Introduction

Animals continuously encounter diverse social and environmental cues and must rapidly evaluate them to guide appropriate behaviour. In humans and other social species, encounters with unfamiliar individuals require rapid appraisal to determine the likely outcome of the interaction, a process that is critical for guiding behaviour and ensuring safety^1,2. Social interactions can be highly beneficial, supporting development, reproduction, and emotional resilience, yet they can also introduce challenges during competition for limited resources, excessive aggression and the establishment of dominance hierarchies^3–8. Importantly, these social encounters are often ambiguous, requiring rapid evaluation under conditions of uncertainty^9–11. Failure to appropriately assess social cues can lead to maladaptive outcomes, either through delayed responses to adverse situations or through unnecessary behavioural and physiological costs associated with overestimation of risk^12,13. Despite the importance of these processes, the neural mechanisms that support rapid social assessment remain poorly understood.

Emerging evidence implicates corticotropin-releasing hormone neurons in the paraventricular nucleus of the hypothalamus (CRH^PVN), classically known for orchestrating endocrine stress responses, in processing behaviourally relevant information beyond canonical stress regulation^14–18. Previous work has shown that CRH^PVN neurons are required for social investigation of stressed conspecifics, a behaviour that facilitates the transmission of affective state between individuals, and that these neurons are engaged by social stimuli¹⁹. However, whether CRH^PVN neurons contribute to the rapid evaluation of conspecifics to guide appropriate behavioral responses during social encounters remains unclear.

Here, using fibre photometry and optogenetics in an ethologically relevant social interaction paradigm, we demonstrate that CRH^PVN neurons exhibit rapid, transient increases in activity that are critical for subsequent investigative behaviours. This activity pattern suggests a role in gating behavioural responses that enable the assessment of conspecifics. Together, our findings identify CRH^PVN neurons as a key component of the neural circuitry underlying rapid social appraisal, linking endocrine regulation to real-time behavioural decision-making.

Results

Increased CRH^PVN neuron activity during social appraisal of an unfamiliar conspecific

Social interactions vary in predictability. Unfamiliar conspecifics introduce social uncertainty demanding rapid evaluation to guide appropriate behavioral responses^9,20. To investigate how CRH^PVN neural activity contributes to this process, we used transgenic mice expressing GCaMP6f selectively in CRH neurons¹⁷ (Fig. 1A). We implanted an optical fibre above the PVN to perform fibre photometry recordings of calcium-dependent fluorescence as a proxy for neuronal population activity. Using a classical resident-intruder test, the neural activity was specifically recorded from the resident mouse during social interaction. Following baseline recording of the resident mouse during a solitary period, a familiar same-sex cage mate or an unfamiliar same-sex, same-strain intruder was introduced into the resident’s home cage (Fig. 1B). Approach toward a familiar intruder elicited a transient increase in CRH^PVN activity, which returned to baseline within seconds and remained around baseline during the remainder of the 3-min interaction period (Fig. 1C). This activity emerged within seconds of intruder introduction across conditions, indicating rapid engagement of CRH^PVN neurons during the initial stages of social evaluation (Fig. 1D-F). Fine-grained temporal analysis of each condition revealed that although the introduction of both familiar and unfamiliar intruders produced rapid CRH^PVN activity, the introduction of an unfamiliar intruder triggered a significantly greater and more sustained early response (Fig. 1G). The peak amplitude, peak area under the curve (AUC) and time to baseline of the averaged CRH^PVN Ca²⁺ responses in the resident mouse were significantly larger when exposed to unfamiliar intruders than to familiar intruders (Fig. 1H-J). Likewise, this differential response occurred in both male and female resident animals (Fig. S1A-B). Together, these findings demonstrate that CRH^PVN neurons are recruited during social appraisal in a manner that scales with familiarity status, supporting a role for these neurons as rapid detectors of ambiguous social information.

Figure 1.

Differential CRH^PVN neuron activity persists in novel contexts and across conspecific strains

To determine whether the heightened CRH^PVN responses to unfamiliar conspecifics reflect social uncertainty rather than context-specific factors, we examined neuronal activity in a neutral non-territorial environment. To this end, we placed mice in a novel cage with fresh bedding and performed fibre photometry recording during exposure to both familiar and unfamiliar intruders (Fig. 2A).

Figure 2.

Consistent with observations made in the home cage, approach toward unfamiliar intruders in a novel context elicited a significantly greater increase in CRH^PVN activity compared to familiar interactions (Fig. 2B, S2A,B). The peak amplitude and AUC of CRH^PVN responses were significantly higher during interactions with unfamiliar compared to familiar mice (Fig. 2C,D), indicating that enhanced CRH^PVN responses persist independently of environmental familiarity. We next asked whether this differential activity generalizes across conspecific types. In the home cage, interactions with either unfamiliar same-strain (C57) or non– same-strain (CD1) intruders elicited significantly greater CRH^PVN activity compared to familiar mice (Fig. 2E,F, S3A,B). Analysis of the first calcium transient revealed increased responses to both unfamiliar C57 and CD1 intruders relative to familiar partners, but no difference between unfamiliar groups (Fig. 2G,H). Together, these findings demonstrate that CRH^PVN neuron activity robustly scales with conspecific unfamiliarity across contexts and strain differences, supporting a role for these neurons in detecting social uncertainty during the initial stages of social appraisal.

Differential CRH^PVN neuron activity is absent during exposure to unfamiliar juvenile intruders and object

To determine whether CRH^PVN activity reflects unfamiliarity or is selectively engaged under social conditions requiring evaluation, we compared responses across interactions with familiar adult, unfamiliar adult, and unfamiliar juvenile conspecifics. While approach toward unfamiliar adult intruders elicited an increase in CRH^PVN calcium activity relative to familiar adult interactions, responses to unfamiliar juvenile intruders were not elevated and were comparable to those observed during familiar adult encounters (Fig. 3A–D, S4A-B). These findings indicate that enhanced CRH^PVN activity is not a general feature of novelty but is selectively recruited during interactions with unfamiliar adult conspecifics.

Figure 3.

We next assessed whether this activity pattern extends to non-social stimuli. Following a 3-day pre-exposure to the familiar object, resident mice were exposed to either a familiar or a novel object^21,22.

During the approach to the familiar and unfamiliar objects, CRH^PVN activity did not differ between conditions (Fig. 3E–H). Together, these results demonstrate that while CRH^PVN neurons are broadly responsive to novelty or unfamiliar stimuli, they are differentially selectively engaged in social contexts that require evaluation of unfamiliar adult conspecifics, supporting a role in context-dependent social appraisal.

Increased social investigation during exposure to unfamiliar conspecifics

The selective increase in CRH^PVN activity towards unfamiliar intruders raises the question of whether this neural signal carries behavioral consequences. To address this, we first sought to determine how conspecific familiarity shapes behavioural engagement during social encounters, we systematically quantified distinct behavioural classes during interactions with familiar and unfamiliar conspecifics. Raster plots revealed the temporal structure of discrete behavioural events, including sniffing, rearing, digging, and grooming, across the interaction period (Fig. 4A). These behaviours provide a comprehensive assessment of distinct social and non-social components of the interaction. Sniffing reflects direct social investigation and active information sampling of conspecifics^23–25, whereas rearing, digging, and grooming provide measures of general exploration, arousal, and self-directed behaviour^26–28. This distinction allowed us to isolate behaviours specifically related to social assessment. Across conditions, sniffing behaviour was initiated immediately upon exposure to the intruders and was predominantly directed from the resident toward the intruder, reflecting active investigation of the social partner (Fig. 4Bi–ii). Likewise, the total duration of resident-to-intruder sniffing was increased during interactions with unfamiliar conspecifics compared to familiar conspecifics (Fig. 4Biii). This enhancement in directed investigation indicates that exposure to unfamiliar conspecifics triggers increased sampling of socially relevant cues, consistent with a greater need for social assessment. In contrast, non-social behaviours, including digging, rearing, and grooming, were not significantly altered between familiar and unfamiliar conditions (Fig. S5A-C), and the investigation of novel objects was reduced compared to familiar objects (Fig. S5D), indicating that the observed differences are specific to social investigation.

Figure 4.

To further resolve the structure of this investigation, we decomposed sniffing into anatomically defined subtypes corresponding to anogenital, head, torso, and tail regions. This breakdown was motivated by the functional relevance of these regions, as different targets provide distinct social information, including identity, reproductive status, threat level, and recent social experience^19,29–35. Temporal analysis revealed that sniffing behaviour is dynamically distributed over the interaction period, with distinct patterns emerging across sniffing subtypes (Fig. 4Ci-iv). Cumulative measures showed that increases in sniffing during exposure to unfamiliar intruders were not uniform but selectively enriched for specific investigative behaviours (Fig. 4Di-iv). Specifically, the investigation of unfamiliar intruders was characterized by increased anogenital and head-directed sniffing. In contrast, torso and tail sniffing did not differ substantially between conditions, indicating that unfamiliarity selectively enhanced behaviours most relevant to social evaluation rather than broadly increasing all forms of contact (Fig. 4Ei-iv). Permutation-based comparison of behavioural transition matrices revealed a significant difference in overall transition architecture between familiar and unfamiliar conditions (Frobenius norm = 0.375, p = 0.0044). The observed statistic exceeded the 95th percentile of the null distribution, indicating that the behavioural structure differed beyond chance levels.

Inspection of the difference matrix (unfamiliar − familiar) showed that this effect was driven by structured changes in specific transitions rather than uniform shifts across the network. The most prominent increase was observed in transitions from torso to anogenital sniffing (+0.25), with additional increases in transitions from tail to anogenital sniffing (+0.11). In contrast, transitions from torso to tail (−0.16) and torso self-transitions (−0.14) were reduced. Edgewise statistical testing with false discovery rate correction identified a subset of significant transitions, all originating from the torso state. These included an increased probability of transitioning from torso to anogenital sniffing and decreased probabilities of transitions from torso to tail and torso itself (Fig. 4Fi-ii, S5E-G). Together, these results indicate that unfamiliar animals exhibit a reorganization of behavioural transitions characterized by a preferential routing of behavioural sequences toward anogenital sniffing, along with suppression of alternative transitions. Our results therefore highlight anogenital sniffing as a key behavior for social assessment.

CRH^PVN activity is necessary for targeted anogenital investigation during social assessment

Our behavioral analysis revealed enhanced investigative sniffing directed towards unfamiliar intruders. We next asked if the rapid, transient increase in the CRH^PVN activity neurons during social approach causally drives the behavioral response. To test this, we expressed Archaerhodopsin 3.0 virus in CRH neurons to optogenetically inhibit CRH^PVN neuron activity (Fig. 5A). Targeted viral expression and fibre implantation enabled selective silencing of CRH^PVN neurons, which was validated by ex vivo slice recordings demonstrating reliable light-driven inhibition of neuronal firing³⁶ (Fig. 5B). During the experiment, optogenetic inhibition of the resident mouse was done during the initial phase of interaction (20s), corresponding to the period when CRH^PVN activity rapidly increased upon intruder exposure. Behavioural quantification was performed to capture the entire 3 min interaction period. We first quantified the overall social investigation by measuring the total sniffing behaviour in Arch-inhibited and mCherry expressing controls. Under control conditions, interactions with unfamiliar conspecifics elicited greater overall sniffing than interactions with familiar conspecifics. However, optogenetic inhibition of CRH^PVN neurons selectively reduced total sniffing during exposure to unfamiliar intruders, without altering investigation of familiar intruders (Fig. 5C).

Figure 5.

To further resolve how CRH^PVN activity shapes social assessment, we decomposed total sniffing into distinct investigative subtypes, including anogenital, head, torso, and tail-directed sniffing. This breakdown is important because these behaviours reflect different components of social investigation^29,31,33,34. In control animals, unfamiliar interactions were characterized by a distinct distribution of these behaviours, indicative of targeted investigation. In contrast, inhibition of CRH^PVN neurons disrupted this organization, leading to an altered distribution of sniffing behaviours specifically during unfamiliar encounters, while leaving familiar interaction patterns intact (Fig. 5Di-ii, S6A-F).

We next examined the temporal dynamics of investigation, focusing on anogenital sniffing as a key index of sustained social evaluation^19,33 (Fig. S5F). In control animals, unfamiliar conspecifics elicited prolonged, temporally structured anogenital investigation. Inhibition of CRH^PVN neurons markedly reduced this sustained engagement, as reflected in both the temporal profile and the sum of cumulative anogenital sniffing events (Fig. 5E–Hi-iii). Analysis of behavioural transitions demonstrated that while CRH^PVN inhibition reduced time spent on anogenital sniffing, no corresponding changes in the sequence of social behaviours were observed. Across both familiar and unfamiliar conditions, we found no differences between mCherry and Arch animals in either global transition structure or individual behavioral transitions. Observed variations were consistent with chance-level fluctuations as determined by permutation testing (Fig. 5Ii-ii). This suggests that inhibition of CRH^PVN neurons alters the overall duration of specific social behaviors without reorganizing the flow of behavioral transitions.

Discussion

Social interactions require animals to rapidly evaluate conspecifics and adjust behaviour accordingly. Here, we identify CRH^PVN neurons as a rapid, transient signal engaged during this process, with activity elevated during encounters with unfamiliar, relative to familiar conspecifics. This response is consistent across environments and strains, but absent in juveniles or non-social objects. Importantly, though transient, this signal is necessary for targeted social investigation. Together, these findings position CRH^PVN neurons as a fast, integrative node for assessing conspecifics when interaction outcomes are uncertain.

Neuroimaging studies across humans and rodents have implicated the hypothalamic nuclei as central mediators of socioemotional integration³⁷. For example, reproductive and mating-related behaviors engage the medial and ventral hypothalamic nuclei, while predation and defensive evasion recruit the lateral hypothalamic to periaqueductal gray circuitry³⁸. These elaborate behaviors, however, are downstream consequences of a more fundamental step: the rapid appraisal of social encounters and the rapid selection of a behavioral response strategy. At this initial decision point, the CRH^PVN neurons are well-positioned to play a pivotal role in integrating sensory inputs with internal states, projecting to specialized socially-related hypothalamic nuclei and regulating the neuroendocrine axis, thereby coordinating adaptive responses to environmental challenges^39–42. Our findings extend this framework by positioning CRH^PVN neurons not solely as stress effectors, but as early-stage evaluators of social information. This mechanism may confer an adaptive advantage by enabling individuals the anticipation of potential challenges and the selection of appropriate responses, thereby linking sensory processing to neuroendocrine and behavioral outputs^43–45.

The behavioural significance of this rapid CRH^PVN neural signal is evident in social investigation. Investigative behaviours, particularly anogenital sniffing, are central to how rodents evaluate conspecifics and extract socially relevant cues^46–49. Our optogenetic inhibition experiments reveal that transient silencing of CRH^PVN activity during the early activity window produces lasting changes in investigative behavior that extend beyond the period of inhibition, indicating that this signal gates subsequent behavioural responses. Specifically, inhibition of CRH^PVN neuron activity reduced the time animals spent engaged in anogenital investigation, without altering the overall structure of behavioural transitions. This dissociation highlights a primary role for CRH^PVN neurons in regulating the intensity and persistence of specific behaviours rather than their sequential organization⁵⁰. Consistent with this interpretation, CRH^PVN neurons exert rapid central effects via their projections, in addition to their slower endocrine functions, thereby recruiting downstream circuits that maintain behavioural responses across timescales^51–53. Given that these neurons also govern the HPA axis^39,54, it will be important to determine whether their influence on persistent social behaviour arises from synaptic signalling, neuroendocrine modulation, or the coordinated interaction of both systems⁵⁵.

The novel social appraisal functionality of CRH^PVN neurons, in addition to their established role in social transmission of stress, defensive flight-fight responses, and anxiety-related behaviors is consistent with the broader mechanistic principles by which PVN operates^14,15,17,19. Molecularly defined PVN populations form combinatorial assemblies rather than discrete labeled-line circuits⁵⁶, which facilitates the generation of complex behaviors. Within this PVN framework, the CRH neurons likely function not as a single-output relay but as a part of a distributed combinatorial computation that simultaneously coordinates endocrine, behavioral, and autonomic arms of survival responses. How this survival circuit machinery therefore integrates information from the nodes of the social decision-making network remains an important open question. Future work aimed at cell-level encoding, unravelling how distinct CRH^PVN neuronal populations govern different aspects of socio-survival processing, will be essential for deepening our mechanistic understanding of social appraisal.

Finally, our study opens new avenues for exploring how CRH^PVN neurons contribute to different aspects of social behavior, such as approach, avoidance, and consummatory interactions. More broadly, these neurons may extend the predictive, feed-forward logic seen in homeostatic regulation^57–59 into the social domain, linking the detection of social uncertainty to the preparation of appropriate internal states. Because disrupted CRH signalling and altered salience processing are key features of stress-related disorders, understanding this mechanism could help guide strategies to restore balanced stress responses and flexible social behaviour, particularly under conditions of social uncertainty.

Methods

Animals

All animal protocols were approved by the University of Calgary Animal Care and Use Committee (AC21-0067). Mice that were the offspring of Crh-IRES-Cre (B6(Cg)-Crh^tm1(cre)Zjh; stock no: 012704, Jackson Laboratories) and Ai14 (B6.Cg-Gt(ROSA)26Sortm14(CAG-TdTomato)Hze/J; stock number 007914, Jackson Laboratories) mice crossed with Ai148 (Ai148(TIT2L-GC6f-ICL-tTA2)-D; stock number 030328, Jackson Laboratories) were utilized for all fibre photometry and optogenetics with behavioural tracking experiments. Mice were group-housed on a 12:12 h light/dark schedule (lights on at 07:00 hours) with ad libitum access to food and water. CD-1 mice (Charles River Laboratory) were used for behavioural alternative strain intruder experiments. Mice were 6–8 weeks old at the time of surgery and virus injection.

Viruses

AAV carrying Arch3.0-mCherry (LotAAV1081, AAV2/9-EF1a-DIO-Arch3.0-mCherry; Canadian Neurophotonics Platform Viral Vector Core Facility (RRID:SCR_016477) or mCherry (LotAAV1088, AAV2/9-EF1a-DIO-mCherry; Canadian Neurophotonics Platform Viral Vector Core Facility (RRID:SCR_016477)) was used for optogenetic manipulations.

Stereotaxic injection and optical fibre implantation

Mice were maintained under isoflurane anesthesia in the stereotaxic apparatus. Mono fibre optic cannula with 400 µm diameter (Doric Lenses, MFC_400/430/0.48_5.5mm_MF2.5_FLT) was implanted dorsal to the PVN for fibre photometry (AP, −0.7 mm; L, −0.2 mm from the dura; DV, −4.85 mm from the dura) and for optogenetics (for Arch3.0: AP: −0.7 mm; L: 0.0 mm from the dura; DV: −4.5 mm) following tracking with a custom-made needle (AP: −0.7 mm; L: −0.2 mm from the dura; DV: −4.2 mm) and bilateral injections (AP: −0.7 mm; L: 0.2 mm from the dura; DV: −4.5 mm) fixed to the skull with METABOND® and dental cement. Mice were given 2 weeks to recover prior to the start of the experiment. Experiments started after an additional two weeks of recovery and handling.

Histology

To verify GCaMP expression and ferrule location, following the experiments, mice were euthanized using isoflurane overdose and a subset of mice were transcardially perfused with phosphate-buffered saline (PBS), followed by 4% paraformaldehyde (PFA) in phosphate buffer (PB, 4 °C). Brains were placed in PFA 24 h followed by 20% sucrose PB. 50 µM coronal brain sections were obtained via cryostat. The other subset of mice was cut into 250 μm sections in ACSF, then fixed in 4% PFA, 1 M PBS, and 30% sucrose for 10 minutes each. Slide-mounted and coverslipped sections were imaged using a confocal microscope (Leica SP8). Images were prepared using ImageJ.

Slice preparation and electrophysiology

Four to eight weeks post AAV-injection, mice were anesthetized via isoflurane inhalation and decapitated. Rapidly dissected brains were immersed in slicing solution (0–4 °C, 95% O2/5% CO2 saturated) containing (in mM): 87 NaCl, 2.5 KCl, 0.5 CaCl2, 7 MgCl2, 25 NaHCO3, 25 D-glucose, 1.25 NaH2PO4, 75 sucrose. A vibratome (Leica) was used to prepare coronal hypothalamic slices (250 μm thickness), which were transferred for 1+ hours before recording in artificial cerebrospinal fluid (32 °C, 95% O2/5% CO2 saturated) containing (in mM): 126 NaCl, 2.5 KCl, 26 NaHCO3, 2.5 CaCl2, 1.5 MgCl2, 1.25 NaH2PO4, 10 glucose. Recordings were performed in artificial cerebrospinal fluid (1 ml min−1 perfusion) at 30–32 °C. PVN neurons were identified using differential interference contrast and epifluorescence optics (UVICO, Rapp Optoelectronics) and a camera (AxioCam MRm) on an upright microscope (Zeiss). Borosilicate electrodes (3–5 mΩ tip) were backfilled with recording solution composed of (in mM) 108 K-gluconate, 2 MgCl2, 8 Na-gluconate, 8 KCl, 1 K2-EGTA, 4 K2-ATP, 0.3 Na3-GTP, 10 mM HEPES, 10 mg ml−1 biocytin. Signals were amplified (Multiclamp 700B, Molecular Devices), low-pass filtered (1 kHz), digitized (10 kHz, Digidata 1322) and recorded (pClamp 9.2) for offline analysis.

Fibre photometry recording

Fibre photometry was used to record calcium transients from CRH neurons in the PVN of freely moving mice. After the recovery period, animals were handled for 5 min a day for 3 days and then habituated to the optic fibre in their home cage (5 min a day) for 3 days. We recorded 20 min of CRH^PVN neuron activity in the homecage immediately before and for 5 min after each test to improve bleaching correction. Doric fibre photometry system consisting of two excitation LEDs (465 nm and 405 nm from Doric) controlled by an LED driver and console, running Doric Studio software (Doric Lenses). The LEDs were modulated at 208.616 Hz (465 nm) and 572.205 Hz (405 nm), and the resulting signal was demodulated using lock-in amplification. Both LEDs were connected to a Doric Mini Cube filter set (FMC6_IE(400-410)_E1(460-490)_F1(500-540)_S)_E2(555-570)_F2(580-680)_S and the excitation light was directed to the animal via a mono fibre optic patch cord (DORIC MFP_400/430/1100-0.48_2m_FC/MF2.5). The LED power was adjusted to 30 µW at the end of the patch cord. The resulting signal was detected with a photoreceiver (Newport model 2151).

Fibre photometry data analysis

Fluorescent signal data were acquired at a sampling rate of about 100Hz (Doric system). The data were then exported to MATLAB (MathWorks) for offline analysis using custom-written scripts (https://github.com/leomol/stimbox). Briefly, raw fluorescence traces were resampled to 20 Hz to reduce data size while retaining calcium signal dynamics. A 0.5 Hz low-pass filter was then applied to suppress high-frequency noise. The 465 nm and 405 nm data were first fit individually with a second-order polynomial, and the resulting fits were subtracted to remove bleaching artifacts. Next, a least-squares linear fit was applied to the 405 nm to align it with the 465 nm channel, and then the change in fluorescence (ΔF) was calculated by subtracting the 405 nm Ca²⁺ independent baseline signal from the 465 nm Ca²⁺ dependent signal at each time point. The output of these transformations yielded a motion-corrected dataset of changes in GCaMP6f fluorescence. For analysis, z-score calculation was performed using the following equation: ΔF/F = (F-F₀)/ σF, where F is the test signal, F₀ is the median signal of the baseline, and σF is the median absolute deviation of the signal.

Optogenetics

Following the recovery period, mice were handled for 5 min on 3 consecutive days. Then, they were habituated to the fibre-optic cable attached to the cannula (without light) for 3 additional days (15-30 min per mouse).

For in vitro recordings, a micromanipulator-attached fibre optic cable (105 μm core diameter) delivered light from a laser (Arch3.0: 593 nm, IKE-593-100-OP, IkeCool Corporation) was placed 1–2 mm away from the target area. Light intensity was calibrated by a Photodiode Power Sensor (Thorlabs). Maximally, 15 mW light was delivered to the tissue.

For in-vivo experiments, the light source (for Arch3.0: 561 nm, LRS-0561-GFO-00100-5, Laserglow Technologies) was connected to the Doric filter cube (FMC6_IE(400-410)_E1(460-490)_F1(500-540)_S)_E2(555-570)_F2(580-680)_S and the laser was directed to the animal via a mono fibre optic patch cord (DORIC MFP_400/430/1100-0.48_2m_FC/MF2.5). For Arch3.0, a yellow light (15 mW laser intensity) was used continuously for 20 seconds as the intruder was introduced to the resident’s home cage.

Behavioural recordings and analysis

Adult male and female mice were used as residents. Resident mice encountered either a familiar or an unfamiliar intruder at each session. A familiar intruder is a conspecific from the same cage, whereas an unfamiliar intruder is from a different cage or a different strain (e.g., CD1 intruder). The test was conducted in the resident’s home cage. For video acquisition, a top-down camera (20 fps, Doric Studio) recorded all sessions under ~30–50 lux of ambient light and was synchronized to the fibre photometry recording. These videos were later analyzed using the Behavioural Observation Research Interactive Software (BORIS, version 9), and the following behaviours were scored: anogenital sniffing (directing the snout toward the anogenital area of conspecific); head sniffing (directing the snout toward the head of conspecific); torso sniffing (directing the snout toward the torso of conspecific); tail sniffing (directing the snout toward the tail of conspecific), digging (use of forepaws and sometimes hind limbs and snout to displace bedding), rearing (standing upright on hind legs with forepaws off the ground.) and grooming (repetitive licking of the body and wiping the face and head with the forepaws). Behaviour during 3 min after the introduction of the intruder was quantified and plotted. Quantification and analysis of behavioural ethograms, sliding time windows, cumulative sums of behavioural events, and transition probabilities were carried out using custom-written Python scripts. Behavioural sequences were converted into transition probability matrices by counting state-to-state transitions between consecutive behavioural events and row-normalizing counts to obtain conditional probabilities. Individual matrices were averaged across animals to generate group-level transition structures for each condition.

Statistics

Where quantification was made, data are represented as mean ± standard error of the mean (s.e.m.). Statistical analysis was performed using GraphPad Prism 10.0. For comparisons, either two-tailed t-tests, ordinary or repeated-measures one-way ANOVA were applied as appropriate. When comparisons involved more than two groups and/or repeated factors, a two-way ANOVA was conducted, followed by Tukey’s, Sidak’s, or Holm-Sidak’s post hoc multiple-comparisons and planned comparisons test. Differences between conditions were assessed using a nonparametric permutation framework. A global test was performed using the Frobenius norm of the difference between condition-averaged matrices. Statistical significance was determined via Monte Carlo permutation (10,000 iterations), in which animal labels were randomly reassigned while preserving group sizes. P-values were computed as the proportion of permuted statistics exceeding the observed value. To identify specific transitions contributing to global differences, edge-wise comparisons were conducted using two-sided Mann–Whitney U tests on per-animal transition probabilities, with p-values corrected for multiple comparisons using the Benjamini–Hochberg false discovery rate procedure (α = 0.05). A significance threshold of p < 0.05 was applied for all analyses.

Data availability

The raw data that support the findings of this study are available from the corresponding author upon request.

Acknowledgements

We thank Mrs. Dinara Baimoukhametova, Ms. Alexis Passmore, Mr. Leonardo Molina, Dr. Jianjun Sun, Mrs. Cheryl Breiteneder, and Mr. Rodney Barasi for expert technical support. We also thank Dr. Matthew Hill for providing us with access to materials. We are grateful for the support of the Cumming School of Medicine Optogenetics Core Facility and the Hotchkiss Brain Institute Advanced Microscopy Platform. This work was supported by an operating grant to J.S.B. from the Canadian Institutes for Health Research and the Brain Canada Neurophotonics Platform. I.A. was supported by the Canadian Institutes for Health Research-REDI-Early career transition award.

Additional information

Author contributions

I.A. and M.P. conducted experiments, analyzed data, and prepared figures. T.S designed experiments and contributed to the paper; T.F. designed experiments, discussed results, and prepared figures. I.A. and J.S.B. designed experiments, discussed results, wrote the paper, and supervised the project.

Funding

Canadian Institutes of Health Research (CIHR)

• Jaideep Bains

• Ibukun Akinrinade

Natural Sciences and Engineering Research Council of Canada (NSERC)

• Ibukun Akinrinade

Azrieli Foundation

• Ibukun Akinrinade

Additional files

Supplementary file 1.