{"id":"96a78ffb-a93e-43d2-82c9-8d32b8a2a298","arxiv_id":"2608.08096","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"An exponential displacement-dependent active feedback in a coupled oscillator chain reproduces cochlear amplification, compression, tuning, and phase, and outperforms the cubic approximation at high stimulus levels.","lead":"This paper builds a minimal computer model of the cochlea as a chain of oscillators whose active amplification fades as the membrane moves farther. The model reproduces several hallmarks of inner-ear responses and suggests why the exact form of nonlinear feedback matters at high sound levels.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The numerical scheme for the pressure-displacement system is not specified as a well-posed procedure, so the reported traveling waves, phase accumulation, and Q10 values are not yet attributable to the continuum model.","rationale":"I read the strongest claim as a qualitative demonstration that the minimal coupled-oscillator model produces cochlear-like nonlinear responses. The least secure condition is not the instantaneous feedback, which the Discussion candidly acknowledges as a future extension and therefore a known scope limit under a qualitative claim; it is the numerical evidence itself. Eqs. (2)-(3) form a constrained system, and the implementation description is internally ambiguous: the pressure field is said to be obtained from the discretized fluid equation and then used to evaluate acceleration, which is not a simultaneous solution of the coupled equations. No code is provided, and an essential parameter (ω_b) is missing, so the simulation cannot be independently regenerated. The reader's CONDITIONAL verdict was partly driven by this reproducibility issue; my attack makes it the primary concern. I recommend UNVERDICTED rather than CONDITIONAL because, until the numerical method is specified and tested, the central claims cannot be checked at all. The Discussion's listed discrepancies (i)-(iii) are limitations rather than internal errors and do not resolve this gap. If the authors supply a derivation that eliminates P or release their code, the concern can be settled by the concrete test above.","tokens_in":14335,"tokens_out":8629,"duration_ms":98375,"concrete_test":"Eliminate P from Eqs. (2)-(3) to obtain one integro-differential equation for η and solve it with an independent spectral method using the Table 1 parameters and a stated ω_b consistent with the reported CF=10 kHz location; compare the resulting CF gain curve, Q10 values, and phase accumulation with Figs. 2-5, with no free parameter adjustment. If the results match, the reported numerics are validated; if they differ substantially, the figures are artifacts of the unspecified pressure update.","verdict_should_be":"UNVERDICTED","load_bearing_attack":"The central claims are entirely numerical, but the manuscript does not define a well-posed numerical procedure for the coupled system (2)-(3). Equation (3), ∂²P/∂x² = (2ρ/H)η¨, is a constraint: with P(0,t) and P(L,t) specified, P is determined by spatial integration of η¨. Removing P from Eq. (2) yields an integro-differential equation for η; a time-marching solution must respect that global coupling at every time step. The text instead says that 'at each time step, the pressure field was obtained from the discretized fluid equation and subsequently used to evaluate the acceleration,' which sounds like an explicit staggered update. Such a procedure would not solve (2)-(3) simultaneously and can introduce numerical dispersion, reflections at the apex boundary, and phase errors—precisely the quantities used to claim agreement with experiment. No code or derivation is provided, and the base frequency ω_b consistent with a CF=10 kHz location is not tabulated, so the simulation cannot be independently regenerated. Until this is resolved, the reported phase accumulation of about 2.6 cycles, the compressive gain curves, and the Q10 values in Table 2 cannot be attributed to the continuum model.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript proposes a one-dimensional cochlear model in which each longitudinal segment is a damped, driven oscillator with a displacement-dependent exponential active feedback term, coupled through the fluid pressure field and through elastic and dissipative longitudinal interactions. The authors solve the coupled system numerically, compare the exponential feedback with a cubic approximation, and study the effect of longitudinal coupling. They report level-dependent compression, traveling-wave propagation, phase accumulation of roughly 2.6 cycles, Q10 values in the experimental range, and level-dependent gain curves at the characteristic frequency. The Discussion candidly lists three quantitative discrepancies with chinchilla data and proposes extensions including feedback delay and a two-dimensional geometry.","tokens_in":14559,"tokens_out":10005,"duration_ms":108121,"significance":"If the numerical results are reliable, the paper is a useful minimal demonstration that a single exponential displacement-modulated velocity-feedback law, combined with fluid and longitudinal coupling, produces several hallmark cochlear nonlinearities without invoking separate nonlinear feedback laws. The systematic comparison between exponential and cubic nonlinearities and the clean separation of the role of longitudinal coupling are valuable, and the authors deserve credit for explicitly cataloging the model's remaining discrepancies rather than overstating agreement. The main limitation is that the central claims are entirely numerical, while the numerical procedure is not specified at the level needed to verify them and no code or convergence analysis is provided. The significance is therefore conditional on the numerical implementation being made precise and reproducible.","major_comments":[{"comment":"The numerical scheme as described is not a well-posed discretization of the coupled system. Equation (3), ∂²P/∂x² = (2ρ/H)η¨, is an elliptic constraint that must be solved together with Eq. (2) at each instant; stating that 'at each time step, the pressure field was obtained from the discretized fluid equation and subsequently used to evaluate the acceleration' suggests an explicit staggered update that does not enforce the constraint simultaneously. The reported phase accumulation, compressive gain curves, and Q10 values are entirely numerical, so the paper must supply the semi-discrete equations, the time integrator, the treatment of both pressure boundary conditions, the steady-state criterion, and convergence tests in N_x and the time step, or make the code available. Without this information the results cannot be attributed to the continuum model.","section":"Numerical Implementation, Eqs. (2)–(3)"},{"comment":"The base frequency ω_b in ω0(x)=ω_b exp(-μx) is not tabulated, even though it determines the location of CF=10 kHz and therefore all frequency-place and Q10 comparisons. In addition, the 30 dB middle-ear gain is mentioned in the text but omitted from Table 1, and it is unclear whether the dB SPL values in Figs. 2–4 refer to ear-canal sound pressure before or after this gain. These omissions make the simulations impossible to reproduce independently.","section":"Table 1 and Numerical Implementation"},{"comment":"The active feedback is assumed to be instantaneous and exactly in phase with velocity, and the Discussion acknowledges that real feedback will generally have a delay. Because the phase accumulation and the level-dependent lead-lag reversal in Fig. 5 are direct consequences of this phase relation, the claim that the model 'captures the observed phase behavior' is contingent on the no-delay assumption. The authors should either report a sensitivity test with a small phase lag or explicitly qualify the phase claim in the Results and Summary.","section":"Discussion, first paragraph and extension (i)"}],"minor_comments":[{"comment":"The phase accumulation is reported as approximately 2.5 cycles in the Results and 2.6 cycles in the Discussion; please make these numbers consistent and specify the stimulus level and frequency at which the value is evaluated.","section":"Results, Phase"},{"comment":"The axis labels appear garbled in the rendering (for example, '7 m/sec' and '#10-7'); please check the fonts, units, and exponent formatting.","section":"Figures 2 and 8"},{"comment":"The Summary states that the model 'captures the gain curve at the CF,' but the Discussion lists as discrepancy (iii) a fivefold underestimate of the low-level CF response. Please soften the Summary claim to 'qualitative form of the gain curve' or otherwise add the qualification.","section":"Summary and Discussion"},{"comment":"The text says the first oscillator at the base was fixed, η(0,t)=0, but Eq. (4) contains second spatial derivatives; please specify the finite-difference stencils used at both boundaries so that the discretization of the longitudinal coupling is unambiguous.","section":"Numerical Implementation"}],"recommendation":"major_revision","confidential_remarks":"The main obstacle is numerical reproducibility: the paper's qualitative conclusions are plausible, but the central results cannot be checked without a precise discretization or code. If the authors provide the actual semi-discrete equations, the time-stepping scheme for the pressure-displacement constraint, and convergence tests, the paper could be acceptable as a qualitative modeling study. The novelty is incremental relative to the Duke–Jülicher and van der Pol–Rayleigh modeling tradition, but the exponential cutoff and the systematic exponential-versus-cubic comparison are useful contributions. The authors' explicit catalog of discrepancies is a strength and should be preserved."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nHere's my read on arXiv:2608.08096. The paper belongs to a long line of Van der Pol/Rayleigh and Hopf-type cochlear models. What's new is the exponential displacement-dependent active damping term, β exp(−(η/η_th)^2) η̇, and the systematic comparison with its cubic truncation. That's a legitimate variant: the exponential form naturally recovers cubic nonlinearity at low amplitude and retains higher-order terms at high SPL, which the authors show preserves tuning and produces a gradual return to passive behavior. The model also includes fluid coupling and longitudinal elastic/dissipative coupling, and the qualitative outcomes—traveling wave, compression, phase accumulation, Q10—are consistent with chinchilla data. The paper is honest about its limitations: it lists three specific quantitative discrepancies, acknowledges the lack of delay in the active force, and notes that parameters are chosen from general considerations rather than detailed fitting.\n\nThe main soft spot is the numerics. The central claims are entirely numerical, but the manuscript does not give a reproducible scheme. Equation (3) is a constraint linking pressure to acceleration; the text says the pressure field was obtained from the discretized fluid equation at each time step and then used to evaluate acceleration. That sounds like an explicit staggered update, which would not solve the coupled system consistently and could introduce dispersion and phase errors—exactly the quantities claimed to match experiment. No code, no convergence analysis, and ω_b (the base frequency needed to fix the tonotopic map) is not in Table 1. These are concrete omissions, not style preferences. The stress-test note is on target here. I'd like to see the actual discretization, a DAE or integro-differential formulation, and at minimum a grid-convergence check.\n\nThat said, the modeling idea is sound and the physics is clearly explained. The comparison between exponential and cubic nonlinearities at high levels is useful, and the discussion of longitudinal coupling is thoughtful. This is not a major breakthrough—it is incremental progress in a mature field—but it is a serious contribution.\n\nWho should read it: anyone working on minimal oscillator models of the cochlea, or on the role of nonlinear damping laws in active systems. It deserves a serious referee; the numerical issues are fixable, but the authors need to supply code or a full scheme before the quantitative claims are trustworthy.\n\nMy recommendation: send it to peer review, with a request for major revision on the numerical reproducibility.","headline":"Useful minimal model of cochlear nonlinearity, but the numerical scheme is under-specified and the quantitative claims are not yet reproducible.","tokens_in":15073,"tokens_out":3563,"would_cite":true,"duration_ms":33010,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper claims that a minimal chain of damped oscillators with exponential, displacement-dependent active feedback can reproduce the cochlea's level-dependent amplification, compressive nonlinearity, frequency selectivity…","keywords":["cochlear mechanics","active process","basilar membrane","nonlinear feedback","exponential active damping","traveling wave","compressive nonlinearity","frequency selectivity"],"falsifier":"A decisive check would be to measure basilar-membrane velocity slopes and phase at the characteristic-frequency location for stimulus levels above 90 dB SPL: the exponential model predicts a partial recovery of the local slope toward unity and a phase accumulation near 2.5-2.6 cycles, while a cubic-only model predicts continued compression and a more basal traveling-wave peak. If experiments instead show continued monotonic compression without slope recovery, or phase accumulation differing substantially from the predicted value, the assumed instantaneous in-phase exponential feedback is not the mechanism at work.","tokens_in":14116,"feed_emoji":"👂","tokens_out":9410,"duration_ms":90094,"temperature":0.7,"pith_summary":"This paper argues that a deliberately small model of the cochlea — a one-dimensional chain of damped oscillators, each carrying an exponential, displacement-dependent active feedback term $\\beta e^{-(\\eta/\\eta_{\\mathrm{th}})^2}\\dot{\\eta}$ that mimics outer-hair-cell force generation — can reproduce the main nonlinear signatures of mammalian hearing. Because the active term is strongest at small displacements and fades as displacement grows, weak sounds are amplified, intense sounds grow compressively, and the response stays sharply tuned over a wide range of levels. Coupled through the cochlear fluid and through nearest-neighbor elastic and dissipative interactions, the chain produces traveling waves with realistic frequency-place mapping, Q10 values, gain curves, gain-bandwidth behavior, and roughly 2.5 cycles of phase accumulation at the characteristic frequency. The paper further claims that the higher-order terms contained in the exponential feedback, beyond the cubic approximation used in many active-oscillator models, are needed above about 60 dB SPL to reproduce the gradual transition toward passive mechanics and the partial recovery of linear growth.","feed_headline":"One feedback law reproduces cochlear amplification and compression","feed_subtitle":"A minimal oscillator chain with level-dependent active feedback captures tuning, phase, and traveling waves.","key_machinery":"The load-bearing object is the displacement-dependent active damping term in the oscillator equation, $\\beta e^{-(\\eta/\\eta_{\\mathrm{th}})^2}\\dot{\\eta}$, which represents outer-hair-cell feedback opposing viscous damping most strongly at small basilar-membrane displacements. Around $\\beta = \\gamma$ the local element undergoes a supercritical Hopf bifurcation, so the chain operates close to criticality, and for weak responses the exponential reduces to the familiar cubic nonlinearity of Van der Pol and Rayleigh-type cochlear oscillator models. The distributed model then couples these oscillators through the cochlear fluid via the pressure-displacement relation $\\frac{2\\rho}{H}\\ddot{\\eta} = \\frac{\\partial^2 P}{\\partial x^2}$, and through longitudinal elastic and dissipative coupling governed by the strengths $\\alpha$ and $\\alpha'$, yielding the one-dimensional field equation that is solved numerically by the method of lines.","core_discovery":"The central claim is that a single physically motivated feedback law, $\\beta e^{-(\\eta/\\eta_{\\mathrm{th}})^2}\\dot{\\eta}$, appended to a forced damped oscillator, is enough to generate the cochlea's hallmark behaviors once the oscillators are distributed tonotopically and coupled through fluid pressure and nearest-neighbor elastic and dissipative interactions. The system operates near a Hopf bifurcation, with the active strength $\\beta$ slightly below the damping coefficient $\\gamma$, so small displacements receive nearly complete compensation of damping while larger displacements progressively lose that boost. In the resulting chain, a traveling wave peaks at the place where the local resonant frequency matches the stimulus, the response at that place grows linearly at low levels and compressively at high levels, and phase accumulates steeply near the characteristic place, reaching about 2.5-2.6 cycles. Comparison with a cubic truncation shows nearly identical behavior up to about 60 dB SPL, after which the cubic model over-suppresses gain, keeps compressing, and shifts the traveling-wave peak basally, while the full exponential model retains a sharper peak and recovers toward passive linear response. Longitudinal coupling broadens the frequency response, lowers peak gain, delays the onset of compression, and increases phase accumulation.","pith_inferences":["Beyond the paper, a sharp experimental test of the exponential-versus-cubic distinction would be to measure basilar-membrane input-output slopes above 90 dB SPL: the exponential model predicts a partial recovery toward linear growth, while the cubic model predicts continued compression.","The paper's own suggested extension of adding a phase lag to the active force is the most direct path to fixing its quantitative underestimate of low-level amplification; a small delay would change phase and gain predictions without altering the qualitative framework.","Because the feedback is displacement-based rather than velocity-based, responses to transient or chirp stimuli, where displacement and velocity envelopes differ, would discriminate this law from velocity-dependent alternatives more sharply than steady-tone gain curves.","Extending the model to two dimensions, as the paper notes, is the most likely way to address the missed high-frequency, high-level basilar-membrane response; the one-dimensional approximation the authors use is known to fail when the excitation wavelength is not small compared to the chamber height."],"forward_implications":["If the exponential feedback law is right, a single active term accounts for level-dependent amplification, compressive growth, and frequency selectivity without a separate saturating mechanism.","The model predicts Q10 values broadly in line with chinchilla basilar-membrane measurements for target response levels from 25 to 400 micrometers per second.","Higher-order nonlinearities beyond the cubic approximation become essential above about 60 dB SPL; cubic-only models mispredict high-level gain, slope recovery, and the location of the traveling-wave peak.","Longitudinal elastic and dissipative coupling broadens tuning, reduces low-level gain, delays the onset of compression, and increases accumulated phase; its influence shrinks as stimulus level rises.","At the characteristic frequency the model produces roughly 2.5-2.6 cycles of phase accumulation and a level-dependent lead-lag reversal around the CF."],"supporting_citations":[{"why":"Supplies the chinchilla basilar-membrane measurements against which the model's Q10 values, gain curves, and phase accumulation are compared.","marker":"[24]"},{"why":"Provides the fluid-pressure-displacement relation and the active-traveling-wave description the distributed model adapts.","marker":"[22]"},{"why":"States the modeling principles that the paper says its formulation satisfies: linked active and nonlinear processes, passive recovery without activity, and spatially localized amplification.","marker":"[14]"},{"why":"Motivates the role of couplings between nonlinear cochlear elements, which the present model incorporates as longitudinal interactions.","marker":"[21]"},{"why":"Provides the physics-of-hearing background that frames the active process and traveling-wave description underlying the model.","marker":"[1]"}],"fun_headline_variants":["Minimal model: one feedback law drives cochlear gain and tuning","Single exponential feedback explains cochlear amplification and compression","One feedback law in an oscillator chain mimics cochlear nonlinearity","Minimal physical model reveals active feedback mechanism in cochlea"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the active feedback force acts instantaneously and exactly in phase with membrane velocity through the term $\\beta e^{-(\\eta/\\eta_{\\mathrm{th}})^2}\\dot{\\eta}$, with no delay; the model's phase accumulation, gain, and high-level behavior all depend on this phase relation, and the authors note that a real active process would generally involve some delay.","fun_headline_variants_meta":{"raw":{"variants":["Minimal model: one feedback law drives cochlear gain and tuning","Single exponential feedback explains cochlear amplification and compression","One feedback law in an oscillator chain mimics cochlear nonlinearity","Minimal physical model reveals active feedback mechanism in cochlea"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000785,"raw_usage":{"total_tokens":3499,"prompt_tokens":1018,"completion_tokens":2481,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":634,"completion_tokens_details":{"reasoning_tokens":2413}},"tokens_in":634,"tokens_out":2481,"duration_ms":19085,"temperature":1.0,"reasoning_tokens":2413,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T00:26:26.857659+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A decisive check would be to measure basilar-membrane velocity slopes and phase at the characteristic-frequency location for stimulus levels above 90 dB SPL: the exponential model predicts a partial recovery of the local slope toward unity and a phase accumulation near 2.5-2.6 cycles, while a cubic-only model predicts continued compression and a more basal traveling-wave peak. If experiments instead show continued monotonic compression without slope recovery, or phase accumulation differing substantially from the predicted value, the assumed instantaneous in-phase exponential feedback is not the mechanism at work.","supporting_citations":[{"cited_title":"The Journal of the Acoustical Society of America , volume=","cited_arxiv_id":null,"evidence_quote":"Supplies the chinchilla basilar-membrane measurements against which the model's Q10 values, gain curves, and phase accumulation are compared."},{"cited_title":"2008 , file =","cited_arxiv_id":null,"evidence_quote":"Provides the fluid-pressure-displacement relation and the active-traveling-wave description the distributed model adapts."},{"cited_title":"Nature , volume=","cited_arxiv_id":null,"evidence_quote":"States the modeling principles that the paper says its formulation satisfies: linked active and nonlinear processes, passive recovery without activity, and spatially localized amplification."},{"cited_title":"Cold Spring Harbor Perspectives in Medicine , volume=","cited_arxiv_id":null,"evidence_quote":"Motivates the role of couplings between nonlinear cochlear elements, which the present model incorporates as longitudinal interactions."},{"cited_title":"Neuron , volume=","cited_arxiv_id":null,"evidence_quote":"Provides the physics-of-hearing background that frames the active process and traveling-wave description underlying the model."}],"review_version":1}