{"id":"4ddffb02-dd69-4bff-b31b-ecc88b23d4c4","arxiv_id":"2411.19664","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"Inhibitory neuron variability in a receiver population controls whether two coupled cortical-like networks switch from delayed to anticipated synchronization via bistability or via zero-lag synchronization.","lead":"This paper simulates two coupled networks of spiking neurons and shows that varying the mix of fast-spiking versus low-threshold-spiking inhibitory cells changes whether the networks lock with a delay, with anticipation, at zero lag, or in a bistable regime. The result gives a concrete model mechanism for how local inhibitory cell diversity might shape communication between distant brain areas.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Sec. III.D mechanism is post-hoc: the uncoupled free-running period relation may not persist under coupling, so the proposed DS-AS route rule remains untested.","rationale":"The reader's weakest_assumption identifies Sec. III.D as the load-bearing concern: the uncoupled free-running period comparison is used to explain the DS-AS transition route, but it is post-hoc, derived from the same simulations, and lacks a derivation or statistical test. I concur that this is the single most critical weakness. The paper's central observational claims—that inhibitory heterogeneity (Xi) modulates phase diversity and that the transition can occur via zero-lag or bistability—are supported by the simulations shown, though the maps lack error bars and code is absent. However, the stronger explanatory claim about the mechanism is not secure. My proposed test directly checks whether the period-relation rule generalizes beyond the chosen parameter lines and whether it survives when the period is measured under coupling rather than at gE = 0. This is a concrete, feasible computational check that would settle whether the mechanism is real or an artifact. Since the reader already assigned CONDITIONAL with the same primary concern, my read does not change the verdict; I set verdict_should_be to UNCHANGED. I also note the secondary reproducibility issue (no code/data) and the lack of error bars in phase maps, but those do not alter the verdict beyond what the reader stated. The paper's prior work on excitatory heterogeneity (Ref. 53) provides context, but the inhibitory-heterogeneity extension is novel and worth reporting if the mechanism claim is either supported or explicitly softened.","tokens_in":14214,"tokens_out":4221,"duration_ms":36212,"concrete_test":"Run a systematic grid over the parameter space shown in Figs. 8-10, e.g., X in [-5, 10] step 1, Xi in [-0.045, 0.045] step 0.005, and gI in [3, 7] nS step 0.5, for both the uncoupled (gE = 0) and coupled (gE = 0.5 nS) cases. For each parameter set, record the free-running periods TS and TR from the gE = 0 simulation, and classify the DS-AS transition route (via bistability or zero-lag) from the coupled simulation using the τ_i histogram criterion illustrated in Fig. 5. Then test the paper's rule: if TR < TS for all points along the DS branch, the route should be via bistability; if TR crosses TS (TR = TS), the route should be via zero-lag. Compute the confusion matrix and balanced accuracy over the full grid, and also repeat the test using the period of the receiver measured in the coupled DS regime (not gE = 0) to see whether the rule holds when the sender's drive is present.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central mechanistic claim is in Sec. III.D: the DS-AS transition occurs via bistability if the free-running receiver is already faster than the sender (TR < TS) in the DS regime, and via zero-lag if the free-running periods cross (TR = TS). This rule is extracted from comparing gE = 0 simulations (Fig. 11) with coupled simulations (Figs. 9, 10, 12) for a few selected parameter lines, with no derivation from the coupled equations and no statistical validation. The load-bearing weakness is that the uncoupled period TR is measured without the sender's excitatory drive (gE = 0), yet in the coupled regime each receiver neuron receives 20 excitatory synapses from the sender, which can slow or speed the receiver and change its effective period. If the period relation measured at gE = 0 does not persist when coupling is turned on, the proposed mechanism for why the FS network transitions via bistability and the LTS network via zero-lag is not established. The raw phase-diversity observations (DS, AS, ZL, BI) could still stand, but the explanatory claim about the route would be unsupported. The maps in Figs. 8-10 also lack error bars, and the 'enlarges the region' claim rests on a single visual comparison (Fig. 9(g), Xi ≈ 0.1 vs Xi = -0.04). No code or data are provided, so the reader cannot independently check the period comparisons.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies a motif of two unidirectionally coupled Izhikevich spiking networks, a sender and a receiver, and investigates how the composition of inhibitory neuron types (fast-spiking vs. low-threshold-spiking) in the receiver affects the phase relationship between the two populations. By varying the inhibitory heterogeneity parameter Xi, the excitatory heterogeneity X, and the inhibitory conductance gI, the simulations produce delayed synchronization (DS), anticipated synchronization (AS), zero-lag synchronization (ZL), and a bistable regime (BI) that alternates between DS and AS. The authors compare homogeneous only-FS and only-LTS receivers, map the regimes in parameter space, and propose in Sec. III.D that the route of the DS-AS transition is determined by the uncoupled free-running period of the receiver relative to the sender: bistability occurs when the receiver is already faster in the DS regime, while zero-lag occurs when the free-running periods cross.","tokens_in":14447,"tokens_out":4010,"duration_ms":33763,"significance":"If the reported observations are robust, the paper makes a useful contribution by showing that local inhibitory cell-type composition, not just coupling strength, can control inter-areal phase diversity in a biologically plausible model class. The extensive time-series examples, delay histograms, and parameter scans provide a clear phenomenology of DS, AS, ZL, and BI regimes, and the comparison between only-FS and only-LTS networks is a natural and instructive design. The strength of the paper is the demonstration that phase diversity, including anticipating synchronization and bistability, arises in a spiking network model with heterogeneous inhibitory neurons, which had not been shown before. The main weakness is that the mechanistic rule proposed in Sec. III.D is post-hoc and untested, and the prominent claim that heterogeneity 'enlarges' the zero-lag and bistability regions rests on a visual comparison without statistical support. With additional quantitative analysis, the contribution could be significant for understanding how local inhibitory variability shapes communication between cortical areas; as it stands, the paper is more descriptive than explanatory.","major_comments":[{"comment":"The proposed mechanism for the DS-AS transition route is not established. The rule that bistability occurs when the free-running receiver period is already shorter than the sender's (TR < TS in the DS regime) and that zero-lag occurs when TR = TS is inferred by comparing gE = 0 simulations with gE = 0.5 nS coupled simulations at a small set of parameter values. However, in the coupled regime each receiver neuron receives 20 excitatory synapses from the sender (Sec. II.C), which can alter the receiver's effective oscillation period. The uncoupled period relation may therefore not persist once coupling is switched on, and the paper provides neither a derivation from the coupled equations nor a systematic test of the rule across the parameter space. The comparison covers only a few horizontal slices (e.g., gI = 4.0 and 6.0 nS, X in a narrow range), and no statistical measure is given for how well the period relation predicts the observed route. Because the concluding remarks state that the mechanism has been demonstrated, this post-hoc, untested rule is load-bearing and requires either a direct test (e.g., computing the effective receiver period under coupling and checking the prediction over the full (X, Xi, gI) grid) or an explicit reframing as a heuristic observation that is not part of the paper's central claims.","section":"Sec. III.D, Figs. 11-12"},{"comment":"The claim that inhibitory heterogeneity 'enlarges the region of zero-lag synchronization and bistability' (Abstract and Sec. III.C) is supported only by a visual comparison of the widths of the green region at Xi ≈ 0.1 and Xi = −0.04 in Fig. 9(g). This is a single qualitative comparison with no error bars, no bootstrapped region boundaries, and no quantitative definition of what constitutes an enlargement. The phase maps in Figs. 8-10 are color-coded without confidence intervals, so it is unclear whether the apparent enlargement is robust to initial conditions, finite simulation length, or noise realizations. Since this claim is prominent in the abstract and in the final summary of findings, it should be supported by a quantitative measure, such as the area of each regime in parameter space as a function of Xi, with error bars computed from repeated simulations or bootstrapping.","section":"Sec. III.C, Fig. 9(g)"}],"minor_comments":[{"comment":"The text says 'only fast-spiking inhibitory neurons ( only-FS network). or only low-threshold spiking inhibitory neurons ( only-FS network )'; the second instance should read 'only-LTS network'.","section":"Sec. II.B"},{"comment":"Fig. 11 refers to 'the parameter controlling the excitatory heterogeneity Xi', but Xi controls the inhibitory heterogeneity, not the excitatory one. The same mislabeling appears in the text immediately above Fig. 11.","section":"Sec. III.D and Fig. 11 caption"},{"comment":"The caption begins 'The man time delay τ as a function of Xi' and should be 'The mean time delay'.","section":"Fig. 10 caption"},{"comment":"No code or data availability statement is provided, and the simulation details (number of realizations, simulation length for each point, random seed handling) are not fully specified, which would be needed for independent reproduction of the parameter maps.","section":"General"},{"comment":"The introduction contains an incomplete citation '[6, 12? ]'; the placeholder should be resolved.","section":"Sec. I"}],"recommendation":"major_revision","confidential_remarks":"The paper's main phenomenological results are likely of interest to the q-bio.NC community, but the current manuscript is stronger as a descriptive study than as a mechanistic one. The Sec. III.D mechanism is presented as a suggestion, but it is then referenced in the concluding remarks as if demonstrated; this mismatch needs to be resolved. The lack of quantitative support for the 'enlargement' claim and the absence of code/data are also concerns in the current reproducibility climate. I would support a major revision that adds a systematic test of the period-ratio rule, quantitative error bars for the phase diagrams, and a clear separation between observed phenomenology and heuristic explanation."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick read for you. This is a simulation study of two unidirectionally coupled Izhikevich populations, asking whether the mix of fast-spiking (FS) and low-threshold-spiking (LTS) inhibitory neurons in the receiver controls the phase relation with the sender. The core observation is new and, as far as I can tell, holds up: only-FS receivers transition from delayed to anticipated synchronization through a bistable regime, while only-LTS receivers do it through zero-lag synchronization, and the heterogeneity parameter Xi moves the system between these routes. The supporting material — time series, per-cycle delay histograms, and the tau versus X and tau versus gI scans — is internally consistent. Credit where due: the distinction between FS-driven bistability and LTS-driven zero-lag is genuinely not in their cited literature, and they are honest about what is observation versus mechanism.\n\nSoft spots, in proportion. Section III.D proposes that the transition route is set by the free-running period relation: TR < TS in the DS regime facilitates bistability, TR = TS gives zero-lag. This is a post-hoc correlation drawn from comparing gE = 0 simulations with coupled runs for a handful of parameter lines. No derivation, no statistical test, and the coupled receiver is phase-locked to the sender by construction, so its coupled period equals the sender's period; the free-running period at gE = 0 is a different quantity, and nothing in the paper shows that relation survives coupling. It is labeled a suggestion, which is honest, but the Conclusion leans on it harder than the evidence supports. Also: no error bars on the phase maps, the 'enlarges the region' claim rests on one visual comparison, and no code or data are provided. There are typos — Sec. II.B mislabels the only-LTS case as 'only-FS', the Xi approximately -0.3 in Sec. III.D should presumably be -0.03, and the Xi approximately 0.1 mentioned for Fig. 9(g) is outside the stated [-0.045, 0.045] range where their parameterization is defined.\n\nWho this is for: anyone working on anticipated synchronization, phase lags between cortical areas, or functional roles of inhibitory subtypes. It deserves a serious referee. My recommendation: send it to review, and let the referee ask for either a real test of the period-ratio mechanism (or removal of the mechanistic framing), error bars on the maps, and a code release. The raw phase-diversity observations are worth reporting regardless.","headline":"Solid and genuinely new observations on how FS/LTS inhibitory composition controls DS-AS transition routes, but the proposed period-ratio mechanism is post-hoc and should be tested or demoted.","tokens_in":15043,"tokens_out":3507,"would_cite":true,"duration_ms":27118,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["87.18.Sn","87.19.ll","87.19.lm"],"model":"deepseek-v4-flash","headline":"The paper argues that the composition of inhibitory neuron types in a receiver population—fast-spiking versus low-threshold-spiking—determines the phase relationship between two unidirectionally coupled cortical-like networks, and that…","keywords":["neuronal heterogeneity","inhibitory neurons","fast-spiking neurons","low-threshold-spiking neurons","phase diversity","anticipated synchronization","phase bistability","zero-lag synchronization"],"falsifier":"Simulate the coupled motif across a dense grid of (Xi, gI, X) and compare the DS-AS transition route with the sign of TR - TS from the corresponding uncoupled simulations; find one parameter point where TR < TS throughout the DS regime yet the transition goes through zero-lag synchronization rather than bistability, or where TR = TS yet the transition is bistable, either of which would contradict the proposed period-relation rule.","tokens_in":13958,"feed_emoji":"🧠","tokens_out":8483,"duration_ms":66689,"temperature":0.7,"pith_summary":"This paper studies two coupled populations of spiking neurons, a sender and a receiver, and asks whether the known diversity of inhibitory neurons changes how the two populations lock in phase. It claims that the local mix of fast-spiking (FS) and low-threshold-spiking (LTS) inhibitory neurons in the receiver is an active control parameter: depending on this mix, the receiver can lag behind the sender (delayed synchronization), fire ahead of the sender (anticipated synchronization), fire at zero lag, or switch unpredictably between lag and anticipation (phase bistability). The paper further claims that inhibitory heterogeneity enlarges the regions of zero-lag and bistable behavior, and that the route from delayed to anticipated synchronization—via zero lag or via bistability—is set by whether the uncoupled receiver's free-running period is already shorter than the sender's or crosses it. If true, this gives a mechanistic role to inhibitory cell-type diversity in determining how distant brain areas communicate through oscillations.","feed_headline":"Inhibitory neuron mix controls phase lag between coupled brain areas","feed_subtitle":"Fast-spiking vs low-threshold-spiking composition sets delayed, anticipated, zero-lag, or bistable phase.","key_machinery":"The central object is a sender-receiver motif of two Izhikevich spiking-neuron populations coupled unidirectionally by excitatory chemical synapses, with local inhibitory feedback conductance gI. Within this motif, neuronal heterogeneity is controlled by the parameter Xi, which shifts the distribution of the Izhikevich parameters (a,b) for inhibitory neurons between fast-spiking (a = 0.10, b = 0.20) and low-threshold-spiking (a = 0.02, b = 0.25) types, and by X, which shifts excitatory types among regular-spiking, chattering, and intrinsically bursting neurons. The observable that carries the argument is the mean time delay τ between the peaks of the two populations' mean membrane potentials: positive τ defines delayed synchronization, negative τ defines anticipated synchronization, near-zero τ defines zero-lag synchronization, and a bi-Gaussian distribution of cycle-to-cycle delays defines phase bistability. The proposed mechanism for the DS-AS route compares the free-running periods TS (sender) and TR (uncoupled receiver, gE = 0): bistability arises when TR < TS already in the DS regime, and zero-lag transitions occur where the period curves cross, TR = TS.","core_discovery":"The central discovery is that changing the distribution of inhibitory neuron types at the receiver—from all fast-spiking to all low-threshold-spiking, and through the intermediate heterogeneous mixtures controlled by the parameter Xi—changes the phase-locking regime between the sender and receiver populations. Both homogeneous inhibitory networks can show delayed synchronization, anticipated synchronization, and phase bistability, and the all-LTS network additionally shows zero-lag synchronization. As Xi is varied, the system can transition from DS to AS through either a bistable phase (at lower inhibitory conductance, gI = 4 nS) or through zero-lag synchronization (at higher gI = 6 nS), and the paper relates this route to the free-running periods: when the receiver remains faster than the sender through the DS regime, the transition is via bistability; when the receiver's free-running period crosses the sender's, the transition is via zero-lag. The paper presents this period-relation rule as a suggested mechanism rather than a derivation from the coupled equations.","pith_inferences":["Beyond the paper's claims: because the period-relation rule is inferred from comparing coupled and uncoupled simulations at selected parameter points, a natural next step is to map the TR = TS surface across the full (Xi, gI, X) parameter space and test whether the zero-lag/bistability boundary tracks that surface exactly; this would turn a suggested correlation into a predictive law.","Beyond the paper's claims: if the FS/LTS ratio acts this way in vivo, optogenetically shifting the inhibitory cell-type balance in one cortical area should measurably change its phase lag relative to a connected area, providing a direct experimental test.","Beyond the paper's claims: the slow alternation between DS and AS in the bistable regime could, if the mechanism generalizes, serve as a substrate for perceptual rivalry and other switch-like cognitive states, because the phase difference itself would carry the switching information.","Beyond the paper's claims: extending the two-population motif to a chain of areas, each node's inhibitory composition could set its phase relation to the next and thereby route information directionally without changing connection strengths."],"forward_implications":["The local proportion of FS versus LTS inhibitory neurons becomes a candidate control parameter for inter-area phase relations, complementing coupling strength and propagation delay.","Changing the inhibitory cell-type composition of a receiver population should shift the system from delayed to anticipated synchronization, with the transition route (zero-lag or bistable) tracking the free-running period relation.","In bistable regimes, cycle-to-cycle phase differences alternate between positive and negative values, so short observation windows would misclassify the regime as purely DS or purely AS.","The enlargement of zero-lag and bistable regions by heterogeneity implies that neuronal variability can promote, rather than merely disrupt, coherent phase relations between coupled populations.","Anticipation times in this model emerge from heterogeneity rather than being imposed by hard-wired delayed self-feedback."],"supporting_citations":[{"why":"Defines the anticipated synchronization solution R(t) = S(t + td) for unidirectionally coupled systems with delayed self-feedback, the regime the paper studies.","marker":"[27]"},{"why":"Shows that an inhibitory loop mediated by chemical synapses can produce anticipated synchronization in neuronal systems, a mechanism the paper builds on.","marker":"[44]"},{"why":"Demonstrates phase diversity and anticipated synchronization in coupled cortical-like populations and provides the phase-locking framework and motivation.","marker":"[48]"},{"why":"Reports phase bistability between delayed and anticipated synchronization, the phenomenon the paper extends to heterogeneous inhibitory networks.","marker":"[52]"},{"why":"Provides the previous study of excitatory heterogeneity in the same motif, whose approach is extended here to inhibitory heterogeneity.","marker":"[53]"},{"why":"Defines the Izhikevich neuron model used for all neurons in the sender and receiver populations.","marker":"[55]"},{"why":"Supplies the random parameter distributions for neuronal types that the paper modifies to control heterogeneity.","marker":"[56]"},{"why":"Shows effects of inhibitory heterogeneity on synchronization in single populations, providing the backdrop for the paper's two-population question.","marker":"[6]"}],"fun_headline_variants":["Inhibitory neuron variability controls phase lag between brain areas","Mix of fast and slow inhibitors sets brain area phase relations","Inhibitory heterogeneity switches sync from delayed to anticipated","Phase diversity in coupled networks tuned by inhibitory mix","How inhibitory variability drives zero-lag and bistable phases"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The paper's mechanistic explanation rests on the premise that the free-running periods measured with the coupling turned off (gE = 0) continue to dictate the transition route once the coupling is turned on, a relation inferred from a limited set of parameter points rather than derived from the coupled equations.","fun_headline_variants_meta":{"raw":{"variants":["Inhibitory neuron variability controls phase lag between brain areas","Mix of fast and slow inhibitors sets brain area phase relations","Inhibitory heterogeneity switches sync from delayed to anticipated","Phase diversity in coupled networks tuned by inhibitory mix","How inhibitory variability drives zero-lag and bistable phases"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000543,"raw_usage":{"total_tokens":2634,"prompt_tokens":1011,"completion_tokens":1623,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":627,"completion_tokens_details":{"reasoning_tokens":1545}},"tokens_in":627,"tokens_out":1623,"duration_ms":11381,"temperature":1.0,"reasoning_tokens":1545,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T05:58:48.679663+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Simulate the coupled motif across a dense grid of (Xi, gI, X) and compare the DS-AS transition route with the sign of TR - TS from the corresponding uncoupled simulations; find one parameter point where TR < TS throughout the DS regime yet the transition goes through zero-lag synchronization rather than bistability, or where TR = TS yet the transition is bistable, either of which would contradict the proposed period-relation rule.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Defines the anticipated synchronization solution R(t) = S(t + td) for unidirectionally coupled systems with delayed self-feedback, the regime the paper studies."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Shows that an inhibitory loop mediated by chemical synapses can produce anticipated synchronization in neuronal systems, a mechanism the paper builds on."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Demonstrates phase diversity and anticipated synchronization in coupled cortical-like populations and provides the phase-locking framework and motivation."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Reports phase bistability between delayed and anticipated synchronization, the phenomenon the paper extends to heterogeneous inhibitory networks."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the previous study of excitatory heterogeneity in the same motif, whose approach is extended here to inhibitory heterogeneity."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Defines the Izhikevich neuron model used for all neurons in the sender and receiver populations."},{"cited_title":"Izhikevich, IEEE Transaction on Neural Networks 14, 1569 (2003)","cited_arxiv_id":null,"evidence_quote":"Supplies the random parameter distributions for neuronal types that the paper modifies to control heterogeneity."},{"cited_title":"Di Volo and A","cited_arxiv_id":null,"evidence_quote":"Shows effects of inhibitory heterogeneity on synchronization in single populations, providing the backdrop for the paper's two-population question."}],"review_version":1}