Social overtures in Corydoradinae catfish mitigate the effects of acute stress

doi:10.21203/rs.3.rs-7217527/v1

Download PDF

Research Article

Social overtures in Corydoradinae catfish mitigate the effects of acute stress

https://doi.org/10.21203/rs.3.rs-7217527/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Stress from a variety of environmental factors has negative effects on individuals across species. In social species, stress may be exacerbated by social isolation and mitigated by social overtures, where individuals communicate with groupmates to fully reap the benefits of social grouping and cope with negative effects of environmental stress. We evaluated the effects of stress on pair coordination and engagement with a foraging task in Osteogaster aeneus, a gregarious catfish that communicates via tactile interactions. We netted pairs of fish from a social housing tank (a known stressor) and placed them in an unfamiliar tank for 1 hour, after which pairs performed a foraging task. We quantified stress via respiration rate (counting opercular movements) and frequency of stress-associated stereotyped behaviors, and propose this combination as a viable method for measuring both short and long-term stress. In the context of ecologically crucial tasks like foraging, this method informs how organisms cope with stress that comes with changing and/or unfamiliar environments. In our study, we found that respiration rate decreased over time and activity levels increased over time, and that a positive association emerged between stereotyped stress behaviors and social interactions, indicating that more stressed individuals initiate more social overtures. In the context of foraging, we found that rates of social overtures, and not the extent of the stress response, was associated with an individual’s willingness to forage in pairs. We conclude that social overtures may mitigate the effects of stress, leading to better foraging outcomes.

Understanding how animals cope with stress is crucial to understanding their ability to survive in changing conditions. In social species, directed social interactions can buffer stress and improve outcomes in ecologically essential tasks. We explore this in a highly social species, Osteogaster aeneus, using a known stressor to study the impacts of stress on a foraging task by measuring opercular movements and stereotyped locomotion as non-invasive indicators of stress. We discovered that this unique species mitigates stress through an increased rate of social overtures, thereby increasing their willingness to forage after a sudden environmental stressor. This work is crucial to expanding our understanding of how social behaviors help animals buffer stress and remain functional in the face of environmental challenges.

Stress causes a variety of physiological and behavioral changes across taxa; these include the release of catecholamines and glucocorticoid steroid hormones, as well as changes in feeding, social behavior, swimming behaviors, metabolic rate, and ventilatory frequency (Keay et al., 2006; Huber et al., 2017; Rix and Cutler 2022; Sopinka et al., 2016). Fish usually respond to a stressor in three stages: a “primary” release of catecholamines and steroid hormones like cortisol, a “secondary” biochemical response such as changes in metabolic rate and blood biochemistry mediated by stress hormones, and a “tertiary” whole-animal response where an increase in energy demand causes the fish to move away from anabolic processes (Iwama 2006). Accordingly, both acute and chronic stresses impose costs on individuals, including reduced growth (Savendal, 2012) and detrimental physiological effects such as increased rates of aging (Yegorov et al, 2020). Understanding and quantifying various indicators of stress can serve as an objective measure of fish welfare (Iwama 2007) and shed light on responses to change in environmental conditions (Sharma 2019). In addition, stress responses have various costs and benefits in natural populations, and understanding how stress manifests, how stress affects behavior, and how individuals recover from the effects of stress is critical to contextualizing both individual life histories and large-scale behavioral evolution.

Beyond effects on individual fish, stress can manifest changes in population-level social dynamics (Adams and Greeley 2000). In the cichlid A. burtoni, cortisol levels determined whether non-dominant males exhibited direct or displaced behavioral responses to aggression, where a direct response allows challenging aggressors and may also allow elevation of social status (Clement et al. 2005). Dominance and social hierarchy may also be determined by stress in the salmonid O. mykiss, where fish exposed to cortisol were more likely to become subordinate in a pair interaction, and low-status fish had lower growth rates and increased mortality as a result of higher sustained levels of cortisol (Gilmour, Dibattista, and Thomas 2005). By contrast, sociality seems to reduce the negative effects of stress in many taxa, an effect that is enhanced by positive relationships within social groups (Takeda et al. 2003). For example, the cichlid N. pulcher showed lower cortisol levels when recovering from a stress event in the presence of a group than those who recovered alone (Culbert, Gilmour, and Balshine 2019), and the lake sturgeon A. fulvescens showed a shortened stress response when in a group versus in isolation (Allen et al. 2009). These studies demonstrate that social factors can enhance or mitigate the effects of stress depending on context, and a more nuanced understanding of how social context influences the stress response would shed light on the evolution of sociality.

Other factors also influence the effects and manifestation of stress, including sex. Sex differences are prevalent across many species of animals, displayed through both physiology and behavioral responses to various stimuli. In the salmonid O. mykiss, sexually mature females exhibited higher stress responsiveness measured by elevated cortisol and adrenocorticotropic hormone levels when compared to males (Pottinger & Carrick, 1999). In addition, the Eurasian perch (Perca fluviatilis), displayed significant length differences between sexes with females being larger (Mooij, Rooij, Wijnhoven, 2011).

In many social species, social interactions may be initiated via overt overtures (i.e. auditory signals, physical contacts, etc.) as part of a broader communication network between individuals in a social group. Tactile social overtures are present in many mammals, especially primates and cetaceans, serving a variety of purposes such as grooming, play, or appeasement (Nakamura and Sakai, 2013). The use of tactile social overtures in dolphin species like S. frontalis and T. aduncus has been associated with joining or departing a pod, with variation between species in frequency of contact arriving/departing their social group (Paulos, Dudzinski, and Kuczaj 2007). Physical contact through social overtures has also been shown in the cleaner wrasse L. dimidiatus,where they use tactile interaction to mitigate conflict with “client” reef fish (Bshary and Würth 2001). Tactile social overtures are also present in our study species, Osteogaster aeneus (formerly Corydoras aeneus).

Osteogaster aeneus is a unique system to investigate questions related to the impacts of induced physiological stress. This highly social neotropical catfish has been shown to facilitate communication and coordination through an easily quantifiable, directed social interaction behavior termed “nudging” (Riley, Gillie, Savage, et al., 2019). Nudging behavior consists of direct and discrete physical contact between individuals, and is reflective of changes in environmental and social conditions. Previous work has shown that nudging serves as a way of coordinating group anti-predatory response to threats in the wild (Riley, Gillie, Savage, et al., 2019), and is deployed by individuals unfamiliar with social partners to improve social coordination (Riley et al., 2019)

As previously mentioned, stress in fish is often characterized by changes in behavior and physiology (Sopinka et al., 2016). This involves locomotion feeding, social interactions, duration, and swimming intensity (Sopinka et al. 2016). In Corydoradinae catfish, stress-related changes in behavior can also be observed through alterations in nudging behavior. Paired with ventilation rates and ‘wall surfing' we hypothesized that nudging serves as a recovery measure for stress. Wall surfing occurs when the fish move off the bottom of the tank and swim up and down the sides and is considered a non-standard behavior and a potential indicator of stress (Martins et al. 2012). Higher rates of respiration would be correlated with higher degrees of stress, quantified via ventilation rates. This was measured by observing the opening and closing of the opercula (the bony plate that protects the gills and facilitates the movement of water over the gills in respiration) via direct observation of videos, similar to the methodology used by White et al., 2008.

This paper examines how stress influences behavior in O. aeneus pairs, and how stress-induced behavioral changes impact exploratory and foraging behaviors. We evaluated stress in pairs in the first hour of being netted from a social housing tank (a known stressor; Jeffrey et al. 2014) and introduced to an unfamiliar tank, then evaluated how this stress manifested in a foraging task, by placing bloodworms (a preferred food) on the opposite end of a divider and measuring the time it took for both fish to find the worms after the divider was lifted. Our basic predictions were an increase in respiration rate immediately after fish are moved into experimental tanks, which slowly decreases as fish become acclimated. We would also expect to find increased rates of activity and social overtures as fish become more acclimated to the experimental tank, manifesting as increased nudging rates (i.e. nudges divided by time active). Along with this, we explored the relationship between stress, wall surfing, and nudge rate to investigate how manifestations of sociality, including social overtures, are associated with stress mitigation and an individual’s willingness to perform social tasks like foraging.

General Husbandry

Our general husbandry procedure closely followed the protocol described in our laboratory husbandry protocol (Chiang, Haine et al., 2024). The adult fish used in this experiment were bred from the F1 generation of wild O. aeneus (CW097 designation) catfish from the southern Peruvian Amazon and were housed in the Biology-Psychology Building at the University of Maryland in College Park, Maryland, USA. All fish were at least one year of age at the time of this experiment.

Experimental Housing

Protocol for experimental housing involved randomly netting O. aeneus catfish from a 586L social housing enclosure. Two separate pairs were determined after random selection based on visual distinguishability and sexing (all pairs were same-sex); pairs were assigned so individual fish could be easily distinguished by size and/or color. The purpose of this selection was to facilitate ease of analysis and tracking of an individual fish.

For the experimental setup, a 76.84x31.75x32.39 cm tank was sectioned off with a divider, in which the fish are placed in this “stress chamber” for examination (See Figure 1). A set of three cameras were used for video analysis: 1 Kasa Smart brand camera (Model: KC410S) placed overtop the tank for a top view, and 2 Akaso Outdoor brand 4k high-definition action cameras (Model: Brave 4) for a front and lateral view of the chamber. The divider created a 20.32x30.48 cm section, with water in the tank set to 22.86 cm tall. The divider was slightly slanted to allow for full camera view for video analysis. The fish were kept in the stress chamber for 1 hour. Following this, bloodworms were placed on the center of the opposing side of the chamber using a PVC pipe. The divider was then lifted, and pairs were filmed until both fish reached the bloodworms. This process was repeated for each experimental group, resulting in 20 trials and 20 total pairs.

Data Collection and Video Analysis

Behavioral video analysis

Our behavioral video analysis focused on three different sections of each video: the first 5 minutes that the fish were in the experimental tank (after the experimenters had left the room; termed “nudging period 1” from here on), the last 5 minutes before the divider was lifted (termed “nudging period 2” from here on), and the portion between when the divider was lifted and the fish found the food on the opposite end of the tank.

For both 5-minute nudging period we scored before the divider was lifted, we collected information on the total nudges (direct physical contact between individuals) performed as outlined in Riley, Gillie, Savage et al. 2019. We also recorded information about activity, which we measured as the amount of time (in seconds) each fish in the pair spent actively swimming during the 5 minute segment. From nudges and activity, we calculated an Activity Normalized Nudging (ANN) variable, which is the number of nudges an individual initiates in a given nudging period divided by the number of seconds the individual was active during that nudging period. Nudges can only be performed when fish are active, and this variable accounts for differences in activity levels to more accurately portray an individual’s social motivation.

In addition, we recorded the amount of time in seconds that we observed “wall surfing behavior.” Wall surfing occurs when the fish move off the bottom of the tank and swim up and down the sides, and is considered a non-standard behavior and an indicator of stress (Martins et al., 2012). As a more specific, distinct measure of activity than just moving versus not moving, wall surfing allows us to quantify more granular information about a fish’s behavior, allowing for better insights on social motivation/cooperation beyond general swimming behavior. As part of collecting this information, we assumed by default that both fish were moving and that wall surfing was not present. This state is their most commonly observed state under natural conditions (Riley, personal observation). Activity and wall surfing was measured with a 3 second threshold, where a non-default behavior (i.e. presence of wall surfing or no activity) was only recorded if it occurred for 3 or more consecutive seconds.

We also recorded the length of each individual fish using pixel measurements from our filming videos. The length of each fish was converted from pixels to inches via a pixel/inch ratio (calculated as the length of the experimental tank in pixels, divided by its length in inches), during a random segment where fish were clearly visible.

After the divider was lifted, we counted the time in seconds that the fish remained frozen (i.e. not moving; we called this variable “freeze time”), ending after one fish moved. We also recorded the amount of time in seconds that it took for the fish to find the bloodworms on the opposite end of the tank after the divider was lifted (which we termed “time to feed” or “foraging time”).

Respiration video analysis

Finally, we recorded respiration by counting opercular movements of each fish during four intervals after fish were placed into the tank: 0 to 5 minutes after experiments left the room (Respiration interval 1), 15 to 20 minutes after leaving (Respiration interval 2), 35 to 45 minutes after (Respiration interval 3), and 55 to 60 minutes after (Respiration interval 4). Movements were counted during a 10 second segment during each interval in which the fish were stationary and on the bottom of the tank.

Statistics

All analyses were conducted in R 4.5.0 (R Core Team, 2025) with the support of tidyverse 2.0.0 (Wickham, et al. 2019) for data wrangling and plotting, mice 3.18.0 (Van Buuren and Groothuis-Oudshoorn 2011) for multiple imputation, lme4 1.1-37 (Bates et al. 2015) and lmerTest 3.1-3 (Kuznetsova et al. 2017) for mixed-effects modelling, glmmTMB 1.1.11 (Brooks et al. 2017) for negative-binomial models, MuMIn 1.48.11 (Bartón 2016) for model ranking, performance 0.14.0 (Lüdecke et al., 2021) for diagnostic checks, and emmeans 1.11.1 (Lenth 2018) for marginal means and contrasts. A full list of packages and reproducible code is available as supplementary material.

Data were imported for 40 fish (20 pairs); one pair was excluded a priori due to exhibiting no activity (i.e. staying frozen for the duration of the trial) and thus never interacting with each other, leaving 38 fish in 19 pairs. Administrative columns were discarded, variables were renamed for clarity, and identifiers (group ID, fish ID, and sex) were converted to factors.

Respiration rate contained 30 missing values due to camera limitations (10 in interval 2, 17 in interval 3, and 3 in interval 4). Missingness was handled with multiple imputation by chained equations using the random-forest algorithm (20 imputations, seed = 2679) in mice (Stef van Buuren, Karin Groothuis-Oudshoorn, 2011) . For figures, imputed draws were averaged to a single value per missing datum, while all inferential analyses combined estimates across imputations using Rubin’s rules. For the three values missing in interval 4, the mean of the 20 imputations was inserted into the main dataset so that the second nudging period respirations variable contained no missing values.

Two derived variables were created. Activity normalized nudging (ANN) was calculated for each nudging period as (nudges ÷ activity) × 60, having units of nudges per minute active, and retaining zeros when activity was zero. Active foraging time was defined as total feed time - freeze time.

To inform distributional choices for count outcomes, each count variable (activity, nudges, and wall surfing) in both periods was fitted to an intercept-only negative-binomial model and evaluated for zero inflation. There were modest extra zeros for activity (ratio = 0.40, p = 0.064 in nudging period 1; ratio = 0.04, p = 0.008 in nudging period 2) and for wall surfing in period 2 (ratio = 0.29, p = 0.008). Nudging counts were well captured (ratio = 0.92 in nudging period 1; ratio = 0.55 in nudging period 2). Because the excess zeros were small and full models later showed satisfactory residual behavior, the negative binomial family was retained without an added zero-inflation term for those responses where it was needed. Specifically, a generalized linear mixed model (GLMM) with a negative binomial error distribution with a log link was used for five of the fifteen final regressions: the intrinsic and behavioral models of activity, and all three models of wall surfing. All other models employed Gaussian errors.

We investigated potential redundancy between activity and wall surfing three ways. Pearson correlations between the two behaviors were calculated separately for each observation period, variance-inflation factors (VIFs) were obtained from linear models with total feeding time regressed on both predictors to gauge multicollinearity, and a two-factor varimax-rotated exploratory factor analysis of activity, wall surfing, and nudging was run within each period to evaluate their loading structure.

Period-to-period differences in each behavior were tested with single-predictor mixed-effects models that contrasted nudging period 2 against nudging period 1. Activity, nudging, and wall-surfing counts were fitted with negative-binomial GLMMs, and respiration rate, which was approximately normal, was analyzed with a Gaussian mixed model. All models included a random intercept for fish ID. To corroborate these tests without distributional assumptions, the same four behaviors were compared with paired Wilcoxon signed-rank tests.

Body length differences between the sexes were checked with a linear model and a confirmatory Wilcoxon rank-sum test.

For each response variable, three a priori mixed-effects models were fitted as outlined in Table 1. Intrinsic models contained sex, nudging period, length, and the sex × period interaction. Behavioral models contained nudging period, and the three contemporaneous behaviors aside from the outcome behavior. Lagged models regressed the second period response behaviors on all of the first nudging period behaviors (activity, ANN, respiration rate, and wall surfing). All models used a random intercept for group ID.

Table 1: Fixed-effect structure of candidate mixed-effects models tested for each behavioral response. The random term (intercept for group ID) is identical in all models.

Response	Model 1 (intrinsic)	Model 2 (behavioral)	Model 3 (lagged)
Activity	Activity² ~ Sex × Nudging Period + Length	Activity² ~ Nudging Period + ANN+ Respiration Rate + Wall Surfing	Activity 2nd¹ ~ Activity 1st + ANN 1st + Respiration Rate 1st + Wall Surfing 1st
Activity Normalized Nudging (ANN)	ANN¹ ~ Sex × Nudging Period + Length	ANN¹ ~ Nudging Period + Activity + Respiration Rate + Wall Surfing	ANN 2nd¹ ~ Activity 1st + ANN 1st + Respiration Rate 1st + Wall Surfing 1st
Respiration Rate	Respiration Rate¹ ~ Sex ×Nudging Period + Length	Respiration Rate¹ ~ Nudging Period + Activity + ANN+ Wall Surfing	Respiration Rate 2nd¹ ~ Activity 1st + ANN 1st + Respiration Rate 1st + Wall Surfing 1st
Wall Surfing	Wall Surfing² ~ Sex × Nudging Period + Length	Wall Surfing² ~ Nudging Period + Activity + ANN + Respiration Rate	Wall Surfing 2nd² ~ Activity 1st + ANN 1st + Respiration Rate 1st + Wall Surfing 1st
¹ Linear mixed-effects model ² Negative-binomial (zero-inflated) generalized mixed model

Fixed-effect terms were assessed with type-III likelihood ratio tests for GLMMs, or with Satterthwaite degrees of freedom for linear mixed models (LMMs). Where the sex × period interaction appeared significant in an intrinsic model, planned pairwise contrasts were extracted with estimated marginal means. P-values for multiple fixed effects within each intrinsic model were adjusted by the Benjamini-Hochberg method. Model diagnostics confirmed that residual dispersion, zero patterns, and random-effect variances were acceptable in every case.

Foraging period analyses focused on total feed time, freeze time, and active foraging time. Each outcome was modelled with an LMM that included a random intercept for group ID. For each outcome variable, a global model was first specified containing sex, body length, activity, ANN, respiration rate, and wall surfing; all variables aside from sex were centered and scaled to have a mean of zero and standard deviation of one to aid convergence and to place regression coefficients on a comparable scale.

We fitted every possible subset of the global model and ranked the alternatives with the small-sample corrected Akaike Information Criterion (AICc). For each outcome variable, the single model with the lowest AICc was selected for inference. The best supported models for freeze time and active foraging time retained both activity and ANN, which share some variance, however variance-inflation factors (VIFs) were below 5^{^[1]}, so no term was removed for collinearity. Likelihood-ratio tests verified the contribution of every fixed effect in the chosen models, and residual checks confirmed homoscedasticity and normality assumptions.

Results are reported as unstandardized regression coefficients (ß) for Gaussian models, incidence-rate ratios (IRR = e^ß) for negative-binomial models, likelihood-ratio test statistics (LRTs), and two-sided p-values. A ß represents the expected additive change in the outcome per one-unit increase in the predictor, whereas an IRR expresses the multiplicative change in the expected count. LRTs compare nested models to assess the contribution of individual terms. Significance was set at p < 0.05 unless stated otherwise.

Males were significantly shorter than females (estimate = -0.88 cm, p < 0.001; Wilcoxon p < 0.001), confirming a strong sex difference in body length. Activity, nudging, wall surfing, and respiration rate all changed markedly between the two nudging periods (See Figure 2). Fish were active for more than twice as long in nudging period 2 than in nudging period 1(IRR = 2.09, 95% CI = 1.24 0 3.52, z = 2.77, p = 0.006; Wilcoxon p < 0.001). The increase in nudging was even stronger, with nudging counts rising 6.3-fold (IRR = 6.26, 95% CI = 3.45 - 11.36, z = 6.04, p < 0.001; Wilcoxon p < 0.001). Time spent wall surfing also grew substantially, being 5.1-times higher in the second nudging period (IRR = 5.10, 95% CI = 2.26 - 11.50, p < 0.001; Wilcoxon p < 0.001). In contrast, respiration rate declined by 19.2 breaths per minute (ß = -19.17, 95% CI -22.43 to -15.91, p < 0.001; Wilcoxon p < 0.001). Taken together, these findings indicate that Corydoradinae become more active and socially engaged, and breathed more slowly during the second nudging period.

Activity

In the intrinsic model, activity rose markedly from nudging period 1 to nudging period 2 (IRR = 2.404, adj-p = 0.001). Sex (IRR = 1.027, adj-p = 0.945), length (IRR = 0.940, adj-p = 0.869; LRT = 0.203), and the sex × period interaction (IRR = 0.755, adj-p = 0.704; LRT = 0.848) were all non-significant, implying that the approximately 2.4-fold increase in activity across nudging periods was uniform across size and sex.

In the behavioral model, wall surfing (IRR = 1.005, p < 0.001; LRT = 13.302) within a given nudging period was the only significant predictor of activity. Nudging period no longer contributed once the concurrent behaviors were included in the model (IRR = 1.220, p = 0.562; LRT = 0.334). ANN showed a non-significant negative trend (IRR = 0.981, p = 0.087; LRT = 2.767), and respiration was unrelated to activity (IRR = 0.999, p = 0.925; LRT = 0.009).

In the lagged model, activity in the second nudging period was not predicted by any behavior measured in the preceding nudging period. Neither first nudging period activity (ß = 0.086, p = 0.182, LRT = 2.258), ANN (ß = 0.027, p = 0.955, LRT = 0.008), respiration rate (ß = -0.722, p = 0.101; LRT = 3.475), nor wall surfing (ß = 0.054, p = 0.533; LRT = 0.452) reached significance, implying that variation in activity during the second nudging period arises largely from factors other than a fish’s behavioral state in the first nudging period.

Activity and wall surfing were only moderately related in the first nudging period (r = 0.495, p = 0.002) but became highly correlated in the second nudging period (r = 0.863, p = 3.055 × 10^-12). Collinearity was not an issue in either nudging period, with a VIF of 1.325 in the first nudging period, and 3.928 in the second nudging period. Factor analysis of the three behaviors in each nudging period showed that, although both variables loaded on to the same latent factor, their loadings were markedly different in the first nudging period (activity = 0.643, wall surfing = 0.247) and converged in the second nudging period (activity = 0.728, wall surfing = 0.664). So the two measures are initially distinct and align over time, indicating that they aren’t redundant proxies of the same behavior.

Activity-Normalized Nudging (ANN)

In the intrinsic mode, ANN differed significantly by both period and sex. For females, ANN rose by about 8.85 nudges per minute active between nudging periods 1 and 2 (ß = 8.851, adj-p < 0.001), whereas males showed no such change, as evidenced by the negative period × sex interaction (ß = -8.872, adj-p = 0.010; LRT = 8.523). Sex on its own (ß = 2.326, adj-p = 0.447) and body length (ß = 0.423, adj-p = 0.719; LRT = 0.139) were non-significant, indicating that the period effect emerged only in females and was unrelated to size.

In the behavioral model, no predictor reached significance. ANN showed no detectable association with activity (ß = -0.011, p = 0.420; LRT = 0.702), respiration rate (ß = -0.009, p = 0.934; LRT = 0.007), or wall surfing (ß = 0.014, p = 0.415; LRT = 0.719). Period also failed to predict ANN once same-period behaviors were included (ß = 4.097, p = 0.181; LRT = 1.931). Thus variation in ANN within a period was not explained by the other measured behaviors.

In the lagged model, there also were no significant carry-over effects from the first nudging period. ANN in the second nudging period was unrelated to activity in the first nudging period (ß = -0.021, p = 0.313; LRT = 1.195), ANN in the first nudging period (ß = 0.106, p = 0.642; LRT = 0.173), respiration rate in the first nudging period (ß = -0.222, p = 0.211; LRT = 1.656), or wall surfing in the first nudging period (ß = 0.009, p = 0.830; LRT = 0.044).

Respirations

Respiration rates declined sharply over the course of the trial (Figure 3). Specifically, respiration in interval 1 was significantly higher than in interval 4 (estimate = 18.955, p = 1.861 × 10^-20). There was no significant difference between intervals 3 and 4 (estimate = 0.588, p = 0.808), suggesting a plateau in respiration rate by the third interval.

In the intrinsic model, respiration rate fell sharply between nudging periods 1 and 2 (ß = -18.318, adj-p < 0.001). Sex (ß = -1.296, adj-p = 0.689), body length (ß = -0.598, adj-p = 0.689; LRT = 0.230), and the sex × period interaction (ß = -1.802, adj-p = 0.689; LRT = 0.335) were all non-significant, indicating that the approximately 18 breaths/minute drop between periods was consistent across size and sex.

In the behavioral model, nudging period remained the dominant predictor, with fish respiring at a rate of about 22 opercular beats per minute more slowly in period 2 than in period 1 (ß = -21.698, p < 0.001; LRT = 65.614). Within periods, wall-surfing showed a modest positive association with respiration (ß = 0.046, p = 0.011; LRT = 6.943), whereas activity (ß = -0.022, p = 0.118; LRT = 2.579) and ANN (ß = -0.002, p = 0.985; LRT < 0.001) were unrelated.

In the lagged model, respiration rate in the second nudging period was largely independent of first-period behavior, with only wall surfing in the first nudging period showing a modest carry-over effect (ß = 0.068, p = 0.049; LRT = 3.928), whereas first nudging period activity (-0.019, p = 0.279; LRT = 1.319), ANN (ß = -0.039, p = 0.832; LRT = 0.050), and respirations (ß = -0.114, p = 0.428; LRT = 0.543) were non-predictive.

Wall Surfing

In the intrinsic model, sex (IRR = 0.811, adj-p = 0.660), the sex-period interaction ( IRR = 1.879, adj-p = 0.232; LRT = 2.326) and body length (IRR = 1.128, adj-p = 0.624; LRT = 0.520) did not influence wall surfing. Period was initially significant, but only suggests a trend after multiplicity correction (IRR = 2.148, unadj-p = 0.016, adj-p = 0.064), suggesting that wall surfing roughly doubled from nudging period 1 to 2.

In the behavioral model, wall surfing was strongly related to concurrent behavior. Within a period, activity was a 0.07% increase in wall surfing time (IRR = 1.007, p < 0.001; LRT = 16.503), and each extra breath per minute increase in respiration rate was associated with a 2.3% increase in wall surfing (IRR = 1.023, p = 0.044; LRT = 3.949). Period remained important, with fish wall surfing about 2.7-fold more in the second nudging period compared to the first (IRR = 2.734, p = 0.007; LRT = 6.524). ANN was unrelated to time spent wall surfing (IRR = 1.001, p = 0.968; LRT = 0.002)

In the lagged model, earlier behavior predicted later wall surfing in the opposite direction, with higher first nudging period ANN predicting 9% less surfing in the next period (IRR = 0.909, p < 0.001; LRT = 7.102), and faster first nudging period respiration rate was associated with a decrease in wall surfing in the second nudging period (IRR = 0.979, p = 0.024; LRT = 16.348). Activity in the first nudging period (IRR = 1.001, p = 0.666; LRT = 0.187), and wall surfing in the first nudging period (IRR = 1.002, p = 0.192; LRT = 1.707) had no carry-over effect, indicating that subsequent wall surfing is shaped more by earlier ANN and respiration state than by prior levels of wall surfing or general activity.

Foraging Performance

When considering total time spent finding food, activity in the second period was the sole variable retained by model selection. Each extra second of activity predicted a 4 second reduction in time spent finding food (ß = -4.066, p = 0.015; LRT = 6.182; See Figure 5).

For active foraging time, the best supported mixed-effects model included activity in the second period, nudging rate weighted by activity, and fish length. Each extra second of activity in the second period predicted a 2.4 second reduction in active foraging time (ß = -2.385, p = 0.026; LRT = 5.315; See Figure 7). Each extra nudge per minute of activity in the second period was associated with 15.9 extra seconds spent actively foraging (ß = 15.893, p = 0.039; LRT = 4.943). Although retained in the model, fish length did not explain a significant proportion of variation in active foraging time (ß = -122.364 p = 0.123; LRT = 2.427).

For freeze time, the best supported mixed-effects model included activity in the second nudging period, ANN in the second nudging period, and fish length. Each extra nudge per minute of activity in the second period was associated with a 21.6 second reduction in freeze time (See Figure 6). Although they were retained in the model, neither activity in the second period (ß = -2.351, p = 0.067; LRT = 3.948), nor fish length (ß = 123.586, p = 0.092; LRT = 3.044) explained a significant proportion of variation in freeze time.

Activity in the first period significantly corresponded to nudging in the second period, and there were significantly more nudges in the second period compared to the first. There was no sex difference in ANN in the first nudging period, but in the second nudging period, ANN was higher in females than in the first nudging period. Respiration rates were significantly different between the first and last intervals, and respiration rate in the first interval was negatively associated with nudging in the second nudging period, indicating that higher acute stress leads to reduced nudging. While activity was generally not a significant predictor of nudging, wall surfing and nudging rates were positively correlated in both sexes for the second nudging period, though only among female fish in the first period. However, activity in the second nudging period and nudging rate were significant predictors of an individual’s willingness to forage (more activity and higher nudging rates meant more time spent foraging after the barrier was released).

Nudging rate and activity during the second nudging period were the best predictors of the willingness to forage in Corydoradinae pairs. Individuals that nudged more frequently in the second nudging period spent less time frozen (i.e. exhibited a lower extent of the freeze stress response) following the removal of the barrier; individuals that nudged more frequently in the second nudging period also required less active foraging time to find the food.

Fish that had higher activity and nudging rates before the removal of the divider (i.e. an environmental change that can cause a stress response) had better foraging outcomes: they spent less time frozen, a common stress response in this species, and spent more time attempting to find food as compared to less active fish and fish with low nudging rates. The relationship between activity immediately prior to the foraging task and the time to find food is logical in that fish that were moving more frequently immediately prior to the foraging task will tend to move across the tank and find the food more quickly. The fact that high nudging rate led to more active foraging time suggests that nudging can help manage and mitigate the effects of stress to be able to perform social tasks like foraging. Higher nudging rates may therefore mitigate stress, leading to the presence of more foraging in active pairs. Previous work in this system has shown that nudging is used to coordinate group responses to a threat and predicts recovery from a threat event (Riley, 2019), and our results suggest that foraging success may increase as nudging rate increases (via mitigating individual stress and increasing the amount of time fish actually spend foraging). Future work may measure the extent of high nudge rates and activity in combination with physiological data like respiration rate and cortisol levels to evaluate the lasting effects of stress over longer timescales (especially in the context of performing a task, like the foraging measured in this study), as part of a larger study of biological stress and social coordination.

ANN during the second period was the only variable where there was an effect of sex, where females exhibited significantly more nudges than males. While female Corydoradinae are usually larger (in both length and midsection) than males (Riley, 2012), this is unlikely to account for the observed sex differences as length was not a significant predictor of nudging behavior. A closer examination into sex differences in nudging in the future would help to elucidate what causes the difference, and would continue to inform on previous work showing that sex differences in fish can result in different physiological and behavioral response differences to stimuli (King 2013, Eliason 2020).

Our results also showed a significant decrease in respiration rates between the first and last intervals. We observed that sex and size did not significantly influence respiration rates, which indicates that this observed pattern was a shared physiological response across Osteogaster aeneus individuals, and not induced by specific physical characteristics. An individual’s amount of time active was also independent of respiration rate, suggesting that differences in respiration rate are not because of differences in metabolic rate or high/low activity. These findings demonstrate that O. aeneus exhibited a pronounced stress response indicated by elevated respiration rates, which significantly decreased as they adjusted to the initial stressor.

This pattern aligns with existing research in stress physiology in fish where respiration rates are commonly used as a proxy for stress (Sopinka et al. 2016). This is due to an influx of catecholamines from chromaffin cells resulting in acute stress due to factors such as handling and transport (Bonga 1997). In our study, the initial stress response was triggered by handling and netting the fish from their tank of origin into test tanks. This technique of inducing stress has been validated through previous studies, such as the work by Jeffrey et al. 2014.

Our results highlight both the utility and limitations of using respiration rate as a proxy for stress. One potential difficulty of measuring respiration via direct observation of ventilation of the gills is that movement of the operculum becomes more difficult to see as breathing rate slows down, which could make it harder to obtain results for longer-term stress evaluation. In addition, while frequency of opercular movements is useful as a measure to detect disturbances from stressors, it is limited in its ability to measure the extent to which a stimulus acts as a stressor (Barreto and Volpato 2004). While ventilation rate has limitations as a proxy to quantify the degree of stress in individual animals compared to chemical methods like measuring cortisol, it is still a viable tool for identifying the initial onset of stress, and can be used to show immediate measurable manifestations in physiological stress, as well as the abatement of stress, as seen in the results reported above. It is also a noninvasive technique for assessing physiological stress, minimizing the possibility that researcher interventions can affect the assessment of the stress response, which is common with techniques that measure cortisol directly (Lemos et al., 2023). It seems to be especially useful for examining the short-term manifestation of stress immediately after a stressful event, and when paired with additional measures of stress like activity or wall surfing, can give valuable insights about recovery from stressful events.

Along with ventilation, wall surfing is also a behavior that reflects a stress response. Our results suggest that wall surfing is indicative of the stress response beyond its relation to activity. While wall surfing is a manifestation of activity (i.e. fish that are wall surfing are also active), it is a more nuanced measure of stress and sociality, since it correlates with nudging where activity in general does not. This means that fish that are more active do not necessarily nudge more frequently, but fish that spend more time wall surfing do nudge more frequently. Previous work in O. aeneus has shown that wall surfing can be used as a quantitative measure of stress (Siddiqui et al., 2024), but the correlation we see in this study with nudging suggests that wall surfing is associated with an increase in social interactions following a stressor, as pairs that exhibit higher rates of wall surfing also exhibit higher frequencies of nudging. Therefore, wall surfing in Corydoradinae represents a stereotyped behavior similar to behaviors other animals display to cope with stress (Shepherdson 2013, Sun 2023), but with a potentially novel effect on social systems. In our study, wall surfing is associated with a higher ANN, suggesting that wall surfing may provide benefits in mitigating the effects of stress.

Our findings have been able to contribute to a broader understanding of how animals respond to and mitigate stress through social interactions. The observed patterns of nudging, wall surfing, and ventilation rates in Osteogaster aeneus suggest that social overtures can enhance stress regulation and allow individuals to return to homeostasis. Among mammals, M. fuscata utilizes overtures to maintain group cohesion similar to how O. orca can maintain stable social groups for hunting and survival through distinct vocalizations (Tyak, 2018). Through initiating and receiving social overtures, social animals can reduce the negative effects of stress, improve social coordination, and ultimately maximize the benefits of sociality. This sheds light on an important consequence of social selection: if selection pressures lead social interactions to reduce stress, then we should expect to see further selection for social motivation.

This emphasizes the importance of communication mechanisms for social animals, and how initiating social interactions can facilitate coordination following a stressful event (Riley, 2019). As animals have to face a huge variety of severe stressors in natural environments, mechanisms to maintain social coordination despite stress play a crucial role in the life histories of social animals. In the case of our study system, social overtures between partners not only improved coordination and minimized the effects of stress, but facilitated partners to more-willingly forage following a stressful event.

The willingness to forage is essential to the survival of social species such as Osteogaster aeneus. O. aeneus are native to murky, sediment-rich freshwater habitats within the Amazon, which forces them to navigate low-visibility environments (Nijssen in Lambourne, 1995; Riley, 2012). These challenges are compounded by their naturally poor eyesight (Riley et al, 2019, Kohda et al., 2002). In wild and captive settings, fish can be seen constantly scouring their habitats with their sensitive barbels to locate food, indicating that active exploration of feeding patches is integral to these fishes’ foraging ecology (Riley 2019, Riley 2012). Therefore, withdrawal from a foraging opportunity following a stressor could result in missed feeding opportunities. Accordingly, the fishes’ use of social overtures allows them to resume their foraging behavior and minimize costly cessation of foraging attempts, which is essential for survival in the face of environmental stressors.

This represents a fundamental adaptation in social animals: the association between social interactions and lowered stress. As sociality is instrumental in so many species, animals must be motivated to seek out social partners and coordinate with them in order to reap the benefits of group living. Receiving additional benefits from social overtures, such as mitigation of stress, would intensify the selection pressures for social and prosocial behaviors. Vocal overtures, for example, are associated with the formation of social bonds in a variety of taxa (Tyack, 2008), and social bonds improve fitness in horses (Cameron et al., 2009), baboons (McFarland et al., 2017), and macaques (Schulke, 2010). Social bonds also have the effect of mitigating stress for individuals in proximity to social affiliates (McNeal et al., 2017; Culbert, Gilmour, and Balshine 2019; Allen at al. 2009). In this way, social animals that receive stress-lowering benefits not only from social proximity, but upregulated social overtures, would experience enhanced selection on social behaviors because of these benefits. Corydoradinae catfish are a system with high potential for elucidating the behavioral ecology and evolutionary underpinnings of social selection in the context of stress and social coordination.

Acknowledgements

We would like to thank Dr. Kimberly Paczolt for feedback throughout the process of this work, Dr. Gerald Wilkinson for facilities support, Dr. James Savage for thoughtful comments on the manuscript, Ms. Elizabeth Haynes for input on the manuscript and logistical support, and Mr. John Riley for logistical support in husbandry.

Ethics Statement

The protocols delineated above were approved by the appropriate committees associated with ethics in animal research at the University of Maryland (UMD). These methods were approved as part of our IACUC protocol at UMD (protocol no.: 2123596-1).

Funding Declaration

This work was funded by the NSF AGEP Promise Academy fellowship held by RJR

Author Contribution

E.L. co-conceived of the experiment, led and coordinated the execution of the experiment, co-directed the analysis of videos to obtain data from the experimental trials, assisted in data analysis, and co-led the writing of the main manuscript text. M.S. contributed to experimental design, co-coordinated the execution of the experiment, co-directed the analysis of videos to obtain data from the experimental trials, assisted in data analysis, and co-led the writing of the main manuscript text. S.S.S.H contributed to experimental design, participated in the execution of the experiment, assisted in video analysis, and provided input on the main manuscript textN.T. contributed to experimental design and was in charge of camera set-up, participated in the execution of the experiment, assisted in video analysis, and provided input on the main manuscript text E.D.C provided input in video analysis, led the data analysis, and contributed to the writing of the main manuscript text, including the production of all tables and figures R.J.R. co-conceived of the experiment, co-led and coordinated the execution of the experiment, contributed to video analysis, contributed to and coordinated data analysis, and co-led the writing of the main manuscript text.

Data Availability

Data is provided within the supplementary information

Allen, P. J., C. C. Barth, S. J. Peake, M. V. Abrahams, and W. G. Anderson. 2009. “Cohesive Social Behaviour Shortens the Stress Response: The Effects of Conspecifics on the Stress Response in Lake Sturgeon Acipenser Fulvescens.” Journal of Fish Biology 74 (1): 90–104. https://doi.org/10.1111/j.1095-8649.2008.02112.x.
Barreto, Rodrigo Egydio, and Gilson Luiz Volpato. 2004. “Caution for Using Ventilatory Frequency as an Indicator of Stress in Fish.” Behavioural Processes 66 (1): 43–51. https://doi.org/10.1016/j.beproc.2004.01.001.
Barton, Kamil. "Package ‘MuMIn’." CRAN Repos at https://cran. r-project. org/web/packages/MuMIn/MuMIn. pdf (2016).
Brooks, Mollie E., Kasper Kristensen, Koen J. Van Benthem, Arni Magnusson, Casper W. Berg, Anders Nielsen, Hans J. Skaug, Martin Mächler, and Benjamin M. Bolker. "glmmTMB balances speed and flexibility among packages for zero-inflated generalized linear mixed modeling." (2017).
Bshary, Redouan, and Manuela Würth. "Cleaner fish Labroides dimidiatus manipulate client reef fish by providing tactile stimulation." Proceedings of the Royal Society of London. Series B: Biological Sciences 268, no. 1475 (2001): 1495-1501.
Bates, Douglas, Martin Maechler, Ben Bolker, Steven Walker, Rune Haubo Bojesen Christensen, Henrik Singmann, Bin Dai, Gabor Grothendieck, Peter Green, and Maintainer Ben Bolker. "Package ‘lme4’." convergence 12, no. 1 (2015): 2.
Cameron, E. Z., T. H. Setsaas, and W. L. Linklater. 2009. “Social Bonds between Unrelated Females Increase Reproductive Success in Feral Horses.” Proceedings of the National Academy of Sciences 106 (33): 13850–53. https://doi.org/10.1073/pnas.0900639106.
Culbert, Brett M., Kathleen M. Gilmour, and Sigal Balshine. 2019. “Social Buffering of Stress in a Group-Living Fish.” Proceedings of the Royal Society B: Biological Sciences 286 (1910): 20191626. https://doi.org/10.1098/rspb.2019.1626.
Eliason, Erika J., Melissa Dick, David A. Patterson, Kendra A. Robinson, Jeremy Lotto, Scott G. Hinch, and Steven J. Cooke. "Sex-specific differences in physiological recovery and short-term behaviour following fisheries capture in adult sockeye salmon (Oncorhynchus nerka)." Canadian Journal of Fisheries and Aquatic Sciences 77, no. 11 (2020): 1749-1757.
Gilmour, Kathleen M., and Brittany Bard. 2022. “Social Buffering of the Stress Response: Insights from Fishes.” Biology Letters 18 (10): 20220332. https://doi.org/10.1098/rsbl.2022.0332.
Huber, Nikolaus, Leonida Fusani, Andrea Ferretti, Katharina Mahr, and Virginie Canoine. 2017. “Measuring Short-Term Stress in Birds: Comparing Different Endpoints of the Endocrine-Immune Interface.” Physiology & Behavior 182 (December):46–53. https://doi.org/10.1016/j.physbeh.2017.09.017.
Iwama, George K. 2007. “The Welfare of Fish.” Diseases of Aquatic Organisms 75 (2): 155–58. https://doi.org/10.3354/dao075155.
James, G., Witten, D., Hastie, T., & Tibshirani, R. (2013). An Introduction to Statistical Learning: With Applications in R. Springer New York, p 102
Keay, Jessica M., Jatinder Singh, Matthew C. Gaunt, and Taranjit Kaur. 2006. “FECAL GLUCOCORTICOIDS AND THEIR METABOLITES AS INDICATORS OF STRESS IN VARIOUS MAMMALIAN SPECIES: A LITERATURE REVIEW.” Journal of Zoo and Wildlife Medicine 37 (3): 234–44. https://doi.org/10.1638/05-050.1.
King, Andrew J., Ines Fürtbauer, Diamanto Mamuneas, Charlotte James, and Andrea Manica. "Sex-differences and temporal consistency in stickleback fish boldness." PLoS One 8, no. 12 (2013): e81116.
Kohda, Masanori, Kanako Yonebayashi, Miyako Nakamura, Nobuhiro Ohnishi, Satoko Seki, Daisuke Takahashi, and Tomohiro Takeyama. "Male reproductive success in a promiscuous armoured catfish Corydoras aeneus (Callichthyidae)." Environmental Biology of Fishes 63 (2002): 281-287.
Kuznetsova, Alexandra, Per B. Brockhoff, and Rune HB Christensen. "lmerTest package: tests in linear mixed effects models." Journal of statistical software 82 (2017): 1-26. Lambourne, Derek. Corydoras and Catfish: An Aquarist's Handbook. Blandford Press, 1995.
Lemos LS, Angarica LM, Hauser-Davis RA, Quinete N. Cortisol as a Stress Indicator in Fish: Sampling Methods, Analytical Techniques, and Organic Pollutant Exposure Assessments. Int J Environ Res Public Health. 2023 Jun 27;20(13):6237. doi: 10.3390/ijerph20136237. PMID: 37444085; PMCID: PMC10341563.
Lenth, Russell, and Maintainer Russell Lenth. "Package ‘lsmeans’." The American Statistician 34, no. 4 (2018): 216-221.
Lüdecke, Daniel, Mattan S. Ben-Shachar, Indrajeet Patil, Philip Waggoner, and Dominique Makowski. "performance: An R package for assessment, comparison and testing of statistical models." Journal of open source software 6, no. 60 (2021).
McFarland, Richard, Derek Murphy, David Lusseau, S. Peter Henzi, Jessica L. Parker, Thomas V. Pollet, and Louise Barrett. 2017. “The ‘Strength of Weak Ties’ among Female Baboons: Fitness-Related Benefits of Social Bonds.” Animal Behaviour 126 (April):101–6. https://doi.org/10.1016/j.anbehav.2017.02.002.
McNeal, Neal, Katherine M. Appleton, Alan Kim Johnson, Melissa-Ann L. Scotti, Joshua Wardwell, Rachel Murphy, Christina Bishop, Alison Knecht, and Angela J. Grippo. "The protective effects of social bonding on behavioral and pituitary-adrenal axis reactivity to chronic mild stress in prairie voles." Stress 20, no. 2 (2017): 175-182. 10.1080/10253890.2017.1295444
Mooij, W. M., J. M. Van Rooij, and S. Wijnhoven. 1999. “Analysis and Comparison of Fish Growth from Small Samples of Length-at-Age Data: Detection of Sexual Dimorphism in Eurasian Perch as an Example.” Transactions of the American Fisheries Society 128 (3): 483–90. https://doi.org/10.1577/1548-8659(1999)128<0483:AACOFG>2.0.CO;2.
Nakamura, Michio, and Mai Sakai. "Social touch in apes and dolphins." In Primates and cetaceans: Field research and conservation of complex mammalian societies, pp. 355-383. Tokyo: Springer Japan, 2013.
Pottinger, T. G., and T. R. Carrucj. 2000. “Contrasting Seasonal Modulation of the Stress Response in Male and Female Rainbow Trout.” Journal of Fish Biology 56 (3): 667–75. https://doi.org/10.1111/j.1095-8649.2000.tb00764.x.
Paulos, Robin D., Kathleen M. Dudzinski, and Stan A. Kuczaj. "The role of touch in select social interactions of Atlantic spotted dolphin (Stenella frontalis) and Indo-Pacific bottlenose dolphin (Tursiops aduncus)." Journal of Ethology 26 (2008): 153-164.
Riley, Riva. "Watching the watchmen: Evidence for mixed-species associations and a novel social behavior in Corydoras." Harvard University, 2012.
Riley, Riva Jyoti. "Keeping it together: the effect of familiarity, personality, and active interactions on group coordination." PhD diss., 2019.
Riley, Riva J., Elizabeth R. Gillie, James L. Savage, Neeltje J. Boogert, Andrea Manica, and Arne Jungwirth. "The role of tactile interactions in flight responses in the Bronze Cory catfish (Corydoras aeneus)." Ethology 125, no. 11 (2019): 810-820. https://doi.org/10.1111/eth.12935
Rix, Rachel R., and G. Christopher Cutler. 2022. “Review of Molecular and Biochemical Responses during Stress Induced Stimulation and Hormesis in Insects.” Science of The Total Environment 827 (June):154085. https://doi.org/10.1016/j.scitotenv.2022.154085.
Sävendahl, Lars. "The effect of acute and chronic stress on growth." Science Signaling 5, no. 247 (2012): pt9-pt9.
Schaerf, Timothy M., Peter W. Dillingham, and Ashley J. W. Ward. 2017. “The Effects of External Cues on Individual and Collective Behavior of Shoaling Fish.” Science Advances 3 (6): e1603201. https://doi.org/10.1126/sciadv.1603201.
Schülke, Oliver, Jyotsna Bhagavatula, Linda Vigilant, and Julia Ostner. 2010. “Social Bonds Enhance Reproductive Success in Male Macaques.” Current Biology 20 (24): 2207–10. https://doi.org/10.1016/j.cub.2010.10.058.
Sharma, Madhu. 2019. “Behavioural Responses in Effect to Chemical Stress in Fish: A Review.”
Shepherdson, David, Karen D. Lewis, Kathy Carlstead, Joan Bauman, and Nancy Perrin. "Individual and environmental factors associated with stereotypic behavior and fecal glucocorticoid metabolite levels in zoo housed polar bears." Applied Animal Behaviour Science 147, no. 3-4 (2013): 268-277.
Siddiqui, Munir, Austin Chiang, Ethan Lac, Jesse Kern, Gerald Wilkinson, Arne Jungwirth, James Allen, and Riva J. Riley. "Developmental Social Experience Changes Behavior in a Threatening Environment in Corydoras Catfish." Ecology and Evolution 14, no. 10 (2024): e70391. https://doi.org/10.1002/ece3.70391
Sopinka, Natalie M., Michael R. Donaldson, Constance M. O’Connor, Cory D. Suski, and Steven J. Cooke. 2016. “11 - Stress Indicators in Fish.” In Fish Physiology, edited by Carl B. Schreck, Lluis Tort, Anthony P. Farrell, and Colin J. Brauner, 35:405–62. Biology of Stress in Fish. Academic Press. https://doi.org/10.1016/B978-0-12-802728-8.00011-4.
Sun, Fuyu, Xiaoyang Chen, Yongfeng Li, Guangyong Zhao, and Xianhong Gu. "Evaluation of Holstein cows with different tongue-rolling frequencies: stress immunity, rumen environment and general behavioural activity." Journal of Animal Science and Biotechnology 14, no. 1 (2023): 104.
Takeda, Ken-ichi, Shusuke Sato, and Kazuo Sugawara. "Familiarity and group size affect emotional stress in Japanese Black heifers." Applied Animal Behaviour Science 82, no. 1 (2003): 1-11.
Tyack, Peter L. "Convergence of calls as animals form social bonds, active compensation for noisy communication channels, and the evolution of vocal learning in mammals." Journal of Comparative Psychology 122, no. 3 (2008): 319.
Van Buuren, Stef, and Karin Groothuis-Oudshoorn. "mice: Multivariate imputation by chained equations in R." Journal of statistical software 45 (2011): 1-67.
Wickham, Hadley, Mara Averick, Jennifer Bryan, Winston Chang, Lucy D'Agostino McGowan, Romain François, Garrett Grolemund et al. "Welcome to the Tidyverse." Journal of open source software 4, no. 43 (2019): 1686.
White, Alaina J., Jason F. Schreer, and Steven J. Cooke. 2008. “Behavioral and Physiological Responses of the Congeneric Largemouth (Micropterus Salmoides) and Smallmouth Bass (M. Dolomieu) to Various Exercise and Air Exposure Durations.” Fisheries Research 89 (1): 9–16. https://doi.org/10.1016/j.fishres.2007.08.008.
Yegorov, Yegor E., Anastasia V. Poznyak, Nikita G. Nikiforov, Igor A. Sobenin, and Alexander N. Orekhov. "The link between chronic stress and accelerated aging." Biomedicines 8, no. 7 (2020): 198

A VIF value that exceeds 5 or 10 indicates a problematic amount of collinearity (James et al. 2013).

No competing interests reported.

Lacetal2025data.xlsx

Download PDF

Reviews received at journal
08 Nov, 2025
Reviewers agreed at journal
28 Oct, 2025
Reviewers agreed at journal
24 Oct, 2025
Reviewers agreed at journal
14 Aug, 2025
Reviewers invited by journal
06 Aug, 2025
Editor assigned by journal
30 Jul, 2025
Submission checks completed at journal
29 Jul, 2025
First submitted to journal
25 Jul, 2025

You are reading this latest preprint version

Social overtures in Corydoradinae catfish mitigate the effects of acute stress

Status:

Version 1

Abstract

Figures

Significant Statement

INTRODUCTION

METHODS

RESULTS

DISCUSSION

Declarations

Author Contribution

Data Availability

References

Footnotes

Additional Declarations

Supplementary Files

Status:

Version 1