Wiki

Wiki: Basics of verification infrastructure

Before celebrating a score, fix the ruler, the route card, and the failure log.

Mind Uploading Research Project

Public Page Updated: 2026-04-03 Beginner guide (updated with the human proxy family-internal split)

How to use this page

Read this first to avoid getting lost

This page explains from the beginning why Mind-Upload focuses on standards, benchmarks, preregistration, route cards, and audits. The goal is to show why technical progress becomes comparable only when the benchmark object, evidence pathway, and any human proxy bundle or sequential bridge are disclosed rather than left implicit.

  • Data alone does not become comparable progress.
  • Benchmark meaning still depends on split regime, metric bundle, and current benchmark rules rather than the task name alone.
  • Standards, benchmarks, preregistration, route cards, and audits each solve a different failure mode.
  • Modern neuroscience results often need claim-specific route cards, not only a generic score sheet.
  • For EEG source imaging, the route card now also has to separate validation class, source regime, inverse family / uncertainty object, montage / coverage policy, and focal-centre versus source-extent benchmark object rather than stopping at one localization score.
  • For tractography, the route card now has to type acquisition / harmonization, cortical endpoint assignment, graph construction, uncertainty, and external calibration rather than only the modality name.
  • For effective connectivity, the route card now also has to disclose observed-subsystem closure / latent-confound audit, node-definition policy, and sampling / transformation sensitivity rather than stopping at candidate-model family.
  • Proxy-rich human evidence still needs a Human Proxy Composition Card before it is read as same-subject state closure.
  • Human proxy bundles now need three axes first: proxy class, operational maturity, and calibrator role.
  • On the human side, those three axes now start only after the family-internal comparison family is fixed: deuterium absolute-quantification versus kinetic-rate imaging, BBB versus blood-CSF-barrier / choroid-plexus routes, astrocyte PET route-role families, and neuroimmune PET target families are different beginner rows.
  • Even after those axes are logged, a bundle rises only if it passes robustness, common-driver / quantity-bridge, and increment gates.
  • Same-subject or same-brain workflows still need a State-Continuity Bridge Card before they are read as same-state evidence.
Helpful next pages
Verification platform Casework Datasets and Hands-On Work
Best for
People who want to understand the logic of the Verification Commons from the beginning
Reading time
14-18 minutes
Accuracy note
The analogies here are for orientation only. Always return to the public specification pages to see the actual operational rules.

Relatively clear at this stage

What we know now

  • Comparable progress requires aligned inputs, evaluation, rules, and records.
  • Without preregistration and auditing, it becomes too easy to promote only favorable conditions.
  • Recent primary literature shows that decode, tractography, effective-connectivity, thermodynamic, closed-loop, human-proxy-composition, and bridge claims fail in different ways and therefore need different cards.
  • ESI claims still need named validation class, source regime, inverse family / uncertainty object, montage / coverage policy, and target object disclosed separately; a focal-centre result does not automatically settle source extent or propagation-rich reconstruction.
  • A tractography-derived connectome can still move with endpoint bias, graph construction, voxel size, q-space scheme, uncertainty routing, and calibration route, so comparison requires object typing before interpretation.
  • Effective-connectivity claims still need observed-subsystem closure / latent-confound audit, node-definition policy, and sampling / transformation sensitivity logged separately from model comparison and validation.
  • Proxy class, operational maturity, and calibrator role are different questions; a real human route may still calibrate only one bounded hidden-state family.
  • Human proxy bundles are not typed safely from family labels alone; deuterium, blood-CSF-barrier / choroid-plexus MRI, astrocyte PET, and neuroimmune PET each already split into different comparison families before proxy class, operational maturity, and calibrator role are assigned.
  • A proxy bundle still needs repeatability / transfer, shared-driver / quantity-bridge, and increment disclosure before it rises above the strongest single row.
  • Same-subject or same-brain wording can secure specimen identity while still leaving state continuity unresolved.

Still unresolved beyond this point

What we still do not know

  • A complete WBE benchmark stack does not yet exist in finished public form.
  • Which public ESI board should become the default beginner comparison for focal-centre versus source-extent targets under matched montage / geometry controls remains unresolved.
  • Which combination of human proxy rows and bridge validations could ever support stronger same-subject continuity claims remains unresolved.
  • Which additional requirements would be sufficient for future L4 or L5 claims remains unresolved.

Learn the basics

Check the basics in the wiki

What the wiki is for

The wiki is a learning aid. For the project's official current synthesis, success criteria, and operating rules, always return to the public pages.

Why the ruler comes first

If two groups use different inputs, different split rules, different metric bundles, different hidden assumptions, different human proxy combinations, and different bridge assumptions, comparing their scores does not tell you who made more progress. The Verification Commons exists so that progress claims can survive comparison rather than only presentation.

2026-03-19 beginner update: generic audits were no longer enough

The older beginner explanation correctly said that standards and audits matter, but it still sounded as if one generic score sheet could cover every kind of neuroscience result. That is no longer safe enough. Different claim families now need different route cards because they fail for different scientific reasons.

2026-03-25 beginner tightening: proxy bundles and bridges are not side details

The current site had also become stricter in two ways that this beginner page still underplayed. First, several living-human proxy rows can all be real while still measuring different quantity types, spatial units, timescales, and model burdens. Second, same-subject or same-brain wording can still hide a sequential bridge across fixation, deformation, behavior, sleep/wake regime, or elapsed time. Those are scientific stop lines, not administrative add-ons.

2026-03-26 beginner tightening: proxy bundles now need three axes and three gates

The remaining weakness was subtler. Even after telling readers that proxy bundles matter, this page still left too much room to think that listing several human rows side by side was already most of the work. The current site rule is stricter: a bundle must first disclose proxy class, operational maturity, and calibrator role, and then show that it passes robustness, common-driver / quantity-bridge, and increment gates. Without that, proxy-rich evidence remains scientifically meaningful but still below same-subject state closure.

2026-03-27 correction: route-family names were still too coarse

This beginner page also had to become stricter inside the human-bundle example itself. On the current site, 1H-MRSI similarity, 31P metabolite / pH balance, 31P MT exchange-flux, 31P NAD-content mapping, localized functional 31P NAD-dynamics, deuterium metabolite-mapping / absolute-quantification, and deuterium kinetic-rate imaging are different spectroscopy rows; myelin-water / MT-family / bilayer / qT1 remyelination-sensitive routes are not one myelin meter; BBB water-exchange MRI is not tracer-specific BBB PET transport; and MAO-B astrocyte PET is not I2BS astrocyte PET. If those splits are hidden, the bundle looks much closer to one state meter than the primary literature allows.

2026-03-28 correction: tractography route cards are about object typing, not only caution

This beginner page still treated tractography too generically for the current site rule. Gajwani et al. (2023) showed hub-location variability across 40 pipelines and 44 group reconstructions, He et al. (2024) showed filtering-dependent laterality shifts, McMaster et al. (2025) and Bramati et al. (2026) showed voxel-size and q-space dependence even before interpretation, and Manzano-Patrón et al. (2025) plus Zhu et al. (2025) showed that uncertainty propagation and MRI-microscopy fusion improve only parts of the tractography chain. Therefore, a tractography route card on this site now has to name acquisition / harmonization, cortical endpoint assignment, graph construction, instability / uncertainty, and external calibration, not just say that tractography is "coarse."

2026-03-29 correction: effective-connectivity route cards are about closure and observation conditions, not only model family

The remaining weakness in this beginner verification page was that its effective-connectivity examples still sounded too close to candidate-model comparison plus reliability. The primary literature does not support that shortcut. Smith et al. (2011) showed that lag-based fMRI approaches perform poorly and that functionally inaccurate ROIs are especially damaging to network estimation. Barnett & Seth (2017) showed that subsampling can create detectability black spots and sweet spots. Vink et al. (2020) showed that resting-state EEG functional connectivity explains less than 10% of TMS-evoked propagation variance. Villaverde et al. (2019) foregrounded full input-state-parameter observability rather than silently treating the observed subsystem as closed, Novelli et al. (2025) showed that slow BOLD sampling can still induce spurious Granger-causal inference, Jafarian et al. (2024) showed that reliability can be strong only under tightly matched MEG sessions, and Yan et al. (2026) showed that latent confounders remain an active reconstruction problem. Therefore, this beginner page now treats the effective-connectivity route card as requiring observed-subsystem closure / latent-confound audit, node-definition policy, sampling / transformation sensitivity, validation, reliability window, and abstention, not just a modeling family name.

2026-03-30 correction: ESI route cards also need validation-class, montage, and target-object disclosure

This beginner page still had one technical gap after the recent site-wide ESI updates. It still left too much room to think that improved source imaging could be summarized by one localization score or one density label. The current primary literature does not support that shortcut. Horrillo-Maysonnial et al. (2023) showed that a targeted 33-36 electrode montage can reach 54/58 sublobar concordance against an 83-electrode HD montage while still degrading for tangential generators. Rong et al. (2025) showed that DeepSIF can remain comparatively stable from 75 to 16 electrodes, but that does not erase route dependence. Unnwongse et al. (2023) and Hao et al. (2025) showed that direct validation still depends on coverage geometry, conductivity assumptions, source depth, and source power. Pascarella et al. (2023) compared ten ESI methods on an in-vivo focal benchmark and showed that method ranking still depends on input parameters, while Feng et al. (2025) targeted extended-source reconstruction, which is not the same benchmark object as focal-centre localization. Therefore, on this site, an ESI route card now has to name validation class, source regime, same-geometry controls including montage / coverage policy, and whether the benchmark scores centre, extent, overlap, or propagation.

2026-03-31 correction: ESI route cards also need inverse-family and uncertainty-object disclosure

One more beginner shortcut still remained after the target-object update. The current literature does not support reading all inverse methods as if they were estimating one common hidden object with one generic uncertainty term. Luria et al. (2024) describe a probabilistic focal-support family, Tong et al. (2025) describe a sparse debiased-inference family, and Feng et al. (2025) target extended-source reconstruction with uncertainty quantification. Those routes do not return the same target object or the same uncertainty object. At the same time, Vorwerk et al. (2024) and Vorwerk et al. (2026) show that forward-model uncertainty remains a separate audit, while Jahromi et al. (2026) add a deep-source pediatric phantom board rather than one more interchangeable validation checkbox. Therefore, on this site, an ESI route card now also has to name the inverse family, the target object, the uncertainty object, the forward-model uncertainty route, and the named validation board before a comparison is treated as reusable progress.

2026-04-03 correction: family-internal comparison family now comes before the three axes

One more beginner shortcut remained. This page already said that human proxy bundles need proxy class, operational maturity, and calibrator role, but it still let readers start from family labels that were already too broad. The current primary literature does not support that shortcut. Karkouri et al. (2026) and Li et al. (2025) are different deuterium routes (absolute quantification versus kinetic-rate imaging), while Bøgh et al. (2024) shows that repeatability is still protocol- and time-point-specific. Zhao et al. (2020), Petitclerc et al. (2021), and Petitclerc et al. (2026) do not report one generic barrier meter; they report choroid-plexus perfusion, blood-to-CSF transport, and simultaneous BBB-versus-BCSFB exchange. Matsuoka et al. (2026) adds a simplified SL25.1188 MAO-B route distinct from older SMBT-1 or I2BS astrocyte-related PET, while Biechele et al. (2023), Ogata et al. (2025), and Yan et al. (2025) show that TSPO, CSF1R, and COX-2 are not one interchangeable neuroimmune PET row. Therefore, on this site, the beginner rule is now stricter: family-internal comparison family first, three axes second, bundle gates third.

The five parts of comparable progress

Part Role Simple analogy
Data standard Aligns the structure and metadata of the input. The same answer sheet format.
Benchmark Aligns the task, split rules, metric bundle, and benchmark governance. The same exam and scoring rubric.
Preregistration / registry Locks success/failure conditions before the result is known. The rules announced before the exam starts.
Route card Discloses the evidence pathway, assumptions, omitted alternatives, and abstention boundary for a specific claim family or bridge class. The answer key that says which reasoning steps were allowed.
Audit / model card / log Records weaknesses, drift, failures, recalibration, benchmark postmortems, and unresolved limits. The report card plus the error log.

What goes wrong when one part is missing

What is missing What goes wrong
Standard You cannot be sure that people are even talking about the same input.
Benchmark Different groups can claim victory on different tasks, split regimes, metric bundles, or organizer-conditioned benchmark versions.
Preregistration It becomes too easy to keep only the convenient conditions and present them as the intended target all along.
Route card A result is overread beyond the evidence class or bridge class that produced it.
Audit Weaknesses, drift, shortcuts, recalibration burden, benchmark revisions, and abstentions disappear behind one score.
Two newer beginner stop lines this page now has to carry

The current site can no longer leave two failure modes to the deep pages only. One is composition failure: several living-human proxy rows can all be valid yet still fail to compose into one same-subject latent-state estimate. The other is bridge failure: the same subject or same brain can still be sampled under a changed preparation, coordinate frame, or physiological regime. A beginner page that omits those stop lines now understates the scientific uncertainty.

Human proxy bundles now need three axes and three gates

The remaining beginner weakness was that saying proxy bundles matter still left too much room for readers to imagine that listing several living-human rows side by side was already most of the work. The current site rule is stricter. Johansen et al. (2024), Lucchetti et al. (2025), Ren et al. (2015), Li et al. (2025), Karkouri et al. (2026), Padrela et al. (2025), Chung et al. (2025), Villemagne et al. (2022), Tyacke et al. (2018), Hirschler et al. (2025), and Dagum et al. (2026) do not all measure the same quantity, do not run at the same burden, and do not safely calibrate the same hidden-state family. Therefore, a beginner page now has to separate what class of proxy the row is, how operationally mature the route is, and what bounded calibrator role it can safely play.

One more beginner rule now comes before those three axes themselves. A family label can already hide several different comparison families. On this site, deuterium absolute-quantification is not deuterium kinetic-rate imaging, BBB water-exchange is not blood-CSF-barrier / choroid-plexus transport, SMBT-1 / SL25.1188 / I2BS are not one astrocyte PET row, and TSPO / CSF1R / COX-2 are not one neuroimmune PET row. Therefore, the beginner bundle workflow is now ordered as follows: family-internal comparison family first, proxy class / operational maturity / calibrator role second, and bundle gates third.

Axis What it asks What goes wrong if omitted
Proxy class What is directly observed: density proxy, biochemical similarity scaffold, high-resolution metabolite-distribution map, 31P metabolite / pH balance route, 31P MT exchange-flux route, 31P NAD-content map, localized functional 31P NAD-dynamics route, deuterium absolute-quantification or kinetic-rate route, quantity-defined myelin route, BBB water-exchange route, tracer-specific BBB transport route, blood-CSF-barrier / choroid-plexus perfusion or transport route, astrocyte PET target / route-role family, neuroimmune PET target family, support-state / mobility proxy, model-based efflux route, or destructive local scaffold? Different quantity types are silently compressed into one progress bar.
Operational maturity How specialized, small-cohort, model-heavy, or centre-bound is the route at its actual operating point? A proof-of-principle route can be mistaken for portable whole-brain measurement.
Calibrator role Which hidden-state family does the route safely constrain, and what still remains latent? A real route can be overread as broadly calibrating maintenance-state completeness.

Even that three-axis typing is still not enough. Finnema et al. (2018) showed route-specific SV2A PET test-retest reproducibility of 3-9% for regional VT, Holiga et al. (2018) showed fMRI reliability ranging from poor to excellent, and Wirsich et al. (2021) showed that some simultaneous EEG-fMRI connectome relationships can reproduce across 72 subjects from four centres spanning 1.5T to 7T. Amiri et al. (2023) then showed that a multimodal acute-DoC dataset can shrink from 87 enrolled patients to 63 with both EEG and fMRI and 48 for direct same-feature comparison, while Manasova et al. (2026) showed that adding modalities can improve prediction while still increasing inter-modality disagreement in clinically important groups. Finally, Vafaii et al. (2024) showed both common and divergent cross-modal organization, and Epp et al. (2025) showed that significant task BOLD changes can coexist with opposite oxygen-metabolism changes. Therefore, a human proxy bundle now rises only if it passes three separate gates.

Gate What must be shown Ceiling if it is missing
Robustness Show row-level repeatability at the actual operating point, distinguish route-local repeatability from cross-centre transfer, and disclose the real complete-case slice. The bundle remains setup-bound, centre-bound, or complete-case-bound evidence.
Common-driver / quantity-bridge Show that the rows are not only synchronized, but interpretable on an explicitly named biological axis after shared-driver audit. Cross-row agreement remains proxy-rich correlation or shared-factor evidence rather than one validated state variable.
Increment over the strongest single row Show what the bundle adds beyond the best individual row under matched cohort, condition, and reading rule. Row diversity remains richer description, not same-subject state closure.

If the bundle is sequential rather than same-session, the State-Continuity Bridge Card is added on top. Same-subject still solves specimen identity, not state continuity.

Why modern neuroscience needs claim-specific route cards

Claim family Why a generic score sheet is too weak Card this site now asks for
Decode / biomarker / speech High scores can come from subject/session fingerprint, task priors, metadata leakage, or language-model support rather than the target neural variable alone. Neural Contribution Card plus Specificity & Shortcut Card.
EEG source imaging / inverse reconstruction A single localization score is too weak because field formation, validation class, source regime, inverse family / uncertainty object, montage / coverage policy, and focal-centre versus source-extent benchmark object can all change what the result means. Observability Budget plus Inverse-Solver Agreement Log.
Tractography / connectome Hub maps and connectome metrics can shift with acquisition / harmonization, cortical endpoint assignment, graph construction, instability / uncertainty, and external calibration; the tractography graph is not one fixed object by default. Tractography route card.
Effective connectivity / DCM The output still depends on candidate model space, observed-subsystem closure / latent-confound audit, node-definition policy, sampling / transformation sensitivity, recovery, validation, and reliability window. Effective-connectivity route card.
Thermodynamic irreversibility Different papers compute different quantities from different signal routes and coarse-grainings. Irreversibility / thermodynamic route card.
Closed loop / BCI Latency alone does not tell you which sensory, motor, and interoceptive loops were preserved or omitted. Intervention Card plus Body / Environment Boundary Card.
Living-human proxy bundle SV2A PET comparison families, 1H-MRSI similarity versus high-resolution metabolite-distribution mapping, 31P balance / exchange-flux / NAD-content / functional NAD-dynamics, deuterium absolute-quantification versus kinetic-rate imaging plus operating-point burden, quantity-defined myelin MRI, BBB water-exchange versus tracer-specific BBB transport versus blood-CSF-barrier / choroid-plexus routes, astrocyte PET target / route-role families, neuroimmune PET target families, and clearance-support routes measure different quantity / target / transport types, spatial units, timescales, model burdens, and safe calibrator roles; even a well-typed bundle still needs robustness, shared-driver / quantity-bridge, and increment disclosure. Human Proxy Composition Card.
Sequential same-subject / same-brain bridge Specimen identity does not by itself fix state continuity across fixation, deformation, behavior, sleep/wake regime, or elapsed time. State-Continuity Bridge Card, plus Temporal Validity Card when the bridge crosses hours to days.

Why this matters especially for WBE

WBE is especially vulnerable to level substitution. It is easy to describe an L1 decoding result as if it were approaching L4 continuity, or to describe one route-conditioned measurement as if it had solved a whole class of hidden states. Verification infrastructure is therefore not administrative overhead. It is part of the scientific content because it prevents the evidence class from changing silently after the result is known.

Representative primary-literature reasons

Chaibub Neto et al. (2019) and Di et al. (2021) show why decode scores need shortcut audits. Horrillo-Maysonnial et al. (2023), Rong et al. (2025), Unnwongse et al. (2023), Hao et al. (2025), Pascarella et al. (2023), and Feng et al. (2025) show why ESI claims need a route card that types validation class, source regime, montage / coverage policy, and benchmark object rather than only a localization headline. Thomas et al. (2014) and Maier-Hein et al. (2017) show why tractography claims need route disclosure, while Gajwani et al. (2023), He et al. (2024), McMaster et al. (2025), Bramati et al. (2026), Manzano-Patrón et al. (2025), and Zhu et al. (2025) show why the tractography graph itself is still pipeline-, protocol-, uncertainty-, and calibration-conditioned. Penny et al. (2004) and Rosa et al. (2012) show why candidate-model disclosure matters for effective connectivity, but Smith et al. (2011), Barnett & Seth (2017), Vink et al. (2020), Villaverde et al. (2019), Novelli et al. (2025), Jafarian et al. (2024), and Yan et al. (2026) show why route cards still have to type observed-subsystem closure / latent-confound audit, node-definition policy, sampling / transformation sensitivity, and a reliability window before a directed graph is promoted beyond a model-conditioned causal hypothesis. Lynn et al. (2021) and Ishihara & Shimazaki (2025) show why thermodynamic language hides multiple estimator families. Musall et al. (2019) and Flesher et al. (2021) show why closed-loop results need a boundary card rather than latency alone. Johansen et al. (2024), Lucchetti et al. (2025), Ren et al. (2015), Li et al. (2025), Karkouri et al. (2026), Padrela et al. (2025), Chung et al. (2025), Villemagne et al. (2022), Tyacke et al. (2018), Finnema et al. (2018), Holiga et al. (2018), Wirsich et al. (2021), Amiri et al. (2023), Vafaii et al. (2024), Epp et al. (2025), and Manasova et al. (2026) show why current living-human proxy rows do not all measure the same object, do not share one robustness level, and do not automatically add up to same-subject state closure. Lu et al. (2023), Idziak et al. (2023), Benisty et al. (2024), and Egger et al. (2024) show why same-subject or same-brain sequencing still needs bridge validation.

References

  1. Chaibub Neto, E., Pratap, A., Perumal, T. M., et al. (2019). Detecting the impact of subject characteristics on machine learning-based diagnostic applications. npj Digital Medicine, 2, 99. doi:10.1038/s41746-019-0178-x
  2. Di, Y., An, X., Zhong, W., Liu, S., & Ming, D. (2021). The time-robustness analysis of individual identification based on resting-state EEG. Frontiers in Human Neuroscience, 15, 672946. doi:10.3389/fnhum.2021.672946
  3. Horrillo-Maysonnial, A., Avigdor, T., Abdallah, C., et al. (2023). Targeted density electrode placement achieves high concordance with traditional high-density EEG for electrical source imaging in epilepsy. Clinical Neurophysiology, 156, 262-271. doi:10.1016/j.clinph.2023.08.009
  4. Rong, J., Sun, R., Joseph, B., Worrell, G., & He, B. (2025). Deep learning-based EEG source imaging is robust under varying electrode configurations. Clinical Neurophysiology, 175, 2010730. doi:10.1016/j.clinph.2025.04.009
  5. Unnwongse, K., Van Klink, N., Tousseyn, S., et al. (2023). Validating EEG source imaging using intracranial electrical stimulation. Brain Communications, 5(1), fcad023. doi:10.1093/braincomms/fcad023
  6. Hao, S., Zhao, H., Feng, Z., et al. (2025). HD-EEG source imaging with simultaneous SEEG recording in drug-resistant epilepsy. Epilepsia, 66(11), 4451-4464. doi:10.1111/epi.18552
  7. Pascarella, A., Mikulan, E., Sciacchitano, F., et al. (2023). An in-vivo validation of ESI methods with focal sources. NeuroImage, 277, 120219. doi:10.1016/j.neuroimage.2023.120219
  8. Feng, Z., Guan, C., & Sun, Y. (2025). Block-Champagne: A novel Bayesian framework for imaging extended E/MEG source. IEEE Transactions on Medical Imaging. doi:10.1109/TMI.2025.3642620
  9. Vorwerk, J., Wolters, C. H., & Baumgarten, D. (2024). Global sensitivity of EEG source analysis to tissue conductivity uncertainties. Frontiers in Human Neuroscience, 18, 1335212. doi:10.3389/fnhum.2024.1335212
  10. Luria, G., Viani, S., Pascarella, A., et al. (2024). The SESAMEEG package: a probabilistic tool for source localization and uncertainty quantification in M/EEG. Frontiers in Human Neuroscience, 18, 1359753. doi:10.3389/fnhum.2024.1359753
  11. Tong, P. F., Yang, H., Ding, X., et al. (2025). Debiased estimation and inference for spatial-temporal EEG/MEG source imaging. IEEE Transactions on Medical Imaging. doi:10.1109/TMI.2024.3506596
  12. Vorwerk, J., Rampp, S., Wolters, C. H., & Baumgarten, D. (2026). Potential of EEG and EEG/MEG skull conductivity estimation to improve source analysis in presurgical evaluation of epilepsy. Journal of Neural Engineering, 23(1). doi:10.1088/1741-2552/ae2f01
  13. Jahromi, S., Sdoukopoulou, G., Chikara, R. K., et al. (2026). 3D printed pediatric head phantom for assessing deep epileptic sources localization. Computers in Biology and Medicine, 182, 111449. doi:10.1016/j.compbiomed.2026.111449
  14. Thomas, C., Ye, F. Q., Irfanoglu, M. O., et al. (2014). Anatomical accuracy of brain connections derived from diffusion MRI tractography is inherently limited. PNAS, 111(46), 16574-16579. doi:10.1073/pnas.1405672111
  15. Maier-Hein, K. H., Neher, P. F., Houde, J.-C., et al. (2017). The challenge of mapping the human connectome based on diffusion tractography. Nature Communications, 8, 1349. doi:10.1038/s41467-017-01285-x
  16. Gajwani, M., Oldham, S., Pang, J. C., Arnatkevičiūtė, A., Tiego, J., Bellgrove, M. A., & Fornito, A. (2023). Can hubs of the human connectome be identified consistently with diffusion MRI? Network Neuroscience, 7(4), 1277-1304. doi:10.1162/netn_a_00324
  17. He, Y., Hong, Y., Wu, Y., et al. (2024). Spherical-deconvolution informed filtering of tractograms changes laterality of structural connectome. NeuroImage, 297, 120904. doi:10.1016/j.neuroimage.2024.120904
  18. McMaster, E. M., Newlin, N. R., Rudravaram, G., et al. (2025). Harmonized connectome resampling for variance in voxel sizes. Magnetic Resonance Imaging, 122, 110424. doi:10.1016/j.mri.2025.110424
  19. Manzano-Patrón, J. P., Deistler, M., Schröder, C., et al. (2025). Uncertainty mapping and probabilistic tractography using Simulation-based Inference in diffusion MRI: A comparison with classical Bayes. Medical Image Analysis, 103, 103580. doi:10.1016/j.media.2025.103580
  20. Zhu, S., Huszar, I. N., Cottaar, M., et al. (2025). Imaging the structural connectome with hybrid MRI-microscopy tractography. Medical Image Analysis, 102, 103498. doi:10.1016/j.media.2025.103498
  21. Bramati, I. B., Szczupak, D., Carneiro Monteiro, M., Meireles, F., Menezes Guimarães, D., Dean, R. J., Paul, L. K., & Tovar-Moll, F. (2026). Diffusion MRI sampling schemes bias diffusion metrics and tractography. Frontiers in Neuroimaging, 5, 1670604. doi:10.3389/fnimg.2026.1670604
  22. Penny, W. D., Stephan, K. E., Mechelli, A., & Friston, K. J. (2004). Comparing dynamic causal models. NeuroImage, 22(3), 1157-1172. doi:10.1016/j.neuroimage.2004.03.026
  23. Rosa, M. J., Friston, K., & Penny, W. (2012). Post-hoc selection of dynamic causal models. Journal of Neuroscience Methods, 208(1), 66-78. doi:10.1016/j.jneumeth.2012.04.013
  24. Smith, S. M., Miller, K. L., Salimi-Khorshidi, G., Webster, M., Beckmann, C. F., Nichols, T. E., Ramsey, J. D., & Woolrich, M. W. (2011). Network modelling methods for FMRI. NeuroImage, 54(2), 875-891. doi:10.1016/j.neuroimage.2010.08.063
  25. Barnett, L., & Seth, A. K. (2017). Detectability of Granger causality for subsampled continuous-time neurophysiological processes. Journal of Neuroscience Methods, 275, 93-121. doi:10.1016/j.jneumeth.2016.10.016
  26. Villaverde, A. F., Tsiantis, N., & Banga, J. R. (2019). Full observability and estimation of unknown inputs, states and parameters of nonlinear biological models. Journal of the Royal Society Interface, 16(156), 20190043. doi:10.1098/rsif.2019.0043
  27. Vink, J. J. T., Klooster, D. C. W., Ozdemir, R. A., Westover, M. B., Pascual-Leone, A., & Shafi, M. M. (2020). EEG Functional Connectivity is a Weak Predictor of Causal Brain Interactions. Brain Topography, 33(2), 221-237. doi:10.1007/s10548-020-00757-6
  28. Jafarian, A., Karadag Assem, M., Kocagoncu, E., et al. (2024). Reliability of dynamic causal modelling of resting-state magnetoencephalography. Human Brain Mapping, 45(10), e26782. doi:10.1002/hbm.26782
  29. Novelli, L., Barnett, L., Seth, A. K., & Razi, A. (2025). Minimum-Phase Property of the Hemodynamic Response Function, and Implications for Granger Causality in fMRI. Human Brain Mapping, 46(10), e70285. doi:10.1002/hbm.70285
  30. Yan, J., Zhang, S.-W., Zhang, C., Huang, W., Shi, J., & Chen, L. (2026). Dynamical Causality under Latent Confounders for Biological Network Reconstruction. IEEE Transactions on Pattern Analysis and Machine Intelligence. doi:10.1109/TPAMI.2026.3658839
  31. Lynn, C. W., Cornblath, E. J., Papadopoulos, L., et al. (2021). Broken detailed balance and entropy production in the human brain. PNAS, 118(47), e2109889118. doi:10.1073/pnas.2109889118
  32. Ishihara, K., & Shimazaki, H. (2025). State-space kinetic Ising model reveals task-dependent entropy flow in sparsely active nonequilibrium neuronal dynamics. Nature Communications, 16, 10852. doi:10.1038/s41467-025-66669-w
  33. Musall, S., Kaufman, M. T., Juavinett, A. L., Gluf, S., & Churchland, A. K. (2019). Single-trial neural dynamics are dominated by richly varied movements. Nature Neuroscience, 22, 1677-1686. doi:10.1038/s41593-019-0502-4
  34. Flesher, S. N., Downey, J. E., Weiss, J. M., et al. (2021). A brain-computer interface that evokes tactile sensations improves robotic arm control. Science, 372(6544), 831-836. doi:10.1126/science.abd0380
  35. Lu, X., Han, X., Meirovitch, Y., et al. (2023). Preserving extracellular space for high-quality optical and ultrastructural studies of whole mammalian brains. Cell Reports Methods. doi:10.1016/j.crmeth.2023.100520
  36. Idziak, A., Inavalli, V. V. G. K., Bancelin, S., Arizono, M., & Nagerl, U. V. (2023). The Impact of Chemical Fixation on the Microanatomy of Mouse Organotypic Hippocampal Slices. eNeuro. doi:10.1523/ENEURO.0104-23.2023
  37. Johansen, A., Beliveau, V., Colliander, E., et al. (2024). An in vivo high-resolution human brain atlas of synaptic density. Journal of Neuroscience. doi:10.1523/JNEUROSCI.1750-23.2024
  38. Lucchetti, F., Céléreau, E., Steullet, P., et al. (2025). Constructing the human brain metabolic connectome with MR spectroscopic imaging reveals cerebral biochemical organization. Nature Communications. doi:10.1038/s41467-025-66124-w
  39. Ren, J., Sherry, A. D., & Malloy, C. R. (2015). 31P-MRS of healthy human brain: ATP synthesis, metabolite concentrations, pH, and T1 relaxation times. NMR in Biomedicine. doi:10.1002/nbm.3384
  40. Li, X., Zhu, X.-H., Li, Y., et al. (2025). Quantitative mapping of key glucose metabolic rates in the human brain using dynamic deuterium magnetic resonance spectroscopic imaging. PNAS Nexus. doi:10.1093/pnasnexus/pgaf072
  41. Karkouri, J., Novoselova, M., Rodgers, C. T., et al. (2026). Absolute Quantification of Brain Deuterium Metabolic Imaging in Healthy Volunteers and Glioblastoma Patients at 7T. Magnetic Resonance in Medicine. doi:10.1002/mrm.70308
  42. Baadsvik, E. L., Weiger, M., Froidevaux, R., et al. (2024). Myelin bilayer mapping in the human brain in vivo. Magnetic Resonance in Medicine. doi:10.1002/mrm.29998
  43. Padrela, B. E., Slivka, M., Sneve, M. H., et al. (2025). Blood-brain barrier water permeability across the adult lifespan: A multi-echo ASL study. Neurobiology of Aging. doi:10.1016/j.neurobiolaging.2024.12.012
  44. Chung, K. J., Abdelhafez, Y. G., Spencer, B. A., et al. (2025). Quantitative PET imaging and modeling of molecular blood-brain barrier permeability. Nature Communications. doi:10.1038/s41467-025-58356-7
  45. Villemagne, V. L., Harada, R., Dore, V., et al. (2022). Assessing reactive astrogliosis with 18F-SMBT-1 across the Alzheimer disease spectrum. Journal of Nuclear Medicine. doi:10.2967/jnumed.121.263255
  46. Tyacke, R. J., Myers, J. F. M., Venkataraman, A., et al. (2018). Evaluation of 11C-BU99008, a PET Ligand for the Imidazoline2 Binding Site in Human Brain. Journal of Nuclear Medicine. doi:10.2967/jnumed.118.208009
  47. Hirschler, L., Runderkamp, B. A., Decker, A., et al. (2025). Region-specific drivers of CSF mobility measured with MRI in humans. Nature Neuroscience. doi:10.1038/s41593-025-02073-3
  48. Benisty, H., Barson, D., Moberly, A. H., et al. (2024). Rapid fluctuations in functional connectivity of cortical networks encode spontaneous behavior. Nature Neuroscience. doi:10.1038/s41593-023-01498-y
  49. Egger, A., Bayon, C., d'Almeida, J., et al. (2024). Chrono-EEG dynamics influencing hand gesture decoding: a 10-hour study. Scientific Reports, 14, 20247. doi:10.1038/s41598-024-70609-x
  50. Dagum, P., Elbert, D. L., Giovangrandi, L., et al. (2026). The glymphatic system clears amyloid beta and tau from brain to plasma in humans. Nature Communications. doi:10.1038/s41467-026-68374-8
  51. Finnema, S. J., Nabulsi, N. B., Mercier, J., et al. (2018). Kinetic evaluation and test-retest reproducibility of [11C]UCB-J, a novel radioligand for positron emission tomography imaging of synaptic vesicle glycoprotein 2A in humans. Journal of Cerebral Blood Flow & Metabolism, 38(11), 2041-2052. doi:10.1177/0271678X17724947
  52. Holiga, S., Sambataro, F., Luzy, C., et al. (2018). Test-retest reliability of task-based and resting-state blood oxygen level dependence and cerebral blood flow measures. PLOS ONE, 13(11), e0206583. doi:10.1371/journal.pone.0206583
  53. Wirsich, J., Jorge, J., Iannotti, G. R., et al. (2021). The relationship between EEG and fMRI connectomes is reproducible across simultaneous EEG-fMRI studies from 1.5T to 7T. NeuroImage, 231, 117864. doi:10.1016/j.neuroimage.2021.117864
  54. Amiri, M., Fisher, P. M., Raimondo, F., et al. (2023). Multimodal prediction of residual consciousness in the intensive care unit: the CONNECT-ME study. Brain, 146(1), 50-64. doi:10.1093/brain/awac335
  55. Vafaii, H., Mandino, F., Desrosiers-Grégoire, G., et al. (2024). Multimodal measures of spontaneous brain activity reveal both common and divergent patterns of cortical functional organization. Nature Communications, 15, 229. doi:10.1038/s41467-023-44363-z
  56. Epp, S. M., Halani, S., Paquette, M., et al. (2025). BOLD signal changes can oppose oxygen metabolism across the human cortex. Nature Neuroscience. doi:10.1038/s41593-025-02132-9
  57. Manasova, V., et al. (2026). Multimodal multicentre investigation of diagnostic and prognostic markers in disorders of consciousness. Brain. doi:10.1093/brain/awaf412
  58. Bøgh, N., Vaeggemose, M., Schulte, R. F., et al. (2024). Repeatability of deuterium metabolic imaging of healthy volunteers at 3 T. European Radiology Experimental, 8, 21. doi:10.1186/s41747-024-00426-4
  59. Zhao, L., Taso, M., Dai, W., Press, D. Z., & Alsop, D. C. (2020). Non-invasive measurement of choroid plexus apparent blood flow with arterial spin labeling. Fluids and Barriers of the CNS, 17, 58. doi:10.1186/s12987-020-00218-z
  60. Petitclerc, L., Hirschler, L., Wells, J. A., et al. (2021). Ultra-long-TE arterial spin labeling reveals rapid and brain-wide blood-to-CSF water transport in humans. NeuroImage, 245, 118755. doi:10.1016/j.neuroimage.2021.118755
  61. Petitclerc, L., Durrant, H., Hirschler, L., Václavů, L., & van Osch, M. J. P. (2026). Simultaneous measurement of water transport across the blood-brain and blood-CSF barrier in the human brain with arterial spin labeling MRI. Journal of Cerebral Blood Flow & Metabolism. doi:10.1177/0271678X261429042
  62. Matsuoka, K., Matsuura, K., Kikuchi, T., et al. (2026). Quantification of monoamine oxidase B expression with 11C-SL25.1188 for imaging reactive astrocytes in patients with Alzheimer's disease. European Journal of Nuclear Medicine and Molecular Imaging. doi:10.1007/s00259-025-07542-2
  63. Biechele, G., Wind, K., Blume, T., et al. (2023). Translocator protein is a marker of activated microglia in rodent models but not human neurodegenerative diseases. Nature Communications, 14, 6040. doi:10.1038/s41467-023-40937-z
  64. Ogata, A., Ikenuma, H., Yasuno, F., et al. (2025). First-in-Human Study of [11C]NCGG401 for Imaging Colony-Stimulating Factor 1 Receptors in the Brain. Journal of Nuclear Medicine. doi:10.2967/jnumed.124.268699
  65. Yan, X., Noergaard, M., Morse, C. L., et al. (2025). PET Quantification in Healthy Humans of Cyclooxygenase-2, a Potential Biomarker of Neuroinflammation. Journal of Nuclear Medicine. doi:10.2967/jnumed.124.268525