{"id":"fa51607f-9714-414b-a055-b304e34c6759","arxiv_id":"2601.03478","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"high","formal_verification":"none","parameter_count":4,"one_line_summary":"After dance training, inter-person EEG synchronization increased while hand-motion synchronization decreased, a dissociation the authors call the coupling-decoupling paradox.","lead":"A small study of dancers moving together found that after a generative-dance training program, their brain signals synchronized more across partners while their hand movements synchronized less. The finding suggests group togetherness may live in shared neural intention rather than mirrored motion, but the study's small, non-independent samples and lack of controls leave the result provisional.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"EEG 'relative analysis' is undefined and direct pre-post EEG effects are non-significant; the claimed brain increase, and hence the dissociation, is not established.","rationale":"The reader's weakest assumption (non-independence, lack of control) is a valid concern, but I find a more fundamental issue: the neural side of the dissociation is not demonstrably present in the direct pre-post comparisons. Section 4 reports F(1,8)=0.216, p=.654 for the direct EEG comparison, and Table 1 shows no task-level EEG effect. The only significant EEG result is an undefined 'relative analysis' contrast, which the abstract then reinterprets as 'inter-brain synchronization increased.' This is not a reproducible or interpretable claim as written. The motor decrease is direct and significant, so if the brain increase disappears once the relative analysis is properly defined and modeled, the core paradox disappears. The reader's statistical concern about inflated degrees of freedom is real but secondary; even if corrected, the undefined contrast must first be clarified. I therefore agree with the CONDITIONAL verdict: the paper should be accepted only after the authors provide the missing definition, show the effect survives mixed-model treatment, and make the analysis transparent. Since my concern does not move the verdict away from CONDITIONAL, I mark it UNCHANGED.","tokens_in":15033,"tokens_out":7686,"duration_ms":70327,"concrete_test":"Obtain from the authors the exact definition of the 'relative EEG analysis' (e.g., (EEG_FM_POS - EEG_DM_POS) - (EEG_FM_PRE - EEG_DM_PRE), or similar). Re-analyze the EEG data with a mixed-effects model using dyad as a random intercept, testing (a) the direct pre-post effect on a synchronization index (mean IF or edge incidence) for FM and RM separately, and (b) the defined relative contrast. If neither (a) nor (b) shows a significant, interpretable increase in synchronization after GDAM, the central dissociation claim fails.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim — that GDAM increased inter-brain synchronization while decreasing motor synchrony — is undercut by the reported EEG statistics. Section 4 shows the direct pre/post EEG comparison is non-significant: F(1,8)=0.216, p=.654. The only significant EEG result comes from an unnamed 'relative EEG analysis' (F(1,8)=10.787, p=.011), with a post-hoc contrast for FM vs DM (Table 1: DM,FM, p=.01). However, the Methods never define this 'relative analysis' — no formula, no dependent variable specification, no random effects. Table 1 shows no individual task reaches significance for pre/post EEG (FM p=.36; DM p=.37; RM p=.20). Thus the abstract's assertion that 'inter-brain synchronization increased, particularly within the frontal lobe' is an inference from a task-relative contrast, not a demonstrated pre-post increase. The motor decrease, by contrast, is a direct pre-post effect for FM and RM. The purported dissociation therefore compares a direct motor effect with an undefined, task-relative EEG effect. Until the relative analysis is explicitly defined and shown to correspond to a real increase in synchronization, the 'coupling-decoupling paradox' lacks empirical support. This is compounded by the small number of EEG dyads (5) and non-independence, but the missing definition is the primary logical gap.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports a pre-post study of four-person collaborative dance improvisation, with dual motion capture and EEG hyperscanning (recorded from two of four dancers per session). Motor behavior is represented by time-resolved α-exponents from Movement Element Decomposition, and synchrony is quantified with motif synchronization; neural synchrony is assessed with incidence-fidelity multilayer time-varying graphs. The authors claim that after a Generative Dance through Agent Modelling (GDAM) program, inter-brain synchronization increased—particularly in frontal regions—while interpersonal motor synchrony decreased, which they interpret as a 'coupling-decoupling paradox.' The reported statistics show a direct pre/post decrease in motor synchrony for the Free Movement (FM) and Rule-Based Movement (RM) tasks, but the EEG increase relies on an undefined 'relative EEG analysis'; the direct EEG pre/post comparison is non-significant. The paper's central claim is therefore not yet supported by the evidence as presented.","tokens_in":15377,"tokens_out":5036,"duration_ms":53785,"significance":"If the dissociation were robustly established, this would be a valuable contribution to the neuroscience of joint action and dance, offering a concrete example where neural alignment and motor divergence coexist and suggesting that togetherness may be indexed neurally rather than motorically. The methodological toolkit—motif synchronization with lead-lag structure, MED-derived α time series, multilayer TVG, randomized-edge thresholding, and the reported robustness check over Sc—is creative and potentially reusable for other dyadic and group coordination studies. However, the current manuscript does not provide sufficient evidence for its headline claim: the EEG half of the dissociation depends on an analysis that is not defined, and the statistical treatment of non-independent dyadic data inflates confidence. The paper's novelty and interdisciplinary ambition are clear, but the empirical foundation needs substantial work.","major_comments":[{"comment":"The load-bearing EEG result is a 'relative EEG analysis' that is never defined—no equation, no dependent variable, no model specification, no random effects. The direct pre/post comparison is non-significant (F(1,8)=0.216, p=.654), and none of the task-wise EEG contrasts reaches significance (FM p=.36; DM p=.37; RM p=.20). Therefore the abstract's assertion that training 'increased inter-brain synchronization, particularly within the frontal lobe' is not supported by any reported main effect. Please define the relative analysis explicitly, state what it computes (e.g., a task × session interaction or a normalized change score), and show that it actually corresponds to an increase in synchronization rather than a difference between task changes. As written, the dissociation compares a direct motor pre/post effect with an undefined, task-relative EEG effect.","section":"§4, Table 1"},{"comment":"The statistical tests treat non-independent observations as independent. Motor synchronization uses 42 PRE and 30 POS dyads with F(1,70), but each participant contributed to multiple dyads, so dyad-level observations are not independent. EEG analyses report F(1,8) with only 5 dyads per condition, which appears to treat PRE and POS as independent groups rather than paired observations (a paired/repeated-measures analysis would have fewer error degrees of freedom). The Mann-Whitney tests on edge-incidence distributions (U values with hundreds of thousands of cases) treat every edge-time instance as an independent sample, massively inflating the effective N. Please reanalyze using mixed-effects models with random intercepts for session, dyad, and/or participant, or use permutation tests that respect the dyadic structure, and report effect sizes with variance estimates that reflect the true","section":"§4–5, Table 1"},{"comment":"There is no no-training control group. The pre/post design cannot distinguish the effect of GDAM from task familiarization, repeated exposure, or other time-related confounds. This is particularly relevant because the DM task showed no significant change in either modality, which is consistent with a practice/familiarization account. The causal language in the abstract ('training produced an intriguing dissociation') is stronger than the design supports. Either add a control condition or explicitly acknowledge this as a limitation and temper the causal claims throughout.","section":"§2, §4–5"},{"comment":"The dissociation is supported only for FM and RM motor decreases; the DM task showed no significant change in motor or neural synchrony. The abstract and conclusion generalize to 'collaborative dance improvisation' as a whole. Please either present the result as task-specific (FM/RM) or provide a substantive justification—not just a post-hoc narrative—for why DM's null result is consistent with the proposed coupling-decoupling paradox. At minimum, the claims should be scaled to the tasks that actually produced the effects.","section":"Abstract, §5, Table 1"}],"minor_comments":[{"comment":"There are two sections numbered '3' ('Biomechanical data segmentation and processing' and 'Motif Synchronization'). Renumber the sections sequentially.","section":"General structure"},{"comment":"The text says 'See Fig. 3.' when referring to the distribution of edge Incidence values; the correct cross-reference appears to be Fig. 7. Please correct.","section":"§4"},{"comment":"The robustness analysis for Sc ∈ [0.5, 0.9] is asserted but no results, figures, or quantitative summaries are provided. Please include a supplementary figure or table demonstrating that the global topology and the reported effects remain stable across this range.","section":"§3 (Motif Synchronization)"},{"comment":"The text states a 64-channel system but records 28 channels per individual. Clarify whether the remaining channels were not used, or how the 64 channels were split across the two EEG caps.","section":"§2 (EEG)"},{"comment":"The rows labeled 'Relative increase in EEG' and 'Relative increase in biomechanics' are not defined anywhere in the Methods. Define the formula for these change scores and specify how the contrasts were computed.","section":"Table 1"},{"comment":"Some citations are ambiguous or incomplete: 'Ramos et al. (2025a, 2025b)' are not clearly distinguished in the text, and several references (e.g., Chauvigné & Brown, 2018; Chauvigné et al., 2018) appear in the reference list but are not annotated consistently. A thorough reference cleanup is needed.","section":"References"},{"comment":"The statement 'Data will be made available on request' is weaker than current reproducibility standards. Please consider depositing anonymized data and analysis code in a public repository.","section":"Data availability"}],"recommendation":"major_revision","confidential_remarks":"The central claim may be salvageable if the authors can define the 'relative EEG analysis' and show that it genuinely corresponds to an increase in inter-brain synchronization, and if the non-independence issues are addressed with appropriate mixed-effects or permutation methods. As it stands, the headline dissociation is not empirically demonstrated. The manuscript also leans heavily on self-cited methods (Rosário et al., 2015; Sousa et al., 2024; Miranda et al., 2018; Ramos et al., 2025a,b), which is not a problem per se, but the undefined 'relative EEG analysis' cannot be traced to those references. If a reanalysis with proper dependence structure invalidates the dissociation, the paper should be rejected; if the significant effects survive, it could become a useful contribution. I recommend major revision with a strong request for a statistician’s involvement."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a useful multimodal protocol with a direct motor finding, but the headline dissociation doesn't hold up as reported. The motor decrease after GDAM is a straightforward pre-post effect for FM and RM. The EEG increase, however, comes only from an unnamed 'relative EEG analysis' — the direct pre-post EEG comparison is non-significant (F(1,8)=0.216, p=.654). The abstract says 'inter-brain synchronization increased,' but the only significant EEG result is a task-relative contrast (FM vs DM, p=.01) in a method the paper never defines. No formula, no dependent variable, no random effects. That's load-bearing, because the entire 'coupling-decoupling paradox' depends on a brain increase that is never directly demonstrated.\n\nWhat's genuinely new: the combination of EEG hyperscanning and motion capture in a dance improvisation group before/after a training program. I haven't seen that exact protocol. The motor result — reduced interpersonal motor synchrony after training in free and rule-based tasks — is a direct and plausible effect. The theoretical framing (togetherness as neural alignment with motor divergence) is interesting, and the robustness check on the motif threshold is a plus. The methods themselves are from the authors' prior work, which is fine; the contribution is the application.\n\nSoft spots, in rough order of severity:\n- The undefined 'relative EEG analysis' is the main problem. It needs to be explicitly defined and shown to correspond to a real increase in synchronization. As it stands, the abstract overstates the evidence.\n- The statistics ignore non-independence. Dyads overlap, EEG was recorded from two of four dancers per session, and n is 5 pairs per condition. The F-tests with df=8 are optimistic. Mixed models or permutation tests that respect the structure are needed.\n- No control group. Pre-post changes could reflect familiarization with the tasks, not GDAM specifically.\n- The Mann-Whitney tests on edge-incidence distributions treat thousands of non-independent edge values as independent; those p-values are not trustworthy.\n- No code or data (only 'on request'), which makes the undefined analysis harder to verify.\n\nWho this is for: researchers working on joint action, dance neuroscience, and social coordination. They'll find the protocol worth knowing and the motor effect worth testing. But the neural claim, as written, is not supported. The paper deserves a serious referee, not because it's true but because the protocol is novel and the motor finding is real; a revision that defines the contrast, models the dependencies, and tempers the claims could make this a solid contribution.","headline":"The motor effect is real; the neural increase isn't shown — the abstract overclaims a dissociation built on an undefined EEG contrast.","tokens_in":15899,"tokens_out":2207,"would_cite":false,"duration_ms":21480,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Training in generative dance makes improvising partners' brains align more closely while their hand movements diverge, and the paper argues that this paradox is what togetherness actually is.","keywords":["collaborative improvisation","hyperscanning EEG","inter-brain synchronization","motor synchrony","generative dance","cellular automata","motif synchronization","degrees of freedom"],"falsifier":"A replication with mixed-effects models that include dancer and dyad as random factors and a no-training control group, or even just a re-analysis of the present data with such models; if the frontal inter-brain increase or the motor synchrony decrease no longer reaches significance once non-independence is accounted for, the claimed dissociation is not supported.","tokens_in":14942,"feed_emoji":"🧠","tokens_out":2540,"duration_ms":28681,"temperature":0.7,"pith_summary":"The paper studies four dancers improvising together before and after a program of generative dance based on cellular-automata rules, recording both hand movements and EEG from two dancers at a time. It finds that after training, in free and rule-based improvisation, the dancers' frontal brain activity becomes more synchronized with each other while their hand-movement patterns become less synchronized. The authors argue that this coupling–decoupling paradox shows that togetherness in group improvisation is carried by shared neural intentionality rather than by mirroring each other's movements, and that expanding individual motor freedom actually supports collective coordination. The dependent-movement task showed no significant change in either channel, which the authors interpret as evidence that explicit interaction reduces the need for shared executive planning.","feed_headline":"Brain sync rises as dancers' moves diverge","feed_subtitle":"In group improvisation, togetherness may live in shared neural intention, not matching movements.","key_machinery":"The coupling–decoupling paradox is the central claim: increased inter-brain synchronization and decreased interpersonal motor synchrony after training. Methodologically, the paper quantifies motor dynamics with a time-resolved α-exponent from Movement Element Decomposition, which indexes energetic strategy and degrees of freedom in a sliding window, and then measures motor synchrony with Motif-Synchronization applied to those α time series. Neural synchrony is measured with multilayer Time-Varying Graphs, using the Incidence-Fidelity index to detect simultaneous edge occurrences between two participants' dynamic brain networks while excluding chance-level co-occurrence.","core_discovery":"After a program of Generative Dance through Agent Modelling, pairs of co-improvisers show increased inter-brain synchronization in frontal regions during free and rule-based improvisation, while simultaneously showing decreased interpersonal motor synchrony in the same tasks. The dissociation is statistically significant in both modalities: EEG relative analysis yields F(1,8)=10.787, p=.011, and biomechanical post-hoc tests show reduced motor synchrony for FM (p=.04) and RM (p=.02), whereas the dependent-movement task shows no change in either measure. The authors conclude that trained improvisers achieve togetherness through neural alignment of intentional and executive processes, not throu","pith_inferences":["A direct testable extension is to compare self-reported sense of togetherness with frontal inter-brain synchrony and motor synchrony in the same sessions; the paper's account predicts a positive correlation with the former and a negative correlation with the latter after training.","The authors interpret the frontal effect as shared intentionality, but it could also reflect increased individual cognitive load from rule-based improvisation; a control task with matched individual executive demand but no social interaction would disambiguate.","Because there was no no-training control group, familiarization with the task or with the EEG setup could plausibly explain the pre-post changes; a crossover design with a waitlist control would strengthen the causal claim.","The statistical treatment of dyads as independent observations is fragile; re-analysis with mixed models that treat dancer and dyad as random effects could change the p-values, so the dissociation's robustness is not yet established."],"forward_implications":["In improvisational joint action, togetherness should be indexed by neural synchrony (especially frontal) rather than by motor mimicry, and a lack of movement similarity does not indicate a lack of coordination.","Training protocols like generative dance can increase inter-brain coupling without forcing performers into identical movement patterns, suggesting a route to enhance group creativity without constraining individuals.","The time-resolved α-exponent offers a single scalar per time window that captures fluctuations in motor degrees of freedom, enabling pairwise synchrony analysis of whole-body (here, two-handed) movement.","The absence of an effect in the dependent-movement task indicates that the neural increase is specific to conditions requiring autonomous strategy generation, not to all social interaction.","If the dissociation generalizes, studies of team coordination and dance therapy that measure only movement similarity may be missing the relevant neural coupling."],"fun_headline_variants":["Dancers' brains sync up as bodies branch out","In improv dance, brain links grow while moves uncouple","Training makes dancers' brains align, movements diverge","Coupling paradox: brains align, bodies explore","Shared intention: brain sync up, motor sync down"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The statistical evidence for the neural–motor dissociation rests on treating each participant pair as an independent observation even though the same dancers improvised together repeatedly and brain data came from only two of the four dancers per session, so the reported degrees of freedom and p-values are likely inflated.","fun_headline_variants_meta":{"raw":{"variants":["Dancers' brains sync up as bodies branch out","In improv dance, brain links grow while moves uncouple","Training makes dancers' brains align, movements diverge","Coupling paradox: brains align, bodies explore","Shared intention: brain sync up, motor sync down"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000192,"raw_usage":{"total_tokens":1207,"prompt_tokens":794,"completion_tokens":413,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":538,"completion_tokens_details":{"reasoning_tokens":337}},"tokens_in":538,"tokens_out":413,"duration_ms":4141,"temperature":1.0,"reasoning_tokens":337,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T12:16:44.364679+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A replication with mixed-effects models that include dancer and dyad as random factors and a no-training control group, or even just a re-analysis of the present data with such models; if the frontal inter-brain increase or the motor synchrony decrease no longer reaches significance once non-independence is accounted for, the claimed dissociation is not supported.","supporting_citations":[],"review_version":1}