Play Caging Benefits the Behavior of Singly Housed Laboratory Rhesus Macaques (Macaca mulatta)

Abstract

This study addresses a recommendation in The Guide for the Care and Use of Laboratory Animals to provide singly housed nonhuman primates with intermittent access to large, enriched (play) caging. Research on the potential benefits of this type of caging is limited. The present study examines the effects of play caging on behavior, activity, and enrichment use. Singly housed, adult male, rhesus macaques (n = 10) underwent a baseline phase in their home cages, a 2-wk treatment phase with housing in play cages, and a posttreatment phase after returning to their home cages. Each subject underwent focal behavioral observations (n = 10; duration 30 min each) during each study phase, for a total of 150 h of data collection. Results showed increases in locomotion and enrichment use and a trend toward decreased abnormal behavior while subjects were in the play cage, with the durations of these behaviors returning to baseline levels after treatment. Anxiety-related behaviors decreased between the treatment and posttreatment phases but not between baseline and treatment, suggesting that outside factors may have influenced the decline. During the treatment phase, subjects spent more time in the upper quadrants of the play caging and preferred a mirror and forage boards as forms of enrichment. The greatest behavioral improvement occurred during the first week in the play cage. This study provides evidence to support the benefits of play caging for singly housed rhesus macaques.

The present study examined the use of play cages for singly housed rhesus macaques (Macaca mulatta) as a response to a recent recommendation in The Guide for the Care and Use of Laboratory Animals.¹⁵ The Guide is a widely used manual for institutions in the United States housing animals for the purposes of research, teaching, and testing. In this latest edition, it is stated that “if single animals are housed in small enclosures, an opportunity for periodic release into larger enclosures with additional enrichment items should be considered, particularly for animals housed singly for extended periods of time” (p 60). Providing this type of caging would entail considerable financial costs as well as limit the space available for standard-sized caging. Quantitative research providing scientific evidence on the possible advantages of play caging for singly housed nonhuman primates will allow facilities to make more informed decisions about the costs and benefits of this kind of investment. The present study was intended to provide such evidence through the examination of the effects of play caging on abnormal behavior, anxiety, locomotion, and enrichment use of adult rhesus monkeys.

Gaining an understanding of the efficacious treatments for abnormal behavior is vital, not only because of ethical concerns and federal regulations for humane care⁴¹ but also because a psychologically compromised animal makes a poor model for scientific research.^39,42 A survey of 22 nonhuman primate (NHP) laboratories revealed that about 27% of the 35,000 NHP were housed alone.³ Single housing is a major contributing factor in the development of abnormal behavior in primates.^2,5,25 Of particular concern are self-biting and other potentially injurious abnormal behaviors. A study of singly housed rhesus macaques reported that 89% of subjects displayed at least one type of abnormal behavior; of these, 25% were identified as self-biters.²⁷ Other abnormal behaviors that have been correlated with the development of self-injurious behavior (SIB) include bizarre posture, floating limb, eye-directed, and self-oral behaviors.^6,7,27,30,33 These behaviors are hereafter referred to as SIB precursor behaviors.

We have a particular interest in finding effective enrichment interventions for NHP exhibiting or at risk for developing SIB. Social enrichment is often not an option for these animals due to study assignment, and inanimate enrichment alone has not proven effective in ameliorating SIB.^23,30,31 In light of the important link between single housing and SIB, we decided to focus on subjects that exhibited self-biting without wounding or SIB precursor behaviors during routine behavioral observations conducted by our institution's Behavioral Management Unit.

Anxiety- or tension-related behaviors are exhibited in times of stress or uncertainty, and those commonly noted in NHP are self-scratching, yawning, body shaking, and teeth grinding.^1,7,9,13,29 For NHP with a tendency to exhibit SIB, anxiety can elicit this behavior.²⁶ Severe or prolonged anxiety can lead to a decline in overall welfare and, therefore, should be reduced in the laboratory.^14,35,42

Activity level and enrichment use were also of interest in the current study because they are important indicators of wellbeing. Increased physical activity has the potential to improve physical health, whereas increased use of environmental enrichment provides NHP with opportunities to exhibit behaviors that are comparable to those of their wild conspecifics, such as foraging for food. Increased time spent exhibiting species-appropriate behaviors is interpreted as one type of evidence of improved wellbeing.^16,17

Numerous studies using singly housed macaques have involved manipulations to cage size without adding enrichment to evaluate changes in behavior. Research on multiple cage size conditions showed no effect on abnormal behavior^10,11,24 or activity level²⁴ and a decrease in locomotion in the smallest cage sizes.^10,11 In one study in which subjects were moved from their home cage to the next largest cage size appropriate for their body size, hair-plucking actually increased, whereas there was no effect on other types of abnormal behavior.⁴ Additional studies have shown a decrease in stereotypic locomotion as cage size was increased,^12,32 with this behavior being completely eliminated in the largest test cage condition.¹² However, other studies involving a substantial increase in enclosure size showed no effect on abnormal behavior.^18,19 One of these studies also showed a decrease in anxiety and, surprisingly, a decrease in locomotion with movement into a larger space.¹⁹ Overall, the literature suggests that small increases in cage size without adding enrichment has little to no benefit.

Alternatively, studies involving increases in both cage size and enrichment for singly housed NHP showed positive behavioral effects. This type of caging has been referred to as play, exercise, or recreation caging. Studies on the intermittent use of play caging have shown decreased abnormal behavior and increased locomotion and enrichment use in cynomolgus macaques,⁸ Japanese macaques,⁴⁰ baboons,^21,22 and vervet monkeys.³⁶ These studies observed subjects inside the play cage for periods ranging from 1 h to 2 d. The present study sought to expand on past research by using rhesus macaques, one of the most prevalent NHP in laboratories,³ and by observing subjects in the play cage for a period of 2 consecutive weeks. Collection of behavioral data over this extended time period provided the opportunity to assess trends in those behaviors of interest.

The present study sought to explore whether allowing a singly housed NHP intermittent access to a much larger and more environmentally stimulating cage led to a decrease in abnormal and anxiety behavior, an increase in species-appropriate behavior, and an increase in overall activity. In addition, analyses were conducted on the uses of space and enrichment devices to gauge how subjects used the increased space and enrichment in the play cage and to determine whether preferences were apparent.

Materials and Methods

Subjects were 10 male Indian-origin rhesus macaques housed at the Yerkes National Primate Research Center (Atlanta, GA). Subjects ranged from 4 to 8 y of age and weighed between 6.7 and 14.5 kg at the start of the study. All were mother-reared in social groups and had been singly housed an average of 3 y prior to involvement in the study. Subjects were identified through a combination of routine behavioral observations and live focal animal observations conducted by Behavioral Management staff. Subjects were selected if they exhibited either self-biting without wounding or SIB precursor behaviors (for example, bizarre posture, floating limb, eye poking, and digit sucking) on a minimum of 3 occasions during the 6-mo period preceding their involvement in the study. All 10 of the subjects exhibited at least one form of SIB precursor behavior, and 8 of the 10 macaques exhibited self-biting prior to their study involvement.

Study approval.

The Behavioral Management Unit oversees the psychologic wellbeing of the NHP colony and regularly conducts noninvasive behavioral research in an ongoing effort to provide the highest quality of care. As such, treatment interventions are evaluated for their effectiveness. The IACUC of Emory University determined that this study fell within the scope of the current approval for behavioral management research, and a separate IACUC approval was not required.

Housing.

The subjects occupied a home cage appropriate for their size and weight in accordance with federal animal welfare regulations; 6 occupied 4.3-ft² cages, 3 occupied 6.0-ft² cages, and 1 occupied a 9.0-ft² cage. The play cage consisted of 4 interconnected 4.3-ft² cages (total of 51.2 ft³) with moveable walls and bottoms (Figure 1). Both interior walls and one floor pan were removed, allowing subjects access to the entire space throughout the treatment phase. The left upper-tier floor pan was left in place to facilitate temporary confinement into one quadrant (the lower left) for clinical or research procedures and to provide additional floor space in the upper (left) tier of the play cage for the animal to use, if desired. The upper and lower quadrants of the right side of the play cage were vertically open.

Figure 1.

Play cage, with enrichment provided to the subjects.

For the treatment phase, the play cage was located in the subjects’ animal room. It was placed in the same general location as the subjects’ home cage, to minimize the effect on social dynamics within the room. Not all subjects occupied the same room.

Enrichment.

The home cage was furnished with stainless-steel perch bars, a chew toy, and at least one enrichment device attached to the exterior, to provide opportunities for foraging and object manipulation. Enrichment was kept constant throughout the baseline and posttreatment phases.

Increased enrichment was provided during the treatment phase to accommodate the larger space and emphasize the recreational aspect of the play caging. The play cage was furnished with 3 chew toys and multiple perches, in that each quadrant contained a double row of stainless-steel perch bars. The same types of enrichment devices that were attached to the home cage were attached to the exterior of the play cage. Items included puzzle balls, forage boards, a mirror, a paint roller, and several other manipulable hanging objects. Because these devices are commonly rotated throughout all standard caging at the facility, they were considered familiar to the subjects.

In addition, several types of novel enrichment were attached to the interior of the play cage; items included a 6-ft manzanita wood log, a swinging cable perch, a taut fire hose perch, several trapeze rings, and a water tub. The facility's enrichment review committee approved all interior enrichment items on the bases of safety, durability, and effectiveness of sanitization prior to the study.

Each type of enrichment (foraging, manipulable, and structural) was represented in all quadrants. The location, types, and quantity of enrichment were kept constant throughout the treatment phase for each subject as well as across subjects. The water tub was the only enrichment item that did not stay in the play cage at all times. As a result, the tub was excluded from subsequent analyses. The tub was a 3-gal heavy-duty molded rubber bucket, which was filled with water and affixed to the inside of the upper left quadrant via an eye bolt and a padlock. Due to concerns about sanitation and destruction of materials, the tub was placed inside the play cage for a maximum of 3 h at a time on 5 separate occasions during the treatment phase.

Procedures and data collection.

No feeding restrictions were in effect during the study, and therefore, subjects received primate chow, a variety of produce, and foraging items in accordance with the institution's standard operating procedures. During the treatment phase, animal care technicians were instructed to scatter food among the different quadrants to encourage the subjects’ movement within the play cage and to prevent eliciting any preference for location.

Data were collected during ten 30-min focal animal behavioral observations per subject per phase, for a total of 150 h. Behavioral Management staff regularly conduct live behavioral observations of all monkeys in the colony and, as such, it was expected that human presence would have minimal effect on the subjects’ behavior. Data collection occurred 7 d a week during normal operating hours. All data were collected live, outside of research or veterinary access times, feeding and enrichment distribution times, and cage sanitization of neighboring rooms. Every effort was made to collect data in consistent, quiet conditions within each room and across all phases and subjects.

Each subject was enrolled in the study for 7 to 8 wk, with the baseline and treatment phases lasting 2 wk each and the posttreatment phase lasting 3 to 4 wk. Transfer of subjects between the home cage and play cage coincided with cage sanitization procedures, which occurred on a biweekly basis.

Measures and instrumentation.

A single observer used Observer 5.0 software (Noldus, Leesburg, VA) on a handheld personal computer (iPAQ hx2755, Hewlett-Packard, Palo Alto, CA) to collect data via continuous sampling, allowing for the recording of duration of all behaviors. Data were coded by using a comprehensive ethogram of operationally defined behaviors currently in use by the Behavioral Management Unit. The ethogram included both normal and abnormal behaviors, including the manipulation of objects (the object manipulated was recorded and classified as an enrichment item or other item), social behaviors (including aggressive, nonaggressive, and play) with and without physical contact, solitary play, self-grooming, anxiety (including yawning, scratching, body shaking, and teeth grinding), fear, abnormal behaviors (including hair plucking, self-biting, other self-injurious behaviors, floating limb, bizarre posture, coprophagy or urophagy, and stereotypic behaviors), locomotion, eating and drinking, vocalizing, and inactivity.

During the treatment phase, space-use data were collected simultaneously with behavioral data via point sampling with a 30-s interval. This practice provided a measure of where the subject spent proportions of time.

Analyses.

Statistical analyses were conducted by using SPSS 17.0 (SPSS, Chicago, IL). The target behaviors (abnormal behaviors, anxiety behaviors, locomotion, and enrichment use) were compared across treatment phases by using repeated-measures ANOVA, with significance defined as a P level of less than 0.05. The measure for each behavior was the average percentage of time that a subject devoted to that behavior across observations. Paired t tests with Bonferroni correction (α = 0.05/2 = 0.025) were used to compare behavior between the baseline phase and the treatment or posttreatment phases for those dependent variables that yielded significant ANOVA findings. Although we had a relatively small sample size, visual inspection of quantile–quantile plots suggested that our data were generally normally distributed. Mauchly tests of sphericity were nonsignificant (P > 0.05) for all dependent variables. Behavioral and activity durations were plotted for trend analyses. Space use was evaluated through χ² analysis, with significance defined as a P value of less than 0.05.

Trending analysis was conducted by using a power curve fit for all collected data and was broken down by phase. The purpose of the trending analysis was to identify optimal timeframes for the treatment phase within the prescribed study time period.

Results

Species-inappropriate behaviors.

A significant difference in the duration (as measured by percentage of observed time) of abnormal behavior was shown across all phases of the study (F = 4.54, P = 0.025; Figure 2). Comparing the baseline (mean ± 1 SD, 21.5% ± 9.6%) and the treatment (14.3% ± 9.37%) phases indicated a trend (t = 2.44, P = 0.038) toward lower duration of abnormal behavior during the treatment phase when corrected for potential family-wise error by using a P < 0.025. The duration of abnormal behavior returned to baseline (t = 0.922, P = 0.381) during the posttreatment phase (19.5% ± 11.2%). Trending analysis (Figure 3) suggests that this reduction in abnormal behavior dissipated after approximately 1 wk in the play cage.

Figure 2.

Species-appropriate and -inappropriate behaviors across study phases.

Figure 3.

Trending analysis of abnormal behavior.

The abnormal behavior category was divided into individual abnormal behaviors (Table 1). This breakdown was limited to those behaviors exhibited by a minimum of 7 subjects in the baseline phase, because not all abnormal behaviors were observed. Of particular note, subjects that had a history of self-biting without wounding (8 of 10) self-bit during all phases of the study, with the exception of one animal that did not self-bite during the treatment phase. In addition, there appeared to be a decrease in stereotypic locomotion during the treatment phase as compared with baseline. Although there appeared to be differences between the phases, the sample size was insufficient to support exhaustive statistical analysis by individual behavior.

Table 1.

Summary statistics (mean ± 1 SD) for percentage duration of individual abnormal behaviors according to phase

	Baseline	During treatment	After treatment
Stereotypic locomotion (n = 9)	4.3 ± 9.3	0.4 ± 1.1	4.0 ± 9.0
Self-bite (n = 8)	0.9 ± 1.4	0.5 ± 1.0	0.5 ± 0.7
Other self-directed (n = 10)	11.4 ± 17.3	9.2 ± 16.7	12.3 ± 19.2
Hair pluck (n = 9)	1.6 ± 4.6	2.4 ± 8.6	0.7 ± 3.7
Floating limb (n = 10)	3.2 ± 9.3	2.1 ± 7.9	2.3 ± 6.5
Bizarre posture (n = 7)	1.0 ± 2.3	0.1 ± 0.3	0.3 ± 1.5

The duration of anxiety behavior was shown to vary significantly across phases (F = 6.49, P = 0.008; Figure 2). However, the comparison of the baseline (3.0% ± 1.8%) and treatment (2.3% ± 1.2%) phases did not indicate a significant difference (t = 1.51, P = 0.165). Duration of anxiety was shorter (t = 2.92, P = 0.017) during the posttreatment phase (1.5% ± 0.6%) when compared with the baseline phase (3.0% ± 1.8%). Trending analysis suggests a steady yet gradual decline across the treatment and posttreatment phases (Figure 4).

Figure 4.

Trending analysis of anxiety behavior.

Species-appropriate behaviors.

The duration of locomotive behavior across all 3 phases was significantly different (F = 11.94, P = 0.001; Figure 2). Comparison of the baseline (3.0% ± 2.9%) and treatment (4.9% ± 2.5%) phases indicated a much higher (t = −3.34, P = 0.009) duration of locomotion during the treatment period. The duration of locomotion was not significantly different (t = 1.76, P = 0.112) in the posttreatment phase (2.0% ± 1.6%) when compared with the baseline phase. Trending analysis (Figure 5) suggests this increase peaked within the subjects’ first few days or week in the play cage and rapidly declined upon their return to the home cage.

Figure 5.

Trending analysis of locomotion.

The duration of enrichment use (active manipulation of enrichment items) was significantly different across all phases (F = 19.78, P < 0.001; Figure 2). The secondary analysis of the baseline (5.5% ± 5.6%) and treatment (17.7% ± 7.8%) phases showed an increase (t = −4.81, P = 0.001) in enrichment use during the treatment phase. The duration of enrichment use was not significantly different (t = 1.00, P = 0.344) in the posttreatment phase (3.5% ± 1.6%) when compared with the baseline phase (5.5% ± 5.6%). Trending analysis shows that manipulation of enrichment items peaked during the first week of the treatment phase and continued at a reduced rate throughout the second week (Figure 6).

Figure 6.

Trending analysis of enrichment use.

Enrichment device use was compared to determine whether the macaques showed preferences for particular types of enrichment. Use of the enrichment devices varied significantly (χ²[df = 11] = 3.97, P < 0.001), although all types of enrichment were used by the subjects. In general, the mirror and forage boards had the highest percentage duration of use (Figure 7).

Figure 7.

Duration of enrichment item manipulation within the play cage.

Space use.

An analysis of space use during the treatment phase showed that time spent in each quadrant varied significantly (χ² (df = 3) = 57.68, P < 0.001). Subjects spent 83% of time in the upper quadrants of the play cage (upper left, 53.7% ± 33.4%; upper right, 28.3% ± 29.2%) and only 17% of time in the lower quadrants (lower left, 3.8% ± 8.0%; lower right, 13.1% ± 24.0%). The upper left quadrant had the highest percentage of occupancy (54%), and the lower left had the least (4%). Of note, the upper left quadrant had a floor, thereby creating a ceiling for the lower left quadrant, whereas the right quadrants were vertically open.

Discussion

Results of the current study indicated that the play cage had statistically significant and beneficial effects on locomotion and enrichment use, with a trend toward a reduction in abnormal behavior that did not achieve significance when Bonferroni correction was applied. In general, our rhesus macaques spent less time exhibiting abnormal behavior and more time exhibiting species-typical behaviors such as locomotion and manipulation of enrichment items when they were in the play cage as compared with their home cages.

Contrary to expectations, the play cage did not have a significant effect on anxiety behavior. Although there was a significant decrease between the baseline and posttreatment phases, this reduced level of anxiety during the posttreatment period cannot be explained by the treatment (the play cage). This reduction may be due to various environmental factors such as activities of animal care personnel or moves of other animals into or out of the room, or the subjects may have been showing a general adjustment to their situation, with an accompanying reduction in anxiety over time.

Power curves were generated to demonstrate general trending of behaviors within study phases. These analyses are not intended to be predictive (because confidence intervals are not included), but they do provide some insight into potential optimization of the treatment time period. This information is important in designing the optimal method for facilities to use limited resources (for example, space, available play cages). The data suggest that subjects reaped the maximum behavioral benefit within their first week in the play cage, indicating that a schedule of weekly rotation would allow the largest number of monkeys to profit from the experience. However, cagewash scheduling, general animal population stress due to cage rotation, and cost–benefit calculations should be considered when determining a play-cage rotation schedule. We found it more practical to coordinate play cage rotation at cagewash time (biweekly), which prevented extraneous stress on the animals due to caging moves without incurring additional cleaning time for animal care personnel.

Play cages were composed of a standard rack of 4 cages with removable inner walls and floor pans. Similar cage configurations vary in cost by manufacturer. Retrofitting existing or outdated caging to accommodate a similar configuration has been shown to be cost-effective.³⁷ The only additional cost incurred in the present study was the enrichment items used to furnish the play cage. Standard enrichment for a home cage at our facility consists of a single toy and a foraging device, totaling approximately US$45. The total cost of enrichment items for each play cage used in the current study amounted to approximately US$330.

Space-use data showed that the entire area of the play cage was used by each subject, although not in equal proportions. Subjects showed a strong preference for the upper quadrants of the play cage over the lower quadrants. This preference for the highest tier is consistent with the literature on a variety of species of NHP.^20,28,34,38 Subjects spent the most time in the upper left and the least time in the lower left quadrant of the play cage. This pattern may have been influenced by the configuration of the play cage; a floor pan separated the upper left quadrant from the lower left, providing additional floor space for the upper tier and facilitation of confinement in the lower tier, if needed, for research or clinical procedures. In addition, the placement of enrichment items was kept constant. Therefore, we cannot rule out whether any space-use preferences were linked to the type of enrichment that was offered in any particular quadrant.

Subjects used all enrichment items provided in the play cage, although with strong preferences for the mirror and forage boards. Enrichment preferences might have been tied to preferences for using certain portions of the play cage. For example, the high use of the mirror may have been due in part to its placement in one of the upper quadrants.

Subjects had no exposure to the play cage prior to being involved in this study. Therefore, an influence of novelty regarding both the caging itself, as well as the enrichment items inside it, cannot be ruled out. In addition, subjects were studied through only one rotation into the play cage. Future research could assess the same subjects rotating through the play cage multiple times to counter the effects of novelty and to allow for the determination of any long-term changes in behavior after repeated exposure. This information is important because the present study showed only a short-term increase in wellbeing associated with the play cage, and this improvement did not generalize to the posttreatment phase.

Future research might also examine in detail the location preferences within the play cage. If enrichment items are moved around among the 4 quadrants, space-use preferences can be determined without the possible effect of enrichment item preferences. Alternatively, play-cage use might include an additional phase during which subjects have access only to identical types and quantities of enrichment as provided in the home cage.

In conclusion, our results suggest that the provision of play cages for singly housed, laboratory rhesus macaques exhibiting SIB or various precursor behaviors is beneficial to their psychologic wellbeing as measured by behavior. Although studies that assessed an increase in cage size only have shown little or no behavioral benefit to NHP,^{4,10,11,18,19,24} the present study provides evidence that a combination of increased space with increased enrichment increases activity and enrichment use and may reduce abnormal behavior. The data provide justification for primate research facilities to invest in larger and more enriched caging as an effective intervention for singly housed animals that exhibit species-inappropriate behaviors, and our findings support the use of play caging as an important part of behavioral management programs.