Nonsocially Housed Rats (Ratus norvegicus) Seek Social Interactions and Social Novelty More Than Socially Housed Counterparts

Abstract

Sociability is the act or quality of social interaction and can be quantified by determining the number and duration of interactions with conspecifics. The purpose of this study was to examine the extent to which sustained social contact, as achieved by constant social living conditions, influenced social behavior. Beginning in juvenility, 19 male Long-Evans rats were housed in enriched environments, with half living socially in a large group and half living individually. After several months in these housing conditions, rats were tested on a sociality test and a social novelty preference test. Nonsocially housed rats exhibited more social behavior than socially housed rats. In the sociality test, nonsocially housed rats engaged with an unfamiliar rat more than socially housed rats. Similarly, in the social novelty test, nonsocially housed rats visited a novel stranger more than the now-familiar rat (from the sociality test) as compared with the socially housed rats. It is unlikely that general anxiety factors can account for between-groups social effects, as there were no group differences in behavior on the elevated zero maze and open field test. Furthermore, socially and nonsocially housed rats were matched in spontaneous object exploration and novelty preference in a novel object recognition test, eliminating the possibility that general exploratory behavior or novelty preference accounted for group differences in the sociability tasks. These results suggest that lack of social interaction in nonsocially housed rats may be more powerful for social motivation than the consistent opportunity for social contact afforded by social living conditions.

Among the many subfields of comparative psychology, social cognition is one of the most burgeoning and controversial. This is due, in part, to the cognitive revolution, the growing interest in the evolution of human sociality, and the emergence of social neuroscience (Shettleworth, 2010). Sociability, one aspect of social cognition, is defined as a propensity for social interaction. A reasonable approach for studying sociability in the laboratory is to compare behaviors of animals reared in social housing conditions with behaviors of those reared in individual housing conditions. However, a number of important considerations have been overlooked in previous studies.

Rats are among the most widely used animal models in research, and like humans, are highly social and live in groups (Grant & Mackintosh, 1963; Wang, Kessels, & Hu, 2014). Surprisingly, little is known about rats’ basic levels of sociability. For many studies, social or nonsocial housing is simply a living condition used for all subjects rather than a variable of interest. Because social housing is not typically considered an independent variable, there is no consensus on whether social housing of laboratory rats is optimal when the experimental conditions permit. In some studies, however, housing condition is used to provide a primary construct of interest, for example, isolation or social stress in the study of anxiety (Andrade & Guimarães, 2003; Lukkes, Mokin, Scholl, & Forster, 2009; Ruis et al., 1999; Westenbroek et al., 2003). Sociability tasks have even been used as direct measures of anxiety and fear behavior (Lukkes et al., 2009; Ramos, Berton, Mormède, & Chaouloff, 1997).

Social housing is increasingly recognized as an important form of enrichment (Elliott & Grunberg, 2005; Patterson-Kane, Hunt, & Harper, 2002; Schapiro, 2002). In fact, social housing as a form of enrichment is used to study neurogenesis (Westenbroek, Den Boer, Veenhuis, & Ter Horst, 2004) as well as learning and plasticity (Hullinger, O’Riordan, & Burger, 2015). Studies that use social housing as a form of enrichment, however, do not reliably control for the effects of nonsocial enrichment. This is important because rats exposed to nonsocial enrichment show enhancements in learning and memory (Simpson & Kelly, 2011), exploratory behavior (Hellemans, Benge, & Olmstead, 2004; Jenks et al., 2013), and sociality (Jenks et al., 2013; Westenbroek et al., 2003), as compared with rats living in nonenriched conditions. The failure to dissociate social and nonsocial enrichment is a limitation for translational fields that address developmental and neuropsychiatric disorders, such as autism and schizophrenia, in which both social behavior and enrichment may be relevant for diagnosis or treatment (Crawley, 2004; Flagstad et al., 2004; Niesink & Van Ree, 1989; Pletnikov, Rubin, Vasudevan, Moran, & Carbone, 1999; Sun & Cade, 1999).

Another shortcoming of many translational studies using rodent models of sociability is that previous social experience of the rodent subjects is not considered. Individual differences in propensities for social interaction in nonhuman animals may result both innately and from the nature of social experiences in early life and throughout the life span (Argyle, 2013; Capitanio, 2002; Gosling & John, 1999; Niesink & Van Ree, 1982). Thus, to understand the roots of sociability, proximate factors that affect social interaction must be measured across an animal’s lifetime. Such factors include the physiological and behavioral mechanisms of interaction, ontogeny, and social opportunity (Beach & Jaynes, 1954; Hemelrijk & De kogel, 1989; Hofmann et al., 2014). In addition, it is unclear whether another factor, motivation for social behavior, is influenced by previous consistent social experience.

Regarding motivation, social behavior can be divided into appetitive and consummatory behaviors. Appetitive behavior is defined as motivation for social interaction, and consummatory behavior is defined as social interaction itself (Diergaarde, Spruijt, Wolterink-Donselaar, Gerrits, & Van Ree, 2005). In this context, appetitive behavior may be quantified by measuring rats’ approach response to a conspecific (frequency of visits), whereas consummatory behavior can be measured by investigatory or contact behavior (sniff duration and frequency). One advantage of these classifications is that they might capture important aspects of social interaction that could otherwise be overlooked. For example, approach behavior may not always correlate with sniff behavior, as shown in Diergaarde et al. (2005). Another open question is whether appetitive and consummatory social behaviors are differentially affected by social enrichment. Thus, it is important to include these classifications in research on the effects of housing and other forms of enrichment on sociability in rats.

We addressed these open questions and limitations in three experiments in which subjects were either individually housed or group housed. Both housing conditions incorporated nonsocial enrichment. In Experiment 1, using the sociality and social novelty preference paradigm (Crawley, 2004; Jenks et al., 2013; Moy et al., 2004), sometimes referred to as the sociability paradigm or Crawley’s paradigm (Kaidanovich-Beilin, Lipina, Vukobradovic, Roder, & Woodgett, 2011), we measured the total number and duration of subject interactions and sniff bouts with an unknown rat versus an empty pen in Phase 1 (Sociality) and a familiar rat versus a new unknown rat in Phase 2 (Social Novelty). Previous studies on the effect of nonsocial housing on sociability in rats show conflicting results. In some cases, nonsocial housing has been shown to increase social interaction (Latane, Nesbitt, Eck-man, & Rodin, 1972; Niesink & Van Ree, 1982), whereas in others, it decreased interaction (Lukkes et al., 2009). Because of the lack of consensus in this understudied question, we present a two-tailed hypothesis: If isolated housing increases sociability (or social housing decreases sociability), individually housed rats should show higher rates of interaction in the presence of a conspecific than socially housed rats. Alternatively, if individual housing decreases sociability (or social housing increases sociability), individually housed rats should show lower rates of interaction than socially housed rats.

In Experiment 2, to dissociate sociability, which may include social anxiety, and nonsocial anxiety, rats were tested on two highly validated behavioral measures of general anxiety. In the open field test (OFT), subjects explore an open arena where locomotive behavior and anxiety behavior can be quantified by measuring the amount of time subjects spend near the perimeter’s walls as compared with the center of the arena (Díaz-Morán et al., 2014; Gould, Dao, & Kovacsics, 2009; Gregus, Wintink, Davis, & Kalynchuk, 2005; Ramos et al., 1997). More anxious animals are more likely to remain close to more secured walls of the arena, called thigmotaxis, where less anxious animals are more likely to venture to the unsecured center of the arena. In the elevated zero maze (EZM), subjects travel between open and walled sections of a circular maze (Braun & Clarke, 2006; Pellow, Chopin, File, & Briley, 1985; Rodgers, Cao, Dalvi, & Holmes, 1997; Shepherd, Grewal, Fletcher, Bill, & Dourish, 1994). Similar to the OFT, a relatively high propensity to refrain from exploring open areas and remain near the walled areas is considered indicative of anxious behavior. In Experiment 3, we used the novel object recognition (NOR) test to examine whether differences in sociality and social novelty could be explained by general (nonsocial) differences in exploration, memory, and novelty preference (Anderson, 2006; Fantz, 1964).

Experiment 1: Sociality and Social Novelty

Subjects were housed in enriched housing conditions either socially or individually for an extended period (14 months) at the time of Experiment 1. This allowed us to examine the effect of social living on sociability divorced from general environmental enrichment in a seminaturalistic setting.

Experiment 1 was divided into two separate phases to measure sociality, or propensity to interact with an unfamiliar rat, and preference for novel social interaction. In Phase 1, subjects traveled freely between chambers that housed either an unfamiliar rat or an empty pen. In Phase 2, subjects again traveled between the same two chambers but were presented with a now familiar rat (from Phase 1) and a novel, unfamiliar rat. If housing conditions affect subjects’ preference for sociality and social novelty, we expect socially and nonsocially housed rats to interact with unfamiliar and familiar rats at different rates.

Method

Subjects and housing.

Subjects were 19 male Long-Evans rats (Rattus norvegicus) that arrived from Charles River Laboratories on postnatal day (PND) 21. The supplier reported with reasonably high certainty that all animals came from separate litters. Upon arrival, 20 rats were randomly assigned to one of two groups: socially housed (SH) or nonsocially housed (NSH). One SH rat died of natural causes (PND 264), leaving nine rats in the SH group. Initially rats were housed in Plexiglas cages: NSH rats individually in standard shoebox cages (43.2 × 25.4 × 821.6 cm) and SH rats in larger cages (31.2 × 36.6 × 20.3 cm) in two groups of five, until subjects grew large enough to be placed permanently in enriched housing without escape (~PND 35). Permanent housing for both groups consisted of large, metal cages with wire meshing and multiple platforms (Ferret Nation, Amazon) providing opportunity for exercise and enrichment via running wheels, plastic igloos, and wooden chew toys. SH rats were housed together in one large cage (91.4 × 61 × 160 cm) where they had continual access to cage mates, and NSH rats were housed in large individual cages (43.2 × 32.4 × 61 cm). All subjects were housed in the same colony room, allowing NSH rats to smell, hear, and see other rats, but never physically interact. The colony room was maintained at 21.6°C and on a 12:12 reversed light: dark cycle, with light offset at 8 a.m. and light onset at 8 p.m. Subjects had access to water ad libitum, and rat chow (LabDiet, St. Louis, MO) was provided ad libitum until rats reached adulthood (~PND 60), after which point, weighed rations of chow were provided daily. Chow rations were given only after testing for each day was complete. Subjects were weighed weekly and chow rations were adjusted accordingly to maintain body weight at approximately 90% of their estimated free-feeding weight. In SH rats, if individual rat weights fell beneath standard body weight, they were temporarily removed from their group feeding regimen and fed their ration separately (30–60 min in isolation) until weight was restored. No NSH rats ever fell beneath standard body weight, but two SH rats did and were fed separately until their weights matched the average body weight of all other rats.

At the time of testing described in this article, subjects were not experimentally naive and had previously participated in unrelated behavioral experiments, including various memory and spatial navigation tasks. NSH rats never had physical access to other rats, except during the sociability task described here. All procedures described in the current study were approved by the Institutional Animal Care and Use Committee of Providence College.

Apparatus.

All training and testing sessions occurred inside a three-chamber, social interaction box (Noldus Technologies, Sturbridge, MA; 76.2 × 114.3 cm). Chambers (25.4 × 114.3 cm each) were separated by two plastic opaque doors. In the left and right chambers were two cylindrical holding pens (diameter = 21.6 cm, height = 38 cm) with vertical plastic bars where nonsubject, demonstrator rats (Stranger 1 and Stranger 2) could be placed. A circular weight was placed on top of each holding pen to prevent demonstrator rats from escaping.

Two unfamiliar male Long-Evans rats (PND 220–222) were used in Experiment 1 and are referred to as demonstrator rats. Demonstrator rats were pair-housed and had neither been involved in any previous experiments nor had previous physical contact with the test subjects or any other animals. For the purpose of this study, these nonsubject, demonstrator rats are identified as Stranger 1 and Stranger 2.

Habituation.

Rats were placed in the center chamber and were allowed to freely explore for 5 min. Doors to left and right chambers remained closed. No demonstrator rats were present during habituation.

Test procedure.

Subjects were tested across 3 days (PND 391–393), with testing cohorts consisting of five SH and five NSH rats each day. Each rat completed Phase 1 and Phase 2 of testing in 1 day, where each phase comprised one trial. Testing was conducted under normal light conditions, despite taking place during the dark portion of the rat’s light cycle, to facilitate videotaping and video-tracking. EthoVision XT (Noldus Technologies, Sturbridge, MA) was used for data collection and quantification. Before each trial, demonstrator rats were placed in the barred holding pens that permitted visual and olfactory contact but allowed minimal physical contact. Bars were approximately 2 cm apart, making reaching paws or noses into or out of the holding pen the maximum physical contact possible. After each trial, the apparatus and holding pens were wiped down with 70% isopropyl alcohol to control for odor cues.

In Phase 1, the cylindrical holding pens were placed at the end of the left and right chambers. The left chamber contained a rat unfamiliar to the test subject (Stranger 1) and the right remained empty (Figure 1). Subjects were placed in the center chamber and doors to the left and right chambers were removed to allow exploration of the apparatus and interaction with Stranger 1 for 10 min. At the end of each trial, subjects were placed back in the center chamber and doors to the left and right were closed. Once subjects were isolated in the center chamber, the experimenter placed Stranger 2 in the empty pen, while Stranger 1 remained in position, in preparation for Phase 2 (Figure 1). Doors were removed and again subjects were free to explore for 10 min during Phase 2.

Figure 1.

Experimental setup of Experiment 1. Subject is displayed in the center chamber of the testing arena. (A) Phase 1: Sociality. The holding pen in the left chamber contains Stranger 1 and the holding pen in the right chamber is empty. (B) Phase 2: Social novelty. The left chamber contains the holding pen with Stranger 1 (from Phase 1) and the right chamber contains the holding pen with Stranger 2, a novel conspecific.

Data analysis.

Social behavior was measured in two ways: (a) EthoVision XT automatically generated cumulative durations and total number of visits to relevant half of the left and right chambers where the holding pens were located (10 × 22.5 in.; designated “zones”), and (b) experimenters blind to the experimental conditions coded durations of sniffing and total number of sniff bouts from videos of each phase. Sniffs were defined as any instance when a subject directed its nose less than 2.5 cm from the holding pen. Sitting or standing next to the pen was not considered a sniff. Heat maps were also generated by EthoVision XT, with color representing the average relative time spent by the rats at all locations of the arena, with the maximum and minimum defined across SH and NSH rats but separately within Phase 1 and within Phase 2 (Figure 2, representative rats). This allowed visual comparison of relative durations spent in different zones across SH and NSH rats within a testing phase (e.g., long durations are shown in warmer red colors). Outlier analyses were conducted on all measures before reporting where any value above or below 2 SD from the mean were replaced with a score one unit less or more extreme than the closest data point in the distribution of the measure (Tabachnick & Fidell, 1983; Wiig & Burwell, 1998). Effect sizes for the analysis of variance (ANOVA) and t tests are reported as partial eta-squared and Cohen’s d, respectively. Independent samples t tests were two-tailed, and a p < .05 was considered statistically significant.

Figure 2.

Representative spatial heat maps reflecting total time spent within a testing session in all locations of the arena in Experiment 1. Nonsocially housed rat in (A) Phase 1 and (B) Phase 2. Socially housed rat in (C) Phase 1 and (D) Phase 2. See the online article for the color version of this figure.

Results and Discussion

Phase 1: Sociality.

A repeated-measures ANOVA on time spent in different zones revealed that there was a main effect of conspecific (stranger vs. empty pen), a main effect of group (SH vs. NSH), and an interaction, conspecific: F(1, 17) = 114.403, p < .001, ${\eta }_{\text{p}}^2$ = 0.871; group: F(1, 17) = 8.289, p = .010, ${\eta }_{\text{p}}^2$ = 0.328; Conspecific × Group: F(1, 17) = 20.327, p < .001, ${\eta }_{\text{p}}^2$= 0.545 (Figure 3A). This indicates that both SH and NSH rats spent more time with the demonstrator rat, Stranger 1, than the empty pen, but that NSH rats spent more time with the demonstrator rat than the SH rats. An analogous repeated-measures ANOVA was performed on total number of visits to different zones and revealed both groups also visited Stranger 1 significantly more than the empty pen, but no differences across housing group were found, conspecific: F(1, 17) = 67.369, p < .001, ${\eta }_{\text{p}}^2$ = 0.335; group: F(1, 17) = 0.625, p = .440, ${\eta }_{\text{p}}^2$ = 0.029; Conspecific × Group: F(1, 17) = 0.328, p = .574, ${\eta }_{\text{p}}^2$ = 0.003 (Figure 3C). Regarding sniff behavior, a repeated-measures ANOVA was performed on duration of sniffing and revealed a main effect of conspecific, main effect of group, and an interaction, conspecific: F(1, 17) = 135.523, p < .001, ${\eta }_{\text{p}}^2$ = 0.889; group: F(1, 17) = 10.106, p = .005, ${\eta }_{\text{p}}^2$ = 0.373; Conspecific × Group: F(1, 17) = 20.453, p < .001, ${\eta }_{\text{p}}^2$ = 0.546. An analogous repeated-measures ANOVA was performed on total number of sniff bouts and also revealed a main effect of conspecific, main effect of group, and an interaction, conspecific: F(1, 17) = 43.544, p < .001, ${\eta }_{\text{p}}^2$ = 0.719; group: F(1, 17) = 6.841, p = .018, ${\eta }_{\text{p}}^2$ = 0.287; Conspecific × Group: F(1, 17) = 6.533, p = .020, ${\eta }_{\text{p}}^2$= 0.278 (Figure 4). Follow-up t tests show that NSH rats sniffed Stranger 1 for a longer duration, NSH: M = 196.2, SD = 35.04, SH: M = 112.44, SD = 54.83, t(17) = 4.012, p < .001, d = 1.82 (Figure 4A), and more frequently than SH rats, NSH: M = 14.7, SD = 4.19, SH: M = 9.56, SD = 3.61, t(17) = 2.850, p = .011, d = 1.32 (Figure 4C). Duration of sniffing and total number of sniff bouts to the empty pen were also calculated, revealing no significant group differences, duration: NSH: M = 30.7, SD = 7.56, SH: M = 39.56, SD = 21.28, t(17) = –1.24, p = .223, d = 0.55; total sniff bouts: NSH: M = 6.9, SD = 1.37, SH: M = 6.11, SD = 2.37, t(17) = .901, p = .380, d = 0.41. These results indicate little variability between groups in nonsocial sniffing but a markedly greater motivation for social contact by NSH rats as compared with SH rats.

Figure 3.

Zone results for Experiment 1. Time spent with (A) Stranger 1 chamber versus empty chamber for Phase 1 and (B) Stranger 1 versus Stranger 2 for Phase 2. Total number of visits to (C) Stranger 1 chamber versus empty chamber in Phase 1 and (D) Stranger 1 chamber versus Stranger 2 chamber in Phase 2. Error bars depict 95% confidence intervals. * p < .05 in two-tailed paired samples t tests/no * = ns. See the online article for the color version of this figure.

Figure 4.

Sniff results for Experiment 1. Duration of sniffs with (A) Stranger 1 in Phase 1 and (B) Stranger 1 versus Stranger 2 in Phase 2. Total number of sniff bouts with (C) Stranger 1 for Phase 1 and (D) Stranger 1 versus Stranger 2 for Phase 2. Error bars depict 95% confidence intervals. * p < .05 in two-tailed paired samples t tests/no * = ns. See the online article for the color version of this figure.

Phase 2: Social novelty.

A repeated-measures ANOVA on zone duration revealed no significant main effects of conspecific novelty, group, or interaction between the two variables, conspecific novelty: F(1, 17) = 0.615, p = .444, ${\eta }_{\text{p}}^2$= 0.035; group: F(1, 17) = 1.077, p = .314, ${\eta }_{\text{p}}^2$= 0.060; Conspecific × Group: F(1, 17) = 0.287, p = .599, ${\eta }_{\text{p}}^2$= 0.017 (Figure 3B). Results from an analogous repeated-measures ANOVA on total zone visits revealed no main effects of conspecific novelty or group, but there was an interaction between conspecific novelty and group; con-specific novelty: F(1, 17) = 1.194, p = .290, ${\eta }_{\text{p}}^2$= 0.066; group: F(1, 17) = 2.028; p = .173, ${\eta }_{\text{p}}^2$ = 0.107; Conspecific Novelty × Group: F(1, 17) = 11.765, p = .003, ${\eta }_{\text{p}}^2$ = 0.409 (Figure 3D). A follow-up t test revealed that NSH rats visited Stranger 2 significantly more than SH rats, NSH: M = 31.2, SD = 11.03, SH: M = 19.11, SD = 8.37, t(17) = 2.665, p = .016, d = 1.23. All rats sniffed Stranger 2 significantly longer than Stranger 1, conspecific novelty: F(1, 17) = 5.899, p = .027, ${\eta }_{\text{p}}^2$= 0.254 (Figure 4B), and there was a trend for NSH rats to sniff Stranger 2 longer than SH rats, group: F(1, 17) = 3.458, p = .080, ${\eta }_{\text{p}}^2$= 0.158, but no interaction between variables was found, Group × Conspecific: F(1, 17) = 0.001, p = .970, ${\eta }_{\text{p}}^2$= 0.004. Analogous repeated-measures ANOVAs were performed on total sniff bouts, revealing that NSH rats sniffed the novel rat more than the familiar rat as compared with SH rats, conspecific novelty: F(1, 17) = 2.365, p = .147, ${\eta }_{\text{p}}^2$= 0.087; group: F(1, 17) = 7.994, p = .012, ${\eta }_{\text{p}}^2$ = 0.381; Conspecific Novelty X Group: F(1, 17) = 0.697, p = .415, ${\eta }_{\text{p}}^2$= 0.027 (Figure 4D). These results suggest that both SH and NSH rats were equally motivated to spend time investigating each of the two rats but that NSH rats visited the novel rat more than the SH rats.

Taken together, results demonstrate that when given the choice between an unfamiliar rat and an empty pen in Phase 1, all subjects, regardless of solitary or group living conditions, preferred to interact with an unknown conspecific rather than being alone. At the same time NSH rats spent more time interacting with the novel rat than the SH rats, as seen through zone duration and total sniff bouts and sniff duration. Interestingly, when given the choice between a familiar and novel rat in Phase 2, no overwhelming conspecific preference was shown for all subjects, yet NSH rats interacted with the novel rat, Stranger 2, more than SH rats did. Although results may be explained by a difference in general sociality and a social novelty preference, alternative explanations were addressed in Experiments 2 and 3.

Experiment 2: Elevated Zero Maze and Open Field Test

Experiment 2 was conducted to determine if results from Experiment 1 were due to potential group differences in general anxiety divorced from social anxiety, neophobic tendencies, or locomotor behavior, rather than sociability per se. In Experiment 2A, which occurred on PND 464, testing was conducted on an EZM (122 cm diameter, 51 cm height off ground) divided into four sections of walled (two) and open (two) sections. Rats were tested for one trial on the EZM where they could freely explore the open and walled sections of the maze for 5 min. Relatively anxious animals should spend more time in the closed areas of the maze, whereas less anxious animals should be more likely to venture into the unsecured open areas of the maze. In Experiment 2B, which occurred on PND 560, testing occurred inside a square arena, an open-topped wooden box (81 × 81 × 51 cm) sealed with black waterproof paint. Rats were individually tested for one trial of the OFT each where they were allowed to freely explore the arena for 5 min. To initiate a trial, subjects were placed along a wall of the apparatus and never forced into the center area by experimenters. Similar to the EZM predictions, relatively anxious rats should engage in more thigmotaxis, whereas less anxious animals should be more likely to venture to the center of the arena.

If NSH rats spend more time in the open areas than in the more secured, walled areas of the elevated maze or the arena than SH rats, and/or traveled greater distances, this would suggest that NSH rats engage in less anxiety-related behaviors than SH rats. Such a finding would reveal that putative differences in sociability found in Experiment 1 may have in fact been due to differences in anxiety or increased general exploratory or locomotor behavior by NSH rats, as NSH spent more time exploring the novel rat in Experiment 1. If, however, NSH rats engage in equivalent or more anxiety-related behaviors as compared with SH rats, this would indicate that general anxiety and locomotor behavior are not responsible for differences in sociability found Experiment 1. It should be noted that although, under some circumstances, it may be plausible that increased anxiety would increase sociality, here we assume that increased anxiety would most likely decrease sociality based on previous literature (Burman, Owen, Aboulsmail, & Mendl, 2008; Prichett-Corning, 2015; Pritchard, Van Kempen, & Zimmerberg, 2013; Tanas, Ostaszewski, & Iwan, 2015).

Method

Subjects and housing were the same as described in Experiment 1. Data from both tasks were collected and automatically coded via EthoVision XT (Noldus Technologies), and statistical analysis procedures were the same as described in Experiment 1 (e.g., independent samples t tests were two-tailed and p < .05 was considered statistically significant).

Results and Discussion

Overall, no group differences were found in Experiment 2A or 2B. In the EZM, NSH and SH rats spent an equal amount of time in the open areas of the apparatus, NSH: M = 103.36, SD = 25.91, SH: M = 99.18, SD = 32.51, t(17) = 0.312, p = .759, d = 0.14 (Figure 5A), and traveled equal distances, NSH: M = 2,241.85, SD = 529.30, SH: M = 2,057.85, SD = 665.07, t(17) = 0.671 p = .511, d = 0.31 (Figure 5C). In the OFT, NSH rats and SH rats spent an equal amount of time in the center arena, NSH: M = 19.32, SD = 15.48, SH: M = 23.96, SD = 16.53, t(17) = −0.631, p = .536, d = 0.29 (Figure 5B), and traveled a similar distance during the trial, NSH: M = 1,399.61, SD = 529.38, SH: M = 1,694.94, SD = 388.97, t(17) = −1.372, p = .188, d = 0.64 (Figure 5D).

Figure 5.

Anxiety results for Experiment 2. Time spent in an open area during the (A) elevated zero maze (EZM) and (B) open field test (OFT). Distance traveled during the (C) EZM and (D) OFT. Error bars depict 95% confidence intervals. p > .05 in two-tailed independent samples t-tests for each comparison = ns. See the online article for the color version of this figure.

Experiment 2 measured potential anxiety differences between housing groups through widely used measures of anxiety, the EZM and OFT (Braun & Clarke, 2006; Díaz-Morán et al., 2014; Gould et al., 2009; Gregus et al., 2005; Pellow et al., 1985; Ramos et al., 1997; Rodgers et al., 1997; Shepherd et al., 1994). No differences between groups in duration in the open areas away from the wall or distance traveled for were found in either task, indicating that rats’ housing condition had no behavioral effect on their anxiety-related behaviors or locomotor behavior. This finding supports the hypothesis that differences in sociality and social novelty between the rats in the two housing conditions are most likely the result of actual differences in sociability.

Experiment 3: Novel Object Recognition

In Experiment 2, we were able to eliminate the possibility that anxiety or locomotor behavior was a confounding variable that could have explained differences found in rats’ sociability in Experiment 1. Although the conclusion that social versus nonsocial housing conditions produced differing levels of sociability is thus strengthened, it remained possible that there were general differences in novelty preference and perhaps exploration. Mirroring the experimental design of Experiment 1, Experiment 3 was divided into two phases where general exploration and novelty preference were examined, respectively, in the NOR test (Antunes & Biala, 2012; Bevins & Besheer, 2006; Reger, Hovda, & Giza, 2009).

In Phase 1, rats explored two identical objects (A and A) until cumulative exploration time reached 20 s. As NSH rats demonstrated greater sociality with a novel rat in Phase 1 of Experiment 1, it is possible that this result is simply a reflection of greater general exploration or less neophobia—without a social component, in NSH rats as compared with SH rats. If this were the case, here we would expect to see that NSH rats would reach the 20-s criterion of exploration faster than SH rats. If differences in sociability cannot be explained by general exploration, we would expect to see equivalent levels of exploration across groups or perhaps an increase in time to criterion in the NSH rats as compared with the SH rats.

In light of preference for social novelty found by NSH rats in Phase 2 of Experiment 1, Phase 2 of Experiment 3 was analogously used with the goal of assessing potential group differences in novelty preference without a social component and 5-min recognition memory from Phase 1 to Phase 2. If, for any reason, NSH rats have a greater preference for novelty than SH rats, and this is the source of putative differences in social novelty, then NSH rats should explore a novel object (B) when presented with the previous explored object in Phase 1 (A) more than SH rats. If instead differences in social novelty found in Experiment 1 reflect true disparities in social behavior in SH versus NSH rats rather than novelty preference per se, then there should be no differences in exploration of the novel object in Phase 2. Finally, an additional alternative explanation to differences in social novelty is that SH rats might, for any reason that we are not privy to, not remember which rat they encountered 5 min ago in Phase 1 of Experiment 1 as well as NSH rats, causing them to explore the novel rat less than NSH rats. Similar to aforementioned predictions, if this were the case, we would expect to see NSH rats explore a novel object more than an already explored object from Phase 1 as compared with SH rats, mirroring the results with conspecifics in Experiment 1. However, if this is not the case, and recognition memory cannot explain differences in social novelty preference, there should be no difference in exploration of the novel object in SH versus NSH rats.

Method

Testing sessions, consisting of Phase 1 and Phase 2, occurred inside the OFT box described in Experiment 2. Rats received four testing sessions (PND 582, 584, 587, and 589), consisting of one session per day. Four objects were used per session in this experiment: Two identical objects A1 and A2 were presented in Phase 1; a third copy (A3) of the Phase 1 objects was presented alongside a novel object, B1, in Phase 2. In Phase 1, rats were placed along a wall of the box at an equal distance from either object that was secured to the floor with Velcro at a distance of 30 cm apart. Subjects were required to contact either or both objects (e.g., A1 or A2) with their nose for a cumulative 20 s before Phase 1 was terminated. If rats did not reach this criterion in 10 min, they were removed from the box and not included in data analysis; however, no rats exceeded this maximum allowed time.

Rats were then placed in their transport cage for a delay of 5 min before Phase 2 began. During the 5-min delay preceding Phase 2, objects from Phase 1 were replaced by one identical object (e.g., A3, which denotes identical but nonencountered object as explored in Phase 1) and one novel object (e.g., B1) in the same position as before. Commencing Phase 2, rats were again placed along the wall equidistant from either object and allowed to freely explore for 5 min. With each new session, rats received objects never experienced before, resulting in four sets of objects in total. Objects, approximately 8 × 8 × 12 cm, were constructed of LEGO Duplo blocks (Amazon). However, due to concern that Duplo structures were not adequately stimulating, the final two testing sessions consisted of more complex, yet still similarly sized, toys such as snow globes, animal figurines, hourglasses, and decorative jars. At the end of each trial, rats were removed and all materials were cleaned with 70% isopropyl alcohol.

Exploration was evaluated through time to criterion in Phase 1 and duration of novel object exploration in Phase 2. Time to criterion in Phase 1 was defined as the total time required for the subject to contact either or both objects with their nose for a cumulative 20 s. Duration of contact with the novel object was manually recorded by an experimenter with a stopwatch. Exploration of the familiar and novel objects (e.g., A3 and B1, respectively) in Phase 2 was quantified via manual video coding in parallel to conspecific exploration in Experiment 1: Duration of sniffing and total sniff bouts were defined as nose contact less than 2.5 cm from the object.

Results and Discussion

In Phase 1, no group differences were found in time to criterion, NSH: M = 149.51, SD = 40.63, SH: M = 147.56, SD = 47.18, t(17) = 0.097, p = .924, d = 0.044 (Figure 6A), suggesting that housing conditions had no effect on rats’ exploration in a nonsocial context. In Phase 2, no group differences were found in duration or total number of interactions with the novel object, duration: NSH: M = 46.93, SD = 11.47, SH: M = 39.53, SD = 11.37, t(17) = 1.41, p = .177, d = 0.65 (Figure 6B); total sniff bouts: NSH: M = 6.675, SD = 0.87, SH: M = 7.00, SD = 1.54, t(17) = −0.568, p = .578, d = 0.26, suggesting that housing conditions did not affect rats’ general novelty preference and that 5-min memory was equivalent across groups.

Figure 6.

Exploration and novelty preference results for Experiment 3. (A) Time to 20 s contact criterion in Phase 1. (B) Time spent exploring novel object in Phase 2. Error bars depict 95% confidence intervals. p > .05 in two-tailed independent samples t-tests for each comparison = ns. See the online article for the color version of this figure.

NOR was tested to compare novelty preference with social novelty preference in Experiment 1. Based on previous literature, it is standard to test sociability on one trial, as testing on more trials would require multiple demonstrator rats (Crawley, 2004; Jenks et al., 2013; Moy et al., 2004). However, it is also standard for NOR tests to include more than one trial, typically about four trials with different objects (Antunes & Biala, 2012; Bevins & Besheer, 2006; Reger et al., 2009). This discrepancy opens up the possibility of differences in habituation across groups in Experiment 1 versus Experiment 3. To address this potential confound, we analyzed individual NOR trials separately, and no group differences were found for either phase on Trial 1, 2, 3, or 4 (see online supplemental materials), eliminating the possibility that there was a difference in habituation across trials for NSH rats and SH rats. Taken together, these findings indicate enhanced sociality and social novelty preference by NSH rats found in Experiment 1 are not due to nonsocial sources.

General Discussion

Results presented here are consistent with the hypothesis that nonsocially housed male rats are more sociable and have higher social novelty preferences than their socially housed counterparts. These differences between rats housed socially or nonsocially are due to social experience and not environmental enrichment per se, nor can they be explained by levels of anxiety or general preference for novelty. Related studies examining sociality found that when given the option between a conspecific and inanimate object, normal rodents prefer the social opportunity (Burwell & Templer, 2017; Moy et al., 2004; Van Loo, Van de Weerd, Van Zutphen, & Baumans, 2004).

Our findings are consistent with those of two other studies reporting that social isolation in rats led to increased sociability when presented with a conspecific (Latane et al., 1972; Niesink & Van Ree, 1982). These studies differ from ours in several ways, however. First, these previous studies did not examine the effect of social housing when animals were in enriched social versus nonsocial conditions for long periods, which better approximates sustained social contact experienced in nature. Second, in previous studies of socially isolated and socially housed subjects, both subject groups were exposed to a conspecific from the socially housed subjects’ social group, leaving sociability confounded with attraction to social novelty in the nonsocially housed rats. Correspondingly, previous social interaction with the conspecific was a confound in the socially housed rats. Another study reported that individual housing decreased social interaction (Lukkes et al., 2009). In that study, individually housed rats were returned to social housing and socially housed rats were assigned to a new housing group after 3 weeks. Thus, sociability may have been confounded with social stress. The present study avoided these limitations in an effort to determine the extent to which social housing affects sociality and social novelty as separate constructs. Given the complexity of social interactions, especially during previous group housing social interactions (Latane et al., 1972; Niesink & Van Ree, 1982), our aim was to examine motivation for social interaction at its most basic level. We therefore did not examine social behavior of the demonstrator rat or allow the rats to fully interact with each other, as done in the study by Niesink and Van Ree (1982). Accordingly, rats were tested on sociality and social novelty preference, rats remained in the same housing conditions for the duration of the study, and both SH and NSH rats were presented with never-before-seen rats in sociality and social novelty tests.

In Phase 1 of Experiment 1, all rats, regardless of housing condition, preferred interacting with an unfamiliar conspecific rather than visiting an empty pen. No differences between groups in appetitive social behavior were found, as demonstrated by equal numbers of visits to the conspecific by NSH and SH rats. NSH rats did show higher levels of consummatory social behavior as compared with SH rats, as demonstrated by time spent in proximity to the unfamiliar rat and number and duration of sniff bouts. We are unable to speculate about why this dissociation in social behavior occurred, except that appetitive and consummatory social behavior represent separate approach versus contact aspects of motivated social behavior, respectively (Diergaarde et al., 2005; Ikemoto & Panksepp, 1996; Varlinskaya, Spear, & Spear, 1999). In addition, although differences in sociality between NSH and SH rats found here are strong, they may not be robust across all measures when considering sociability as one construct. Because all measures of consummatory social behavior were higher in NSH rats as compared with SH rats, we can conclude that when housing conditions were enriched, isolation increased motivation for social interaction. This is in line with only a limited number of existing findings (Ferdman, Murmu, Bock, Braun, & Leshem, 2007; Latane et al., 1972; Niesink & Van Ree, 1982). Those studies, however, did not dissociate long-term social enrichment and nonsocial enrichment, nor did they exclude differences in anxiety and general novelty preference.

In Phase 2 of Experiment 1, subjects were exposed to the now familiar rat, Stranger 1, and a novel rat, Stranger 2, to test if housing conditions influence social novelty preference. Unlike in Phase 1, NSH rats showed a greater tendency toward appetitive social behavior than SH rats, as seen in total number of visits to the novel conspecific. Interestingly, there were no differences between groups in time spent with the novel conspecific, but all other measures of consummatory social behavior—duration and total number of sniffs bouts— did suggest that NSH rats spent more time interacting with the novel rat than SH rats did. Taken together, these findings indicate that long-term social isolation increases rat’s affinity for novel social opportunity in both appetitive and consummatory aspects of social novelty preference.

As in Phase 1 of Experiment 1, three out of four measures of social behavior pointed to the conclusion that NSH rats were more sociable than SH rats. Variability between groups was highest in results for total zone visits and time spent in zones, but we have no reason to suspect a dissociation in approach versus contact behavior in sociality versus social novelty preference. Sniff behavior, a consummatory behavior, more consistently demonstrated increased sociability by NSH rats across measures in both Phase 1 and Phase 2 (including, however, only a trending group difference in sniff duration in Phase 2, Figure 4B). This may be because sniff bouts are more reliable measures of social interaction; some research groups that employ the social interaction paradigm use sniff bouts as the sole measure of social exploration (Molas et al., 2017).

In Experiment 2, we found no group differences in behavioral measures of anxiety, allowing us to eliminate anxiety as a possible alternative explanation for group differences in sociability found in Experiment 1. Yet, because anxiety and exploration levels are known to affect animals’ behavior on a number of unrelated behavioral measures (Clément, Calatayud, & Belzung, 2002; Rodgers et al., 1997), we further examined exploration in Experiment 3. To evaluate the possibility that purported differences in sociability in Experiment 1 were due to exploration (Phase 1), general novelty preference, or recognition memory (Phase 2), we used a NOR test in Experiment 3. The experimental design for Experiment 3 mirrored that of Experiment 1, where Phase 1 involved spontaneous (social or object) exploration and Phase 2 involved preference for novel (social or object) interaction with a 5-min delay between phases. Because no group differences emerged in either phase of Experiment 3, we can reject the alternative hypotheses that NSH rats were simply more exploratory, had better recognition memory, or had greater novelty preference than SH rats. Further, because measures in Experiment 3 comprised nose contact behavior, and no group differences were found, we can eliminate the possibility that sniff results from Experiment 1 were due to SH rats habituating faster to chamber rats’ odors, given their increased previous odor exposure in their home cage. Along the same vein, whereas SH rats may have been exposed to more odors in proximity, NSH rats were exposed to olfactory cues in the same colony room, as housing was not isolative. Indistinguishable behavioral performance by SH and NSH rats in Experiment 3 further reinforces the conclusion that enhanced sociality and social novelty preference in NSH rats is due to actual differences in sociability caused by social housing conditions.

Normal rodents have a preference for both social novelty and general object novelty (Burwell & Templer, 2017), as seen in Experiment 3. Both SH and NSH rats displayed social novelty preference to some extent, suggesting that both lack of and/or constant social interaction did not cause a significant aberration in normal social behavior. It may be impossible to determine with certainty if lack of social opportunity caused an increase in sociability or constant social interaction caused a decrease in sociability. Though predicted and obtained results are congruent across both hypothetical mechanisms, future studies should focus on determining which housing condition should be considered the experimental versus control group. Importantly, the consensus that rodents prefer social novelty may be weighted by studies in which rodents are socially isolated, even though this housing condition is not an examined independent variable. Social isolation may therefore increase rats’ sociality and social novelty preference, so that rats’ baseline levels of sociability are actually artificially enhanced by housing them nonsocially.

There are a few limitations to our study. First, a decoy rat or otherwise interesting object was not included in the sociality phase of Experiment 1. This may have led to a difference in visual and olfactory interest in the two chambers. However, two of the three studies we modeled our methodology after did not include a decoy rat (Crawley, 2004; Jenks et al., 2013; Moy et al., 2004). Second, we did not change the location of chambers in Phase 1 or Phase 2. Counterbalancing the demonstrator rat and novel rat in both phases, respectively, may have been ideal, but we were unable to do this because of difficulty it would create in automated video-tracking that ensured optimal objectivity in our design. If rats did have a side bias for any reason, we would have no reason to suspect that those biases would be unequally distributed between groups. Although both of these factors could be improved upon in future studies, we do not see a reason why either factor could lead to the group differences we observed.

Finally, to keep the demonstrator rats as docile and acclimated to the chamber as possible, the sociability apparatus was not cleaned between Phase 1 and Phase 2 of Experiment 1 but only between subjects. It is possible that subjects could have scent-marked the chamber holding Stranger 1 in Phase 1 and used their odor as a cue during Phase 2; however, we are unable to identify a reason why this could account for group differences shown in sociability, as isolated and socially housed male rats do not mark at different rates when presented with an unfamiliar rat’s odor or their own odor (Brown, 1985). Further, this potential confound could only affect novelty preference results in Phase 2, and not general sociality differences seen in Phase 1. Results of increased sociability by NSH rats in Phase 2 replicate those of Phase 1, decreasing the possibility that scent-marking could account for group differences observed. However, as social behavior is complex and obviously intimately related to sophisticated odor perception in rodents, it is entirely possible that odor cues did play a role in sociability, which would be interesting in its own right. Our results indicate an increase in sociability by NSH rats as compared with SH rats, but we have not attempted to determine the mechanism of the difference besides the housing conditions themselves. We look forward to future studies that attempt to determine what top-down and bottom-up processes guide social interactive behavior. Such work may expand upon recent rich literature that has identified the neural bases of social and nonsocial novelty preference that are driven by familiarity signals (Molas et al., 2017).

In light of our results and previous findings that nonsocial housing does not increase anxiety in rats if they are provided with object enrichment (Brenes, Rodríguez, & Fornaguera, 2008; Hellemans et al., 2004; Simpson & Kelly, 2011; Sparling, Mahoney, Baker, & Bielajew, 2010), we offer a hypothesis that sociability is affected by cage enrichment, which may moderate anxiety, even in novel contexts (Fernandez-Duque & Huntington, 2002; Hender-shott, Cronin, Langella, McGuinness, & Basu, 2016; Jenks et al., 2013; Simpson & Kelly, 2011). If rats are not given enrichment but are socially isolated, this might lead to increases in anxiety, which is seen in typical anxiety measures like behavior in the OFT and EZM (Lukkes et al., 2009; Sáenz, Villagra, & Fornaguera Trías, 2006; Westenbroek et al., 2003). Personality measures such as boldness and anxiety have indeed been shown to be moderated by home cage and natural environments (Rödel & Meyer, 2011; Žampachová, Kaftanová, Šimánková, Landová, & Frynta, 2017). These anxiety differences would presumably also manifest in NSH rats displaying less sociability than their SH counterparts. This may explain conflicting results in the literature with regard to the extent to which and direction in which social housing affects anxiety (Andrade & Guimarães, 2003; Arndt et al., 2009; Beery & Kaufer, 2015; Lukkes et al., 2009; Sáenz et al., 2006) and sociability (Beery & Kaufer, 2015; Crawley, 2004; Hol, Van den Berg, Van Ree, & Spruijt, 1999; Jenks et al., 2013; Moy et al., 2004). However, if anxiety levels are behaviorally equivalent across groups, which could be achieved by enriched housing conditions (Brenes et al., 2008), this could result in increased sociability by NSH rats as found here. Future studies should examine this hypothesis by housing rats socially and nonsocially in both enriched conditions and nonenriched conditions. To our knowledge, this question of how housing enrichment and reductions in anxiety might interact with social versus nonsocial housing has not been asked, but results presented here and results by Rödel and Meyer (2011) suggesting that litter size can both increase and decrease anxiety should directly inform such investigations.

Another possibility is that conflicting findings about the effects of social housing on anxiety are due to inconsistent definitions of nonsocial housing or social isolation. It would therefore also be informative to determine if more extreme social isolation, such as with no visual or olfactory contact, rather than simply no physical contact, as used here, would produce differences in social behavior. It is possible that more extreme housing conditions would increase anxiety, much like has been found when housing conditions are nonenriched, as explained earlier. In addition, it would be beneficial for future studies to determine if the same results would occur if female rats were tested and if differential housing conditions were maintained for shorter amounts of time than the long durations used here.

We conclude that when housing was enriched across groups, long-term nonsocial housing increased rats’ sociality and preference for novel social interactions. Our findings should inform general comparative psychology research in which social housing of rodents is often not considered a relevant variable, especially in tasks in which social behavior is measured. Results presented here should also have implications for the common use of the sociability task in modeling neuropsychiatric profiles, such as autism and schizophrenia, in which social novelty and familiarity signals are dysregulated (Burwell & Templer, 2017; Flagstad et al., 2004; Niesink & Van Ree, 1982; Pletnikov et al., 1999; Sun & Cade, 1999).