` + +--- + +## SOMA | CONTEXT: AP + +``` +scope DAY | context AP: + + if branch_Vm > AP_threshold: + AP_fired = True + + // Fast trace: nuclear Ca²⁺ + soma_fast_trace += nuclear_Ca_influx() // FAST_TRACE — will not reach NIGHT + soma_fast_trace *= decay(τ = seconds) + + // Tag: set when fast trace AND neuromodulator context coincide + if soma_fast_trace > CREB_threshold and dopamine_level > D1_threshold: + soma_tag = SET // TAG — persists to NIGHT + // soma_fast_trace NOT modified + // two variables, two lifetimes + + soma_budget -= AP_generation_cost +``` + +## SOMA | CONTEXT: NOT_AP + +``` +scope DAY | context NOT_AP: + + branch_Vm = integrate(DEND.branch_Vm, all_branches) + + // Ship resources to tagged branches — tag status determines priority + for branch in branches_ranked_by(dend_tag == SET): + dend_budget[branch] += soma_budget × shipping_fraction + soma_budget -= shipping_cost + + // Neuromodulator broadcast: amplifies post_fast_trace at tagged spines + // Does NOT set tags — that requires the Ca²⁺ coincidence in POST + if dopamine_level > D1_threshold or NE_level > β_threshold: + for spine where post_tag == SET: + post_fast_trace *= PKA_amplifier // FAST_TRACE boosted — lowers commit threshold +``` + +--- + +## ASTRO | CONTEXT: CONTINUOUS + +``` +scope DAY | context CONTINUOUS: + + // Clearance — rate set by astro_structure (READ) + glutamate -= astro_structure.clearance_rate × Δt + astro_budget -= clearance_cost + + // Fast trace: local Ca²⁺ from overflow + if glutamate > spillover_threshold: + astro_fast_trace += mGluR5_Ca_influx() // FAST_TRACE — will not reach NIGHT + astro_fast_trace *= decay(τ = astro_Ca_τ) + + // D-serine proportional to fast trace — budget-limited + D_serine = min(proportional_to(astro_fast_trace), astro_budget × Ds_fraction) + astro_budget -= D_serine_cost + + // Tag: set when overflow is sustained above territory threshold + if astro_fast_trace > astro_tag_threshold: + astro_tag = SET // TAG — persists to NIGHT + // astro_fast_trace NOT modified + + // Global overload: fast trace spike triggers lockdown + if astro_fast_trace > OVERLOAD_threshold: + trigger(shockwave_lockdown) // uses astro_fast_trace, not astro_tag + + // Fuel pipeline — capped at vascular_glucose_supply (FIXED) + astro_lactate = min(glycolysis(vascular_glucose_supply), astro_budget × lactate_fraction) + astro_budget += glycolysis(vascular_glucose_supply) × Δt + deliver(astro_lactate → pre_budget, post_budget, dend_budget) +``` + +--- +--- + +# SCOPE: NIGHT +# TAGs evaluated. STRUCTUREs written. Budgets replenished. All traces cleared. + +--- + +## Step 1 — Replenish Budgets + +``` +scope NIGHT | step 1: + + // soma_tag magnitude determines how much synthesis runs + soma_structure.synthesis_rate = CREB_expression(soma_tag) // TAG READ here + astro_budget += overnight_synthesis() × Δt_night + soma_budget += overnight_mitochondria() × Δt_night + dend_budget += soma_budget × dend_replenish_fraction + post_budget += soma_budget × post_replenish_fraction + pre_budget += soma_budget × pre_replenish_fraction +``` + +--- + +## Step 2 — Structural Commit + +``` +scope NIGHT | step 2: + + // Three-layer filter — uses TAGs, not FAST_TRACEs + event_strong = post_tag == SET // TAG evaluated + overflow_seen = astro_tag == SET // TAG evaluated + context_valid = pre_tag == SET // TAG evaluated + and soma_tag == SET // TAG evaluated + + if event_strong and overflow_seen and context_valid: + + // All four compartments commit simultaneously — budget-gated + ΔAZ = min(AZ_cost, pre_budget × pre_structural_fraction) + ΔAMPA = min(AMPA_cost, post_budget × post_structural_fraction) + ΔDEND = min(dend_cost, dend_budget × dend_structural_fraction) + ΔASTRO = min(astro_cost, astro_budget × astro_structural_fraction) + + pre_structure += ΔAZ // STRUCTURE WRITTEN — larger active zone + post_structure += ΔAMPA // STRUCTURE WRITTEN — more receptors, larger spine + dend_structure += ΔDEND // STRUCTURE WRITTEN — denser mitochondria, more flux + astro_structure -= ΔASTRO // STRUCTURE WRITTEN — walls IN, ECM sealed, D-serine tonic ↑ + + pre_budget -= ΔAZ × structural_cost + post_budget -= ΔAMPA × structural_cost + dend_budget -= ΔDEND × structural_cost + astro_budget -= ΔASTRO × structural_cost + + // Partial grants: deficit queued for next NIGHT + if ΔAZ < AZ_cost: queue(pre_deficit → next NIGHT) + if ΔAMPA < AMPA_cost: queue(post_deficit → next NIGHT) + + return "potentiated" + + elif event_strong and not context_valid: + return "temporary — no commit" // tags expire without structural write + + elif post_tag == SET and post_fast_trace < Ca_LOW: + // LTD: resources returned to budgets + pre_structure -= LTD_pre_amount + post_structure -= LTD_post_amount + astro_structure += LTD_astro_amount // walls OUT, ECM dissolved + pre_budget += recovered_AZ_cost + post_budget += recovered_AMPA_cost + astro_budget += recovered_process_cost × recycling_fraction + return "depressed" + + else: + return "baseline — no change" +``` + +--- + +## Step 3 — Homeostatic Scaling + +``` +scope NIGHT | step 3: + + if soma_tag.magnitude > homeostatic_ceiling: + for each synapse: + post_structure *= homeostatic_scale_down // STRUCTURE WRITTEN + pre_structure *= homeostatic_scale_down // STRUCTURE WRITTEN + + for each branch where dend_tag == CLEAR and disuse_duration > disuse_threshold: + dend_structure -= pruning_rate × Δt_night // STRUCTURE WRITTEN + dend_budget += recovered_branch_cost +``` + +--- + +## Step 4 — Clear All Traces + +``` +scope NIGHT | step 4: + + // FAST_TRACEs: already decayed during DAY — confirm zero + pre_fast_trace = 0 + post_fast_trace = 0 + dend_fast_trace = 0 + soma_fast_trace = 0 + astro_fast_trace = 0 + + // TAGs: explicitly cleared after commit or expiry + if pre_tag < tag_expiry_threshold: pre_tag = CLEAR + if post_tag < tag_expiry_threshold: post_tag = CLEAR + if dend_tag < tag_expiry_threshold: dend_tag = CLEAR + if soma_tag < tag_expiry_threshold: soma_tag = CLEAR + if astro_tag < tag_expiry_threshold: astro_tag = CLEAR + // Tags above threshold: persist to next NIGHT (multi-night consolidation) +``` + +--- + +## Key Structural Asymmetry + +``` +// astro_structure controls both diffusion geometry and D-serine tonic simultaneously: +// LTP commit → astro_structure decreases (walls IN) +// → clearance_rate ↓ + D_serine_tonic ↑ +// → every future DAY event easier to potentiate +// LTD commit → astro_structure increases (walls OUT) +// → clearance_rate ↑ + D_serine_tonic = 0 +// → every future DAY event easier to depress +// Self-reinforcing in both directions. + +// Budget conservation: +// LTP draws from shared budgets → neighbors passively impoverished +// LTD returns resources to budgets → neighbors passively enriched +// Hard ceiling: vascular_glucose_supply (FIXED) — cannot be exceeded + +// The FAST_TRACE / TAG split enforces the core loop precisely: +// FAST_TRACE: behavior leaves a trace that biases the next behavior (within DAY) +// TAG: behavior leaves a trace that gates resource allocation (DAY → NIGHT) +// STRUCTURE: resource allocation produces permanent change (written in NIGHT) +// BUDGET: all of the above is constrained by available energy +``` + + +--- +--- + + +# Additions + +## Neuromodulators + +These are produced by small, anatomically concentrated nuclei that broadcast widely across the brain: + +- dopamine_level // "save button" — validates LTP +- norepinephrine_level // arousal / signal-to-noise gain +- acetylcholine_level // attention — lowers LTP threshold + +### Dopamine + +Dopamine is produced primarily by neurons in the Substantia Nigra pars compacta (projecting to the striatum, relevant for motor learning and habit formation) and the Ventral Tegmental Area (VTA) (projecting to the prefrontal cortex and limbic system via the mesolimbic and mesocortical pathways, relevant for reward, motivation, and the "save button" function in your model). + +### Norepinephrine +Norepinephrine is produced almost exclusively by the Locus Coeruleus, a tiny nucleus in the brainstem pons. Despite its small size it projects diffusely across virtually the entire brain — cortex, hippocampus, cerebellum, spinal cord. It's essentially the brain's arousal and signal-to-noise broadcaster, firing tonically at low rates during calm wakefulness and phasically during novel or stressful events. + +### Acetylcholine +Acetylcholine has two main sources: the basal forebrain nuclei (including the nucleus basalis of Meynert) projecting to the cortex and hippocampus — relevant for attention and learning gating — and the medial septum projecting specifically to the hippocampus, where it strongly modulates theta rhythms and memory encoding. + +What's striking in the context of your model is that all three systems share the same architectural logic: a tiny, localized cell population broadcasts a global contextual signal that shifts the operational threshold of millions of synapses simultaneously — none of them carrying specific content, all of them modulating how content gets written. + +### Simple organisms + +Excellent point. The hippocampal replay model is a vertebrate solution to a specific problem — how to consolidate many parallel experiences quickly without catastrophic interference. But the underlying molecular logic of the commit function is far more ancient and appears in organisms that have no hippocampus, no sleep architecture in the vertebrate sense, and sometimes no centralized nervous system at all. + +**In invertebrates with simple ganglia** — *Aplysia*, *C. elegans*, *Drosophila* — the slow-scale consolidation still requires protein synthesis and still uses CREB as the nuclear transcription factor. The same PKA→CREB axis that validates LTP in the mammalian hippocampus was actually first characterized in *Aplysia* gill-withdrawal reflex studies by Kandel. What differs is the trigger and the timing. Without a hippocampus to compress and replay experiences during a rest phase, consolidation in these organisms appears to depend simply on **repetition and spacing of the stimulus itself**. A single strong shock to the siphon produces short-term sensitization lasting minutes — the intermediate scale running without commitment. Four or five spaced shocks over hours produce long-term sensitization lasting days — the slow-scale commit running because repeated PKA activation eventually crosses the threshold needed to drive CREB-dependent gene expression. The spacing matters because cAMP degrades between stimuli, and spaced repetition keeps re-elevating it above the threshold for nuclear translocation, whereas massed repetition saturates and desensitizes the cascade. + +**In *C. elegans***, which has exactly 302 neurons and no sleep in any recognizable sense, consolidation-like phenomena still occur through the same molecular logic. What substitutes for the validation signal is less clear, but there is evidence that **neuromodulatory interneurons releasing serotonin and dopamine** gate whether a repeated experience gets written into long-term behavioral change — functionally identical to the save button role, just implemented in a three-neuron circuit rather than a diffuse broadcast system. + +**In organisms without a nervous system at all** — *Physarum polycephalum*, the slime mould — something that looks functionally like consolidation occurs through purely biochemical means. Repeated exposure to a stimulus that causes contraction leaves a persistent change in the cytoskeletal and biochemical state of the organism that biases future responses. There is no synapse, no calcium influx through NMDA receptors, no PKA — but the logic of the three-layer filter is still present in a stripped-down form: the organism detects that something happened, detects that it was significant enough to exceed a threshold, and commits a physical change to its internal architecture that persists. The molecular substrate is different but the computational structure is the same. + +**What this suggests about consolidation in general** is that the slow-scale commit function does not require sleep or replay — those are vertebrate optimizations on top of a more primitive mechanism. The core requirements appear to be just three things that evolution has solved repeatedly with different molecular tools: + +The first is a **threshold mechanism** that prevents every transient event from being written permanently — in neurons this is the calcium amplitude and rise-speed filter, in slime moulds it is a chemical concentration threshold, but the logic is the same. + +The second is a **context validation signal** that asks whether the event was behaviorally meaningful — in vertebrates this is the neuromodulatory broadcast from VTA and locus coeruleus, in *Aplysia* it is facilitatory interneurons releasing serotonin, in slime moulds it may simply be the magnitude of the perturbation itself serving as its own validation. + +The third is a **protein synthesis requirement** that enforces a time cost on permanent storage — in all organisms where long-term behavioral change has been studied, blocking protein synthesis blocks consolidation, suggesting that the metabolic cost of building new structural proteins is a universal gate that prevents trivial events from consuming permanent resources. + +Sleep and hippocampal replay are therefore best understood not as the mechanism of consolidation but as a **vertebrate solution to the scaling problem** — how to consolidate thousands of experiences per day across billions of synapses without running the protein synthesis machinery continuously at full cost during waking. Simpler organisms consolidate fewer experiences, face less interference, and can afford to let repetition and spacing do the work that sleep does in more complex nervous systems. + +## Calcium in pre, post, astro + +**In the postsynapse**, the calcium amplitude and rise-speed filter works because CaMKII and the phosphatases PP1/PP2B have different sensitivities to calcium-bound calmodulin, and calmodulin itself has different binding kinetics depending on how fast calcium rises. + +Calmodulin has four calcium binding sites and its activation is highly cooperative — it does not activate linearly with calcium concentration but switches sharply above a threshold. When calcium rises fast and high, as during a strong high-frequency burst, calmodulin saturates quickly and activates CaMKII. CaMKII then autophosphorylates at Thr286, which is the critical step — once autophosphorylated it remains active even after calcium falls back to baseline, effectively converting a transient calcium event into a sustained kinase signal that outlasts the trigger. This persistence is what gives CaMKII its memory-like property and is what drives AMPA receptor insertion. + +When calcium rises slowly and to a lower amplitude, as during weak low-frequency input, calmodulin activates preferentially the phosphatases PP2B (calcineurin) and downstream PP1 instead, because these enzymes have higher affinity for calcium-calmodulin complexes at lower occupancy. PP1 then dephosphorylates AMPA receptors, triggering their internalization and driving LTD. + +So the filter is not a simple threshold — it is a **kinetic competition** between two enzyme systems with different calcium-calmodulin affinities. Fast large rise activates the low-affinity high-gain system (CaMKII). Slow small rise activates the high-affinity low-gain system (PP2B/PP1). The same calcium messenger routes to opposite outcomes depending purely on its dynamics. + +**In the presynapse**, the calcium filter is structurally simpler but operates on a different principle — **proximity and timing** rather than kinetic competition. Calcium enters through VGCCs clustered directly beneath the active zone, and the vesicles docked at that zone sit within nanometers of the channel mouth. The local calcium concentration at the release site reaches extremely high values — estimated at hundreds of micromolar — for a very brief window of microseconds before diffusing away. Synaptotagmin, the calcium sensor on the vesicle membrane, has a low affinity but fast on-rate, meaning it only fires in response to this extremely high local transient, not to the diffuse residual calcium that lingers afterward. + +The residual calcium that accumulates with repeated spikes — the pre_Ca_residual in the pseudocode — acts on a completely different target: Munc13 and RIM proteins at the active zone, which have higher affinity for calcium but slower kinetics. These proteins respond to the sustained low-level residual and increase the size of the readily-releasable pool and the probability of release — this is facilitation. So the presynaptic filter distinguishes between the sharp local transient (triggers release via synaptotagmin) and the slow diffuse residual (modulates future release probability via Munc13/RIM). Two calcium signals, two sensors, two time scales, within the same compartment. + +**In the astrocyte**, the calcium filter is the least understood of the three but operates through IP3 receptor gating. IP3 receptors on the endoplasmic reticulum have a bell-shaped calcium dependence — they open in response to rising calcium but are inhibited at very high calcium concentrations. This means the astrocyte's internal calcium release is self-limiting: a moderate IP3 signal produces a local calcium rise that drives D-serine release, but an excessive signal triggers the global soma wave that activates the circuit-breaker response instead. + +The key filter here is therefore the **spatial containment of the IP3 signal**. Under normal high-frequency activity, IP3 production is local to the perisynaptic process and the calcium rise stays local — driving D-serine release proportionally. Only when multiple neighboring synapses fire simultaneously does IP3 accumulate enough to propagate as a regenerative wave across the entire astrocyte via gap junctions to adjacent astrocytes, triggering the global alarm. The astrocyte is therefore filtering not just amplitude but **spatial coherence** — a single strong synapse produces a local response, but coordinated overactivity across a territory produces a qualitatively different global response. + +The common thread across all three compartments is that none of them use a simple concentration threshold. Each uses a molecular mechanism that is sensitive to the **dynamics** of the calcium signal — its speed, its spatial spread, its duration, or its subcellular location — rather than just its peak value. This is what allows the same ion to encode fundamentally different instructions depending on the context in which it arrives. + +## Resource budget + +Exactly right. This is a resource allocation problem, and it is one of the most important constraints the system operates under. Each compartment has a finite physical budget, and potentiation at one synapse necessarily draws from a shared pool that serves many others. + +**In the presynapse**, the axon has many en passant boutons — synaptic release sites distributed along its length, sometimes hundreds of them. The total vesicle pool, the mitochondrial capacity to run the VATPase pumps that refill vesicles, and the cytoskeletal machinery that mobilizes reserve pools are all shared across the entire axonal arbor. When one bouton undergoes LTP and expands its active zone and increases its RRP capacity, it is drawing on the same pool of synaptic proteins — RIM, Munc13, VGCC subunits — that all other boutons on that axon compete for. There is evidence for a **synaptic tagging and capture** mechanism here: a potentiated bouton plants a molecular tag that allows it to capture plasticity-related proteins drifting along the axon, effectively pulling resources away from untagged boutons. This means strong potentiation at one site can passively deplete neighboring sites — a form of competitive resource allocation baked into the axonal architecture. + +**In the postsynapse**, the dendrite hosts thousands of spines, and the situation is even more constrained. The soma produces plasticity-related proteins — new AMPA receptor subunits, CaMKII, scaffolding proteins like PSD-95 — at a rate determined by CREB-driven gene expression, and these proteins must be shipped out along the dendritic arbor to wherever they are needed. The same synaptic tagging logic applies on the postsynaptic side: a spine that has been tagged by early LTP can capture these drifting proteins when they pass, but the total production rate is finite. There is also a **spine morphology budget** — actin polymerization drives spine head enlargement, but the actin machinery and the small GTPases (Rac1, RhoA) that regulate it are shared across the dendritic segment. Potentiating many spines simultaneously on the same dendritic branch would require more actin remodeling machinery than is locally available, meaning strong potentiation at a cluster of nearby spines may physically constrain how much each individual spine can grow. + +Additionally, the postsynapse has a **receptor recycling pool** — a finite intracellular reserve of AMPA receptors held in endosomes near the spine that can be rapidly inserted during early LTP. This pool is local to a dendritic segment and is not immediately replenished. If multiple nearby spines are potentiated in rapid succession, they compete for the same local receptor reserve before new receptors synthesized in the soma can arrive. + +**In the astrocyte**, the budget constraint is the most spatially explicit of the three because a single astrocyte wraps somewhere between 100,000 and 2,000,000 synapses in humans — an enormous territory. Its finite resources include the total EAAT transporter protein available for glutamate clearance, the D-serine synthesis capacity which depends on serine racemase enzyme levels, the ECM protein production capacity for Glypicans and Thrombospondins, and critically the ATP budget for running all of these simultaneously. When many synapses in its territory are active simultaneously, the astrocyte faces a genuine allocation problem: it cannot maximally support all of them at once. There is evidence that the perisynaptic process — the fine astrocytic extension that wraps individual synapses — is itself a dynamic structure that the astrocyte extends and retracts selectively, suggesting it physically prioritizes which synapses receive close wrapping and therefore which ones benefit from tighter D-serine delivery and glutamate containment. + +**The deeper implication** is that potentiation is not just a local bilateral negotiation between one presynapse and one postsynapse. It is a **network-level resource competition** in which: + +A strongly potentiated synapse draws proteins and structural resources away from its neighbors on the same axon and dendrite. The astrocyte must allocate its clearance, co-agonist, and structural support capacity across its entire territory. The metabolic pipeline — glucose from blood vessels, lactate to neurons — has a ceiling determined by capillary supply, meaning a highly active region of the astrocyte territory can become fuel-limited if demand outpaces vascular supply. + +This is probably why **heterosynaptic LTD** exists — the observation that strong LTP at one synapse is often accompanied by spontaneous depression at neighboring synapses on the same dendrite even without those synapses being directly stimulated. It may be less a deliberate regulatory mechanism and more a direct consequence of resource depletion: the potentiated synapse captured the available receptors, actin machinery, and astrocytic support, leaving neighboring synapses passively impoverished. + +The pseudocode should therefore include a budget layer that sits above the individual synapse functions — something like a shared resource pool per axonal arbor, per dendritic segment, and per astrocyte territory, from which each commit function draws and against which each structural expansion is checked. Potentiation that would exceed the available budget either fails to consolidate fully, triggers compensatory depression at neighboring synapses, or waits for new protein synthesis to replenish the pool before completing.