{"id":"cf99bcc7-3c32-454a-a020-3b4ee6153c5e","arxiv_id":"2411.19917","paper_version":1,"verdict":"REJECT","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Traction force microscopy is cast as a parameter identification inverse problem, with a new nonlinear 2D hyperelastic model, a computed Fréchet derivative and adjoint, and numerical reconstructions.","lead":"Traction force microscopy, a standard technique for measuring the forces cells exert on soft gels, is reformulated as a mathematical inverse problem. The authors develop theory for linear and nonlinear elastic substrates and test stable reconstruction algorithms on simulated and experimental data.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Nonlinear forward operator is not established: Theorem 3.3 gives only W^{1,p} minimizers, while Theorem 3.4 and Remark 3.6 require W^{2,p} admissibility; Section 3.2.2 admits the missing regularity argument.","rationale":"The reader's weakest_assumption identifies exactly the same gap: the W^{1,p} minimizer is assumed to be W^{2,p} admissible. I agree with that assessment. The paper is self-aware about the missing step in Section 3.2.2, but Remark 3.6 treats it as resolved, and all downstream theorems (3.7 and 3.8) invoke Theorem 3.4. I considered other potential weaknesses: the polyconvexity proof of (20), the identification of effective thickness h, the domain of self-adjointness of S'(T), and the numerical comparison with FTTC. None of these is as load-bearing as the regularity gap, because without W^{2,p} admissibility the forward operator may not be single-valued and the Fréchet derivative theory has no foundation. The numerical experiments and released code are genuine contributions, and the paper would be substantially repaired by a valid regularity theorem. In its current form, however, the central theoretical claim is unsupported, so the verdict should remain REJECT.","tokens_in":21930,"tokens_out":9348,"duration_ms":86462,"concrete_test":"Prove or disprove the regularity bridge: for W in (20) and T∈L^p(Ω,R^2), show whether every energy minimizer of (18) lies in W^{2,p}(Ω,R^2) with det(I+∇u) bounded away from 0. A specific analytical check is to write the Euler–Lagrange equation −div σ(u)=T with σ from (26) and determine whether F^{-1}=(I+∇u)^{-1} is bounded in L∞; this requires inf det(I+∇u)>0. A numerical probe on the same question: solve the displacement BVP for a discontinuous right-hand side, e.g. T a scaled characteristic function of a disk, on successively refined meshes and monitor min det(I+∇u_h) and ∥u_h∥_{W^{2,p}}. If det(I+∇u_h) approaches 0 or the W^{2,p} norm fails to stabilize, Theorem 3.4's hypotheses are not met for that T, and Remark 3.6's function space setting collapses.","verdict_should_be":"REJECT","load_bearing_attack":"The paper's central theoretical claim is that S:X→Y in (13) is well-defined, locally unique, and Fréchet-differentiable (Theorems 3.4, 3.7, 3.8) with X=L^p and Y=W^{2,p}∩W^{1,p}_0 (Remark 3.6). This requires the energy minimizer from Theorem 3.3 to be admissible in W^{2,p}. The paper explicitly flags the obstacle in Section 3.2.2: \"Since the existence result from Theorem 3.3 just gives a solution in the space W^{1,p}(Ω,R^2), the regularity of this solution has to be improved ... This might be done in a similar manner as described in [37]\" and notes that \"no deformation state u∈W^{1,p}(Ω,R^2) can be admissible.\" Remark 3.6 then asserts, without proof, that the conditions of Theorem 3.5 for p=s>2 imply \"at least one solution u∈W^{2,p}∩W^{1,p}_0.\" That implication does not follow from polyconvexity and coercivity; higher regularity of minimizers for polyconvex energies is not supplied by the cited results, and the Morrey regularity reference is not applied to this problem. The Fréchet derivative and adjoint computations inherit the gap because they are built on the Implicit Function Theorem in W^{2,p}. If the minimizer is only W^{1,p}, uniqueness can fail for some T, S is not single-valued, and the parameter identification problem may be ill-posed in the stated spaces. The gap is internal and acknowledged, not a mismatch with external consensus, but it is load-bearing for the main theoretical contribution.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper reformulates traction force microscopy as a parameter identification problem. For linear 2.5D TFM, it defines the forward operator A from boundary traction stresses to interior displacements via the weak mixed boundary value problem (8)-(9), proves well-posedness, and derives the adjoint (Lemma 2.2). For nonlinear pure 2D TFM, it introduces the stored energy function (20) from the polyconvex Ogden family, proves coercivity and polyconvexity under the parameter condition lambda > 2mu/(e-1) (Theorem 3.5), states an existence result for minimizers in W^{1,p} (Theorem 3.3), and states a local uniqueness theorem requiring admissible W^{2,p} solutions (Theorem 3.4). It then asserts that the parameter-to-state map S in (13) is well-defined, locally unique, and Frechet-differentiable with a computable self-adjoint derivative (Theorems 3.7 and 3.8), and it reports numerical experiments on simulated and experimental data, including a comparison with the standard FTTC method.","tokens_in":22331,"tokens_out":10395,"duration_ms":89158,"significance":"If the theoretical claims were fully established, the paper would provide a useful functional-analytic framework for nonlinear TFM, with explicit derivative and adjoint formulas that allow the use of standard regularization algorithms, and an open implementation. The linear 2.5D analysis is sound, the constitutive derivation in Theorem 3.5 is careful and explicit, and the Frechet derivative and adjoint computations in Theorems 3.7 and 3.8 are detailed and implementable. The availability of code and data is a clear strength. However, the central well-posedness claim for the nonlinear forward operator is not supported because of the regularity gap described in the major comments; the stress-test concern lands. Since this gap is load-bearing for the main theoretical contribution, the paper in its current form cannot be accepted as a rigorous mathematical treatment of nonlinear TFM.","major_comments":[{"comment":"The existence result Theorem 3.3 supplies minimizers only in W^{1,p}(Omega,R^2), whereas Theorem 3.4 and the subsequent analysis require a solution in W^{2,p}(Omega) intersect W^{1,p}_0(Omega). The manuscript itself states in Section 3.2.2 that 'the regularity of this solution has to be improved ... This might be done in a similar manner as described in [37]' and that 'no deformation state u in W^{1,p}(Omega,R^2) can be admissible.' Remark 3.6 nevertheless asserts, without proof, that the conditions of Theorem 3.5 for p=s>2 imply 'at least one solution u in W^{2,p}(Omega,R^2) intersect W^{1,p}_0(Omega,R^2).' Polyconvexity and coercivity alone do not yield this higher regularity, and the cited Morrey regularity result is not applied to the present minimization problem. Because S in (13) is defined through this solution and because Theorems 3.7 and 3.8 both rely on Theorem 3.4 via the Implicit Function Theorem, the well-definedness and differentiability of S are not established. This is a load-bearing gap in the paper's central theoretical claim.","section":"Sections 3.2.2 and 3.3 (Remark 3.6)"},{"comment":"Even if a W^{2,p} solution were known to exist for every T, Theorem 3.4 only provides local uniqueness in a neighborhood V(bar u) x W(bar T) of an admissible pair. Remark 3.6 then selects X=L^p(Omega,R^2) (or H^1_0(Omega,R^2)) without restricting the parameter domain to such a neighborhood. Without a global uniqueness statement, the map S in (13) may be multi-valued on the chosen space X, and the inverse problem as posed on the full space may be ill-posed. The paper should either prove global uniqueness, or formulate S and the derivative theorems as local statements on an explicitly defined open set of admissible data.","section":"Section 3.2.2, Theorem 3.4"}],"minor_comments":[{"comment":"The notation 'V(bar u) in W^{2,p}(Omega) intersect W^{1,p}_0(Omega)' and 'W(bar T) in L^p(Omega)' should read 'V(bar u) subset W^{2,p}(Omega) intersect W^{1,p}_0(Omega)' and 'W(bar T) subset L^p(Omega),' since these are neighborhoods, not elements of the spaces.","section":"Theorem 3.4"},{"comment":"The statement that 'no deformation state u in W^{1,p}(Omega,R^2) can be admissible' is surprising in view of the Sobolev embedding W^{1,p}(Omega,R^2) into C^0 for p>2 used in the same paragraph; the precise notion of admissibility and the required regularity threshold should be clarified.","section":"Section 3.2.2"},{"comment":"The conclusions about the L^2-penalty being better for low noise and the H^1_0-penalty being better for high noise are based on only two force fields and two noise levels; in Table 3 the L^2 reconstruction at 15.63% noise has an error of 82.93%, which is close to a trivial reconstruction, so the comparative claim should be tempered.","section":"Section 4.3, Tables 2 and 3"},{"comment":"The statement that S'(T) is self-adjoint in L^2 should specify the domain of the unbounded operator (for example, the natural domain inherited from L^p(Omega,R^2) or W^{2,p}(Omega,R^2) intersect W^{1,p}_0(Omega,R^2)) and should note that the integration-by-parts argument is justified under the regularity assumptions; otherwise the term 'self-adjoint' is ambiguous.","section":"Theorem 3.8"},{"comment":"The heading 'Proof of Lemma 3.3' should refer to Theorem 3.3, since the statement being proved is labeled as a theorem in the main text.","section":"Appendix A"},{"comment":"There are minor typographical issues, such as the missing space in 'Lame constants lambda' before equation (5), the awkward piecewise formatting of equation (31), and the use of Euler's number e in condition (23) without an explicit definition; these are easy to correct.","section":"Throughout"}],"recommendation":"reject","confidential_remarks":"The linear 2.5D part and the constitutive calculations are solid, and the derivative/adjoint formulas are useful. However, the nonlinear forward operator theory has an acknowledged and load-bearing regularity gap: Theorem 3.3 gives only W^{1,p} minimizers, while Theorem 3.4 and Remark 3.6 require W^{2,p} solutions without proving their existence. A revision would require either a genuine regularity argument for the polyconvex energy (20) or a fundamental restatement of the nonlinear theorems as conditional on an admissibility hypothesis, with the parameter-to-state map restricted to a local neighborhood. As written, the central theoretical contribution is not supported."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things to know before you read this one. First, the nonlinear half of the paper is a genuine mathematical contribution: the authors construct a 2D polyconvex stored energy (Eq. 20) with an explicit coercivity condition (23), prove the Fréchet derivative formula (27), and show it is self-adjoint (Thm 3.8). The linear 2.5D analysis is clean, and the numerical experiments—including a comparison with FTTC on real cells—are honest and reproducible, with code on GitLab. Second, the central claim—that the nonlinear forward operator S is a well-defined, differentiable map between the stated Banach spaces—is not actually established. The existence theorem (3.3) only gives a minimizer in W^{1,p}, while the uniqueness and differentiability theorem (3.4) requires an admissible solution in W^{2,p}. Section 3.2.2 explicitly says the regularity has to be improved and points to Morrey [37]; Remark 3.6 then simply asserts that the conditions of Thm 3.5 give a W^{2,p} solution. That step does not follow from polyconvexity and coercivity alone. If the bridge fails, S may not be single-valued, and the Fréchet derivative theory collapses. The gap is internal and acknowledged, but it is load-bearing.\n\nThe rest of the paper is solid: the adjoint computations check out, the coercivity condition (23) is carefully derived, and the Newton-CG implementation is standard. The effective-thickness calibration in Section 4.4 is a little circular (it is inferred by matching the FTTC solution), and the experimental comparison lacks error bars, but those are minor. My main worry remains the regularity gap. It is specific and probably repairable—one needs a regularity theorem for minimizers of this particular polyconvex energy, or a weaker statement about the forward operator. But as written, the paper overclaims in Remark 3.6.\n\nWho should read it: people working on inverse problems for nonlinear elasticity, and mechanobiologists who want a mathematically grounded alternative to the standard Boussinesq-based TFM. It deserves a serious referee: the gap is real but the contribution is substantial, and a careful review could push the authors to close it. I would not desk-reject; I'd send it out and ask for a major revision, then re-check the regularity step carefully.","headline":"Solid nonlinear TFM framework with a real regularity gap in the well-posedness claim; worth refereeing, not desk-rejecting.","tokens_in":22896,"tokens_out":2751,"would_cite":false,"duration_ms":23831,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["65J20","35Q74","74B20","74G75","65N21"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper shows that traction force microscopy can be formulated as a parameter identification inverse problem whose nonlinear forward operator is Fréchet-differentiable with a self-adjoint derivative, enabling stable regularized…","keywords":["traction force microscopy","inverse problems","parameter identification","nonlinear elasticity","hyperelasticity","polyconvex stored energy","Fréchet derivative","regularization"],"falsifier":"Compute, on a domain with a nonsmooth corner, all energy-minimizing solutions of the displacement problem with $W$ from (20) for a fixed $T \\in L^p$; finding two distinct minimizers, or a minimizer whose second derivatives fail to be integrable, would break the single-valuedness and differentiability of the nonlinear forward operator.","tokens_in":21657,"feed_emoji":"🔬","tokens_out":9690,"duration_ms":81874,"temperature":0.7,"pith_summary":"This paper aims to put traction force microscopy (TFM) on a rigorous inverse-problem footing. It treats linear 2.5D TFM and nonlinear pure 2D TFM as parameter identification problems, where the unknown cell force is recovered from measured substrate displacement. The central claim is that the nonlinear force-to-displacement map is well defined on suitable Banach spaces, Fréchet-differentiable, and invertible by regularized Newton-CG, provided the substrate follows a specifically chosen polyconvex energy law. A working numerical method follows, including reconstructions from experimental data, which matters because nonlinear substrates under large deformation cannot be handled by the analytical Boussinesq-based methods used in standard TFM.","feed_headline":"Traction force microscopy becomes a provably stable inverse problem","feed_subtitle":"A polyconvex material law makes the force-to-displacement map differentiable, enabling regularized Newton-CG inversions.","key_machinery":"The load-bearing object is the forward operator $S$ together with the stored energy function (20), a two-dimensional polyconvex Ogden-type hyperelastic law $W(F) = \\frac{\\mu}{2}|F|^2 + \\frac{\\lambda}{4}(\\det F)^2 - (\\mu + \\frac{\\lambda}{2})\\ln(\\det F) - \\frac{3\\mu}{2} - \\frac{\\lambda}{4}$. Polyconvexity and coercivity allow the displacement boundary value problem to be solved by minimizing the energy $G(u) = \\int_\\Omega W(I + \\nabla u)\\,dx - \\int_\\Omega T u\\,dx$, and the same law is engineered so that its linearization matches Hooke's law near a natural state. The Fréchet derivative $S'(T)h = v$ solves the linearized elasticity system in (27), and its $L^2$-self-adjointness means the adjoint can be evaluated by solving the same type of boundary value problem, which makes the adjoint-based Newton-CG iteration practical.","core_discovery":"For the nonlinear pure 2D model, the paper claims that the parameter-to-state map $S: T \\mapsto u$ defined by the boundary value problem $-\\operatorname{div}(\\sigma(u)) = T$ in $\\Omega$, $u = 0$ on $\\partial\\Omega$, with stress derived from the stored energy function $W(F) = \\frac{\\mu}{2}|F|^2 + \\frac{\\lambda}{4}(\\det F)^2 - (\\mu + \\frac{\\lambda}{2})\\ln(\\det F) - \\frac{3\\mu}{2} - \\frac{\\lambda}{4}$, is a well-defined forward operator. The paper proves existence of energy minimizers (Theorem 3.3), local uniqueness of admissible solutions and bounded invertibility of the linearized operator (Theorem 3.4), polyconvexity and coercivity of $W$ (Theorem 3.5), and an explicit Fréchet derivative (Theorem 3.7) that is self-adjoint in $L^2$ (Theorem 3.8). These results justify applying discrepancy-principle-truncated Newton-CG with an adjoint-based inner iteration to recover force densities. For the linear 2.5D case, the forward operator is a bounded linear map whose adjoint is obtained through an auxiliary elasticity problem, and the paper demonstrates reconstructions from simulated and experimental displacement data; for a high-force phantom, using the linear reconstruction as an initial guess reduces the nonlinear reconstruction error from roughly 7% to below 3%. The nonlinear theory depends on a regularity bridge: the energy minimizer from Theorem 3.3 lies in $W^{1,p}$, while the uniqueness and derivative theorems require a $W^{2,p}$ admissible state, a step the paper states is still needed and points to a known regularity route.","pith_inferences":["One consequence the paper leaves implicit is that the same adjoint-based machinery would allow joint estimation of the effective thickness $h$ and the traction field $T$, since the conversion $T = t/h$ is linear in $t$ and $h$ enters only as a scalar; the reported experimental value $h \\approx 1\\,\\mu$m indicates the data carry information about it.","If Morrey-type regularity could be established for the specific energy (20), the local character of Theorem 3.4 would become global on a full neighborhood of admissible data, and the conditional stability constants of the linearized operator would likely determine convergence rates for the Newton-CG iteration, which the paper does not compute.","The reported norm-selection behavior suggests a practical stopping-rule design: monitor the discrepancy-principle iterate under both $L^2$ and $H^1_0$ penalties and prefer the $H^1_0$ reconstruction when the noise estimate exceeds a few percent; the paper documents the effect but does not propose this rule."],"forward_implications":["Nonlinear TFM with hyperelastic substrates under large deformation can be solved by iterative regularization rather than by numerical differentiation of noisy images.","The same framework supplies a linear pure 2D solver; applying it to measured fibroblast data yields traction maps in the same range as the standard Fourier method and suggests an effective substrate thickness near $1\\,\\mu$m.","Because the derivative is self-adjoint, computing the adjoint for the Newton-CG inner iteration costs one linearized elasticity solve, so the extra expense of nonlinear inversion over linear inversion is moderate.","Choosing $H^1_0$ as the parameter space smooths high-noise reconstructions, whereas $L^2$ gives lower errors on cleaner data, providing a practical norm-selection rule.","The analytical framework extends to other polyconvex stored energy functions, so material laws can be swapped without redoing the inverse-problem architecture."],"supporting_citations":[{"why":"Supplies the direct-method existence theorem for hyperelastic boundary value problems that Theorem 3.3 adapts to the 2D displacement setting.","marker":"[5]"},{"why":"Provides the Fourier transform traction cytometry method with Tikhonov regularization used as the comparison baseline for experimental data.","marker":"[9]"},{"why":"Provides the polyconvexity and coercivity definitions, the natural-state expansion, and the admissible-solution uniqueness framework used throughout Section 3.","marker":"[14]"},{"why":"Introduces the thin-substrate thickness-averaged 2D TFM model on which the nonlinear pure 2D boundary value problem is built.","marker":"[15]"},{"why":"Supplies the truncated Newton-CG algorithm whose regularizing properties justify the inner/outer iteration used for the nonlinear inverse problem.","marker":"[22]"},{"why":"Provides the earlier mathematical framework for TFM and the effective-thickness parameter $h$ that converts traction stress to force density.","marker":"[36]"},{"why":"Cited as the route for improving $W^{1,p}$ minimizers to the $W^{2,p}$ regularity required by the uniqueness and differentiability theorems.","marker":"[37]"},{"why":"Supplies the standard TFM workflow, material parameters, and the Fourier-based reconstruction approach used as the experimental reference.","marker":"[47]"},{"why":"Provides the similar 2.5D linear reconstruction approach whose adjoint-based method is adapted for the linear 2.5D problem.","marker":"[49]"},{"why":"Contains the local existence and uniqueness theory for admissible deformation states used in Theorem 3.4.","marker":"[57]"}],"fun_headline_variants":["Provably stable inversion for nonlinear traction forces","Cell force recovery gains rigorous inversion theory","Polyconvex elasticity enables stable traction force inversion","Newton-CG inversion for traction microscopy with proven stability"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The nonlinear theory stands on the unproven bridge that the energy-minimizing displacement, which exists in the Sobolev space $W^{1,p}$, is actually regular enough (in $W^{2,p}$ and admissible) for the local-uniqueness and derivative theorems to apply; the paper says this regularity still has to be improved and cites a possible route.","fun_headline_variants_meta":{"raw":{"variants":["Provably stable inversion for nonlinear traction forces","Cell force recovery gains rigorous inversion theory","Polyconvex elasticity enables stable traction force inversion","Newton-CG inversion for traction microscopy with proven stability"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000637,"raw_usage":{"total_tokens":2992,"prompt_tokens":1061,"completion_tokens":1931,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":677,"completion_tokens_details":{"reasoning_tokens":1874}},"tokens_in":677,"tokens_out":1931,"duration_ms":14718,"temperature":1.0,"reasoning_tokens":1874,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T05:41:55.865202+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute, on a domain with a nonsmooth corner, all energy-minimizing solutions of the displacement problem with $W$ from (20) for a fixed $T \\in L^p$; finding two distinct minimizers, or a minimizer whose second derivatives fail to be integrable, would break the single-valuedness and differentiability of the nonlinear forward operator.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the direct-method existence theorem for hyperelastic boundary value problems that Theorem 3.3 adapts to the 2D displacement setting."},{"cited_title":"Blumberg and U.S","cited_arxiv_id":null,"evidence_quote":"Provides the Fourier transform traction cytometry method with Tikhonov regularization used as the comparison baseline for experimental data."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the polyconvexity and coercivity definitions, the natural-state expansion, and the admissible-solution uniqueness framework used throughout Section 3."},{"cited_title":"Dembo, T","cited_arxiv_id":null,"evidence_quote":"Introduces the thin-substrate thickness-averaged 2D TFM model on which the nonlinear pure 2D boundary value problem is built."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the truncated Newton-CG algorithm whose regularizing properties justify the inner/outer iteration used for the nonlinear inverse problem."},{"cited_title":"Maskarinec, C","cited_arxiv_id":null,"evidence_quote":"Provides the earlier mathematical framework for TFM and the effective-thickness parameter $h$ that converts traction stress to force density."},{"cited_title":"Michel, V","cited_arxiv_id":null,"evidence_quote":"Cited as the route for improving $W^{1,p}$ minimizers to the $W^{2,p}$ regularity required by the uniqueness and differentiability theorems."},{"cited_title":"Schwarz, N.Q","cited_arxiv_id":null,"evidence_quote":"Supplies the standard TFM workflow, material parameters, and the Fourier-based reconstruction approach used as the experimental reference."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the similar 2.5D linear reconstruction approach whose adjoint-based method is adapted for the linear 2.5D problem."},{"cited_title":"Toyjanova, E","cited_arxiv_id":null,"evidence_quote":"Contains the local existence and uniqueness theory for admissible deformation states used in Theorem 3.4."}],"review_version":1}