Social transmission of freezing behavior has been conceived of as a one-way phenomenon in which an observer "catches" the fear of another. Here, we use a paradigm in which an observer rat witnesses another rat receiving electroshocks. Bayesian model comparison and Granger causality show that rats exchange information about danger in both directions: how the observer reacts to the demonstrator's distress also influences how the demonstrator responds to the danger. This was true to a similar extent across highly familiar and entirely unfamiliar rats but is stronger in animals preexposed to shocks. Injecting muscimol in the anterior cingulate of observers reduced freezing in the observers and in the demonstrators receiving the shocks. Using simulations, we support the notion that the coupling of freezing across rats could be selected for to more efficiently detect dangers in a group, in a way similar to cross-species eavesdropping.
rats exchange information about danger in both directions: how the observer reacts to the demonstrator's distress also influences how the demonstrator responds to the danger
convergent predictive analyses and ACC perturbation support feedback, without fully identifying behavioral causation
This was true to a similar extent across highly familiar and entirely unfamiliar rats
overlapping coupling estimates and mostly null familiarity tests, with a small strain-specific temporal effect
but is stronger in animals preexposed to shocks
convergent observer-freezing and coupling analyses from a small nine-dyad comparison
Injecting muscimol in the anterior cingulate of observers reduced freezing in the observers and in the demonstrators receiving the shocks
randomised pharmacological comparison with small groups and an inconsistent reported ANOVA denominator
Derived from the full evaluation — not a separate score.
Strengths
Bayesian model comparisons and second-by-second Granger analyses both identify coupling in each direction across the individual- and strain-familiarity experiments. This convergence makes the bidirectional pattern more credible than either analysis alone.
↳ Results, Effect of Familiarity and Feedback; Results, Moment-to-Moment Coupling—Granger Causality
Muscimol injection reduced observer freezing and demonstrator freezing, with blinded behavioral scoring and histological confirmation of ACC targeting. The locomotor analysis also argues against simple sedation as the explanation for reduced observer freezing.
↳ Results, Role of the ACC in Information Transfer; Fig 9; Fig 10
The Discussion directly compares the limited rat familiarity effect with stronger mouse findings and tests whether strain discrimination could explain the result. Fig 6 indicates that rats could distinguish the strains under the experimental lighting.
↳ Discussion, familiarity paragraphs; Fig 6
Limitations
Bayesian and Granger analyses establish conditional and temporal prediction within a shared shock context, not the precise causal pathway from observer freezing to demonstrator freezing. ACC inactivation supports feedback but can also alter observer outputs other than the measured freezing response.
↳ Results, Moment-to-Moment Coupling—Granger Causality; Results, Role of the ACC in Information Transfer
The preexposure comparison uses nine nonpreexposed dyads, while the ACC analysis retains six muscimol and eight saline dyads. These samples limit certainty about effect magnitude and secondary mechanistic interpretations.
↳ Materials and Methods, Experiment 3—Preexposure Effect; Experiment 4—ACC Deactivation
Experiments 1 and 2 exclude multiple dyads for technical failure, including four of fourteen in one Experiment 1 condition, without assessing whether these exclusions altered the reported group pattern.
↳ Materials and Methods, Experimental Procedures, Experiments 1 and 2
The central contribution is supported by convergent Bayesian and Granger analyses across two familiarity experiments, with ACC inactivation adding experimental evidence that altering the observer changes both members of the dyad. The methodology is competently reported, including random allocation, scoring reliability, blinding, locomotor checks, and histological verification. The score is moderated because shared shock-related causes are not fully separated from behavioral feedback, technical exclusions receive no sensitivity analysis, and the preexposure and ACC experiments have small final groups. The paper handles mouse familiarity findings and several alternative explanations constructively, but its causal language should be narrowed to match what the combined analyses identify.
Nabu’s assessment, alongside the field’s view.
Are you an author of this paper?
Sound3.5
Confidence mediumThe score reflects a meaningful reframing of rodent social freezing as bidirectional coupling, supported across Bayesian, time-series, and ACC-inactivation analyses. The advance remains bounded to an established shock-observation paradigm.
“Conceiving of the influence as mutual has the conceptual advantage of integrating what has been considered emotional contagion and social buffering.”
Random allocation, reliable behavioral scoring, blinded Experiment 4 scoring, histological verification, model comparison, and temporal analysis support competent execution. Small secondary samples, unexplored technical exclusions, possible muscimol diffusion, and incomplete separation of feedback from shared causes lower certainty.
“A total of 8 dyads were excluded because of technical failure.”
The manuscript follows a coherent progression from behavioral coupling through temporal analysis, perturbation, and simulation. Its headline use of “influences” is stronger than the predictive analyses alone establish, although the ACC experiment supplies partial causal support.
“how the observer reacts to the demonstrator's distress also influences how the demonstrator responds to the danger”
The discussion engages contrasting mouse familiarity findings, tests a strain-discrimination alternative, and distinguishes behavioral mimicry from affective contagion. It does not fully trace shared-context causal uncertainty or small-study constraints into the strength of its feedback conclusions.
“Whether we also have an alignment of affective state… remains unclear.”
Caveats4 of 4 checks
The reported programme is broadly coherent, but an ANOVA denominator inconsistency in Experiment 4 impedes full auditability and small secondary samples temper interpretation. These issues warrant caution without indicating that the reported effects are fabricated or numerically impossible.
Animal ethics approvals, welfare procedures, blinded scoring for Experiment 4, and histological exclusions are declared. No supplied-text evidence indicates a research-conduct concern.
Flags: 1 declared / 5 total
73 of 73 checkable references verified
73 references in manuscript citation details diverge from the cited records for 1 reference the reference list may be incompletely parsed; counts are indicative
References not verified automatically are resolved manually within 24 hours.
No retraction or expression of concern found in Crossref (includes Retraction Watch).
Sources: Crossref (incl. Retraction Watch) ✓
Where this paper’s evidence sits on the path from initial observation to real-world use.
The work characterises behavioral and ACC-related mechanisms under controlled laboratory conditions rather than testing an intervention or deployment pathway. Substantial additional validation would be needed before practical use.
AI-generated, human-governed. Something look off? Contact us to request a review.
Keep reading
Other papers Nabu has evaluated blind, starting with the same field.
John Jumper, Richard Evans, Alexander Pritzel, et al. · 2021 · Computer and information sciences
first computational method that can regularly predict protein structures with atomic accuracy even in cases in which no similar structure is known
Well-supported2026 · Biological Sciences
Applying TEgment to CDK9-inhibited AML revealed that TEs are predominantly expressed as readthrough, 3’UTR-embedded, and lncRNA-embedded transcripts rather than autonomous units, challenging prevailing assumptions about TE reactivation.
TentativeDustin Sokolowski, Alexander Nater, Mihai Miclăuș, et al. · 2026 · Biological sciences
Pms1, a MutL homolog, has six intact copies in the NMR, arising from a series of nested tandem duplications on chromosome 3 that also include Fsip2, Osgepl1, Ankar and Fkbp8, but a single copy in B.sui and F.ans.
Well-supported