Mind-Upload

The public pages stay at the orientation level. The current one-question-at-a-time deepening route, with bounded EEG claims and fixed dataset anchors, is kept in the wiki so the front page does not overstate progress. Start from Verification here, then move to the wiki only when you need the active deep-dive route.

After the March 2026 re-audit of primary literature, this site no longer accepts the reading that "if we have the wiring diagram and cell types, the rest is almost filled in." Sleep/wake-dependent renormalization, myelination and oligodendrocyte support, local bioenergetic / mitochondrial state, neurovascular-unit / BBB / pericyte state, glial metabolism, and active maintenance under molecular turnover remain separate variables. Accordingly, this site does not treat connectome-complete as equivalent to emulation-complete. It also keeps measurement-side vascular-state / CVR audit separate from maintenance-side neurovascular / BBB state. For a short explanation, see the hidden-state section in WBE 101; for the evidence structure, see Wiki: Why a Connectome Is Not Enough and Wiki: Homeostatic Plasticity and Maintenance State.

This site now blocks that shortcut too. Reveley et al. (2015) showed that superficial white matter can hide long-range cortical connections from roughly half of the cortical surface, Schilling et al. (2018) showed that tractography endpoints are biased toward gyral crowns across deterministic and probabilistic algorithms and even high-resolution data, Gajwani et al. (2023) showed across 40 pipelines and 44 group-representative reconstructions that hub location is highly variable and that hub connectivity correlates with regional surface area in 69% of assessed pipelines, McMaster et al. (2025) showed that graph measures shift significantly across voxel-resolution changes and recommended resampling to 1 mm isotropic for robust comparisons, and Bramati et al. (2026) showed on the same 3 T scanner with uniform processing that common diffusion-sampling schemes still change tractography outputs. Therefore, this site does not read a tractography-derived graph as the connectome itself. It reads it as an acquisition- and pipeline-conditioned macro pathway prior whose endpoint assignment, harmonization route, parcellation, weighting, and abstention boundary still need to be disclosed. The shortest route is WBE 101: hidden state at the entry point, Verification: Observability Budget, and Wiki: tractography route card.

This site blocks that shortcut too. Bosch et al. (2022) showed that live physiology to ultrastructure correlation is a multistage landmark-based bridge, and MICrONS Consortium et al. (2025) extended that route to a same-brain local dataset with about 75,000 neurons linked to a later EM reconstruction of more than 200,000 cells and 0.5 billion synapses, not a simultaneous whole-state sample. Ding et al. (2025) then added a validated stimulus-conditioned response model while also cautioning that model internal representations still need careful interpretation, and Gamlin et al. (2025) still assigned transcriptomic identity through morphology-based predicted labels rather than direct transcriptomic assay inside the EM volume. The remaining state wall is also real: Holler et al. (2021) showed that structure-function correspondence still leaves synaptic-strength structure unresolved, Molnár et al. (2016) showed that human synapses can contain multiple docked vesicles and multivesicular release, Sakamoto et al. (2018) showed that Munc13-1 supramolecular assemblies set independent release sites, Dürst et al. (2022) showed that release probability strongly sets individual synaptic strength, Emperador-Melero et al. (2024) showed that Ca_V2 clustering and vesicle priming are executed by distinct active-zone machineries, and Mittermaier et al. (2024) showed that membrane-potential state gates synaptic consolidation. Beiran & Litwin-Kumar (2025) then showed that connectome-constrained dynamics can still remain degenerate until extra recordings narrow the solution family. Therefore, this site reads same-brain functional connectomics as a sequential local structure-function scaffold or task-bounded conditional predictor, not as direct transcriptomic truth, not as a readout of release-site number, docked-vesicle architecture, active-zone nanostructure / priming-site assembly, or current release competence, and not as a solved local twin. The shortest route is WBE 101: hidden state at the entry point, Wiki: Why a Connectome Is Not Enough, Wiki: Observability and Claim Ceiling by Measurement Stack, and Wiki: State-Continuity Bridge.

The March 2026 update added Verification's Observability Budget so claim ceilings can be enforced in normal page operations. For the table that prevents "multimodal" from being misread as "state-complete," see Wiki: Observability and Claim Ceilings by Measurement Stack.

This site now blocks that shortcut too. Massonis & Villaverde (2020) showed that structural unidentifiability can arise from symmetry and may require a symmetry-breaking observable or reformulation, not just more fitting. White et al. (2016) showed that seemingly optimal extra experiments can instead expose omitted mechanisms and increase model discrepancy, while Beiran & Litwin-Kumar (2025) showed that even connectome-constrained recurrent networks remain degenerate until a small targeted recording set is added. Therefore, a richer stack is not read here as near-unique internal-state recovery unless the Identifiability Card names which ambiguity class survived and why the next condition should actually break it. The longer explanation is in Wiki: From observation to estimation.

The measurement-stack wiki now pulls together the direct intracranial-validation literature for EEG / MEG and the vascular / autonomic literature for hemodynamic signals. If you want the numbers rather than the slogan, start with EEG / MEG visibility, inverse, and validation wall and hemodynamic transfer wall.

This site now blocks that shortcut too. Vafaii et al. (2024) and Chen et al. (2025) showed that simultaneous multimodal recordings keep both shared and modality-specific structure, while Bolt et al. (2025) and Özbay et al. (2019) showed that low-frequency/global fMRI-linked components can also carry autonomic physiology. Therefore, a coupled common factor is not read here as the target neural variable unless the Fusion Card discloses shared-vs-specific decomposition and physiology-side calibration or abstention.

This site blocks that shortcut too. Rohaut et al. (2024) showed that adding modalities in acute brain injury can reduce prognostic uncertainty and improve prediction, which is a real bundle-level gain, while also warning that multimodal approaches can increase discrepancies across markers that lead to choice paralysis or biased decisions. Amiri et al. (2023) showed that direct same-sample multimodal comparison can shrink to a 48-patient complete-feature subset, and Manasova et al. (2026) showed that second-level multimodal classifiers can rely on missing-value substitution while still facing cross-centre transfer and higher pairwise disagreement in minimally conscious or improving patients. Therefore, a multimodal gain is read here as bundle-performance evidence under a declared availability, transfer, and discordance-handling regime, not as automatic same-subject cross-stack state identification. The shortest route is Wiki: multimodal integration basics, then Verification: Fusion Card and Verification: Human Proxy Composition Card.

The front door was still too coarse if it jumped from a vague label such as human proxy evidence straight to how close to readout the paper sounded. The current entrance rule is now explicit. Route family comes first. Lucchetti et al. (2025) define a five-metabolite 1H-MRSI similarity scaffold, whereas Li et al. (2025) define a 7 T dynamic deuterium kinetic route with blood-input modeling. Zhao et al. (2020), Petitclerc et al. (2021), and Petitclerc et al. (2026) split choroid-plexus perfusion, blood-to-CSF transport, and joint BBB-versus-BCSFB exchange. Villemagne et al. (2022), Hiraoka et al. (2025), and Tyacke et al. (2018) split SMBT-1 target-validation / quantification from I2BS PET. Biechele et al. (2023), Ogata et al. (2025), and Yan et al. (2025) split TSPO, CSF1R, and COX-2. Only after that split does the second question become meaningful: what human-proxy role is the paper actually playing? On this site, Johansen et al. (2024) is a healthy atlas / cohort-prior route, Finnema et al. (2018) is a same-subject baseline / repeatability route, Smart et al. (2021) is a within-subject activation-change boundary, and Holmes et al. (2022) is a 24 h intervention-response boundary. Only after route family and route role are typed do proxy class, operational maturity, calibrator role, composition, and state continuity become meaningful. The shortest route is WBE 101: human observability ladder, then Wiki: measurement-stack observability and claim ceilings, then Wiki: Human Proxy Composition and Route Maturity. It also keeps visible the still-missing whole-brain in vivo routes for current presynaptic release-machinery / active-zone nanostructure state, current transcription / chromatin state, current phospho-signaling / second-messenger state, ECM / PNN gate state, branch-local proteostasis, branch- or bouton-specific cargo delivery, current chloride set point, and local mitochondrial positioning. In other words, "a human nanoscale paper exists" is not silently rephrased here as "living-human state-complete measurement is close and already field-ready."

The remaining weakness at the front door was that the warning above was still mostly verbal. A reader moving quickly could still see a long list of human advances and silently compress it into one composite progress bar. The current primary literature is sharper than that. Even representative front-rank rows already differ in direct observable, time axis, and operating burden, so they still stop at different ceilings.

The remaining front-door weakness was that recent barrier-side MRI / PET, clearance transport MRI, astrocyte-related PET, and target-defined neuroimmune PET papers could still sit close enough to look like one convergent human maintenance-controller meter. The primary literature does not support that shortcut. Petitclerc et al. (2026) separate BBB-versus-BCSFB water transport in a small healthy-volunteer MRI study, while Chung et al. (2025) quantify tracer-specific BBB permeability but explicitly leave human ground truth and test-retest as future work. Hiraoka et al. (2025) show that [¹⁸F]SMBT-1 quantification still depends on scan window, reference region, and model choice, while Tyacke et al. (2018) established an I₂BS route with a different pharmacological profile. Biechele et al. (2023), Ogata et al. (2025), and Yan et al. (2025) then keep TSPO, CSF1R, and COX-2 apart as different neuroimmune targets rather than one reusable inflammation scalar. Finally, Hirschler et al. (2025) and Dagum et al. (2026) remain transport-side mobility / efflux routes with their own imaging or model burden. Therefore, this site now keeps the four lanes below separate before any bundle claim is read strongly.

This site now blocks that shortcut too. Terceros et al. (2026), Dewa et al. (2025), and Bukalo et al. (2026) strengthen local rodent causal routes for transcriptional stabilization and astrocyte-supported memory representations. But the strongest current human-side routes answer different questions: Villemagne et al. (2022) and Tyacke et al. (2018) are target-defined astrocyte-related PET routes, Lim et al. (2025) is a respiration-conditioned CSF net-flow route, Yoo et al. (2025) is an exercise-conditioned contrast-influx / meningeal-lymphatic-flow route, Hirschler et al. (2025) is a CSF-mobility MRI route, and Dagum et al. (2026) is a model-based overnight biomarker-efflux route. These papers do not share one species, one spatial unit, one direct observable, one intervention regime, or one controller identity. Therefore, this site does not add "causal relevance in rodents" and "a human proxy improved" and call the responsible controller measured in humans. The shortest route is WBE 101: causal relevance versus human observability plus Wiki: Human Proxy Composition.

This site now blocks that shortcut too. Yiu et al. (2014) constrains relative excitability as an allocation-bias route, Hadzibegovic et al. (2025) constrains early neocortical engram-excitability plasticity, Benoit et al. (2025) constrains axon-initial-segment dynamics during associative fear learning, and Hengen et al. (2016) constrains firing-rate set-point / recovery control across sleep and wake. Tallman et al. (2025) adds a human hippocampal single-unit allocation-linked route in epilepsy patients, while Huber et al. (2013), Kuhn et al. (2016), Khatri et al. (2025), Fehér et al. (2026), plus Zrenner et al. (2018) remain living-human perturbation-conditioned proxy routes rather than direct whole-brain readouts of AIS geometry, ion-channel distribution, or cell-specific recovery controllers. Therefore, one excitability paper is not read here as if the current excitability landscape or the recovery controller had already been measured, and even a human local-unit result is not read as a whole-brain controller map. The shortest route is WBE 101: hidden state at the entry point, WBE 101: human observability ladder, and Wiki: intrinsic excitability / homeostatic-set-point route card.

This site now blocks that shortcut too. Ngo et al. (2013) constrains phase-locked auditory slow-oscillation stimulation, Baxter et al. (2023) show that strong SO / spindle effects can coexist with no added memory benefit when stimulation disturbs sleep continuity, Whitmore et al. (2022) show that cue benefit depends on ample and undisturbed N3 sleep, Schreiner et al. (2021) constrains endogenous scalp decoding around aggregated slow-oscillation / spindle events, Schreiner et al. (2023) adds a respiration-linked physiology gate for SO-spindle coupling and reactivation strength, Geva-Sagiv et al. (2023) constrains an intracranial closed-loop synchrony intervention in epilepsy patients whose benefit depends on precise timing, Schreiner et al. (2024) constrains spindle-locked ripple support for human memory reactivation, Whitmore et al. (2024) show that cue-induced sleep disruption does not have the same consequence for week-old memories as for newly formed ones, Jourde et al. (2025) constrains thalamocortical spindle-phase dependence of auditory stimulation, Deng et al. (2025) constrains a time-structured NREM window, and Duan et al. (2025) plus Shin et al. (2025) constrain item-level and difficulty-selective variability rather than uniform overnight strengthening. Therefore, sleep happened, oscillations increased, a cue was delivered, overnight memory changed, and replay-coupling matched are not one claim on this site. The route now has to disclose sleep-integrity / disturbance burden, NREM substate / physiology gate, event class, and memory subset / age, not only cue timing. The shortest route is Verification: maintenance-state error budget plus Wiki: sleep replay route card.

This site now blocks that shortcut too. Naganawa et al. (2021) showed that SV2A PET depends on tracer and quantification route, Johansen et al. (2024) is a healthy-human atlas, Shatalina et al. (2024) is a task/cognition association study, Smart et al. (2021) showed that brief visual activation does not measurably change [¹¹C]UCB-J binding, and Holmes et al. (2022) found no measurable overall SV2A change 24 h after ketamine despite symptom improvement. Meanwhile, Molnár et al. (2016), Sakamoto et al. (2018), and Emperador-Melero et al. (2024) show that release-site number, docked-vesicle architecture, and active-zone nanostructure / priming-site assembly are separate presynaptic objects. Therefore, a synaptic-density PET result is not read here as a direct meter of current synaptic efficacy, release-site number, docked-vesicle architecture, active-zone nanostructure / priming-site assembly, current release competence, or rapid plasticity. The shortest route is Wiki: PET measurement-model caution plus Wiki: SV2A / synaptic-density PET route card.

This site now blocks that shortcut too. A myelin paper can be about learning-dependent oligodendrogenesis, node / internode / periaxonal timing control, developmental plasticity brake, remyelination-linked functional recovery, a tract-scale transmission-speed estimate, or a myelin-sensitive / tissue-health-sensitive MRI ratio proxy, and those are not the same inferential object. The human side is also not one meter. Arshad et al. (2017) showed that calibrated T₁w/T₂w can be reliable while still having low criterion validity against myelin water fraction, Hagiwara et al. (2018) showed stronger agreement between SyMRI and MT_sat than with T₁w/T₂w, van Blooijs et al. (2023) estimated tract-scale transmission speed from tractography plus qT₁ and found heterogeneous delays across cortico-cortical pathways, Baadsvik et al. (2024) demonstrated myelin-bilayer mapping only in two healthy volunteers on high-performance hardware, Chen et al. (2025) showed that conventional MT remains orientation-dependent, Galbusera et al. (2025) showed that qT1 but not MWF or MTR distinguished demyelinated from remyelinated cortical lesions in a histology-linked design, and Colaes et al. (2026) showed that T₁w/FLAIR had only weak associations with MWF and supports a broader tissue-health reading rather than a myelin-specific one. Therefore, a human myelin map, tract-speed estimate, or T₁w/FLAIR ratio is not read here as one interchangeable myelin-state meter, not read as per-axon timing-state ground truth, and not read as proof that healthy myelin-state was completely restored. The shortest route is WBE 101: human observability ladder plus Wiki: myelin / oligodendrocyte route card.

Recent primary papers require a narrower reading here as well. Rangaraju et al. (2014) and Underwood et al. (2023) constrain presynaptic ATP-linked support / respiration, Rangaraju et al. (2019), Divakaruni et al. (2018), Bapat et al. (2024), and Hu et al. (2025) constrain dendritic positioning / fission and synaptic nano-organization, Vishwanath et al. (2026) constrains mitochondrial Ca²⁺-efflux tuning of long-term memory, while human ³¹P routes already split further: Ren et al. (2015) constrains resting ATP / metabolite / pH balance, Ren et al. (2017) constrains MT-based CK and Pi→ATP exchange flux, Guo et al. (2024) constrains whole-brain intracellular NAD content, Kaiser et al. (2026) constrains task-evoked NAD⁺ dynamics in a visual-cortex voxel, and Li et al. (2025) remains a deuterium kinetic-rate route. Therefore, a ³¹P or deuterium result is not read here as local mitochondrial-state ground truth, and an animal mitochondrial-plasticity paper is not read here as solved human observability. The shortest route is WBE 101: human observability ladder plus Wiki: bioenergetic / mitochondrial route card.

This site now blocks that shortcut too. Lucchetti et al. (2025) defined a five-metabolite gray-matter parcel-similarity graph, Guo et al. (2025) showed that high-resolution 1H-MRSI metabolite-distribution mapping is a separate object with its own acquisition / reconstruction burden, Ren et al. (2015) quantified resting ATP synthesis, phosphorus metabolites, and intra-/extracellular pH with ³¹P-MRS in 12 resting human brains, Ren et al. (2017) estimated PCr→ATP and Pi→ATP exchange fluxes with a band-inversion / MT 5-pool model, Guo et al. (2024) mapped whole-brain intracellular NAD content at 7 T, Kaiser et al. (2026) detected task-evoked NAD⁺ dynamics in a functionally localized occipital voxel, Li et al. (2025) used 7 T dynamic DMRSI plus blood-input and kinetic modeling to map CMR_Glc, CMR_Lac, V_TCA, and T_max in five healthy participants under a repeated same-brain / same-setup / same-parameter operating point, and Karkouri et al. (2026) produced absolute HDO / Glc / Glx / Lac maps and rate maps at 7 T in a mixed 12-healthy / 5-glioblastoma workflow whose abstracted post-glucose healthy slice was only two volunteers. Ahmadian et al. (2025) then showed that [6,6'-²H₂]glucose dose materially changes downstream metabolite visibility, and Bøgh et al. (2024) showed that repeatability depends on protocol and time point, with the strongest whole-brain repeatability at 120 min in that 3 T clinical DMI setup. Therefore, one spectroscopy paper is not read here as if biochemical similarity, high-resolution 1H-MRSI metabolite distribution, 31P metabolite / pH balance, 31P exchange-flux, 31P NAD-content mapping, localized functional 31P NAD-dynamics, deuterium absolute metabolite mapping / quantification, and kinetic rate imaging were one solved route, and even the deuterium rows are not treated as dose-, timing-, or protocol-invariant by default. The shortest route is WBE 101: human observability ladder, Verification: Observability Budget, and Wiki: human maintenance-state ladder.

This site now blocks that shortcut too. Pizzorusso et al. (2002) constrains adult plasticity-window reopening, Frischknecht et al. (2009) constrains AMPA-receptor mobility and short-term plasticity under matrix constraint, Nguyen et al. (2020) constrains microglia-driven ECM remodeling for memory consolidation, Alexander et al. (2025) constrains cell-type-specific CA2 versus PV PNN roles across hippocampal memory tasks, Mehak et al. (2025) constrains an age-linked CA2 rescue route, and Lehner et al. (2024) remains human ex vivo hippocampal histology. Therefore, a ChABC rescue, a CA2-specific aggrecan paper, or a human pathology map is not read here as direct ground truth of the current whole-brain ECM plasticity gate. The shortest route is WBE 101: hidden state at the entry point plus Wiki: ECM / PNN route card.

This site now blocks that shortcut too. Santoni et al. (2024) constrains chromatin-linked allocation eligibility before memory-trace formation, Traunmüller et al. (2025) constrains a time-resolved, region- and cell-type-specific chromatin / gene-expression response after novel-environment exposure, Guan et al. (2009) constrains a histone-acetylation / HDAC route for memory formation, Gulmez Karaca et al. (2020) constrains an engram-ensemble DNA-methylation route for consolidation stability, Bharadwaj et al. (2014) constrains higher-order chromatin looping linked to neuronal gene expression, Terceros et al. (2026) constrains a thalamocortical transcriptional cascade with distinct roles across the memory-stabilization window, and Coda et al. (2025) constrains locus-specific epigenetic editing of memory expression in defined engram cells. Sun et al. (2024), Mukamel & Yu (2025), and Sun et al. (2025) further show that memory-related transcriptomic signatures remain sensitive to animal-level independence and analysis choices. These papers do not share one direct observable, one persistence timescale, or one human observability ceiling: chromatin accessibility, histone chemistry, DNA-methylation control, higher-order looping, and locus-specific editing are different object families on this site. Therefore, a cell-type atlas, one-shot DEG list, spatial transcriptomic cluster, histone-acetylation intervention, methylation result, or chromatin-loop paper is not read here as if it already fixed the current plasticity-competent program or the memory-stabilization controller. The shortest route is WBE 101: hidden state at the entry point plus Wiki: transcription / chromatin route card.

This site now blocks that shortcut too. Wang et al. (2015) is a splice-isoform route whose downstream object is chromatin / transcriptional control, Dai et al. (2019) is a splice-dependent transsynaptic receptor-balance route, Shi et al. (2018) and Li et al. (2025) are different m6A translation-versus-degradation routes, Peterson et al. (2025) is an RNA-editing route for homeostatic AMPAR composition, and Joglekar et al. (2024) remains an atlas ceiling rather than a living-human in vivo readout. Therefore, a DEG list, bulk RNA count, or long-read atlas is not read here as if it already fixed the current whole-brain RNA controller. The shortest route is WBE 101: hidden state at the entry point plus Wiki: post-transcriptional RNA route card.

Recent primary papers require a narrower reading here as well. Lee et al. (2003) and Tomita et al. (2005) constrain phosphosite-specific plasticity gates, Havekes et al. (2016) and Vierra et al. (2023) constrain compartmentalized second-messenger routing, Altas et al. (2024) constrains region-specific phosphorylation with synapse-type relocalization in mouse and human samples, Rodriguez et al. (2025) constrains a single-site phospho-mutant causal memory intervention, and Biswas et al. (2023) remains a human ex vivo phosphoproteome atlas. Therefore, a transcriptome, proteome, or region-level atlas is not read here as if it already fixed the active phospho-controller or the current whole-brain signaling nanodomain. The shortest route is WBE 101: human observability ladder plus Wiki: phospho-signaling route card.

Recent primary papers require a narrower reading here as well. Frey & Morris (1997) and Shires et al. (2012) constrain tag / capture eligibility, Govindarajan et al. (2011) constrains branch-level integration of protein-synthesis-dependent LTP, Fonseca et al. (2006) and Parker et al. (2025) constrain synthesis-degradation / proteasome-capacity balance with memory consequences, Pandey et al. (2021) and Chang et al. (2024) constrain autophagy-linked plasticity routes, and Lee et al. (2022) plus Thomas et al. (2025) constrain turnover-resistant persistence or a candidate tag substrate. Therefore, a current weight estimate, a local translation clue, or a generic proteostasis sentence is not read here as if late stabilization, reconsolidation, or maintenance completeness were already fixed. The shortest route is WBE 101: hidden state at the entry point plus Wiki: local proteostasis / synaptic-tagging route card.

Recent primary papers require a narrower reading here as well. Park et al. (2006) and Correia et al. (2008) constrain postsynaptic AMPAR / recycling-endosome delivery during LTP, Maas et al. (2009), Uchida et al. (2014), and Wong et al. (2024) constrain transport-path state and local vesicle confinement, Nakayama et al. (2017), Liau et al. (2023), and Espadas et al. (2024) constrain dendritic / synaptic RNA-granule organization and spine-targeted RNA support, de Queiroz et al. (2025) constrains a distinct axonal RNA localization route in a mature in vivo memory circuit, and Aiken & Holzbaur (2024) constrains presynaptic cargo delivery patterned by axonal microtubule dynamics in human neurons. Therefore, a synaptic RNA-granule paper is not read here as proof of axonal RNA targeting, and a local translation paper, a weight estimate, or a cargo image in one compartment is not read here as proof that the correct receptors, RNA cargoes, or presynaptic components reached the correct branch, spine, or bouton. The shortest route is WBE 101: hidden state at the entry point plus Wiki: cargo-transport / cytoskeletal trafficking route card.

Recent primary papers require a narrower reading here as well. Glykys et al. (2014) constrains local chloride set point, Heubl et al. (2017) constrains activity-dependent KCC2 regulation, Byvaltsev et al. (2023) constrains perisynaptic K⁺ clearance by reverse-mode KCC2, Alfonsa et al. (2025) constrains sleep-wake-history-dependent E_GABAA shifts and LTP induction, and Forsberg et al. (2022) constrains a human CSF ion route. Human sodium MRI is also not one meter: Qian et al. (2012) is mm-class tissue-sodium mapping, Fleysher et al. (2013) derived ISMF / ISC / ISVF from combined SQ+TQF imaging, Rodriguez et al. (2022) reported a repeatable normalized sodium density-weighted route, Azilinon et al. (2023) showed that TSC and the short-component fraction f can diverge across epileptogenic and noninvolved tissue, and Qian et al. (2025) separated mono-/bi-T₂ sodium signals under a specialized multi-echo model. Therefore, a CSF ion assay or sodium MRI result is not read here as one interchangeable ionic-state meter, not read as local chloride-state ground truth, and not read as proof that the intra- versus extracellular sodium partition is already routine whole-brain observability. The shortest route is WBE 101: hidden state at the entry point plus Wiki: ionic / chloride route card.

This site now blocks that shortcut too. Galarreta & Hestrin (1999) constrains fast-spiking interneuron gap-junction networks, Anastassiou et al. (2011) constrains endogenous field-driven spike-timing bias, Graydon et al. (2014) constrains local extracellular-volume-fraction geometry and neurotransmitter dilution, Kilb et al. (2006) and Lauderdale et al. (2015) constrain osmotic extracellular-space contraction / edema-linked excitability shifts, Burman et al. (2023) constrains state-dependent inhibitory driving-force regime in active cortex, Yang et al. (2024) constrains activity-dependent electrical-synapse deployment that fuels persistent oscillations, Selfe et al. (2024) constrains local inhibitory driving force but only with a specialized local optical route, and Xie et al. (2013) constrains sleep-linked interstitial-space expansion in mice. Human evidence remains bounded and route-split: Voldsbekk et al. (2020) gives a wakefulness-related diffusion-MRI clue about extra-axonal versus extracellular volume, Örzsik et al. (2023) gives a sleep-conditioned higher-order diffusion / glymphatic clue under a within-subject sleep-deprivation-plus-zolpidem regime, and Feld et al. (2026) is useful as a perturbation clue that electrical coupling can matter for spindle-to-slow-oscillation coordination. Therefore, a chemical connectome, nominal inhibitory edge list, or human diffusion / sleep-memory perturbation result is not read here as ground truth of local electrical coupling, extracellular-space geometry / diffusion barrier, or electrotonic regime. The shortest route is WBE 101: hidden state at the entry point plus Verification and Wiki: electrical-state route card.

The 2026-03-31 recheck tightened one more point at the front door: proxy-rich is not the same as same-subject state identification. Johansen et al. (2024) is a 33-person SV2A atlas, Lucchetti et al. (2025) is a 51-adolescent five-metabolite similarity graph with 13-person site replication, Li et al. (2025) is a 7 T dynamic kinetic route in five healthy participants, Baadsvik et al. (2024) is a two-volunteer myelin proof-of-principle, Hirschler et al. (2025) is a 7 T CSF-mobility route with 20-person whole-brain rest maps, and Dagum et al. (2026) is a 39-participant randomized crossover trial interpreted through a compartmental model. These are not interchangeable pieces of one already field-ready whole-brain state meter. The current primary literature also forces a sharper operational rule. Bøgh et al. (2024), Finnema et al. (2018), and Wirsich et al. (2021) show why route-local repeatability and cross-centre portability are different questions. A second correction is now explicit at the front door as well: the same named quantity is not yet the same validated row. Morgan et al. (2024) showed that DP-ASL and ME-ASL give materially different BBB water-exchange estimates with inconsistent age dependence in the same cohort, and Bøgh et al. (2024) showed that a repeatable 3 T DMI operating point still remains a route-local result rather than automatic equivalence to specialized 7 T deuterium rate or absolute-quantification routes. Vafaii et al. (2024), Chen et al. (2025), Bolt et al. (2025), and Epp et al. (2025) then show why even same-session rows can still mix common and divergent structure, autonomic coupling, and even opposite signs. Rohaut et al. (2024) and Manasova et al. (2026) then show why average multimodal gain can coexist with marker discrepancies or higher pairwise disagreement in harder groups. This site therefore requires a Human Proxy Composition Card before several living-human proxy rows may be promoted together, including separate disclosure of proxy class, operational maturity, and calibrator role, plus a robustness gate that now explicitly includes method-family non-equivalence, a shared-driver / effective-window gate, and an increment-plus-disagreement gate. A second operational split is now explicit at the front door as well: Johansen et al. (2024) is a healthy atlas / cohort-prior route, Snellman et al. (2024) is a cross-sectional risk-contrast route, Finnema et al. (2018) is a same-subject baseline / repeatability route, Smart et al. (2021) is a within-subject activation-change boundary, and Holmes et al. (2022) is a 24 h intervention-response boundary. Those do not define one reusable synaptic-density row or one interchangeable bundle role. The shortest explanations are in WBE 101: human-proxy composition rule and Wiki: Human Proxy Composition and Route Maturity.

This site now blocks that shortcut more narrowly. A result can be same-subject or same-brain and still fail to be one state sample if the bridge crosses from live measurement to later fixation / ex vivo follow-up or from one day / regime to another. Bosch et al. (2022) and MICrONS Consortium et al. (2025) show that correlative same-brain workflows carry local landmarks, targeted subvolumes, or local structure-function correspondences rather than one global state object. Gallego et al. (2020), Noda et al. (2025), Van De Ville et al. (2021), and Di et al. (2021) then show that different population-level witnesses such as latent manifolds, representational maps, and fingerprint features can remain stable on different timescales while raw units or amplitudes drift. Karpowicz et al. (2025), Wilson et al. (2025), and Wairagkar et al. (2025) further show that stable use across time can depend on alignment, recalibration, or only a short fixed-decoder horizon. Therefore, same-subject wording is not enough here unless the paper also discloses a State-Continuity Bridge Card that names the carried object / witness, the tolerance / failure rule, acquisition order, elapsed time, regime continuity, coordinate transfer, rescue route, and residual drift ceiling. The shortest route is WBE 101: human observability ladder, Verification: State-Continuity Bridge Card, and Wiki: State-Continuity Bridge.

This site now keeps mixed arousal proxies such as pupil / HRV, local transmitter sensors, receptor / transporter atlas priors, selected occupancy PET routes, and challenge-limited displacement / release PET routes on separate rungs. The primary literature makes that separation necessary: Reimer et al. (2016) is a mixed-proxy route, Neyhart et al. (2024) is a local ACh route with explicit axon-activity and clearance constraints, Hansen et al. (2022) is a receptor / transporter atlas prior, Wong et al. (2013) is an occupancy route, and Koepp et al. (1998) plus later displacement studies are challenge-limited release proxies. That separation is there so "some neuromodulatory evidence exists" is not silently rephrased as "the current whole-brain transmitter state was measured." Start with WBE 101: human observability ladder, then see Wiki: neuromodulatory proxy ladder and Wiki: neuromodulatory route card.

This site now treats vascular transfer state as a separate audit item for hemodynamic modalities. In other words, a BOLD or fNIRS amplitude difference is not read here as a neural difference by default unless the paper also reports a vascular-state / cerebrovascular-reactivity calibration route or abstains explicitly. The shortest route is Verification: Observability Budget, then Wiki: Observability and Claim Ceilings by Measurement Stack and Wiki: Basics of Multimodal Integration.

This site does not treat wPLI, source-space connectivity, Granger-style directed metrics, or transfer-entropy labels as leak-proof communication maps or causal circuits by name alone. Volume conduction, source leakage, ghost interactions, pipeline dependence, and observation-only limits remain separate audit items. The shortest route is FAQ: how to read connectivity claims, then EEG 101 and Wiki: from observation to estimation.

This site blocks that shortcut too. Luria et al. (2024) expose posterior support and alternative configurations for focal-source hypotheses, Tong et al. (2025) expose debiased estimation and inference for sparse spatial-temporal sources, and Feng et al. (2025) expose empirical-Bayesian uncertainty for extended-source extent. Upstream physics remains separate: Vorwerk et al. (2024) showed that EEG localization shifts with tissue-conductivity uncertainty, and Vorwerk et al. (2026) showed that conductivity estimation can reduce uncertainty for many epilepsy-style source-analysis cases without erasing source-location exceptions at the brain base. Validation boards remain split as well: Pascarella et al. (2023), Unnwongse et al. (2023), and Hao et al. (2025) validate focal or clinical operating regimes, not one universal board for focal, sparse, extended, and spontaneous sources. Therefore, this site does not read new inverse method as one generic truth upgrade. Before an anatomical source claim is raised, the paper must disclose source regime / target object, uncertainty object, forward-model uncertainty route, and named validation board / operating regime. The shortest route is Wiki: from observation to estimation and Verification: Observability Budget.

This site blocks that shortcut too. Smith et al. (2011) showed that lag-based fMRI methods perform poorly and that functionally inaccurate ROIs are extremely damaging to network estimation, Barnett & Seth (2017) showed detectability black spots under subsampling, Vink et al. (2020) showed that resting-state EEG functional connectivity explains less than 10% of TMS-evoked propagation variance, Novelli et al. (2025) showed that slow BOLD sampling can still induce spurious Granger-causal inference even when realistic HRF variability alone need not do so, and Yan et al. (2026) showed that latent confounders remain an active challenge in biological network reconstruction. Therefore, on this site, a DCM or effective-connectivity result is not promoted beyond a model-conditioned causal hypothesis unless it also discloses observed-subsystem closure / latent-confound audit, node-definition policy, sampling / transformation sensitivity, and perturbation or external validation. The shortest route is FAQ: how to read DCM / effective-connectivity claims, then Wiki: effective-connectivity route card and Verification: Observability Budget.

This site now separates Landauer lower bounds, tissue-level energy budgets, irreversibility of coarse-grained neural time series, and model-based entropy-flow estimates. So a paper that reports entropy production, arrow-of-time, or irreversibility is not read here as a direct measurement of microscopic dissipation, whole-brain energy cost, or WBE readiness unless it also discloses its signal route, coarse-graining / hidden-degree risk, estimator family, null / surrogate control, quantity type, stability / nuisance sensitivity, physiology-bridge quality when energetic language is used, and cost isolation. Poudel et al. (2024) showed that small motion can materially alter visibility-graph structure and that only low-motion subsets reached moderate-to-high test-retest reliability for selected metrics, while Chen et al. (2025) showed with simultaneous EEG-PET-MRI that temporal coupling across modalities can coexist with non-identical spatial organization and state trajectories. A zero current or a small asymmetry score is therefore not read here as near-equilibrium, operationally stable, or energetically grounded unless hidden-cycle, memory-order, robustness, and bridge-quality risk are also disclosed. The shortest route is FAQ: how to read thermodynamic claims, then Wiki: thermodynamic grounding basics.

The public pages, including this one, are information portals for quickly seeing what is currently known, what remains unresolved, and where to read next. If you want to learn from the background upward, follow the wiki links at the top of each page.

If you want a site-wide view organized into three modes, getting the overview, learning from the basics, and actually fixing/contributing, see Wiki: Three Ways To Use This Site.

The Summary Booklet condenses the main public pages into a short briefing format. GitHub Actions also generates an A4 PDF from it.

Verification, Roadmap, Perspective, WBE 101, and Datasets may sound similar, but they serve different roles. If you want to decide which public page should come first, see Wiki: Public Page Reading Guide.

This site's public pages separate what can be asserted from what must still remain provisional. If you want a one-page guide to reading known/unknown sections, accuracy assumptions, and external dependencies, see Wiki: How To Read "What Is Known / Not Yet Known".

Each public page begins with blocks such as "how to read this page," "who it is for," "accuracy assumptions," "what is currently known," and "check the basics in the wiki." If you want a one-page explanation of how to use those blocks in order, see Wiki: How To Read Public Page Headers.

If you want only the theory-side distinctions among WBE 101, Perspective, Framework, and Roadmap, see Wiki: Theory Page Reading Guide.

If you want only the practical distinctions among Verification, Datasets, the L0 practice section in Datasets, the casework section in Verification, and the proposal integration section in Issue, see Wiki: Practical Page Reading Guide.

If you want a fixed first set of 3-4 pages depending on whether you start with overview, theory, practice, literature, or participation, see Wiki: The First 30 Minutes By Goal.

For an up-to-date external resource hub, see the Carbon Copies Foundation. As of March 16, 2026, their home page highlights the December 2025 newsletter, the Brain Emulation Challenge overview, the February 22, 2025 Brain Emulation Challenge workshop, the Memory Decoding Journal Club, and the broader BrainGenix stack for WBE-oriented simulation and validation work.

After checking what is directly observed in the Observability Budget, move next to the Verification: latent-state error budget. That section fixes which still-unobserved states continue to dominate present error, including intrinsic excitability, current synaptic efficacy, delay / myelin, neuromodulatory specificity, glial / slow-state variables, and chronic unit identity. It also explains how to read the difference between a connectome-only baseline and an augmentation claim.

After the March 2026 audit, this site's answer is no longer "just add more modalities." The next bottleneck is experiment design that collapses competing internal-state solutions. Villaverde (2019) distinguished observability from identifiability, Prinz et al. (2004) showed that similar activity can arise from disparate parameters, Beiran & Litwin-Kumar (2025) showed that a small targeted recording set can collapse connectome-conditioned degeneracy, Langdon & Engel (2025) showed that models preserving causal interactions among task variables can recover behaviorally relevant computation that correlation-only reductions miss, White et al. (2016) showed that complementary experiments can expose omitted mechanisms rather than solve them, and Gevertz & Kareva (2024) showed that identifiability analysis can derive a minimally sufficient schedule instead of open-ended data accumulation. Therefore, this site now treats not only which extra measurement or perturbation would rule out the remaining alternatives, but also which identifiability objective selected it, whether it exposed model discrepancy, and what minimum-sufficiency design would have been enough, as first-class design questions rather than post hoc appendices. Start with Verification: Identifiability Card and then return to WBE 101: what to build next.

The new Verification: Consciousness Readout Card forces each submission to disclose whether it actually passed construct-validity control, perturbation logging, same-cohort calibration, or deployability checks. That closes a remaining shortcut where labels such as IIT, PCI, criticality, or multimodal could sound stronger than the exposed evidence supports.

The Verification: Temporal Validity Card was added so longitudinal claims are not overread. It separates fixed decoder interval, state annotation, interface / decoder drift, recalibration burden, and transfer ceiling, so a same-day fit is not silently extended to cross-day stability or long-term deployability. The background logic is summarized in Wiki: state, trait, and drift.

The Verification: maintenance-state error budget separates temporal success from evidence about maintenance routes. It reports controller state, sleep / wake history, sleep-integrity / disturbance burden, NREM substate / physiology gate, sleep architecture / replay-coupling state, timing support, thermal-state, bioenergetic / mitochondrial support, neurovascular-unit / BBB / pericyte support, glial substrate-routing, astrocyte-state, and clearance / immune proxy class in separate fields, so a same-day fit or a cross-day hold is not automatically read as a maintenance-consistent or remote-memory-relevant claim. Baxter et al. (2023), Whitmore et al. (2022), Schreiner et al. (2023), Geva-Sagiv et al. (2023), Schreiner et al. (2024), and Deng et al. (2025) together show why a night with sleep, preserved sleep continuity, and consolidation-ready oscillatory / physiological coupling are different claims. The background is summarized in Wiki: homeostatic plasticity and maintenance state.

Recent primary papers require a narrower reading here too. Suzuki et al. (2011) is about lactate-shuttle support, Silva et al. (2022) is about glial ketone-body export under starvation, Pavlowsky et al. (2025) is about intensive-learning glia-to-neuron fatty-acid flux, and Greda et al. (2025) is about apoE / sortilin-dependent neuronal lipid uptake and fuel-choice gating when glucose is limited. These rows do not share one fuel object, one supplier/sink pair, one regime trigger, one transport route, or one human observability ceiling. On this site, glial substrate routing therefore stays separate from both neuronal mitochondrial state and astrocyte ensemble state. The shortest route is Verification: maintenance-state error budget plus Wiki: glial substrate-routing route card.

Recent primary papers require a narrower reading. Cahill et al. (2024) is about minute-scale cortical astrocyte-network encoding, Williamson et al. (2025) is about hippocampal ensemble recall, Dewa et al. (2025) is about multiday stabilization, Bukalo et al. (2026) is about amygdala fear-state representation, Villemagne et al. (2022) is the first-in-human SMBT-1 paper that established an MAO-B-selective astrocyte-related PET route with pharmacological blockade, Villemagne et al. (2022) then measured reactive astrogliosis across the Alzheimer disease spectrum within that tracer family, Hiraoka et al. (2025) showed that SMBT-1 quantification still depends on named scan-window / reference-region choices relative to kinetic modeling, Mesfin et al. (2026) added a whole-body biodistribution burden for the same tracer family, Matsuoka et al. (2026) showed that SL25.1188 in AD also depends on its own simplified arterial-free quantification route, Tyacke et al. (2018) plus Livingston et al. (2022) show that an I₂BS route behaves differently and can vary with region and impairment stage, and Best et al. (2026) shows that even SL25.1188 MAO-B binding can shift with cohort severity and daily cigarette use. On this site, astrocyte-state therefore means more than support background, but these papers still do not share one direct observable, one molecular target, one tracer family, one quantification route, or one claim ceiling, and they still do not license a jump to direct human whole-brain memory readout. The shortest route is Verification: maintenance-state error budget plus Wiki: astrocyte route card.

Recent primary papers require a narrower reading here as well. Louveau et al. (2015) and Ahn et al. (2019) established meningeal-lymphatic drainage routes, Kim et al. (2025) showed that a meningeal-lymphatics-microglia axis can regulate synaptic physiology, and a second human lane is now target-defined neuroimmune PET rather than transport alone: Biechele et al. (2023) showed why TSPO is not a universal human microglial activation-state meter, Wijesinghe et al. (2025) then validated TSPO PET as a microglial biomarker in PSP, Horti et al. (2022) and Ogata et al. (2025) established first-in-human CSF1R PET routes, and Yan et al. (2025) quantified COX-2 in healthy human brain with celecoxib blockade. Fultz et al. (2019) then measured macroscopic CSF oscillations during human NREM sleep, Kim, Huang, & Liu (2025) measured parenchyma-CSF water exchange with MT spin labeling, Eide & Ringstad (2021) showed that sleep deprivation impairs molecular clearance in humans, Eide et al. (2023) linked intrathecal gadobutrol retention and population-pharmacokinetic CSF-to-blood clearance variables to different plasma biomarker patterns, Hirschler et al. (2025) measured region-specific CSF-mobility drivers, Lim et al. (2025) reported respiration-conditioned CSF net flow in awake humans while also cautioning that plane-specific 2D PC-MRI net flow does not by itself represent whole-brain bulk circulation, and Dagum et al. (2026) linked sleep-active physiology to overnight Aβ / tau clearance to plasma in healthy older adults using an investigational device and multicompartment model. On this site, drainage anatomy, microglia-related synaptic control, TSPO disease-context / validation-bounded PET, CSF1R route-setting PET, COX-2 enzyme-defined PET, macroscopic CSF oscillation, parenchyma-CSF water exchange, respiration-conditioned net-flow MRI, intrathecal tracer retention / CSF-to-blood clearance, human CSF-mobility MRI, and model-based human biomarker efflux are not treated as one solved row. Therefore, clearance / immune support is not passive cleanup, and the current human-side gains now split between transport-side observables and target-defined neuroimmune PET, but neither lane identifies a local immune-controller or synapse-specific maintenance mechanism. The shortest route is Verification: maintenance-state error budget plus Wiki: clearance / immune route card.

This site now blocks that shortcut too. Cahill et al. (2024), Williamson et al. (2025), Dewa et al. (2025), and Bukalo et al. (2026) are local rodent causal routes; Villemagne et al. (2022) is a first-in-human SMBT-1 target-validation route, Villemagne et al. (2022) is an AD-spectrum pathology-context route within that tracer family, Hiraoka et al. (2025) is a quantification-route paper, Matsuoka et al. (2026) is a simplified arterial-free SL25.1188 AD quantification route, Tyacke et al. (2018) plus Livingston et al. (2022) show that a human I₂BS route is a different target class, Best et al. (2026) shows that even SL25.1188 MAO-B binding can shift with cohort severity and cigarette use, Wijesinghe et al. (2025) is a disease-context TSPO PET validation route, Horti et al. (2022) is a first-in-human CSF1R PET route-setting paper, Yan et al. (2025) is an enzyme-defined COX-2 PET route, Hirschler et al. (2025) is a human CSF-mobility MRI route whose direct observable is mobility, and Dagum et al. (2026) is a randomized crossover, multicompartment-model route for overnight biomarker efflux to plasma. Those rows do not answer the same question. Therefore, citing them together is still not evidence that the current human astrocyte / immune controller of a given memory, circuit, or synapse was identified.

This site now blocks that shortcut as well. Bell et al. (2010), Pandey et al. (2023), Swissa et al. (2024), and Mai-Morente et al. (2025) show that pericyte / BBB biology can matter for capillary support, plasticity, and long-term memory. By contrast, human Padrela et al. (2025) is a multi-echo ASL water-exchange route whose apparent gray-matter age effect disappeared after CBF and ATT correction, Morgan et al. (2024) showed that DP-ASL and ME-ASL can return markedly different BBB water-exchange values and even inconsistent age dependence, mouse work by Ohene et al. (2019) showed that multi-TE ASL exchange time is sensitive to AQP4 loss at the blood-brain interface, Padrela et al. (2026) showed lower Tex in SCD / MCI and moderate WMH burden while amyloid-group differences did not survive age / sex adjustment, and Chung et al. (2025) is a tracer-specific dynamic PET permeability-surface-area route under kinetic-model assumptions. A distinct human blood-CSF-barrier lane also exists, and it is not one internal quantity either: Zhao et al. (2020) and Sun et al. (2024) constrain choroid-plexus perfusion, Petitclerc et al. (2021) constrains blood-to-CSF water transport, Anderson et al. (2022) constrains choroid-plexus water cycling, Wu et al. (2026) constrains apparent BCSFB exchange with scan-rescan repeatability, and Petitclerc et al. (2026) constrains joint BBB-versus-BCSFB ASL exchange in one acquisition. Those rows do not answer the same question, and they are also different from a measurement-side vascular-state / CVR audit and from downstream clearance-side mobility or efflux routes. Therefore, a clean BOLD / fNIRS nuisance audit is still not evidence that the relevant neurovascular-unit / BBB / pericyte state was matched, and even a human barrier-side proxy must disclose whether it constrained BBB water exchange, tracer-specific BBB transport, choroid-plexus perfusion, blood-to-CSF transport, DCE water cycling, or apparent BCSFB exchange.

The April 2026 deepening pass now treats thermal evidence as more than one class. Local operating-point physiology, field-potential confound, sequence-timing perturbation, device-heating artifact, human passive / task-linked macro thermometry, and human perturbation-conditioned thermal routes do not answer the same question. Therefore, a stable field potential, a same-day fit, or a clean sequence-timing result is not read here as thermal-state-matched unless the claim names its thermal route and what remained latent. The shortest route is Verification: maintenance-state error budget plus Wiki: thermal route card.

The Verification: Neural Contribution Card was added to stop overreading brain-to-text and speech-decode demos. The current primary literature already separates within-subject semantic reconstruction, candidate-bank segment retrieval, known-onset word decoding, and prompt-conditioned generation. Therefore, this card fixes task regime, timing / segmentation regime, language model / prompt / candidate set, no-brain / LM-only / permuted-brain / no-text-prompt baselines, subject route / cooperation, and online vs offline so "a string came out" is not silently rephrased as unrestricted neural reconstruction. For the shortest entry explanation, see FAQ: how to read brain-to-text claims.

The 2026-03-18 addendum reflects the fact that Chaibub Neto et al. (2019) showed identity confounding when repeated measures are not participant-disjoint, Wang et al. (2020) and Di et al. (2021) showed time-robust person identification from resting-state EEG, and Gibson et al. (2022) summarized strong subject-driven EEG variation. For that reason, this landing page now treats subject / session fingerprint as an independent shortcut. Results that let windows from the same raw recording cross train/test, or that can be reproduced from subject / session metadata alone, are not read here as target-specific biomarkers or neural readouts. Start with the Verification: Specificity & Shortcut Card and Roadmap R6: personalization.

This site now blocks a second shortcut at the front door as well. Jiang et al. (2024) already framed mismatched electrodes, unequal sample lengths, varied task design, and low signal-to-noise ratio as core EEG barriers even in LaBraM, and Lee et al. (2025) then found only marginal gains, about 0.5%, over conventional deep baselines despite much larger parameter counts. The newer foundation-model papers then tighten the stop rule further: Han et al. (2025) target channel-permutation equivariance, Chen et al. (2025) target coordinate-based adaptation across heterogeneous devices and more than 150 layouts, and El Ouahidi et al. (2025) push setup-agnostic pretraining to more than 60,000 hours from 92 datasets and 25,000 subjects. Those are real advances in recording-frame compatibility. They are still not proof that different montages, coordinate routes, and reference families already preserve one shared physiology-side representation. Ma et al. (2026) then show that strong EEG foundation models can still generalize poorly when subject-level supervision is limited unless extra adaptation structure is added, while Xiong et al. (2025), Liu et al. (2026), and Lahiri et al. (2026) further show that protocol inconsistency, adaptation regime, and pretraining-population diversity can reverse which model looks strongest. This site also keeps source status visible: peer-reviewed route papers, official benchmark operations / postmortems, and arXiv preprints are not read as one homogeneous evidence pile. The official EEG Challenge (2025) homepage, data page, rules, submission page, and leaderboard then show why benchmark object, downsampling / inference rules, checkpoint disclosure, and later organizer corrections all change what a score means. Therefore, this site does not promote a foundation-model headline beyond qualified transfer / representation evidence unless it also discloses pretraining corpus identity and overlap audit, benchmark object / supervision unit, coordinate route, reference family, omitted-channel policy, adaptation regime or label budget, benchmark provenance / governance, and shortcut resistance. The shortest route is FAQ: foundation-model / leaderboard reading, then EEG 101 and Verification: Pretraining Card.