{"id":"f185c561-7f66-4ab1-9deb-2c15c9799851","arxiv_id":"2608.06321","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":8,"one_line_summary":"An end-to-end simulation framework chains transcranial ultrasound pressure, tissue strain, brain temperature, and a multi-pathway Hodgkin-Huxley neuron to produce anatomy-registered per-voxel firing maps, demonstrated on a human skull model.","lead":"This paper presents an open-source computational pipeline that traces focused ultrasound from the skull surface to predicted per-neuron firing maps, coupling acoustic, elastic, and thermal physics with six candidate biophysical mechanisms. It aims to give the ultrasound neuromodulation field a common tool for comparing mechanisms and defining dose at the cellular level.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Firing-zone headline inherits the uncalibrated K_A·α product, and the paper's own sensitivity table shows the zone is not robust across the literature bracket; the abstract should carry the working-point caveat.","rationale":"The reader's weakest_assumption is the K_A·α strain-to-tension product, and the strongest_claim is the mechanism-discrimination/firing-zone central assertion. My concern is the same soft spot, but stated more precisely as a mismatch between the demonstrated working point (α = 1000) and the unambiguous-mechanism-attribution claim in §4.1. The paper's own supplementary (§S9.4.1) and Table 3 show the absolute firing-zone magnitude and the mechanism ranking at the canonical working point are not robust to the unmeasured gain: α ≲ 460 collapses firing to the focal voxel, α ≳ 2000 saturates, and ±25% on K_A moves spikes by -59%/+89%. So the headline volume and the size (though not the sign) of the S4/S5 mechanism differences are working-point-dependent. This is not an internal inconsistency; the paper flags it in the body ('working point on the literature-bracketed sweep rather than a calibrated absolute') and limitations. It is also not a consensus conflict with the six mechanism candidates. The correct outcome is to keep the CONDITIONAL verdict and require the concrete test before accepting the headline firing-zone number as quantitative evidence. I do not see a separate load-bearing concern beyond this one: the source classification, regression checks, AIS-first validation, and Oh et al. controls are real support, and the deferred in-vivo validation is honestly disclosed. My concrete test is a single re-run at the paper's own bracket endpoints because it directly determines whether the abstract's 8,500 mm³ number is a working-point artifact.","tokens_in":44974,"tokens_out":2376,"duration_ms":24777,"concrete_test":"Re-run the canonical Halle/dACC firing-zone calculation with α swept across the paper's own bracket at fixed K_A = 0.25 N/m (specifically α = 200, 460, 1000, 2000) and report the firing-voxel count and iso-25 volume for each. If the zone shrinks to ~1 voxel below α ≈ 460 and saturates the volume above α ≈ 2000, then the 8,500 mm³ headline is a working-point artifact; if the zone volume varies by less than ±25% across a realistic sub-bracket (e.g., α = 500-1500), the headline is more robust than the current sensitivity table implies.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central demonstration, the 'approximately 8,500 mm³' focal firing zone and the 5,468 firing voxels, collapses when the strain-to-tension gain K_A·α moves within the paper's own stated literature bracket. Eq. 6 sets T = K_A·α·ε_eq with both K_A and α recorded as modelling assumptions. The canonical run fixes α = 1000 (T_max = 6.52 mN/m, 2.42×T_1/2), but §S9.4.1 states that on the same Halle/dACC field α ≲ 460 leaves at most the focal voxel firing and α ≳ 2000 saturates the zone. Since K_A enters only as the product, a joint shift of {K_A, α} within the bracket that halves or doubles the product moves the focal tension from mid-knee to saturation, changing the firing-zone volume by large factors (Table 3 already shows ±25% on K_A alone gives -59%/+89% spike-count changes at fixed α). The claim in §4.1 that mechanism differences can be attributed 'unambiguously to the mechanism' is therefore too strong for the headline firing-zone numbers: the demonstrated mechanism discrimination is itself conducted at a hand-selected point on the steepest part of the Boltzmann curve, exactly where an unmeasured gain has its largest effect. The paper is honest about this in the body and supplementary, but the abstract quotes 8,500 mm³ without the working-point caveat, and the reader's verdict already flags this. The concern is not that unmeasured parameters exist; a framework paper may defer calibration. The load-bearing issue is that the paper's own quantitative headlines are fully determined by that deferred calibration, so the headline numbers do not yet constitute evidence about any mechanism.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports an open-source, end-to-end computational framework that couples nonlinear full-wave acoustic propagation, viscoelastic shear-wave propagation, Pennes bioheat diffusion, a linear strain-to-tension bilayer conversion, and a multi-compartment Hodgkin–Huxley neuron carrying six interchangeable mechanotransduction modules. The pipeline produces anatomy-registered per-voxel firing maps, demonstrated on a theta-burst transcranial sonication through a micro-CT human skull targeting the left dorsal anterior cingulate cortex. The headline demonstration reports a focal firing zone of approximately 8,500 mm^3 (5,468 firing voxels) at a chosen working point α=1000, and a mechanism-comparison sweep over five scenarios is used to compare Piezo1-only, cavitation, calcium/SK, TRP, and combined pathways. All parameters are classified by source, and a ±25% sensitivity analysis identifies the Piezo1 half-activation tension T_1/2 and the strain-to-tension product K_A·α as the dominant uncertainties.","tokens_in":45244,"tokens_out":6260,"duration_ms":63952,"significance":"The framework addresses a real gap: linking transcranial exposure metrics to per-voxel cellular outcomes, with explicit mechanism interchangeability and falsifiable spatial predictions. Strengths include the open-source implementation with unit tests and bit-for-bit regression checks, RK4 convergence verification, a per-parameter source classification, and the focus-to-ring topology under Piezo1 inactivation, which the authors correctly argue is robust to the uncalibrated tension scale. The main weakness is that the headline firing-zone numbers are fully determined by uncalibrated working-point choices (α=1000 and an elevated Piezo1 conductance), and the paper's own sensitivity analysis shows these numbers would collapse or saturate across the literature bracket. The framework is a useful scaffold, but the quantitative demonstration is not yet a calibrated prediction.","major_comments":[{"comment":"The headline claim of 'approximately 8,500 mm^3' firing volume and 5,468 firing voxels is computed at the α=1000 working point of Eq. (6), with K_A·α an uncalibrated modelling-assumption pair. The paper's own supplementary §S9.4.1 states that α≲460 leaves at most the focal voxel firing and α≳2000 saturates the zone. The abstract and conclusion quote the numeric result without the working-point caveat, which is misleading because the number is not robust within the paper's own stated literature bracket. Please re-frame the headline as a working-point illustration with the α-bracket range, or calibrate K_A·α before presenting quantitative predictions.","section":"Abstract, §3.2, Table 2"},{"comment":"The statement that differences in predicted firing between candidates can be attributed 'unambiguously to the mechanism rather than to incidental modelling choices' is too strong. The mechanism-comparison sweep in §3.3 operates at strain levels placing the focal voxel on the steep Boltzmann knee (T≈2.42×T_1/2 at the canonical point), where Table 3 shows ±25% in K_A changes spike count by −59%/+89%. Since α and K_A are unmeasured, the relative ordering and magnitudes of scenario differences are contingent on the chosen working point; the unambiguous-attribution claim is not supported by the presented sensitivity analysis.","section":"§4.1"},{"comment":"The sensitivity analysis reports only ±25% perturbations, but the dominant uncertainty, K_A·α, spans a 10-fold literature bracket (α∈[200,2000], §S9.4.1). A ±25% K_A shift therefore substantially understates the plausible range of firing outcomes. Please include a main-text tabulation (or prominent figure) of firing-voxel count and iso-volume across the full α bracket, so the reader can see the range of the headline numbers rather than a single point.","section":"Table 3, §3.4"},{"comment":"The canonical run elevates Piezo1 conductance density 'above the library defaults...to place the focal voxel at the Boltzmann working point.' This is a second free working-point choice, in addition to K_A·α, that directly sets the absolute firing count. The paper should disclose this tuning explicitly in the main text (not only in the table caption) and report the sensitivity of the firing zone to g_bar within its plausible physiological range, since the headline numbers depend on both choices.","section":"§3.2, Table 2 caption"}],"minor_comments":[{"comment":"The validations in §3.1 are largely directional/qualitative; please state this explicitly in the main text to avoid implying quantitative reproduction of the cited experiments.","section":"§3.1"},{"comment":"In panel (b), the per-panel numeric annotations are described as cumulative spike counts on the displayed slice; consider clarifying in the caption whether these are sums over the slice only, as the reader may otherwise compare them to the whole-volume counts in Table 2.","section":"Figure 8"},{"comment":"The three Heimburg–Jackson rows use a different output metric (dimensionless excitability ξ) than the other rows; add a footnote explaining this difference and the sign convention.","section":"Table 3"},{"comment":"There are several typographical spacing errors (e.g., 'millimetrespatialresolution') that should be corrected.","section":"Abstract"}],"recommendation":"major_revision","confidential_remarks":"The stress-test concern is valid: the paper's own body and supplementary are honest about the K_A·α dependence, but the abstract and conclusion overstate the robustness of the quantitative headline. I recommend a major revision that aligns the abstract/conclusion with the working-point caveats and adds the full-bracket sensitivity analysis to the main text. The paper is within scope for physics.med-ph and the open-source contribution is valuable."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Real talk: this is the first pipeline, as far as I can tell, that carries a transcranial pressure field all the way to per-voxel firing maps with six mechanism modules on a shared multi-compartment neuron, and the per-parameter source classification is a genuinely useful contribution. The code ships with unit tests, bit-for-bit regression, and RK4 convergence, so the numerical discipline is real. The focus-to-ring shape change under Piezo1 inactivation is a nice example of a falsifiable signature that does not depend on the absolute tension scale.\n\nWhere it gets soft: the 8,500 mm^3 firing zone is not a prediction about the brain; it is a consequence of setting alpha=1000 and elevating Piezo1 conductance to place the focal voxel at 2.42x T1/2 on the Boltzmann knee. Table 3 shows +/-25% on K_A alone moves spike counts by -59%/+89%. The supplementary is honest that alpha below ~460 collapses the zone to the focal voxel and above ~2000 saturates it. So the body and S9.4.1 spell this out, but the abstract quotes the number without the working-point caveat. That is the one place the paper overstates. The claim in Section 4.1 that mechanism differences can be attributed 'unambiguously to the mechanism' is too strong for the absolute numbers, though the relative S1-S5 comparison at a fixed working point is still informative.\n\nThe validations in Section 3.1 are mostly reproductions of qualitative directions, not head-to-head quantitative matches; the f-I rheobase mismatch is acknowledged. There are also 'Estimate' parameters in the synaptic/astrocytic pathways, so the mechanism discrimination is more plausibility ranking than proof. None of this is hidden.\n\nVerdict: this deserves serious peer review. The framework is a real step toward a common dose language, and the transparency is above the field norm. What needs to happen before the numbers count as evidence is calibration of the K_A*alpha product against mechano-current or elastography data and at least one head-to-head prediction against extracellular recordings. The abstract must carry the working-point caveat regardless.\n\nI'd bring this to reading group and would cite it for the pipeline design and the parameter taxonomy. Send it to referees; it will need revision, not rejection.","headline":"A genuinely useful framework whose headline firing zone is an uncalibrated working point; the abstract should carry that caveat.","tokens_in":45965,"tokens_out":2996,"would_cite":true,"duration_ms":32942,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"A computational pipeline now carries transcranial ultrasound from skull to per-voxel neuron firing maps, placing six proposed mechanisms on one shared neuron model.","keywords":["transcranial focused ultrasound","neuromodulation","Hodgkin-Huxley","mechanosensitive ion channels","strain-to-tension coupling","per-voxel firing maps","computational framework","ultrasound safety"],"falsifier":"Record high-density extracellular firing from cortex during a 500 kHz, roughly 0.5 MPa transcranial $\\theta$-burst exposure matching the simulated protocol, and compare the spatial extent and per-voxel spike counts against the predicted 5,468 firing voxels and approximately 8,500 $mm^{3}$ zone; if firing appears only near the acoustic focus when $\\alpha$ is low, or saturates across a much larger region when $\\alpha$ is high, the $\\alpha$ = 1000 working point is falsified, and paired shear-wave elastography with single-cell mechano-current recordings would provide the direct calibration needed to replace it.","tokens_in":1836,"feed_emoji":"🧠","tokens_out":3100,"duration_ms":79300,"temperature":0.7,"pith_summary":"This paper presents an end-to-end computational framework that predicts, voxel by voxel in an anatomically registered head volume, which neurons fire during transcranial focused ultrasound and through which biophysical pathway. It chains acoustic propagation, shear-wave displacement, bioheat diffusion, a strain-to-membrane-tension conversion, and a multi-compartment Hodgkin-Huxley neuron whose mechanosensitive, cavitation-coupled, calcium-coupled, thermosensitive, astrocytic, and synaptic modules are interchangeable. The central claim is that this is the first framework to exercise all six candidate mechanisms on a common neuron model with traceable per-parameter source classification, so that differences in predicted firing can be attributed to the mechanism rather than to incidental modelling choices. Applied to a theta-burst sonication through a human skull specimen targeting the dorsal anterior cingulate cortex, it predicts a focal firing zone of approximately 8,500 cubic millimetres and a peak firing rate of 300 Hz, while staying within consensus safety envelopes.","feed_headline":"Simulation maps skull ultrasound to an 8,500 mm3 firing zone","feed_subtitle":"Six proposed mechanisms run on one neuron model, making their firing predictions directly comparable and testable.","key_machinery":"The load-bearing object is the strain-to-tension conversion T = K_A * alpha * epsilon_eq, which turns voxel-scale von Mises equivalent strain into lipid-bilayer membrane tension, feeding a two-state Boltzmann-gated channel population (Piezo1 and K2P channels) embedded in a three-compartment Hodgkin-Huxley neuron with dendrite, soma, and axon-initial-segment compartments. Around this core, six candidate mechanisms are implemented as interchangeable modules on the same neuron, and every numerical parameter is classified by source as literature-anchored, calibrated, assumed, estimated, or derived, so sensitivity of predicted firing to each parameter is traceable.","core_discovery":"The paper's central contribution is a single volumetric spatio-temporal pipeline that maps a transcranial acoustic field to per-voxel neural firing maps, resolving firing jointly with the acoustic, elastic, and thermal field histories that drive it. The demonstration run, at the paper's chosen working point of the multi-scale coupling factor $\\alpha$ = 1000, predicts 5,468 firing brain voxels over a 20 ms ON window, with an iso-25 firing volume of about 8,523 $mm^{3}$, substantially larger than the acoustic -6 dB focal volume of 161 $mm^{3}$, because the strain field elongates the firing zone along the beam axis. On the same real strain field, the framework's mechanism comparison shows that a three-compartment dendrite-soma-AIS neuron is necessary for firing where a single-compartment baseline stays subthreshold, and that the Piezo1-inactivation/calcium/SK pathway reduces firing by about 30-33% at the canonical working point, producing a focus-to-ring spatial signature. The paper also reports that intramembrane cavitation and thermosensor pathways contribute marginally at the sub-MPa, body-temperature regime, while the astrocytic relay predicts a slow, duty-by-on-time accumulating drive that remains below recruitment threshold in this protocol.","pith_inferences":["Editorial inference: if K_A * alpha is calibrated by paired elastography and single-cell mechano-current recordings, the framework would turn the exposure-to-firing relationship into a quantitative fingerprint that could distinguish activation-driven from inactivation-dominated recruitment in living tissue.","Editorial inference: the predicted focus-to-ring redistribution under Piezo1 inactivation is a shape-based signature robust to the uncalibrated tension scale, and would be a stronger experimental test than absolute spike counts because it does not depend on the overall gain.","Editorial inference: the astrocytic relay's distinctive duty-times-on-time accumulation could be tested by comparing firing maps under protocols that hold total energy fixed but vary duty cycle and ON-window duration; the paper's model predicts the relay's contribution grows with cumulative ON time while channel pathways track per-pulse drive.","Editorial inference: because the pipeline accepts any compatible upstream pressure, displacement, and temperature fields, the same cellular stage could be coupled to other acoustic solvers or experimental field maps, making the firing-prediction layer a reusable comparator across the field."],"forward_implications":["Mechanism hypotheses for ultrasound neuromodulation can now be compared on a single neuron model and a single acoustic field, so disagreements between candidates are attributable to the mechanism rather than to different modelling choices.","Per-voxel firing maps give spatially resolved, falsifiable predictions that can be tested against high-density extracellular recordings in the same exposure conditions.","The framework ties acoustic exposure to cellular firing, enabling quantitative safety assessment that includes firing dose alongside the conventional thermal and mechanical indices.","The source-classified sensitivity table identifies the Piezo1 half-activation tension and the strain-to-tension product K_A * alpha as the dominant uncertainties, motivating targeted calibration experiments.","Cell-type-resolved mechanism identification becomes tractable by re-running the same field under different neuron-type profiles and comparing predicted firing topologies."],"supporting_citations":[{"why":"Supplies the heterogeneous nonlinear full-wave acoustic solver that produces the pressure field entering the pipeline.","marker":"[17]"},{"why":"Supplies the reduced viscoelastic shear-FDTD solver that converts radiation force into the tissue displacement and strain field.","marker":"[22]"},{"why":"Provides the cycle-averaged intramembrane-cavitation reduction used for the NICE pathway on the shared neuron.","marker":"[16]"},{"why":"Anchors the Piezo1 half-activation tension, gate area, and inactivation time constants used in the mechanosensitive Boltzmann gates.","marker":"[23]"},{"why":"Supplies the dendrite-to-soma Piezo1 density ratio that makes the three-compartment neuron necessary for firing.","marker":"[38]"},{"why":"Supplies the astrocytic TRPA1-to-glutamate-to-NMDA relay and its knockout/AP5 controls.","marker":"[25]"},{"why":"Supplies the lipid-bilayer area-expansion modulus K_A used in the strain-to-tension conversion.","marker":"[35]"},{"why":"Provides the human dACC theta-burst protocol and transcranial pressure target that define the canonical demonstration exposure.","marker":"[30]"}],"fun_headline_variants":["Open-source tool predicts neuron firing from skull ultrasound","Strain elongates ultrasound firing zone beyond acoustic focus","Six ultrasound mechanisms, one neuron model: direct compare","Ultrasound firing volume 53x acoustic focal volume"],"cache_read_input_tokens":47744,"weakest_assumption_plain":"The entire absolute firing prediction rests on the unmeasured multiplier alpha that converts tissue strain into membrane tension, which the paper brackets from about 200 to 2000 and sets to 1000 for the headline numbers, with the sensitivity analysis showing that plus or minus 25 percent in the companion modulus K_A changes spike count by -59 percent and +89 percent.","fun_headline_variants_meta":{"raw":{"variants":["Open-source tool predicts neuron firing from skull ultrasound","Strain elongates ultrasound firing zone beyond acoustic focus","Six ultrasound mechanisms, one neuron model: direct compare","Ultrasound firing volume 53x acoustic focal volume"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000919,"raw_usage":{"total_tokens":4028,"prompt_tokens":1115,"completion_tokens":2913,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":731,"completion_tokens_details":{"reasoning_tokens":2850}},"tokens_in":731,"tokens_out":2913,"duration_ms":29274,"temperature":1.0,"reasoning_tokens":2850,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T05:27:56.342788+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Record high-density extracellular firing from cortex during a 500 kHz, roughly 0.5 MPa transcranial $\\theta$-burst exposure matching the simulated protocol, and compare the spatial extent and per-voxel spike counts against the predicted 5,468 firing voxels and approximately 8,500 $mm^{3}$ zone; if firing appears only near the acoustic focus when $\\alpha$ is low, or saturates across a much larger region when $\\alpha$ is high, the $\\alpha$ = 1000 working point is falsified, and paired shear-wave elastography with single-cell mechano-current recordings would provide the direct calibration needed to replace it.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the heterogeneous nonlinear full-wave acoustic solver that produces the pressure field entering the pipeline."},{"cited_title":"A reduced viscoelastic FDTD formulation for ultrasound-driven shear wave propagation in soft tissue","cited_arxiv_id":"2607.28414","evidence_quote":"Supplies the reduced viscoelastic shear-FDTD solver that converts radiation force into the tissue displacement and strain field."},{"cited_title":"Lemaire, E","cited_arxiv_id":null,"evidence_quote":"Provides the cycle-averaged intramembrane-cavitation reduction used for the NICE pathway on the shared neuron."},{"cited_title":"Coste, J","cited_arxiv_id":null,"evidence_quote":"Anchors the Piezo1 half-activation tension, gate area, and inactivation time constants used in the mechanosensitive Boltzmann gates."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the dendrite-to-soma Piezo1 density ratio that makes the three-compartment neuron necessary for firing."},{"cited_title":"Justin Lee","cited_arxiv_id":null,"evidence_quote":"Supplies the astrocytic TRPA1-to-glutamate-to-NMDA relay and its knockout/AP5 controls."},{"cited_title":"Rawicz, K","cited_arxiv_id":null,"evidence_quote":"Supplies the lipid-bilayer area-expansion modulus K_A used in the strain-to-tension conversion."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the human dACC theta-burst protocol and transcranial pressure target that define the canonical demonstration exposure."}],"review_version":1}