Introduction

Play behavior occurs across numerous species and is most often seen in juvenile animals. This has been hypothesized as a means by which juveniles practice adult skills such as hunting and fighting for social dominance or resources1,2,3. However, particularly amongst domesticated species, play behaviors can persist into adulthood (behavioral neoteny). For example, amongst domesticated carnivore species (e.g., Felis silvestris catus and Canis lupus familiaris), adults play more frequently than their wild counterparts4.

The domestic ferret (Mustela putorius furo) is a small (≤ 2 kg) carnivorous species related to European polecats. As with canids and felids, the domestic ferret shows behavioral neoteny relative to wild mustelids5. Ferret play behavior is dynamic, characterized by changes in body posture, movement, and vocalization. While ferrets do seem to prefer social play, typified by locomotor and rough-and-tumble interactions1,6,7, they also express behaviors associated with elevated mood and play during solitary activities. The degree and variety of play behaviors expressed by ferrets, as well as their persistence into adulthood, has led to suggestions that these behaviors be used as a metric of physical and/or psychological well-being in a variety of ages and settings, including the laboratory2,8,9,10.

Ferrets are gaining popularity as a model animal for biomedical research. They are naturally susceptible to many human pathogens, including SARS-CoV2 and influenza, making them attractive for immunology and virology research. Ferrets also possess a gyrencephalic brain with elements of cortical organization more similar to humans than rodents6,11. Due to their primate-like cerebroarchitecture, along with their small size, relatively low cost to procure and house, and amenability to behavioral studies6,12,13,14, use of ferrets for neuroscience research has increased. As a result, methods for evaluating neurologic function, including psychological well-being, are critical for the translational value of these models. Rodent assays have been applied in ferrets12,13,14 but, given the broad interspecies differences between rodents (prey species) and ferrets (a predator species), it is possible that this may not be entirely appropriate. This demonstrates the need for species-specific assays, ideally ones that can be executed and analyzed consistently and efficiently.

While play behavior is a convenient and semi-naturalistic potential metric for psychological well-being, no detailed ethograms exist for the systematic evaluation of solitary play in the laboratory setting. This represents a critical gap in ferret brain literature which, to date, has emphasized social play10,15. Social play is informative and supported in the literature, but experimental manipulations often require isolation, making social play assessment inaccessible. In this study, we develop and apply a simple method to assess solitary behaviors associated with play and positive affect in juvenile and young adult laboratory ferrets, ages that are frequently used in neuroscience research12,16,17,18.

Results

Physical well-being

No signs of physical ill-being were noted in any of these animals for the duration of the study. Body weight increased over the course of the experiment and aligned well with established ferret growth curves19. At 3 months of age, ferrets weighed (mean ± SE) 1.41 ± 0.30 kg whereas 6-month-old ferrets weighed 1.73 ± 0.056 kg.

Ethogram

Animals were observed and a literature search was performed to develop a list of behaviors associated with play and elevated mood in ferrets, including the “dook” vocalization, weasel war dance, excitement hissing, and open mouth play face8,20,21. These behaviors occurred in the context of movement (e.g., running, dancing) or object interaction and not during idle time; instances of inactivity were infrequent and very brief. Additional behaviors associated with environmental exploration and “other” behaviors were also described20,22. Definitions were refined to apply to the laboratory setting. See photos in Fig. 1 and the final ethogram in Table 1.

Fig. 1
figure 1

Photographs of ferrets engaging in some of the behaviors on the ethogram including, (A) open mouth play face and weasel war dance, (B) self-groom, (C) object exploration, (D) exiting the carrier via weasel war dance, (E) scent mark, (F) ascend, (G) travel over object, (H) reach, (I) burrow-inside, (J) chase, and (K) burrow-enter.

Table 1 Ferret behavior ethogram.

Play/Positive affect behaviors

Total Play/Positive Affect Behavior was significantly greater in 6-month-old ferrets relative to 3-month-old ferrets (F(1,19) = 6.372, p = 0.021). Given the lower frequency of Excitement Hissing, Open Mouth Play Face, and Weasel War Dance, this effect was mediated almost entirely by Dook Vocalization which just missed significance (p = 0.052). See Table 2.

Table 2 Number of 5 s intervals in which each behavior occurred (mean ± SE). Total Play/Positive Affect and Total Environmental Interaction represent a sum of the frequency of behaviors within their given category. Significant differences between timepoints were noted for Total Play/Positive Affect, Ascend, Burrow- Enter, Object Exploration, and Remain in Carrier. Statistical significance is denoted by asterisk (*p < 0.05; **p < 0.01).

Environmental interactions

While Total Environmental Interaction Behavior did not differ by age (p = 0.932), several behaviors changed over time. Animals aged 3 months explored objects more frequently than those aged 6 months (Object Exploration: F(1,19) = 9.996, p = 0.005). Animals aged 6 months climbed onto objects more frequently than those aged 3 months (Ascend: H(1) = 8.040, p = 0.005) as well as entered burrows more frequently (Burrow-Enter: F(1,19) = 5.857, p = 0.026).

Other behaviors

Interestingly, animals aged 3 months took significantly longer to exit the pet carrier than those aged 6 months (Remain in Carrier: F(1,19) = 5.454, p = 0.031). No additional differences in ‘other’ behaviors were noted.

Discussion

An ethogram was developed to examine solitary behaviors associated with play and positive affect in the pair-housed naïve laboratory ferret and applied in animals that were 3- and 6-months of age. Analyses demonstrated that, at both ages, ferrets engage in a variety of behaviors. Some differences between ages were noted. However, it is possible that the observed differences are simply due to increased familiarity with the behavior suite, given that the same animals were assessed at both ages. For example, increased anxiety could suppress the expression of play behavior and increase the latency to exit the carrier into the behavior suite. Habituation can decrease anxiety-like behavior in response to a stimulus; exposure therapy is a successful treatment for anxiety disorders in humans23.

Extensive habituation was performed prior to the first test at 3 months of age as well as between tests, which likely minimized the effects of spatial novelty/familiarity. Therefore, it is also possible that the observed changes in behavior over time could be linked to a developmental shift in anxiety-like behavior. While a visit to the behavior suite was not intended to be stressful, it was a deviation from the ferrets’ normal routine and likely caused some emotional arousal. In rats, adolescents exhibit higher anxiety-like responses than adults on popular rodent anxiety assays24. While the neural circuits that mediate anxiety are complex, neuroimaging studies in humans suggest that, relative to adults, adolescents are hypersensitive to stress and less capable of emotional regulation25,26, leading to increased reactivity. It is worth noting that, in rodents, anxiety-like behaviors remain relatively constant or can actually increase with repeated exposure to an apparatus, making repeated measures complex to interpret27. Repeated exposure in rodents is associated with decreased exploratory behavior28; this contrasts to what was observed in this study where total environmental interaction remained constant across ages, though the distribution of behaviors shifted (younger animals engaged in more object exploration whereas older animals engaged in more climbing onto/into objects).

Across the literature, the dook vocalization has been described as an indicator of elevated mood, excitement, and/or positive welfare state8,20,29,30. While frequently associated with social play21,31, it is also observed during exploratory behavior20 as well as interaction with enrichment items29, indicating that a social context is not required for this expression of excitement. Interestingly, dook vocalizations occurred most frequently of all play/positive affect behaviors and were strongly influential on Total Play/Positive Affect Behavior. Therefore, the frequency of dook vocalizations alone may be an informative metric of psychological well-being, particularly if the other described behaviors are subject to a floor effect. If this is done, the other behaviors included on the ethogram may still be used to assess the overall activity of laboratory ferrets. Additional studies need to be done to determine the sensitivity of behavioral frequency to neurologic injury and disease.

Limitations

The current study has several limitations. Most notably, only male animals were included. Ferret social play is sexually dimorphic32, suggesting that solitary play could also be influenced by sex and that the results of this study may not generalize to females.

Play behavior is generally associated with positive welfare and elevated mood2,33, giving it good face validity as a potential measure of psychological well-being in ferrets10,34. At present, there are no validated models of depression in ferrets that could be used to evaluate solitary play as a metric of depressive- and/or anxiety-like behavior so, interpretations related to mood should be made with caution. For example, it has been proposed that play behavior could be a compensatory response to an unenriched environment35. While ferrets in this study were pair housed and provided with standard enrichment in the animal facility, it cannot be discounted that inadequate enrichment could influence behaviors. However, it has been shown that ferrets actually increase social play with greater environmental enrichment with no change to object play15, suggesting that this may not be the case. Further study should determine optimal laboratory ferret housing and enrichment conditions [for example, see 10].

In future studies using these methods, differences in play behavior and/or environmental interaction could indicate an effect of treatment on well-being. However, as previously mentioned, the observed changes in behavior due to age may be linked to habituation. As a result, repeated application of these methods to an experimental cohort to gauge recovery may not be appropriate. Rather, they may be more useful to make comparisons between treatment groups, particularly at a single timepoint.

Future directions

The methods and play behavior ethogram developed in the present study are widely applicable to laboratory ferrets, even those who must be isolated due to the presence of dominance behaviors or surgical wounds. For example, we are currently using them to assess putative changes in psychological well-being in a ferret model of chronic traumatic brain injury. Of note, the methods described do not require any specific apparatus or specialized equipment, simply a dedicated space where ferrets may roam at will and a recording device. While the environmental interaction ethogram was developed based on the enrichment items available in our behavior suite, the principles may be applied to create an ethogram for a comprehensive assessment of activity in any space (though note that the inclusion of enrichment items is critical to elicit play behavior29 and environmental interaction). Furthermore, the time sampling method used in this study can be applied to efficiently generate quantitative analyses of behavior without requiring an excessive time commitment, making this a practical assay for biomedical researchers.

In the past decade, ferrets have become a key animal model for infectious and noninfectious disease research as well as development of drugs and vaccines36. Behavioral assessment is crucial in neurologic and psychiatric research but is also an important component of drug development for conditions affecting peripheral organ systems37,38. For example, developmental toxicity assessments determine the effects of substance exposure on the structural and functional characteristics of offspring, including neurologic function. The development of tools to assess functional outcomes after injury/disease or drug treatment is critical for the advancement of the ferret animal model in biomedical research. While some work has already been performed in this arena13, this study represents the first attempt to quantify species-specific solitary play behavior as a metric of psychological well-being in the laboratory ferret. Pharmacologic validation studies are a critical next step to verify interpretations of ferret behavioral outcomes, particularly for depressive- and anxiety-like behavior39,40.

Methods

The animal use protocol was reviewed and approved by the University of Maryland, Baltimore Institutional Animal Care and Use Committee (0620009) and the U.S. Air Force Surgeon General’s Office of Research Oversight and Compliance as protocol number FWR-2020-0015A. These studies were conducted in a facility accredited by AAALAC, in accordance with the Guide for the Care and Use of Laboratory Animals41 and were performed in compliance with DODI 3216.1 and ARRIVE guidelines.

Subjects

The subjects were n = 10 intact male ferrets (Marshall Bioresources, North Rose, NY). Animals arrived at the University of Maryland School of Medicine (UMSOM) at approximately 8 weeks of age. They were housed in pairs with ad libitum access to food and water on a 12 h light: dark cycle. Ferrets received standard enrichment in the vivarium including access to a plastic burrow or cloth hammock, plastic toys, and ferret-specific treats for foraging. Throughout the experiment, animals were visually monitored at least 3 times per week and weighed weekly by laboratory staff as well as checked regularly by UMSOM Veterinary Resources staff. At approximately 9–10 weeks of age, animals were sedated with dexmedetomidine (intraperitoneal injection, 0.03–0.12 mg/kg), anesthetized with isoflurane (4–5% in 30% O2), and subjected to approximately 90 min of magnetic resonance imaging while being maintained under 1–1.5% isoflurane in 30% O2 as part of another experiment. After imaging, an identifying tattoo was applied to their left ear with a lab animal tattoo system and sedation was reversed with atipamezole (intraperitoneal injection, 0.03 mg/kg).

Apparatus

All tests were performed in a dedicated ferret behavior suite (5.9 × 7.2 m) at a similar time of day (1300–1600). The suite was stocked with a set of enrichment objects including boxes/burrows, pipes, balls, and springs that were placed in a standard configuration prior to testing, thus ensuring contextual consistency. Objects and room surfaces were regularly cleaned with 70% ethanol and/or 10% bleach and allowed to air dry. Animals were habituated to the suite five times (30 min each at the following weeks of age: 9, 10, 11, 12, and 14) and had two suite visits between tests (at 4 and 5 months of age) to reduce environmental and experiential novelty.

Experimental procedure

At 3 and 6 months of age (young adulthood in ferrets is estimated to occur at 6 months of age42), animals were transported individually in covered pet carriers on a metal cart from the vivarium to the behavior suite. The carrier was placed on the floor and door opened. A familiar experimenter stood in a neutral location and filmed behavior for 5 min with a GoPro Hero8 (GoPro, San Mateo, CA). The experimenter moved as little as possible to minimize effects on behavior. At the end of the test, animals were given an additional 25 min to exercise in the suite with their cage mate; this behavior was not filmed or analyzed.

Video analysis

An ethogram inventory of play behaviors was developed from a literature search and behavioral observation (see Fig. 1 and Table 1). Behaviors that occurred with low frequency in a minority of videos were excluded from final analysis (e.g., laying down, standing still). The Behavioral Observation Research Interactive Software (BORIS) was used to analyze the videos43, applying the ethogram to quantify the frequency of behaviors. A time sampling method (sometimes referred to as “one-zero” sampling) with a 5 s interval was used to analyze each 5-min video44,45,46. At the end of each interval, the occurrence/absence of all behaviors during that interval was recorded; the total number of intervals in which each behavior occurred across the 5-min testing period was determined. To ensure consistency of data analysis, intra-rater reliability checks were performed throughout this period. Repeat analyses of the same three videos were performed and total frequencies of each behavior category were compared each time, ensuring an R2 value greater or equal to 0.90.

Data analysis

Individual behaviors as well as categorical scores were analyzed using SigmaPlot statistical software (Inpixon, Palo Alto, CA). One-way analysis of variance (ANOVA) was used to determine the effect of age on behaviors. Data that failed the Shapiro–Wilk normality test (p < 0.05) were subjected to the Kruskal–Wallis one-way ANOVA on ranks. A statistically significant difference between ages was defined as p < 0.05.