{"id":"2511d511-1567-4f2d-a6a5-5a264406be5f","arxiv_id":"2502.03610","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":12,"one_line_summary":"A stochastic model combining passive and active ratchet mechanisms, calibrated to tracked astrocyte EVs, reproduces the drift and diffusion differences across Cytochalasin D treated conditions.","lead":"This paper builds a mathematical model of how extracellular vesicles move along neuron surfaces, combining a passive flashing ratchet driven by the cytoskeleton with an active rolling mechanism. The model is calibrated to microscopy tracks and reproduces key differences between control and Cytochalasin D treated conditions.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The experimental evidence for a drifting EV component may be an artifact of sign-flipping each trajectory to make its linear-regression slope positive; the skewness tests anchoring the model comparison need to be rerun on unflipped data.","rationale":"The reader correctly identified the untested treatment-to-mechanism mapping and the inability of the symmetric active potential to generate net drift as weaknesses. However, a more fundamental issue appears earlier in the pipeline: the sign-flipping standardization in Section 2 can manufacture the very skewness that is later used to validate the model. The paper is transparent and the modeling framework is plausible, but its central empirical anchor—the presence of a drifting component in Ctrl and CytoD-EV and its absence in CytoD-HN—rests on a statistical test applied after a transformation that biases the result. The proposed check is cheap and decisive: rerun the skewness analysis on unflipped displacements and run a Brownian null through the same preprocessing. If the concern lands, the model's assignment of a passive directed ratchet to CytoD-EV loses its experimental basis, and the paper would need major revision rather than conditional acceptance. If the concern is refuted, the existing conditional acceptance can stand. I therefore keep the CONDITIONAL verdict but make the reanalysis an explicit condition.","tokens_in":20011,"tokens_out":5841,"duration_ms":59387,"concrete_test":"Recompute the skewness test on the raw, unflipped tangential displacements for Ctrl and CytoD-EV, using the same zero-displacement filtering. If the p-values become non-significant, the directed-transport evidence is an artifact. As a null check, simulate pure Brownian trajectories with the same n, duration, and sampling rate, apply the same sign-flipping rule, and measure how often the skewness test reports p<0.05; if the false-positive rate substantially exceeds 5%, the preprocessing alone explains the reported drift.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"In Section 2 (Data post-processing) the authors state that they 'standardized the preferential direction of the EVs towards positive values by inverting the sign of the displacements in samples where the linear regression slope computed on the tangent displacement was negative.' This per-trajectory sign flip is applied before the skewness test that provides the key evidence for directed transport (Section 3: Ctrl p=0.00296, CytoD-EV p=0.0275, CytoD-HN p=0.731). A finite-length Brownian trajectory almost always has a nonzero regression slope by chance; flipping every trajectory with a negative slope makes the pooled displacement distribution positively biased by construction. The reported drift signal, and therefore the model's success criterion ('only CytoD-HN lacks drift'), may be a preprocessing artifact rather than a physical directed component. This is load-bearing because the central claim is that the model reproduces the presence and absence of drift across conditions. If the drift signal is generated by the sign-flip itself, the calibration target is invalid. Note also that the symmetric active potential (Eq. 4) cannot produce net drift by symmetry, so the only source of drift in the model is the passive asymmetric ratchet; whether CytoD-EV truly requires that ratchet is not independently established.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a data-driven stochastic model for the motion of prion-protein-bearing extracellular vesicles (EVs) on neuronal surfaces. It combines an overdamped Langevin description of the EV with Markov-switching flashing ratchet potentials: an asymmetric sawtooth potential V1 for passive, cytoskeleton-driven transport and a symmetric sawtooth potential V2 for active, actin-based rolling. The model is specialized to three experimental conditions, Ctrl (both mechanisms), CytoD-HN (active only), and CytoD-EV (passive only), and its parameters are calibrated to experimental trajectories from D'Arrigo et al. The authors report qualitative agreement between simulations and experiments in terms of mean squared displacement curves, displacement histograms, kinematic indicators, and a skewness-based classification of which conditions display directed transport, with only CytoD-HN lacking a drifting component.","tokens_in":20365,"tokens_out":5705,"duration_ms":54687,"significance":"If the central claim holds, the paper would provide a useful quantitative framework for decomposing EV motion on neurons into passive and active mechanisms, with potential implications for how prion-like pathogenic cargo spreads in neurodegenerative disease. The manuscript is clearly written in terms of stochastic differential equations and Markov processes, and it has strengths worth acknowledging: the code and data are promised on Zenodo, the one-dimensional reduction is empirically motivated (about 80% of the motion is tangent to the neuron surface), and a sensitivity analysis is included. However, the significance is currently conditional because the experimental evidence for directed transport rests on a per-trajectory sign-flipping procedure that may itself create the reported skewness, and because the model validation is largely in-sample, with mechanism assignments assumed rather than independently tested.","major_comments":[{"comment":"The per-trajectory sign standardization described in Section 2 invalidates the skewness tests used as evidence for directed transport. The authors state that 'we standardized the preferential direction of the EVs towards positive values by inverting the sign of the displacements in samples where the linear regression slope computed on the tangent displacement was negative.' A purely diffusive finite trajectory almost always has a nonzero regression slope by chance, and flipping every trajectory with a negative slope mechanically creates a pooled displacement distribution with positive skew. The p-values reported in Section 3 (Ctrl p=0.00296, CytoD-EV p=0.0275, CytoD-HN p=0.731) therefore cannot be interpreted as evidence of a physical drift component. Because the model's central success criterion is reproducing the presence and absence of drift across conditions, the drift analysis must be rerun on unflipped tangent displacements, or with a reflection-invariant statistical test, and the paper must report whether the skewness result survives this correction.","section":"Section 2 (Data post-processing); Section 3 (Normality and Skewness tests)"},{"comment":"The assignment of transport mechanisms to experimental conditions is assumed rather than tested. The numerical model hard-codes CytoD-HN as purely active and CytoD-EV as purely passive, with the text stating that 'these scenarios are distinguished by the presence of different vesicle transport mechanisms,' while the parameters in Table 4 are calibrated to the same experimental data used for comparison. Consequently the skewness agreement in 'Numerical vs in-vitro results' is partly manufactured: the drift in simulated CytoD-EV arises from the fitted asymmetry alpha=0.2 in V1, and the absence of drift in CytoD-HN is built into the spatially symmetric V2. The paper should validate the mechanism assignments against independent data, for example through the ATP-depletion or myosin-inhibition experiments already mentioned in the Conclusions, or explicitly reframe the results as a conditional demonstration of a hypothesis rather than a confirmation.","section":"Section 2 (Numerical method); Section 3 (Numerical vs in-vitro results)"},{"comment":"No identifiability or uncertainty analysis is provided for the calibrated parameters in Tables 3 and 4. The sensitivity analysis is one-at-a-time, so it does not address whether different parameter combinations could produce similar skewness, MSD, and histogram outputs, nor the effect of the limited sample sizes (n=15, 13, 14 after quality selection). Without confidence intervals, profile likelihoods, or an identifiability check, the mechanistic interpretation attached to individual fitted values, especially alpha, L1, and the switching rates, is not quantitatively supported. Please add an identifiability analysis and report parameter uncertainty, or clearly label the fitted values as representative rather than inferred.","section":"Section 3 (Parameters settings, Sensitivity analysis)"},{"comment":"The active transport equation (3b) is asserted rather than derived from a physical balance; the manuscript itself notes that 'despite its appearance, Eq.(3b) is not a momentum balance equation.' Since active transport is one of the two central mechanisms, the specific form of the symmetric sawtooth potential and the use of the effective friction xi_eff need a mechanistic derivation or, at minimum, a clear statement of which microphysical ingredients (for example, actin filament elongation rates or receptor detachment kinetics) produce each term. Moreover, because V2 is symmetric, the active mechanism cannot generate net drift by construction; the CytoD-HN case is therefore only a model of undirected rolling, and the absence of drift in CytoD-HN does not discriminate this active mechanism from a passive symmetric diffusive process.","section":"Section 2 (Mathematical model, Eq. (3b))"}],"minor_comments":[{"comment":"The unit for net mean velocity is listed as 'microm/s^-1' or 'um^-1' in the printed table; it should be microm/s.","section":"Table 2"},{"comment":"The caption writes 'Cyto-EV' while the text consistently uses 'CytoD-EV'; please unify the terminology.","section":"Figure 1 caption"},{"comment":"Several cross-references appear as unresolved 'Section ??', both in the main text and in the supplementary material; these need to be fixed before publication.","section":"Main text and Supplementary Information"},{"comment":"The quantity Dr in Eq. (3b) is called a 'rotational diffusion coefficient' but it is assigned the translational value k_B T / xi_eff; please correct the terminology or use a distinct symbol.","section":"Section 2 (Mathematical model, Eq. (3b))"},{"comment":"The sensitivity axes for h2 are labeled in units of k_B T (for example, 'h2 = 1e6 kB T') while Table 4 reports h2 in joules; please make the units consistent or clearly state the conversion.","section":"Figure 9"},{"comment":"The entry for the cytosol viscosity, 'x[10 - 1500] eta_w depending on the protein size', is garbled and should be rewritten as a clear range with a supporting reference.","section":"Table 3"}],"recommendation":"major_revision","confidential_remarks":"The sign-flipping issue is the decisive point for this manuscript. If the skewness result does not survive reanalysis on unflipped data, the central claim will require substantial reformulation, so I would make the corrected drift analysis a mandatory part of the revision. The current validation is also largely in-sample, and the mechanistic decomposition would be materially strengthened by any independent experimental condition that isolates the two proposed transport mechanisms."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The new thing here is a concrete stochastic framework: EVs switch between a passive flashing ratchet with an asymmetric sawtooth and an active flashing ratchet with a symmetric sawtooth, and the three CytoD conditions are mapped to those mechanisms. That formalization is genuinely new and the paper is transparent about its assumptions, parameters, and limitations. The authors also say the data and code are on Zenodo, which is real evidence and worth taking seriously. The model is calibrated to the same data it is then compared with, but the paper does not hide that; it labels the parameters as fitted and runs a one-at-a-time sensitivity analysis.\n\nThe soft spots are real, though. The stress-test note about sign-flipping is on target: the authors standardize each trajectory by flipping its sign when the regression slope is negative, and then run the skewness test that produces the p-values proving drift in Ctrl and CytoD-EV. Finite Brownian trajectories almost always have a nonzero regression slope, so flipping all negative slopes biases the pooled displacement distribution toward positive values. That makes the experimental drift signal and therefore the model's success criterion look like a preprocessing artifact. The authors need to rerun the analysis on unflipped data, or justify the sign-flip differently. This is a load-bearing issue, not a minor quibble.\n\nThe other weaknesses are proportionate: Eq. (3b) is explicitly not a momentum balance and is asserted rather than derived, so the active mechanism is a phenomenological model rather than a mechanical one. The symmetric active potential cannot produce net drift by construction, which is consistent with the CytoD-HN result but also means the model can only ever attribute drift to the passive ratchet. The mapping from CytoD treatment to mechanism is taken from earlier work and never independently tested. The sample sizes are small and the dataset was reduced after the fact for video quality, which the authors disclose but which adds fragility. None of these are fatal on their own, and the paper mostly owns them in the discussion.\n\nWho gets value: biophysicists working on vesicle transport or ratchet models, and experimentalists who want a starting framework for interpreting EV trajectories. If the sign-flip issue is fixed, the paper deserves a serious referee. As is, I would send it to review but flag the preprocessing issue as the main thing the referee must check. The central framework is plausible and the writing is honest; it just needs the evidence cleaned up before it can support the strong claim about matching drift across conditions.","headline":"A mostly honest, useful modeling paper for EV transport whose central drift signal may be an artifact of sign-flipping the data before the skewness test.","tokens_in":20843,"tokens_out":851,"would_cite":false,"duration_ms":9357,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["92C37","60H10","82C31"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper claims that extracellular vesicle transport along neuronal surfaces is captured by a flashing Brownian ratchet with an asymmetric sawtooth potential for passive, cytoskeleton-driven motion and a symmetric sawtooth potential for…","keywords":["extracellular vesicles","prion protein","flashing Brownian ratchet","sawtooth potential","cytochalasin D","actin polymerization","mean squared displacement","neurodegenerative disease"],"falsifier":"Track vesicles after treating neurons with Cytochalasin D and simultaneously blocking PrP–receptor interactions with a competing antibody: the passive-ratchet model predicts that the drift seen in CytoD-EV should vanish, and a persistent directed component would falsify the claim. Alternatively, a long-time measurement of CytoD-HN trajectories would falsify the model's assignment if it reveals a net drift, since the symmetric active potential cannot generate one.","tokens_in":19840,"feed_emoji":"🧬","tokens_out":6188,"duration_ms":54019,"temperature":0.7,"pith_summary":"This paper sets out to establish that the movement of prion-bearing extracellular vesicles along neuron surfaces can be described as a combination of two physical ratchet mechanisms: a passive one, in which the neuron's cytoskeleton drags the vesicle through a receptor, and an active one, in which the vesicle's own actin filaments make it roll from receptor to receptor. The authors build a stochastic model of these mechanisms, calibrate it to video tracking data of vesicles under three conditions, and show that the simulations reproduce the key experimental signature: a directed drifting component in control and vesicle-treated samples, and no drift when the neurons' actin is disabled. If correct, the model provides a quantitative framework for interpreting different transport regimes and a basis for asking how vesicle movement contributes to the spread of misfolded proteins in prion-like diseases.","feed_headline":"Two-ratchet model explains prion-vesicle motion on neurons","feed_subtitle":"Simulations reproduce the drift seen in control and vesicle-treated samples, and none when neuron actin is disabled.","key_machinery":"The central object is the flashing Brownian ratchet, a stochastic process that alternates between free diffusion and motion in a periodic potential. For passive transport the potential is an asymmetric sawtooth $V_1$ of period $L_1$, which generates a directed drift; for active transport the potential is a symmetric sawtooth $V_2$, which by construction produces no net drift and models receptor-to-receptor rolling. The vesicle and receptor are coupled by a linear elastic spring, interpreted as the prion protein, and a four-state Markov chain controls switching between bound and unbound states and between passive and active mechanisms. This machinery carries the argument because the asymmetry of $V_1$ is the sole source of directionality in the model, while the symmetric $V_2$ accounts for undirected active motion.","core_discovery":"The central claim is that a data-driven stochastic model, built from a flashing Brownian ratchet with an asymmetric sawtooth potential for passive transport and a flashing ratchet with a symmetric sawtooth potential for active rolling, captures the key features of extracellular vesicle motion on neuronal surfaces. The model assigns each experimental condition a distinct mechanism: untreated control combines passive and active transport; vesicles treated with Cytochalasin D move purely passively; neurons treated with Cytochalasin D support purely active rolling. Under this assignment, simulations reproduce the experimental skewness finding that Ctrl and CytoD-EV displacements contain a drift component while CytoD-HN displacements do not, and the simulated mean squared displacement curves and displacement histograms agree qualitatively with experiment.","pith_inferences":["An implicit consequence is that 'active transport' in this model is not directional: the symmetric active potential can only enhance undirected motion, so any net displacement in the control must be attributed entirely to the passive ratchet.","A testable extension: treating neurons with Cytochalasin D and simultaneously blocking PrP–receptor binding should abolish the drift seen in CytoD-EV; if a directed component survives, the passive mechanism is not the sole source of directionality.","The model's one-mechanism-per-treatment mapping suggests sorting individual vesicle trajectories by kinematic indicators to check whether control trajectories naturally cluster into a passive-like and an active-like subpopulation, as the two-state switching would predict.","Because the control condition requires switching between passive and active states, the fitted transition rates imply a prediction for the statistics of pauses and jumps in single trajectories, which could be checked against longer recordings."],"forward_implications":["If the model is correct, the directed component of vesicle motion on neurons comes from the passive, cytoskeleton-driven ratchet, so treatments that disable the neuronal actin network should remove drift while treatments that disable vesicle actin should not.","The model predicts that increasing the stiffness of the PrP–receptor link increases vesicle displacements, because the receptor can pull the vesicle more effectively against thermal fluctuations.","The model predicts that lowering the surface density of neuronal receptors impairs vesicle motility, because the vesicle must diffuse farther to reach the next binding site.","The model predicts that in the purely active regime, vesicle motion is enhanced diffusion without net direction, consistent with the experimentally observed absence of a drift component in CytoD-HN.","Because the control condition combines both mechanisms, the model attributes the higher zero-velocity rate and lower mobility of control vesicles to the passive state dominating their motion."],"supporting_citations":[{"why":"Supplies the experimental datasets, the Cytochalasin D treatment protocol, and the passive-versus-active transport hypothesis that the model is built to test.","marker":"[12]"},{"why":"Provides the elastic-linker cargo–motor formalism used for the passive transport equations, with the motor pulling the vesicle through the viscoelastic cytosol.","marker":"[22]"},{"why":"Introduces the ligand–receptor contact random walk that justifies receptor stepping and the elastic coupling between vesicle and receptor.","marker":"[34]"},{"why":"Introduces the flashing Brownian ratchet mechanism that generates directed motion from a time-varying periodic potential.","marker":"[2]"},{"why":"Establishes the forced thermal ratchet basis for drift in an asymmetric periodic potential.","marker":"[33]"},{"why":"Demonstrates that zero-average fluctuations of potential barriers can drive a Brownian particle in a nonsymmetric periodic potential, supporting the passive drift mechanism.","marker":"[5]"}],"fun_headline_variants":["Prion vesicles move via dual ratchet mechanism on neurons","Two ratchets, one vesicle: modeling prion-driven transport","Data-driven ratchets explain prion-EV drift on neurons","Passive and active ratchets guide prion vesicles on neurons","Ratchet physics mimics prion-vesicle motion on neuron surfaces"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that Cytochalasin D cleanly separates the two mechanisms—vesicle treatment disabling only the active one and neuron treatment disabling only the passive one—so each fitted parameter keeps its mechanistic meaning; if the drug also perturbs receptor diffusion or adhesion, the agreement between simulation and experiment becomes curve-fitting.","fun_headline_variants_meta":{"raw":{"variants":["Prion vesicles move via dual ratchet mechanism on neurons","Two ratchets, one vesicle: modeling prion-driven transport","Data-driven ratchets explain prion-EV drift on neurons","Passive and active ratchets guide prion vesicles on neurons","Ratchet physics mimics prion-vesicle motion on neuron surfaces"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000709,"raw_usage":{"total_tokens":3212,"prompt_tokens":983,"completion_tokens":2229,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":599,"completion_tokens_details":{"reasoning_tokens":2142}},"tokens_in":599,"tokens_out":2229,"duration_ms":14368,"temperature":1.0,"reasoning_tokens":2142,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-09T04:21:00.433752+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Track vesicles after treating neurons with Cytochalasin D and simultaneously blocking PrP–receptor interactions with a competing antibody: the passive-ratchet model predicts that the drift seen in CytoD-EV should vanish, and a persistent directed component would falsify the claim. Alternatively, a long-time measurement of CytoD-HN trajectories would falsify the model's assignment if it reveals a net drift, since the symmetric active potential cannot generate one.","supporting_citations":[{"cited_title":"D’Arrigo, M","cited_arxiv_id":null,"evidence_quote":"Supplies the experimental datasets, the Cytochalasin D treatment protocol, and the passive-versus-active transport hypothesis that the model is built to test."},{"cited_title":"Goychuk, V","cited_arxiv_id":null,"evidence_quote":"Provides the elastic-linker cargo–motor formalism used for the passive transport equations, with the motor pulling the vesicle through the viscoelastic cytosol."},{"cited_title":"Marbach, J","cited_arxiv_id":null,"evidence_quote":"Introduces the ligand–receptor contact random walk that justifies receptor stepping and the elastic coupling between vesicle and receptor."},{"cited_title":"Ajdari and J","cited_arxiv_id":null,"evidence_quote":"Introduces the flashing Brownian ratchet mechanism that generates directed motion from a time-varying periodic potential."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Establishes the forced thermal ratchet basis for drift in an asymmetric periodic potential."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Demonstrates that zero-average fluctuations of potential barriers can drive a Brownian particle in a nonsymmetric periodic potential, supporting the passive drift mechanism."}],"review_version":1}