{"id":"e4608ddd-66f9-439f-a062-005179155964","arxiv_id":"2506.18229","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Impermeable dendritic spines produce time-dependent diffusion MRI signals that mimic permeative exchange and can bias NEXI/SMEX exchange-time estimates by up to 80%.","lead":"This preprint uses computer simulations and mathematical escape-time theory to show that water moving between dendritic shafts and their tiny spines can create the same time-dependent MRI signal that is normally read as water crossing cell membranes. It then shows how this could bias estimates of membrane water exchange in gray matter by up to 80%.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The up-to-80% bias figure rests on an unvalidated three-compartment Kärger model whose spine-as-dot assumption is acknowledged to potentially overestimate restriction.","rationale":"The paper's qualitative claim—that impermeable spiny dendrites produce time-dependent SDE signals resembling permeative exchange—is well supported by Monte Carlo simulations and the validated two-compartment Kärger fit (Fig. 5). The reader correctly identifies a quantitative weak spot, but the most load-bearing issue is not the Eq. (3) scaling factor; it is the unvalidated extended three-compartment model that generates the headline 80% bias. The three-compartment model's assumption of a zero-diffusivity spine (dot) and neglected neck is explicitly flagged in Section 5.5 as potentially overestimating restriction. Since the 80% figure is a central quantitative claim in the abstract and conclusion, it should be treated as conditional until the model is validated against MC simulations that include both spine diffusion and membrane permeability. The reader's Eq. (3) concern is real but secondary: shaft-to-spine exchange in the three-compartment model is governed by detailed balance (Eq. 9b), so the fitted scaling factor mainly affects the morphology-dependent exchange time distributions in Fig. 7, not the NEXI bias shown in Fig. 9. The proposed concrete test would directly assess whether the 80% bias is robust to the dot-compartment assumption.","tokens_in":43831,"tokens_out":11420,"duration_ms":108040,"concrete_test":"Re-derive Fig. 9D using the extended Kärger model with D_spine = 1 µm²/ms instead of 0, or, more directly, run Monte Carlo simulations of spiny dendrites with permeable membranes (ξ = 1–20 × 10⁻⁶ m/s) and fit NEXI to the resulting signals. If the estimated bias drops below ~40%, the headline 80% claim would need to be revised.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central quantitative claim in the abstract and conclusion—that unaccounted diffusion-mediated exchange introduces up to 80% bias in NEXI/SMEX estimates—is derived from fitting NEXI to signals generated by the extended three-compartment Kärger model (Section 2.5, Fig. 9), not from Monte Carlo simulations. This model assumes D_spine=0 (dot compartment) and neglects the spine neck. The two-compartment modified Kärger model was validated against MC for the parallel signal (Fig. 5), but the three-compartment model with permeative exchange was not. Section 5.5 explicitly states the dot assumption 'might overestimate the restriction effect.' If D_spine is finite, the spine compartment contributes its own diffusion decay, altering the time-dependence and likely reducing the apparent exchange bias. The reader's Eq. (3) scaling-factor concern is real but less central: in the three-compartment model shaft-to-spine rates are set by detailed balance (Eq. 9b), so the fitted scaling factor mainly affects Fig. 7B-D, not the 80% figure.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript investigates whether water diffusion inside impermeable spiny dendrites can mimic the time-dependent single diffusion encoding (SDE) signal that is usually attributed to permeative exchange. Using Monte Carlo simulations on real 3D-EM-derived and synthetic spiny dendrites, the authors derive spine-to-shaft and shaft-to-spine residence times from narrow escape theory (Eqs. 1-3), fit a modified two-compartment Kärger model to simulated parallel signals, and fit NEXI/SMEX to voxel-level signals with and without an extended three-compartment Kärger model. They report exchange times of 3-26 ms that overlap cortical permeative exchange estimates, show that spine density biases NEXI exchange-time estimates by up to 80%, and conclude that time-dependent SDE cannot disentangle diffusion-mediated from permeative exchange.","tokens_in":43940,"tokens_out":8881,"duration_ms":80226,"significance":"If correct, the paper's central qualitative conclusion is important: time-dependent SDE exchange estimates in gray matter may reflect spine-shaft water exchange rather than membrane permeability alone, with implications for NEXI/SMEX interpretation across regions and pathologies. The paper's strengths include the use of independent Monte Carlo simulations against an analytical narrow-escape prediction for spine-to-shaft residence time (agreement better than 8% under narrow pulses), the use of real 3D EM reconstructions to tune synthetic substrates, and the planned release of analysis code. However, the headline 80% bias figure is produced by an extended three-compartment model that is not validated against the Monte Carlo simulator and whose spine-as-dot assumption is acknowledged as potentially overestimating restriction; the shaft-to-spine theory in Eq. (3) is shown to fail by orders of magnitude and is patched with a fitted scaling factor. These issues leave the quantitative claims materially less secure than the qualitative mimicry claim.","major_comments":[{"comment":"Figure 5B reports that the total diffusion-mediated exchange time t_DM estimated from the modified two-compartment Kärger model matches the theoretical prediction, yet Section 5.2 and Supplementary Figure S3A state that the shaft-to-spine residence time τ_shaft→spine predicted by Eq. (3) is orders of magnitude smaller than the Monte Carlo estimates. Because t_DM is defined in Section 2.2 as 1/(k_spine→shaft + k_shaft→spine), the theoretical t_DM must be pulled far below the MC-based value if Eq. (3) is used uncorrected. The authors should state explicitly whether Figure 5B uses the uncorrected or corrected Eq. (3) values; if uncorrected, the claimed validation of t_DM is unsupported, and quantitative exchange-time ranges derived from the total exchange rate need to be revisited.","section":"Section 5.2, Supplementary Figure S3, Figure 5B"},{"comment":"The abstract and conclusion claim that unaccounted diffusion-mediated exchange can introduce up to 80% bias in NEXI/SMEX exchange-time estimates. This number comes from fitting NEXI to signals generated by the extended three-compartment Kärger model (Section 2.5, Figure 9D), not from Monte Carlo simulations. That model assumes D_spine=0 (a 'dot' compartment) and neglects the spine neck; Section 5.5 explicitly acknowledges that this assumption 'might overestimate the restriction effect.' Since the three-compartment model has not been validated against the Monte Carlo simulator that is the paper's principal evidence source, the 80% figure is not established. The authors should either validate the extended model (including finite D_spine) against MC simulations, or present the 80% figure as an illustrative model prediction with appropriate caveats rather than as a headline quantitative result.","section":"Section 5.5, Figure 9, abstract"},{"comment":"The corrected shaft-to-spine residence times in Figure 7B-D are obtained by multiplying Eq. (3) by a scaling factor estimated from linear regression against Monte Carlo simulations (Supplementary Figure S3B). This scaling factor is a free parameter; the paper does not report its value, its uncertainty, or tests of whether the factor is constant across the morphological ranges used in Figure 7. If the factor is morphology-dependent, the degeneracy analysis and the corrected τ_shaft→spine values are not robust. Please report the fitted factor, its confidence interval, and a sensitivity analysis over the morphological distribution used in Figure 7.","section":"Section 4.3, Supplementary Figure S3"}],"minor_comments":[{"comment":"The manuscript contains numerous typographical errors (e.g., 'modikied', 'kirst', 'kitting', 'kindings', 'reklect') that should be corrected by careful copyediting.","section":"Throughout"},{"comment":"The notation for the apparent diffusivities D_1F and D_2F and the initial conditions in Eq. (7) is dense; explicitly listing the definitions of X_shaft^F and X_spine^F as separate labelled equations would improve readability.","section":"Section 2.4, Eq. (7)"},{"comment":"The sentence 'The data points are smoothed with a one-point window' is unclear; please specify the exact smoothing procedure (e.g., boxcar width, kernel, or moving average).","section":"Section 4.4, Figure 8"},{"comment":"The abstract refers to both NEXI and SMEX, but the text in Section 4.4 describes fitting only the NEXI model; please state explicitly whether the same fits apply to SMEX or whether SMEX is treated as equivalent.","section":"Abstract and Section 4.4"}],"recommendation":"major_revision","confidential_remarks":"The central qualitative finding is timely and will be of interest to the diffusion MRI community. The main risk is the quantitative headline: the 80% bias figure is generated by an as-yet-unvalidated three-compartment model, and the internal inconsistency between Figure 5B and Supplementary Figure S3 regarding the shaft-to-spine theory needs to be resolved. The companion preprint (ref. 122) may contain complementary validation; if so, cross-referencing it could strengthen the current manuscript. Self-citations (refs. 46-48, 122) are used as preliminary work, which is appropriate, but the manuscript should clearly state when conclusions depend on the companion work."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Kadir et al. make a solid case that water diffusing inside impermeable spiny dendrites produces time-dependent SDE signals that look like permeative exchange. The Monte Carlo work is careful: they simulate real EM-reconstructed branches and synthetic mushroom-spine substrates, tune the synthetic ones to match particle escape dynamics, and validate spine-to-shaft escape times against narrow escape theory within 8% under narrow pulses. That part is genuinely new—the earlier ISMRM abstracts hypothesized the effect, but here it is quantified with real morphology and a reproducible digital substrate pipeline. The two-compartment modified Kärger model describing the parallel signal is also a useful addition.\n\nThe soft spots are where the quantitative claims outrun the validation. The shaft-to-spine theory (Eq. 3) misses simulation results by orders of magnitude—the paper says so plainly—and the correction is a fitted scaling factor from the same simulations. That is fine as an internal calibration, but the resulting numbers carry no uncertainty and are not independently predictive. More important, the up-to-80% bias in NEXI/SMEX estimates comes from fitting the extended three-compartment Kärger model, which assumes D_spine = 0 and neglects the neck. The authors acknowledge in Section 5.5 that this 'might overestimate the restriction effect.' The stress-test note makes the right point: the three-compartment model was not validated against MC the way the two-compartment model was. If spine water has finite diffusivity, the bias percentage could shrink.\n\nThere are also no error bars on the key fitted exchange times, and code/data are promised 'upon publication' but not yet linked. Those are fixable. The central qualitative claim—that spines confound exchange estimates—holds up. The specific 3-26 ms and 80% numbers should be treated as order-of-magnitude illustrations until the model is tested against MC with finite spine diffusivity and real spine populations.\n\nWho is this for? Anyone interpreting NEXI/SMEX exchange times in gray matter, and method developers working on DDE or free-waveform encoding. It deserves a serious peer review; the flaws are in the quantitative extrapolation, not the core experiment. I would engage with it and cite it, with a caveat on the bias figure.","headline":"A serious simulation study showing dendritic spines can mimic permeative exchange in time-dependent SDE MRI; the qualitative claim holds, but the headline 80% bias figure rests on a model the authors admit may overestimate restriction.","tokens_in":44559,"tokens_out":1990,"would_cite":true,"duration_ms":19867,"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":"Water in dendritic spines can mimic membrane exchange in diffusion MRI.","keywords":["diffusion MRI","dendritic spines","water exchange","time-dependent diffusion","Kärger model","NEXI","SMEX","narrow escape problem"],"falsifier":"Measure time-dependent SDE signals in a physical phantom of impermeable spiny tubes with known spine-volume fraction and fit NEXI or SMEX: if fitted exchange time does not decrease with spine volume fraction, the mimicry claim fails. Alternatively, simulate substrates with neck radius comparable to shaft radius across multiple neck sizes and check whether the fitted scaling factor for shaft-to-spine escape times generalizes; if it does not, the 80% bias estimate loses its quantitative support.","tokens_in":43551,"feed_emoji":"🧠","tokens_out":4119,"duration_ms":39016,"temperature":0.7,"pith_summary":"This paper argues that water diffusing inside impermeable spiny dendrites, moving between dendritic shafts and spines without crossing any membrane, produces time-dependent single-diffusion-encoding (SDE) MRI signals that look like the signals attributed to membrane permeability. Using Monte Carlo simulations and narrow-escape theory, the authors estimate spine to shaft exchange times of 3 to 26 ms, overlapping published cortical exchange times. They show that a modified two-compartment Kärger model fits the simulated signals but yields exchange estimates that reflect total spine volume fraction rather than spine morphology, and that unaccounted diffusion-mediated exchange can bias NEXI and SMEX exchange-time estimates by up to 80%. The paper therefore cautions that time-dependent SDE measurements in gray matter cannot be interpreted as measuring membrane permeability alone.","feed_headline":"Dendritic spines mimic membrane exchange in diffusion MRI","feed_subtitle":"Impermeable spines bias NEXI and SMEX exchange estimates by up to 80 percent.","key_machinery":"The load-bearing machinery is the narrow-escape problem applied to spine shaft geometry, together with Kärger-style compartment exchange models. The spine-to-shaft residence time comes from an asymptotic narrow-escape formula for a spherical head connected to a cylindrical neck, while the shaft-to-spine time uses a pore-area-density formula for escape through a narrow opening in a cylinder. A modified two-compartment Kärger model treats the spine as a zero-diffusivity dot compartment exchanging with the shaft, and an extended three-compartment version adds an isotropic Gaussian extracellular compartment. Monte Carlo simulations on real and synthetic spiny dendrites supply the ground truth against which the analytical and Kärger predictions are tested.","core_discovery":"On the paper's own terms, the central claim is that the time-dependent SDE signal from impermeable spiny dendrites is indistinguishable from the signal arising from permeative exchange. The discovery is supported by matching particle-escape dynamics between real 3D-reconstructed spiny dendrites and tunable synthetic substrates, by fits of a modified two-compartment Kärger model to simulated parallel signals, and by NEXI/SMEX fits that produce exchange times of 10 to 150 ms for typical spine volume fractions, values comparable to in-vivo gray-matter estimates. The paper further proposes an extended three-compartment Kärger model of shaft, spine, and extracellular space that captures both mechanisms, but shows that it cannot uniquely separate membrane permeability from spine volume fraction. The consequence is that previously reported exchange times in gray matter may partly reflect spine density rather than membrane properties.","pith_inferences":["A direct test of the mimicry claim would be to scan a physical phantom containing impermeable spiny tubes with known spine-volume fraction and check whether fitted exchange time tracks spine density with zero permeability.","If the mimicry holds in vivo, studies of conditions that change spine density, such as autism spectrum disorder, aging, or learning paradigms, may need to re-interpret exchange-time changes as potentially microstructural rather than purely permeability-driven.","The fitted scaling factor used to correct the shaft-to-spine narrow-escape formula suggests that a more general first-passage theory relaxing the narrow-neck assumption would place the 80% bias estimate on firmer ground.","The authors point toward a companion DDE and free-waveform study; if published, combining SDE with those encodings could break the degeneracy between spine volume fraction and permeability and give a practical acquisition design."],"forward_implications":["Time-dependent SDE exchange estimates in gray matter cannot be read as pure membrane permeability; spine volume fraction contributes and can dominate the signal.","NEXI and SMEX exchange-time estimates reflect total spine volume fraction rather than specific spine morphology, so morphological inferences from exchange times alone are degenerate.","The proposed three-compartment Kärger model can fit both diffusion-mediated and permeative exchange, but cannot separate the two, meaning a single SDE protocol is insufficient to identify either mechanism.","Regional differences in exchange times across the brain may partly reflect regional spine-density differences rather than membrane-permeability differences.","Wide gradient pulses violate the Kärger model's assumptions and yield underestimated exchange times; narrow-pulse protocols are safer for interpreting these estimates."],"supporting_citations":[{"why":"Supplies the NEXI model whose exchange-time estimates the paper shows are biased by unaccounted spine exchange.","marker":"[14]"},{"why":"Supplies the SMEX model and the gray-matter exchange framework used as the comparison baseline.","marker":"[13]"},{"why":"Provides the narrow-escape formula for spine-to-shaft residence time used as the analytical prediction.","marker":"[93]"},{"why":"Provides the shaft-to-spine exit-rate formula, Eq. (3), that the paper tests and then corrects with a fitted scaling factor.","marker":"[96]"},{"why":"Provides the modified Kärger model formalism used for the two-compartment fits of the simulated parallel signals.","marker":"[12]"},{"why":"Supplies the narrow-escape problem background and asymptotic escape-time theory underlying the exchange-time derivations.","marker":"[87]"},{"why":"Provides the autism spectrum disorder spine-density data used to illustrate how spine-density changes would shift NEXI exchange-time estimates.","marker":"[75]"}],"fun_headline_variants":["Impermeable spines mimic membrane exchange in diffusion MRI","Spine density biases exchange estimates in diffusion MRI","Impermeable spines create permeation-like MRI signals","Spiny dendrites mimic exchange in diffusion MRI"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The quantitative claims rest on the narrow-escape assumption that spine necks are far narrower than the dendritic shaft, which realistic spines violate; the paper patches the discrepancy with a fitted scaling factor, and the central numbers would shift if that correction is not robust.","fun_headline_variants_meta":{"raw":{"variants":["Impermeable spines mimic membrane exchange in diffusion MRI","Spine density biases exchange estimates in diffusion MRI","Impermeable spines create permeation-like MRI signals","Spiny dendrites mimic exchange in diffusion MRI"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000774,"raw_usage":{"total_tokens":3426,"prompt_tokens":949,"completion_tokens":2477,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":565,"completion_tokens_details":{"reasoning_tokens":2418}},"tokens_in":565,"tokens_out":2477,"duration_ms":17574,"temperature":1.0,"reasoning_tokens":2418,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T18:52:49.086037+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure time-dependent SDE signals in a physical phantom of impermeable spiny tubes with known spine-volume fraction and fit NEXI or SMEX: if fitted exchange time does not decrease with spine volume fraction, the mimicry claim fails. Alternatively, simulate substrates with neck radius comparable to shaft radius across multiple neck sizes and check whether the fitted scaling factor for shaft-to-spine escape times generalizes; if it does not, the 80% bias estimate loses its quantitative support.","supporting_citations":[],"review_version":2}