{"id":"ae0474bf-1547-4fe5-a2bc-2cfe084ff7a6","arxiv_id":"1908.06078","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"Active Brownian particles in an obstacle lattice with a density gradient migrate toward lower obstacle density because obstacles shorten their persistence length, producing a drift toward regions of longer persistence.","lead":"This paper shows that tiny self-propelled particles moving through a maze of pillars drift toward the less crowded side, even though nothing pulls them that way. It offers a simple physical explanation for topotaxis, the tendency of migrating cells to follow gradients in their terrain.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Mechanism is not closed: the paper applies the durotaxis persistence-gradient drift law to a pointwise l_eff(d) field without deriving or numerically testing the predicted v_top, so the claim that persistence modulation alone drives topotaxis remains unverified.","rationale":"The abstract's claim that persistent motion alone is sufficient to drive topotaxis is the central novelty. The paper convincingly establishes two ingredients separately: (i) ABPs in uniform lattices have a reduced effective persistence length that increases with obstacle spacing (Fig. 4), and (ii) cited works show that persistent random walkers in a persistence gradient drift toward higher persistence. But a causal explanation requires demonstrating that applying ingredient (ii) to the local l_eff(d(x)) field in the gradient lattice produces the simulated drift. Sec. III B states this as a qualitative conclusion, citing Refs. [35,50,51], without a quantitative prediction of v_top from the measured l_eff(d) and without testing the local-closure assumption. The Fokker-Planck analysis in Sec. III C treats only homogeneous lattices and explicitly invokes an unspecified smoothing of the discontinuous force, so it does not supply the missing gradient-drift derivation. The numerical observation itself is robust, with 10^6 particles and no net y-drift, so the phenomenological result stands; however, the mechanistic claim that persistence modulation alone is responsible remains conditional because it has not been falsifiably connected to the drift data. The proposed obstacle-free-equivalent simulation directly tests whether the effective-PRW description with measured local parameters reproduces the observed topotactic velocity, which would settle the concern.","tokens_in":16099,"tokens_out":8283,"duration_ms":90835,"concrete_test":"Run obstacle-free ABP/PRW simulations with the measured position-dependent effective parameters tau_eff(d(x)) and v_eff(d(x)) from Fig. 4, using the same gradient lattice spacing profile as in Sec. III A; compute the center-of-mass velocity and compare it with v_top in Figs. 2b,d for at least one r (e.g., r=0.07) and lp=5. If the obstacle-free PRW does not reproduce v_top within statistical error, the local l_eff-gradient mechanism is insufficient and an additional obstacle-specific rectification is present.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central explanatory claim in Sec. III B is that topotaxis is caused by the local reduction of persistence by obstacles: l_eff(d) is measured in uniform lattices (Fig. 4d), and the drift law for persistent random walkers in persistence gradients from Refs. [35,50,51] is invoked to conclude that particles drift toward larger d (lower density). This inference requires a local-closure assumption: the gradient lattice at position x behaves like the uniform lattice with spacing d(x), and the durotaxis drift law remains valid when the persistence length changes on a scale comparable to the lattice spacing and persistence length. Neither is established. Sec. III C does not fill the gap: it only derives reduced tau_eff and v_eff in a uniform lattice, and it relies on an unspecified smoothing of a discontinuous force ('one can imagine to smoothen the force ... without altering the qualitative picture'), so it never derives a gradient drift. A direct quantitative check is missing: no predicted v_top computed from the measured l_eff(d) and d(x) is compared with the simulated v_top in Figs. 2b,d. Without that closure, the observed drift could in principle receive a contribution from geometric rectification by the asymmetric obstacle arrangement, rather than purely from the persistence gradient.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript introduces a minimal model of topotaxis: active Brownian particles (ABPs) moving in a two-dimensional square lattice of circular obstacles whose spacing decreases with increasing x, so that obstacle density is higher on the left and lower on the right. Simulations with 10^6 particles show a mean drift in the +x direction, with a topotactic velocity that increases with the density gradient r and with the persistence length lp. In uniform lattices, the authors measure the mean-squared displacement and velocity autocorrelation function, showing that ABPs remain well described by an effective persistent random walk with reduced effective velocity, persistence time, and persistence length when the obstacle spacing is small. They then argue that this local reduction of persistence in dense regions, combined with the known drift of persistent random walkers toward higher persistence (durotaxis literature), explains the observed topotaxis. A Fokker-Planck moment calculation rationalizes the reduced persistence time and velocity in a uniform lattice, but the calculation involves a smoothing assumption for the discontinuous obstacle force.","tokens_in":16296,"tokens_out":12619,"duration_ms":118065,"significance":"If the mechanism claim were fully established, the paper would provide a generic, cell-type-independent physical mechanism for topotaxis, with implications for active colloids and amoeboid cell migration. The strengths of the paper include the high-statistics simulations (10^6 particles) with explicit error bars, the careful extraction of effective persistent-random-walk parameters and their cross-validation between MSD and VACF (Fig. 4), the symmetry check in the y-direction (Appendix C), and the honest discussion of quantitative discrepancies with the cellular experiments of Ref. [15]. The main weakness is that the central mechanistic claim is not quantitatively closed: the predicted topotactic velocity from the measured persistence-length modulation is never computed and compared with simulation.","major_comments":[{"comment":"The central claim that topotaxis is caused by a persistence gradient is asserted rather than demonstrated. The authors measure l_eff(d) in uniform lattices (Fig. 4d) and combine this with the persistence-gradient drift law of Refs. [35,50,51] to conclude that ABPs in the gradient lattice drift toward lower obstacle density. However, no expression for the predicted topotactic velocity is derived from l_eff(d) and the local spacing d(x) of the gradient lattice, and no comparison with the simulated v_top in Figs. 2b,d is made. Such a closure is necessary to rule out that the drift is partially or entirely due to geometric rectification by the asymmetric lattice deformation. I request a quantitative test: either compute v_top from the drift law using the measured l_eff(d) and compare it with Fig. 2b,d, or perform a control simulation in which the persistence-gradient contribution is eliminated (e.g., by position-dependent rotational diffusion tuned to keep l_eff constant) while retaining the geometric density gradient.","section":"Sec. III B (Figs. 2 and 4d)"},{"comment":"The paper assumes a local-closure relation: ABPs in the gradient lattice at position x are taken to behave as in a uniform lattice with spacing d(x), with a pointwise persistence l_eff(d(x)). This is nontrivial because the persistence length can be comparable to the lattice spacing and because the spacing varies continuously across the gradient region; moreover, the drift law of Refs. [35,50,51] was derived for free-space persistent random walkers with a prescribed persistence gradient, not for ABPs in an obstacle lattice, and it assumes a constant speed, whereas here both v_eff and tau_eff vary with d. The validity of this local closure should be tested directly, for example by measuring the local effective persistence length in the gradient lattice, before it is used to explain the origin of topotaxis.","section":"Sec. III B"}],"minor_comments":[{"comment":"The Taylor expansion of the discontinuous hard-wall force (Eq. (4)) is not a controlled approximation; the statement 'one can imagine to smoothen the force ... without altering the qualitative picture' is a heuristic assumption, and the force in Eq. (4) depends on the orientation p, so F(r) is not a function of r alone. Either regularize the obstacle interaction explicitly or label this derivation as heuristic.","section":"Sec. III C, Eqs. (13)-(15)"},{"comment":"The text calls the obstacle spacing a 'linear gradient' and states that the exponential spacing in n 'leads to a linear gradient' in x; this is only true for small r. For the finite values used (e.g., r=0.15 in Fig. 5), d(x) is exponential in x; please clarify or redefine the lattice to make d(x) exactly linear.","section":"Appendix B, Eq. (B3)"},{"comment":"The topotactic velocity is defined as <Delta x>/Delta t averaged over t in [0,30 tau_p]; since the gradient region has finite width and is flanked by uniform lattices, this is a transient average. Please state explicitly whether particles leave the gradient region by t=30 tau_p and whether the drift is stationary over the averaging interval.","section":"Sec. III A, definition of v_top"},{"comment":"There are several typos: 'cell-type-indenpendent' in Sec. I, 'the the persistence time' in Sec. II, 'a exponential fit' in the caption of Fig. 3, and 'alterning' in Sec. III C.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is a clean simulation study of a toy model, but the central mechanism claim is not quantitatively closed. The missing closure test is straightforward and within the scope of a revision; I would be willing to review a revised version. I have no concerns about novelty or citation practices, though Ref. [15] being a bioRxiv preprint by one of the authors is worth noting in the cover letter."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know this paper has one solid, clearly presented numerical result and a plausible but under-supported mechanistic explanation. The result: ABPs moving through a square lattice of obstacles whose spacing varies linearly in x drift toward lower obstacle density, with drift velocity increasing with gradient steepness and persistence length. That's shown with 10^6 particles and error bars; I trust it.\n\nWhat's new: it's the first clean demonstration I know of that topotaxis can emerge from pure persistence modulation in an ABP toy model, without any cell-specific mechanosensing. The paper properly credits the durotaxis literature (Refs [35,50,51]) for the underlying drift-in-persistence-gradient effect, and shows that obstacles locally shorten the effective persistence length. That's a nice, testable connection.\n\nWhere it's soft: the mechanism is inferred, not derived. They measure l_eff(d) in uniform lattices, then invoke the known persistence-gradient drift law to say particles should drift to high persistence (low density). But they never compute a predicted v_top from l_eff(d) and the local spacing gradient, and compare it to the simulated v_top. The stress-test note is right: without that closure, the drift could in principle include contributions from geometric rectification by the gradient lattice, not just persistence modulation. Sec. III C doesn't help; it's a plausibility argument that Taylor-expands a discontinuous hard-wall force with an unspecified smoothing, and it only derives reduced tau_eff and v_eff in uniform lattices, never a gradient drift. These gaps are addressable—a direct simulation with a prescribed l_eff(x) or a comparison of measured and predicted v_top would settle it. The paper also deposits no code/data, which is a shame but minor.\n\nThe paper is honest about its limits: the authors note the topotactic velocity is about 1% of free speed versus 5% in the cell experiments, and they call the effect 'kinesis' rather than 'taxis'. I appreciate that.\n\nWho's it for: active matter physicists modeling transport in complex environments, and experimentalists who want a minimal baseline for interpreting cellular topotaxis. It deserves a serious referee; the main claim is important enough that peer review should push for the quantitative closure.\n\nMy recommendation: send it out, with the request that the authors either provide the closure test or soften the 'persistence modulation alone' claim to 'consistent with persistence modulation.'","headline":"Solid simulation result; the persistence-modulation mechanism is plausible but not quantitatively closed.","tokens_in":16896,"tokens_out":2700,"would_cite":true,"duration_ms":27995,"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":"Persistent motion alone is sufficient to drive topotaxis: an obstacle-density gradient acts as a persistence gradient that pushes active Brownian particles toward lower density.","keywords":["active Brownian particles","topotaxis","persistence length","obstacle lattice","persistent random walk","durotaxis","directed cell migration","Fokker–Planck equation"],"falsifier":"Run the gradient-lattice simulation, measure the local effective persistence length profile $\\ell_{\\mathrm{eff}}(x)$, predict $v_{\\mathrm{top}}$ from the persistence-gradient drift law, and compare with the directly measured drift; disagreement at steep gradients or small spacings would show the local mechanism is incomplete.","tokens_in":15857,"feed_emoji":"🧭","tokens_out":9355,"duration_ms":87149,"temperature":0.7,"pith_summary":"This paper claims that topotaxis—directed motion along a topographical gradient—does not require chemical sensing or mechanosensing machinery. In a toy model of active Brownian particles moving through a square lattice of circular obstacles whose spacing decreases in one direction, the particles drift, on average, toward the region of lower obstacle density. The reason, the authors argue, is that obstacles renormalize the particles' motion: in a uniform lattice the particles still perform a persistent random walk, but with a shorter effective persistence length $\\ell_{\\mathrm{eff}}$ that decreases as the lattice tightens. A spatial gradient in obstacle density is therefore a spatial gradient in persistence, and persistent random walkers are known to drift up persistence gradients. The paper supports this with simulations of millions of particles and a Fokker–Planck analysis of the renormalized transport coefficients.","feed_headline":"Persistent motion alone steers particles toward open space","feed_subtitle":"A toy model shows obstacle gradients shorten persistence, so drift arises with no chemotactic or mechanosensing cues.","key_machinery":"The central object is the effective persistent random walk (PRW) description of an active Brownian particle in a regular obstacle lattice. The velocity autocorrelation function remains exponential, $\\langle \\mathbf{v}(t+\\Delta t)\\cdot\\mathbf{v}(t)\\rangle = v_{\\mathrm{eff}}^2 e^{-\\Delta t/\\tau_{\\mathrm{eff}}}$, and from the mean-squared displacement at long times one extracts the effective diffusion coefficient $D_{\\mathrm{eff}}$; the relations $D_{\\mathrm{eff}}=v_{\\mathrm{eff}}^2\\tau_{\\mathrm{eff}}/2$ and $\\ell_{\\mathrm{eff}}=v_{\\mathrm{eff}}\\tau_{\\mathrm{eff}}$ connect the renormalized transport coefficients. These coefficients are monotone functions of lattice spacing, turning a density gradient into a persistence gradient. The durotactic drift law for persistent random walkers—drift toward larger persistence—then carries the argument; the paper adds a Fokker–Planck calculation showing that obstacle collisions increase effective rotational diffusion at short times.","core_discovery":"In a lattice of obstacles with a linear gradient in lattice spacing, the average particle position $\\langle x\\rangle$ grows linearly in time, defining a topotactic velocity $v_{\\mathrm{top}}$ that is proportional to the gradient parameter $r$ and increases with the free persistence length $\\ell_p$ faster than linearly. The paper's central discovery is the mechanism: uniform obstacle lattices leave the velocity autocorrelation function exponential, so ABPs remain persistent random walkers with effective velocity $v_{\\mathrm{eff}}$ and effective reorientation time $\\tau_{\\mathrm{eff}}$, both smaller than their free-space values and both increasing with lattice spacing. Hence the effective persistence length $\\ell_{\\mathrm{eff}} = v_{\\mathrm{eff}}\\tau_{\\mathrm{eff}}$ is a decreasing function of obstacle density, and the density gradient becomes a persistence gradient. Applying the durotaxis result that persistent random walkers drift toward higher persistence, the authors conclude that the particles move to lower obstacle density purely because persistence is spatially modulated. A short-time Fokker–Planck expansion identifies the effective increase in rotational diffusion caused by obstacle encounters, while the long-time analysis shows why the effective diffusion coefficient is reduced.","pith_inferences":["The paper links the mechanism to the measured $\\ell_{\\mathrm{eff}}(d)$ only qualitatively; a direct test would be to compute the drift velocity predicted by combining $\\ell_{\\mathrm{eff}}(d)$ with the persistence-gradient drift law and compare it with the simulated $v_{\\mathrm{top}}$, especially for steep gradients where local equilibrium may fail.","Because the argument only uses persistence, the same topotactic drift should appear in run-and-tumble or Lévy-walk models with the same persistence length, a prediction that could be tested in the same lattice geometry.","In the deterministic limit of infinite persistence, the mechanism must break down, since reorientation is what lets obstacles randomize directions; this suggests the drift should be non-monotonic in persistence length when $\\ell_p$ greatly exceeds the lattice size.","The cell experiments' fivefold higher efficiency hint that cells may combine this passive persistence effect with active responses; measuring the persistence length of cells on uniform pillar arrays with varying spacing would separate the generic physical contribution from cell-specific sensing."],"forward_implications":["Topotaxis is generic: any self-propelled particle whose orientation persists over a finite time and that moves through a spatially varying obstacle distribution will drift toward the less crowded side, with no need for gradient sensing or biochemical signaling.","The topotactic velocity increases with both the steepness of the obstacle-density gradient and the free persistence length, and because $\\ell_{\\mathrm{eff}}$ depends more strongly on density for large $\\ell_p$, the speed grows superlinearly with persistence.","Inside obstacle lattices, active Brownian motion remains a persistent random walk at long times, so large-scale transport can be summarized by the two renormalized numbers $v_{\\mathrm{eff}}$ and $\\tau_{\\mathrm{eff}}$.","The phenomenon is better classified as kinesis than taxis, since the topographical cue acts by changing local motility rather than by imposing a directional bias on individual steps.","When compared with experiments on amoeboid cells, the toy model reproduces the direction of topotaxis but underestimates its efficiency (about 1% of the intrinsic speed versus about 5% in cells), indicating that cell-specific mechanisms add to this generic persistence effect."],"supporting_citations":[{"why":"It shows persistent random walkers in a persistence gradient drift toward higher persistence, supplying the durotaxis law the mechanism relies on.","marker":"[35]"},{"why":"It is a phenomenological durotaxis model demonstrating the same persistence-gradient drift, supporting the mechanism.","marker":"[50]"},{"why":"It is a random-walk model for durotaxis that derives the net flux toward larger persistence and is used for the kinesis-versus-taxis comparison.","marker":"[51]"},{"why":"It reports the experiments on amoeboid cell topotaxis that motivate the model and provide the quantitative comparison for topotactic velocity.","marker":"[15]"},{"why":"It establishes that ABPs in obstacle lattices behave as an effective persistent random walk, the concept the paper transfers to regular lattices.","marker":"[28]"},{"why":"It supplies the frictionless hard-wall force used for particle–obstacle interactions.","marker":"[19]"},{"why":"It is the companion model for strongly confined ABPs, also used for the wall-force prescription.","marker":"[20]"}],"fun_headline_variants":["Persistence gradients alone drive topotaxis","Obstacle spacing shifts persistence, steering active particles","Active particles navigate by persistence, not cues","Topotaxis emerges from persistence modulation","Drift toward open space: it's all persistence"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument assumes that the typical distance a particle keeps moving in one direction, measured in a uniform obstacle array, still describes its motion at each point in a density gradient, so the known drift toward longer persistence can be applied point by point.","fun_headline_variants_meta":{"raw":{"variants":["Persistence gradients alone drive topotaxis","Obstacle spacing shifts persistence, steering active particles","Active particles navigate by persistence, not cues","Topotaxis emerges from persistence modulation","Drift toward open space: it's all persistence"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000169,"raw_usage":{"total_tokens":1246,"prompt_tokens":907,"completion_tokens":339,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":523,"completion_tokens_details":{"reasoning_tokens":270}},"tokens_in":523,"tokens_out":339,"duration_ms":4188,"temperature":1.0,"reasoning_tokens":270,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T12:56:05.866261+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the gradient-lattice simulation, measure the local effective persistence length profile $\\ell_{\\mathrm{eff}}(x)$, predict $v_{\\mathrm{top}}$ from the persistence-gradient drift law, and compare with the directly measured drift; disagreement at steep gradients or small spacings would show the local mechanism is incomplete.","supporting_citations":[{"cited_title":"Chemotaxis of active janus nanoparticles,","cited_arxiv_id":null,"evidence_quote":"It shows persistent random walkers in a persistence gradient drift toward higher persistence, supplying the durotaxis law the mechanism relies on."},{"cited_title":"Human spermatozoa migration in microchan- nels reveals boundary-following navigation,","cited_arxiv_id":null,"evidence_quote":"It is a phenomenological durotaxis model demonstrating the same persistence-gradient drift, supporting the mechanism."},{"cited_title":"Ciliary contact interactions dominate surface scatter- ing of swimming eukaryotes,","cited_arxiv_id":null,"evidence_quote":"It is a random-walk model for durotaxis that derives the net flux toward larger persistence and is used for the kinesis-versus-taxis comparison."},{"cited_title":"Chemo- taxis and topotaxis add vectorially for amoeboid cell mi- gration,","cited_arxiv_id":null,"evidence_quote":"It reports the experiments on amoeboid cell topotaxis that motivate the model and provide the quantitative comparison for topotactic velocity."},{"cited_title":"kinesis” than as a “taxis","cited_arxiv_id":null,"evidence_quote":"It establishes that ABPs in obstacle lattices behave as an effective persistent random walk, the concept the paper transfers to regular lattices."},{"cited_title":"Dynamics of self-propelled particles under strong confinement,","cited_arxiv_id":null,"evidence_quote":"It supplies the frictionless hard-wall force used for particle–obstacle interactions."},{"cited_title":"Microswimmers in patterned environ- ments,","cited_arxiv_id":null,"evidence_quote":"It is the companion model for strongly confined ABPs, also used for the wall-force prescription."}],"review_version":1}