Online Issues

ILAR Journal Vol 46(2)

Environmental Enrichment for Laboratory Rodents

Eric Hutchinson, Anne Avery, and Sue VandeWoude

Eric Hutchinson, B.A., was employed as an Environmental Enrichment Technician with the Division of Veterinary Resources, OD/ORS/National Institutes of Health/Department of Health and Human Services, Bethesda, Maryland, at the time of writing this manuscript, and is currently a student in the College of Veterinary Medicine & Biomedical Sciences at Colorado State University (CSU), Fort Collins, Colorado. Anne Avery, V.M.D, Ph.D., is an Assistant Professor in the Department of Microbiology, Immunology, and Pathology (MIP) at CSU. Sue VandeWoude, D.V.M., is an Associate Professor of Comparative Medicine, MIP, CSU.

Abstract

Modernization of housing and husbandry techniques for rodents has minimized confounding variables. The result has been vastly improved health maintenance and reproducibility of research findings, advances that have decreased the numbers of animals needed to attain statistically significant results. Even though not all aspects of rodent manipulation have been strictly defined, as housing and handling procedures have become increasingly standardized, many animal care personnel have recognized the lack of complexity of the rodents' environment. Concern for this aspect of animal well-being has led many research facilities to provide "environmental enrichment" for rodents. Additionally, regulatory agencies in the United States and Europe have also been increasingly concerned about this issue relative to laboratory animal husbandry. However, little is known about the influence such husbandry modifications may have on biological parameters. In this article, laws and guidelines relating to rodent enrichment are reviewed, the natural behaviors of select rodent species are discussed, and an overview of widely used types of enrichment in laboratory rodent management is provided. The literature evaluating effects of rodent enrichment is reviewed both in terms of neurological development and as an experimental variable, and results of a study evaluating the effect of enrichment on immune and physiological parameters are reported. Survey data on current enrichment practices in a large multi-institutional organization are presented, and practical aspects requiring consideration when devising a rodent enrichment program are discussed.

Key Words: animal husbandry; behavior; environmental enrichment; immunology; rodents; well-being

Overview

Sequencing of the mouse genome, coupled with technologies that allow sophisticated manipulations for development of mutant strains, has contributed to an explosion in mouse populations in research facilities in recent years (Knight and Abbot 2002). Recent advances in genetic manipulation of rat embryos, completion of the R. norvegicus genome sequence, and the realization that mice do not necessarily provide adequate animal models for many human diseases has also led to a resurgence in the use of this rodent in biomedicine (Abbot 2004). During the last several decades, modernization of housing and husbandry techniques for rodents has primarily related to caging enhancements to ensure biosecurity for animals and personnel. These changes have been made to minimize confounding variables such as infectious diseases, exposure to toxins, or variations in environment. The result has been vastly improved health maintenance and reproducibility of research findings, all advances that have decreased the numbers of animals needed to attain statistically significant results. As physical manipulation of laboratory rodents has become highly standardized, many animal care personnel have recognized or perceived adverse consequences for laboratory rodents with the advent of this technology, in terms of the lack of complexity of the rodents' environment. Concern for this aspect of animal well-being has led many research facilities to provide "environmental enrichment" (EE¹) for rodents as a method to improve quality of life for these species.

Regulatory agencies in the United States and Europe have also been increasingly concerned about this issue relative to laboratory animal husbandry. However, little is known about the influence such husbandry modifications may have on biological parameters. Because introduction of enrichment schemes involves addition of experimental variables, therefore potentially decreasing the power of the experimental design, the implications of altered experimental outcome associated with addition of enrichment to a rodent husbandry scheme must be considered. In addition, the composition of comfort and "enrichment" for rodents may not always be intuitively obvious. Rodent species used in biomedical research range from captive wild animals to strains bred thousands of generations in a laboratory setting, sometimes with spontaneous or induced genetic alterations. Devising enrichment techniques that might be applied safely and easily across such diverse backgrounds is a daunting task. Finally, it is also necessary to factor the practical considerations of rodent enrichment (e.g., ease of use and safety) into the cost:benefit equation. Therefore, what first appears to be a simple and logical act quickly becomes much more problematic when the scientific and practical aspects of providing an enrichment program for rodents are considered.

In this article, we review laws and guidelines relating to rodent enrichment, and discuss the natural behaviors of select rodent species in an effort to define activities that would mimic innate patterns of behavior, thus potentially reducing stress in captive rodents. We provide an overview of widely used types of enrichment in laboratory rodent management and review literature evaluating effects of rodent enrichment, both in terms of neurological development and as an experimental variable. We report results of a study evaluating the effect of enrichment on immune and physiological parameters and review survey data on current enrichment practices in a large multi-institutional organization. We also review practical aspects that require consideration when devising a rodent enrichment program.

Laws and Regulations Governing EE

The Animal Welfare Act (AWA¹) currently contains no provisions for laboratory-reared rats and mice, nor does it explicitly mandate EE standards for other rodents. However, given the attention of other agencies to this topic, and the flurry of recent legal challenges to the AWA such as Alternative Research and Development Foundation v. Veneman (2001), laboratory rodents are likely to be addressed more specifically in this primary regulatory statute at some time in the future. In the Guide for the Care and Use of Laboratory Animals (the Guide¹) (NRC 1996), natural behaviors are highlighted as a crucial measure of success for an animal program. The discussion of proper housing begins with the statement, "Animals should be housed with the goal of maximizing species specific behaviors and minimizing stress induced behaviors" (p. 22). It is also suggested that readers consider housing the animals in "natural environments," and an entire section on behavioral management stresses is devoted to consideration of animals' structural, social, and activity needs. As a result of this suggestion, the Public Health Service and the Association for Assessment and Accreditation of Laboratory Animal Care International more closely emphasize/scrutinize the use of enrichment in laboratory animal facilities for all species, including rodents.

The Canadian Council on Animal Care (CCAC¹) Experimental Animal User Training includes a core topics module in EE (CCAC 2004). The current CCAC Guide to the Care and Use of Experimental Animals (Olfert et al. 1993) includes a section on social and behavioral requirements of experimental animals and the following position statement titled "Social and Behavioral Requirements of Experimental Animals (SBREA)":

"Physical well-being is manifested by a state of clinical health. Behavioural well-being is manifested by behaviour considered to be normal for that species and strain, together with the absence of significantly abnormal behaviour. Behavioural well-being is considered to reflect psychological well-being, and to that extent, the terms are considered to be synonymous in our usage.

In the interest of well-being, a social environment is desired for each animal which will allow basic social contacts and positive social relationships. Social behaviour assists animals to cope with circumstances of confinement. Caging, whether for single animals, pairs, or groups, should be enriched appropriately for the species . . ." (p. 52).

The Council of Europe is currently in the process of revising animal standards in Appendix A of the "European convention for the Protection of Vertebrate Animals Used for Experimental or other Scientific Purposes," also known as European Treaty Series (ETS¹) 123. The European Union's adoption of the final standard would make these guidelines legally binding. The draft document has proposed substantial modification of caging and environment. For example, the draft section on enrichment (4.5.3) includes the following statement:

All animals should be allowed sufficient space of adequate complexity to express a wide range of normal behaviors. They should be provided with a degree of control and choice over their environment to reduce stress-induced behaviors. . . . In addition to social activities, enrichment can be achieved by allowing and promoting physical exercise, foraging, manipulative and cognitive activities, as relevant to the species concerned . . . The enrichment program should be regularly reviewed and updated. Animal care staff should understand the natural behavior . . . of the species . . . They should be aware that all enrichment initiatives are not necessarily to the advantage of the animal and therefore . . . adjust the program as required" (Berge 2003, p. 13).

Controversy surrounds adoption of these standards due to the paucity of data in the scientific literature to support such sweeping changes. Some experts have also raised concerns that cage environments may significantly influence experimental outcomes (Frank 2004). In response to some of these concerns, the Federation of European Laboratory Animal Science Associations (FELASA 2004) has established a working group on "Standardization of Enrichment." This group was assembled to address points of controversy in ETS 123, and is scheduled to issue a report by early 2005 (http://www.felasa.org/Documents/Workinggroups/TermsofReference/stenr-t.html).

Thus, it appears that "expert groups" generally view laboratory animal husbandry, sanitization, and housing standards as having reached a satisfactory level. As a consequence, updated guidelines and regulations will closely address animal welfare in the context of its microenvironment and behavioral management of laboratory animals. It is the responsibility of the laboratory animal community to provide guidance and evidence, and to interject an element of perspective into the legislation currently being crafted.

Natural Behavior as a Clue to Rodent Enrichment Needs

The goal of EE is to provide animals with opportunities to express their full range of species-typical behavioral patterns. To evaluate enrichment strategies, then, one must first understand the natural history and behavioral repertoire of the species in question.

Several different species of mouse, including Mus musculus, Mus domesticus, and Peromyscus spp., are commonly used in biomedical research. Wild counterparts of these species have similar general behavioral profiles. Mice, and most other rodents, are prey species, which shapes much of their behavior. Mice prefer to flee and hide from perceived threats, although they may bite when caught. They are primarily nocturnal and will actively burrow and build nests where they spend their inactive hours (Suckow et al. 2001). Mice are omnivorous and granivorous, are highly social, and communicate primarily through olfactory and auditory cues. They may establish dominance hierarchies, but fighting over territory and social rank still occurs, primarily among males (Jennings et al. 1998).

Mice became established as a common animal model in the early 1900s. Subsequent inbreeding by fanciers and researchers led to the establishment of several well-characterized strains of laboratory mice, including C57BL/6, BALB/c, and DBA. As a result of refined assisted reproductive technologies, transgenic and knockout mouse production, and captured spontaneous mutations, specific strains that emphasize one trait or another have become prolific and have led to countless variations of laboratory mice (Knight and Abbot 2002).

Despite many generations of domestication and directed breeding, the most common strains of laboratory mice retain the bulk of their progenitors' behavioral repertoires, although certain aspects may be exaggerated or de-emphasized. Burrowing and nesting behaviors largely persist. In a study in which nesting behavior was observed, C57BL/6J and DBA/2J mice constructed nests that differed in quality (Bond et al. 2002); and when given access to outside quarters or burrowing boxes, both BALB/cAbg and C57BL/6Abg mice spontaneously constructed burrows, although the complexity of excavation varied (Dudek et al. 1983). The social behaviors of laboratory mice are also largely intact. They have been found to prefer proximity with other mice (Van Loo et al. 2004b), and they will establish dominance hierarchies. Levels of aggression are widely considered to vary by strain, and several investigators have reported that DBA/2 mice were more aggressive than C57BL/6 (Crawley et al. 1997).

Laboratory rats descend from the species Rattus norvegicus. The albino rats used most commonly in research are thought to have been bred originally by fanciers who showed their animals, leading to their domestication and use in experimentation beginning in the 1800s (Sharp and La Regina 1998). Like mice, rats are a prey species, and their basic behavior patterns have thus evolved similarly. They are nocturnal social animals that dig complex burrows (Boice 1977), inhabit dark areas during daylight hours, and generally flee when threatened. Rats are ubiquitous in nearly all parts of the world because of their extremely opportunistic eating patterns; they are able to forage for grains or parasitize the wastes of other organisms. Unlike mice, rats are typically nonaggressive, even among males living in close quarters. They communicate largely through auditory means and are capable of extensive ultrasonic vocalizations (Sharp and La Regina 1998).

Many inbred strains of rats have been selectively bred and established as models of various medical conditions. Despite this inbreeding and extensive domestication, several studies have observed that laboratory rats demonstrate behavioral patterns similar to their wild counterparts. In one definitive study, a group of 10 albino laboratory rats of the commonly used Sprague-Dawley outbred stock were released into an outdoor enclosure and observed as a colony over the next 2 yr. The rats quickly reverted to natural behaviors, digging and residing in burrows immediately after release. Moreover, the animals followed established pathways from one area to another and engaged in social and mating behaviors characteristic of nondomesticated rats (Boice 1977). Although nest building is performed by all ages and sexes of wild rats, adult laboratory rats given nesting material for the first time often will not build nests. This behavior has been shown to be not a deviation from natural behavioral patterns but instead, a consequence of the apparent role that social learning plays in nest building (Van Loo and Baumans 2004).

Guinea pigs (Cavia porcellus) are present in the wild throughout South America. Based on their small size and docility, they have been popular pets since their domestication by Europeans in the 1500s (Terril and Clemons 1998). Like rats and mice, wild guinea pigs are a prey species that flee and hide as primary defenses against threats. In addition, guinea pigs exhibit freezing behavior in response to sounds. They do not dig their own burrows or build nests but instead, inhabit the burrows of other animals or find other pre-existing shelters. They are not nocturnal and have no discernable period of extended sleep, although they are sensitive to intense light and changes or extremes in temperature. Guinea pigs gnaw extensively, are herbivorous, and have a vitamin C requirement unique among rodents. They are very social and use a relatively complex system of distinct vocalizations along with scent marking for communication. Aggression between males may occur over mating privileges, space, and food (Terril and Clemons 1998).

Although certainly not as popular as mice and rats, guinea pigs are a relatively common biomedical research model, with 250,000 to 500,000 used annually between 1998 and 2002 (USDA 1998, 2000, 2002). Due to their long history as companion animals, laboratory guinea pigs are highly domesticated, and further selective breeding has created a number of identified stocks and strains. Nevertheless, research-bred animals have retained the basic behavioral repertoire of their wild counterparts, although the domestication and breeding processes have altered the frequency and circumstances under which they display certain behaviors (Kunzl and Sachser 1999). Kunzl and colleagues (2003) compared the common laboratory guinea pigs with both the first and 30th generation offspring of wild mating pairs housed under standard research conditions. They found that when presented with a novel cage or environment, the research stock were less exploratory and exhibited lower serum cortisol levels than their wild counterparts. Domestic animals also exhibited a decrease in aggression and an increase in positive social interactions compared with both wild groups.

The species of hamster most frequently used in research, Mesocricetus auratus, was imported from Syria and bred in the 1930's, and so has been domesticated recently relative to other common laboratory rodents. Hamsters are nocturnal prey animals that dig burrows and build nests for shelter. In response to shortening days and dropping temperatures, hamsters hoard food and enter pseudohibernation. They are omnivorous, foraging for fruits and plants and transporting food in cheek pouches. When threatened, females may use the cheek pouches to hide their offspring. Hamsters are less social than the other rodents described above, and are only compatible with unfamiliar animals in mating situations (Field and Sibold 1999).

Approximately 180,000 hamsters are used annually in biomedical research (USDA 1998, 2000, 2002). Although to date no published data specifically compare the behavior of laboratory hamsters with their wild counterparts, laboratory hamsters do display many of the basic behavioral characteristics of feral M. auratus, from pseudohibernation in response to lowered temperatures, to aggression between unfamiliar adults. Their relatively recent domestication also suggests fewer differences in behavior relative to other rodents. Studies of artificial domestication in other rodent species support this conclusion. For example, after 30 generations of laboratory housing, the offspring of wild-caught guinea pigs did not differ in behavior from their feral counterparts (Kunzl et al. 2003), and 10 generations of mice reared in either standard laboratory housing or a simulated natural environment differed only in reduced intermale aggression and decreased resistance to capture (Connor 1975).

Defining Enrichment Schemes for Rodents

Generally accepted paradigms support the notion that expanding an animal's options for species-specific behavioral expression can positively affect both physiological and psychological well-being. Given that each species evolves to maximize its chances for survival, innate behaviors would logically favor biological homeostasis and enhance physical health. At the same time, enrichment options can enhance subjects' mental health by providing appropriate avenues for instinctive behaviors and preferences that may otherwise be expressed abnormally or not at all. Static objects and structures not previously encountered may provoke a novelty effect and may tend to elicit decreased behavioral responses with repeated or prolonged exposures. Response-contingent enrichment objects (i.e., items the animals themselves can alter) are often most effective at eliciting a novel interactive response.

Consideration of the concepts described directly above, and the previous description of inherent rodent behavior, allows the characterization of enrichment strategies for laboratory rodents within a few broad categories: structure and substrate, manipulanda, novel foods, social contact, and other less common types. Structure and substrate include any objects or parameters of an animal's enclosure, including its size and shape, that allow an animal to isolate itself into different microenvironments, experience varied textures or materials, or express natural patterns of locomotion. These enrichment strategies are perhaps the most commonly associated with laboratory rodents, and include nesting materials, running wheels, and hiding shelters. Manipulanda include any objects that can be altered by an animal or encourage it to engage in fine motor movements, such as wooden blocks or prefabricated plastic chew toys. Enrichment through diet manipulations may be accomplished by providing novel foods. Selected items (e.g., fruits, vegetables, grains) may differ in taste, texture, or desirability, or may require the animals to forage actively or to process the diet (e.g., spreading small bits of foods through bedding material has been thought to encourage foraging). Social contact can consist of any range of interactions between individual animals, from fully commensurate housing to mere visual contact or auditory communication. Other types of enrichment aim to stimulate senses other than touch or taste but have been used less commonly. Examples include the provision of music or white noise and the use of olfactory stimuli such as vanilla.

Review of the Literature

Historical Background

EE for rodents essentially began as a primary experimental variable with a series of projects in the 1960s that explored the relationship between an animal's surroundings and the basic makeup of its brain. Such research continues today, and still constitutes the bulk of published research in this field (Figure 1). From early conclusions regarding the effects of enrichment on macroscopic parameters such as brain weight (van Praag et al. 2000), increasingly refined studies have progressed to showing effects at the cell and molecular level, including neuronal structure (Fernandez et al. 2003), gliogenesis (Steiner et al. 2004), and synaptic plasticity and gene expression (Farmer et al. 2004). Investigators have also found that enrichment mitigates insults or facilitates repair to the brain in a variety of situations, including developmental lead exposure (Guilarte et al. 2003), stroke and trauma (Dahlqvist et al. 2004; Xerri et al. 2003), prenatal stress (Morley-Fletcher et al. 2003), dark rearing (Bartoletti et al. 2004), and even aging (Frick and Fernandez 2003).

Although many of the reports cited above highlight the effects that cage complexity may have on an animal's behavior and neurocognitive function, the primary goal of these studies is not to enhance their subjects' welfare or to determine preferences. Adding complexity to standard caging as a means of enhancing the welfare of laboratory animals is a relatively recent phenomenon, not having received a great deal of attention until the passage of the Food Security Act in 1985, which amended the AWA. The field of study addressing environmental changes specifically for the welfare of laboratory rodents is in its infancy, representing approximately 20% of the publications involving EE for rodents occurring since 2001 (Figure 1). An overview of this work reveals important themes about use of rodent enrichment, and highlights unanswered questions and future directions for exploration.

Figure 1 Number of citations (N) resulting from a search of Medline Database using Ovid Search Engine (Ovid Medline\R; http://gateway2.ovid.com/) using keywords environmental enrichment and rodent. 1902-June 2004 were searched in total and by decade. Titles and abstracts were reviewed to identify studies that were neurobiological in nature versus studies with a primary objective to identify preferences for or effects of enrichment(Non-Cog).

Types of Experiments

Generally speaking, two types of experiments are performed to evaluate enrichment as a means of enhancing rodent welfare: (1) experiments that examine the desire of the animals for specific objects or environments (preference testing), and (2) those that examine the effects of enrichment on behavioral or physiological parameters. For more information and to complement the very brief description of these experiments below, we urge readers to review the primary literature.

Preference Testing

In preference testing experiments, subjects are given a choice between different conditions, and the environment most frequently chosen is concluded to be the preferred strategy. This type of experiment is appealing because of its simplicity, but it can only measure relative preferences between conditions and cannot establish the strength of the desire for the preference. Motivational testing addresses this problem by measuring the amount of work a subject is willing to perform to gain access to a certain environment or object. This experiment assumes that individuals will work harder to obtain enrichment conditions that fulfill their most basic needs. It must be noted that both preference and motivational tests assume that animals will always pursue their best welfare interests. However, this assumption may be dubious in certain cases, such as when considering most species' preferences for sweet but otherwise nutritionally void foods over more balanced fare.

The natural tendency of mice, rats, and hamsters to dig their own burrows and build nests suggests that providing nesting material may fulfill one of their instinctual behavioral needs. In their excellent comprehensive review of the literature prior to 2002, Olsson and Dahlborn (2002) identify many reports that detail the prevalence of nest-building behavior among both males and females of all examined strains. Preference testing paradigms have shown that mice prefer bedding and nesting materials as a function of their ability to form cohesive nests from them (Ago et al. 2002; Van de Weerd et al. 1997a). It has also been reported extensively that mice prefer nesting material over virtually every other form of enrichment (Olsson and Dahlborn 2002), including prefabricated shelters (Hobbs et al. 1997), mobile and chewable manipulanda (Coviello-Mclaughlin and Starr 1997), and social contact (Van Loo et al. 2004b). In contrast, rats have been shown to prefer paper strip bedding over other materials that could form more cohesive nests (Manser et al. 1998a), and in motivational tests were not willing to work significantly more to access nesting materials than to access an empty cage (Manser et al. 1998b). These findings, however, may be a result of the previously discussed requirement to expose rats to nesting material at an early age (Van Loo and Baumans 2004).

Wild mice, rats, guinea pigs, and hamsters seek natural shelters and burrows, suggesting they may seek hiding spaces or isolated areas within a laboratory environment. Preference testing has demonstrated that research-bred mice demonstrate a preference for opaque or tinted structures over unenriched or larger cages (Olsson and Dahlborn 2002), and they spend more time in opaque tunnels than interacting with manipulanda (Hobbs et al. 1997). As discussed below, however, research-bred mice consistently prefer nesting material over structural complexities. Laboratory-reared rats, in contrast, have been shown to have a strong preference for nest boxes that most limit incoming light (Manser et al. 1998a), and are willing to work significantly more to access a cage containing a nest box compared with a standard cage, regardless of the presence of nesting material (Manser et al. 1998b). However, laboratory-reared rats were found not to prefer small opaque tubes over chewable manipulanda (Chmiel and Noonan 1996). Laboratory-bred guinea pigs, meanwhile, have been shown to express a strong tendency toward thigmotaxia. They have been observed to spend a preponderance of time near the edges of their cages to the near exclusion of the more open cage centers (White et al. 1989), suggesting structures that provide additional surface area may be beneficial to this species.

Behavioral and Physiological Testing Parameters

Experiments designed to examine the effects of different environments on behavior and physiological parameters evaluate how the actual well-being of the animals may be affected by enrichment more specifically than do preference tests. In experiments that examine behavior, it can be concluded that, in accordance with the principles set forth in the Guide (NRC 1996), complexities that reduce abnormal behaviors and increase species-typical behaviors are indeed a benefit to the animals' welfare. In experiments that examine biological parameters, it can generally be concluded that conditions reducing stress indicators (e.g., cortisol) or increasing general indicators of physical health (e.g., body weight and birth rates) enhance the well-being of the subjects. Behavioral and physiological responses to enrichment also provide information regarding potential effects on research measurements.

The most commonly reported physiological effect of nest building is a decrease in food consumption without a consequent decrease in body weight (Olsson and Dahlborn 2002). This effect is commonly attributed to the thermoregulatory properties of a nest; multiple findings correlate nesting behavior with ambient temperature (Sherwin 1997). Van Loo and coworkers have shown that when portions of nesting material are transferred at cage change, intermale aggression is decreased (2003) and long-term stress indicators (urinary corticosterone and thymic weight) were decreased (2004a). Additionally, the majority of studies reviewed by Olsson and Dahlborn (2002) showed no change in other physiological measures or performance in standard behavioral tests. Although nesting material led to decreased body weights and amounts of brown adipose tissue among BALB/c mice in one study (Eskola and Kaliste-Korhonen 1999) and an increase in variability of corticosterone levels in another (Van de Weerd 1997b), no increase in variability was seen in behavioral responses to diazepam treatment among BALB/c and C57BL/6 mice (Augustsson et al. 2003). Thus, it appears that nesting may have minimal impact on most common research variables, even though some isolated consequences to physical health may occur. For example, Bazille and colleagues (2001) have reported an increase in the incidence of conjunctivitis in athymic nude mice. In addition, occasional limb or tail entanglement has been observed when using gauze or commercially available pressed cotton pads (S.V.W., unpublished data).

The exploratory and foraging behaviors of all laboratory rodent species suggest that they may benefit from opportunities to engage in physical activity and manipulation of objects. Although larger cages might be expected to benefit rodents because they allow for more space for exploration, mice generally avoid larger cages unless they include hiding spaces (Olsson and Dahlborn 2002). Even in standard size cages, guinea pigs avoid open space in the center of an enclosure (White et al. 1989). Running wheels may be considered an alternative to larger space. Mice, rats, and hamsters use these objects extensively, and they are used as a measure of periodicity in hamster hibernation and photoperiod research (Pratt and Goldman 1986; Reebs and Maillet 2003). Mice perform more work to access a running wheel than a tunnel system (Sherwin 1998a). However, a running wheel may not provide benefits analogous to natural locomotion (Sherwin 1998b). Furthermore, running wheels have been shown to affect mouse brain development differently from other types of enrichment, increasing proliferation of microglia in several superficial areas of the cortex (Ehninger and Kempermann 2003) but not contributing to the morphological changes in the hippocampus neuron CA1 and the dentate gyrus (Faherty et al. 2003). So although some rodents show preference for wheels, the presence of such items has the potential to skew experimental parameter measurements.

Manipulanda are used by laboratory rodents seemingly as a function of their response-contingency (i.e., the animals'ability to shape or alter the manipulanda). When presented with a marble, nesting material, and an opaque tunnel, mice spent virtually no time manipulating the marble (Hobbs et al. 1997). Similarly, rats prefer chewable over nonchewable manipulanda (Chmiel and Noonan 1996). Use of manipulation and chewing enrichment devices by rats does not affect body weight, food intake, or hematological parameters (Watson 1993). Additional preferences and effects of manipulanda and exercise are difficult to interpret because of the preponderance of studies using a "superenriched" condition (discussed below).

The effects observed as a result of providing structural enrichment in laboratory environments have been varied and have differed by strain. The most common results for mice reported in Olsson and Dahlborn's (2002) review were decreased anxiety in novel situations and increased intermale aggression, found in conjunction with the use of nest boxes, mazes, partitions, and other complexities. The increased aggression is commonly attributed to the introduction of defined "territories" within the enclosure. Other significant responses to structural enrichment of this type are alterations in endocrine and immune parameters, which are not necessarily anticipated. For example, mice provided with a nest box and scaffolding had increased plasma corticosterone and adrenal tyrosine hydroxylase activity levels (Marashi et al. 2003), as well as a reduced resistance to infection with Babesia microti (Barnard et al. 1996). Changes in murine body weight dependent on strain and gender were also reported in response to such structural additions (Olsson and Dahlborn 2002). As with manipulanda and exercise, unfortunately, it is difficult to summarize the effects of shelters on rats because most published reports have used a superenriched condition.

Superenrichment

Superenrichment can be considered any combination of multiple, notably different enrichment strategies, whether simultaneously or in periodic rotation. This condition is useful in highlighting the possible effects that enrichment strategies may have on physiological and psychological parameters and that are most commonly used in the neurological research. General trends from studies of superenrichment conditions include the enhancement of neural plasticity and function (van Praag et al. 2000), decreased inhibition toward novel stimuli (Olsson and Dahlborn 2002), decreased stereotypic behavior (Powell et al. 2000), and increased strain-dependent variability in several common research parameters (Tsai et al. 2002, 2003b). Other findings have shown that environmental factors may affect drug sensitivities (Green et al. 2003; Smith et al. 2003; Zhu et al. 2004), alter breeding performance (Tsai et al. 2003a), and exacerbate or quicken the pace of neurological disorders (Jankowsky et al. 2003; Schridde and Luijtelaar 2004). These results highlight the possibility that the addition of multiple (or sometimes single) enrichment items, or a scheme that produces a novelty effect, has definite potential to affect ongoing research. However, because studies of superenrichment are unable to distinguish between their constituent elements, the results should not be construed as representative of rodent enrichment as a whole. One such study and its proper implications are presented below.

Effects of EE on Murine Immune Responses

A series of studies was performed using either Swiss Webster or BALB/c mice (VandeWoude et al. 2002, 2003). Animals were either purchased from a commercial vendor or bred on site, and were exposed either to highly enriched (ladder and jar) nesting materials or to an unenriched (standard bedded cage) environment. Animals were immunized with ovalbumin and Ribi adjuvant, and various immune response (antibody and cytokine production, thymocyte number) or physiological (body weight, litter size and number) parameters were measured over time.

One mechanism of glucocorticoid-induced immunosuppression results from inhibition of transcription of "pro-inflammatory" cytokines, including interleukin (IL¹) 12. This cytokine initiates or promotes specific immune responses against pathogens or foreign antigens through a complex cascade of intracellular events, including gamma-interferon (IFNγ¹) production (reviewed in Webster et al. 2002). We therefore have hypothesized that mice with enrichment are less stressed, produce less endogenous cortisol, and exhibit higher levels of pro-inflammatory cytokines in response to antigenic challenge. We sought to measure this outcome by evaluating levels of IL-12 and IFNγ in mice with and without enrichment, after immunization.

We further hypothesized that without enrichment, "stressed" mice, with lower levels of pro-inflammatory cytokines following immunization, would consequently exhibit an "anti-inflammatory" cytokine profile, which includes interleukins 4, 5, 13, and the potent anti-inflammatory mediator IL-10. Antigenic challenge dominated by these cytokines (and in particular IL-4) results in antibody isotypes dominated by immunoglobulin (IgG¹) 1 and IgG 2b. In contrast to IL-12 and IFNγ, anti-inflammatory cytokines are associated with a primary humoral versus cell-mediated immune response, resulting in ineffective immune modulation of infectious disease in a variety of mouse models (Webster et al. 2002).

We also have harvested spleen cells from immunized mice exposed to the immunogen, and have assayed the culture fluid for these cytokines using an enzyme-linked immunosorbent assay. Our results, which generally suggest that enriched mice have more of an anti-inflammatory cytokine profile than their non-enriched cohorts, are displayed in Table 1.

Table 1

The most striking finding, severe thymocyte depletion, was noted in both Swiss and BALB/c female mice (Table 2). Thymocytes are extremely sensitive to cortisol-mediated apoptosis, and the most likely explanation for our findings is a stress response leading to cortisol release (Espey and Basile 1999). Phenotypic analysis of thymocytes from enriched BALB/c mice detected a lower percentage of double positive CD8/CD4 expressing T cells (Figure 2). This finding is again consistent with a higher level of endogenous glucocorticoids in the enriched mice (Webster et al. 2002). Furthermore, enriched Swiss Webster female mice demonstrated an increased production of IL-4 and IL-10, consistent with a predominantly "Type II" immune response. Other cytokine and antibody parameters in BALB/c male or female mice were not as consistently reminiscent of a Type 2 versus Type 1 cytokine profile. In fact, thymocyte numbers were higher in male BALB/c enriched mice (p = 0.2). Thus, contrary to our hypothesis, we observed significantly lower thymocyte numbers in enriched female mice in two separate experiments, and witnessed other findings demonstrating a tendency for enriched mice to develop a Type II immune response.

Table 2

Figure 2 Female BALB/c mice with enrichment have fewer immature thymocytes. A. Flow cytometric analysis of thymocytes harvested from an ovalbumin-vaccinated female mouse. CD4-positive T-cells are demonstrated on FL-4; anti-CD8 labeled cells are represented on FL2. Double-positive cells (CD4+CD8+) are shown above baseline scatter on both X and Y axis. B. Unenriched mice demonstrate much greater percentages of double-positive, immature thymocytes than enriched mice, a finding consistent with endogenous corticosteroid release. DP, double-positive; SP, single-positive.

In terms of other physiological alterations noted, both BALB/c and Swiss female mice tended toward a lower mature body weight when exposed to enrichment, and many fewer litters were produced by female BALB/c mice housed in an enriched environment (Table 1). Subsequent breeding trial observations of CB17-Prkdc^scid, B6D2F2, and ICR breeding colonies with and without superenrichment did not result in significant differences in numbers of litters or pups per litter. In addition, female BALB/c parenterally inoculated with malaria (Plasmodium yoelii) did not vary by clinical signs or parasite load relative to the presence of enrichment.

Table 3

Besides thymic atrophy, a second very noteworthy observation is that standard deviations for measured parameters were nearly always greater in enriched mice (Figure 3). In addition to cytokine levels, standard deviations were noted to be higher in enriched mice when other parameters were examined, such as pups born per litter and pups weaned per litter in both CB17-Prkdc^scid and B10D2/nSnJ strains (Table 3). Others (Olsson and Dahlborn 2002; Tsai et al. 2002, 2003b), have noted this effect, which is a logical outcome of less homogeneous experimental conditions. Unfortunately, this unintended consequence of providing EE is not benign. The power of a statistical hypothesis test is very dependent on within-group variability, and any increase in this parameter results in a loss of power. For example, the power of a t-test is primarily a function of its "noncentrality parameter," which is proportional to the following:

where n is the sample size, per group, d is the true difference between group means, and σ is the within-group standard deviation (Rao 1998, p. 368-372). In other words, if σ increases by a factor of x, n must increase by a factor of x² to maintain the same approximate level of power. (Power is also a function of degrees of freedom, 2(n-1), which increases with n, mitigating this requirement slightly.)

Figure 3 Enriched mice have higher standard deviations for cytokine measurements than unenriched mice. BALB/c mice (N = 5-7/group), housed with or without enrichment (described in text) were immunized with ovalbumin per standard protocol. Splenocytes were harvested at necropsy and cultured in the presence of ovalbumin. Cytokine production (interleukin [IL]-2, gamma-interferon [IFN-g], IL-4, and IL-10) was determined by enzyme-linked immunosorbent assay evaluation of supernatants. Comparison of means between groups of male or female mice with or without enrichment did not reveal significant differences. However, in seven of eight comparisons, the standard deviation among measurements appeared to be substantially higher in groups of animals provided with enrichment. These results are represented in this figure as fold differences between ratios of standard deviation/mean for each group. F, female; M, male; UN, unenriched; EN, enriched. Numbers in bold type indicate higher standard deviations in enriched versus unenriched groups.

In our experiments, as illustrated in Figure 3, other variables being equal, IL-4 measurements in female mice provided with enrichment would require 14 times the number of mice to achieve a specific power than mice without enrichment. Although this is the most extreme example, even a 1.5-fold increase in σ requires more than twice the value of n to achieve equivalent power. These findings cause great concern in the context of attempts to reduce animal use in biomedical research.

Evaluation of Current Enrichment Schemes

A critical step in planning EE strategies is first to identify which practices are currently in widespread use. Cage enrichment is not commonly reported in research articles that do not specifically pertain to enrichment, so it is difficult to ascertain its related degree of consensus and use within the laboratory community. At the National Institutes of Health (NIH¹), enrichment strategies vary between institutes.

In an effort to ascertain current practices for EE in use across the campus, the Trans-NIH Enrichment Advisory Panel was organized, with the stated mission of serving as a dynamic resource for the collection and dissemination of information within the NIH Intramural Research Program regarding animal welfare, environmental enrichment, and the behavioral management of laboratory animals. The group sent electronic questionnaires to each NIH animal facility, and respondents were asked to complete a survey for each of the species in their facilities. Sample questions included "About how many of your animals receive structural enrichment?" and "What types of structural enrichment do your animals receive?" A total of 22 separate facilities, which represented 12 institutes, completed and returned surveys. These facilities house mice, rats, and/or guinea pigs, including animals involved in a wide range of study types, from immunology to oncology to genetics.

Among the programs surveyed, mouse enrichment was the most consistently uniform program. All 22 facilities that house mice reported providing structural enrichment to at least some of their animals, and 16 (73%) provided structural enrichment to all of their mice. By far the most widespread form of structural enrichment was nesting material, used in 20 of the facilities, with compressed cotton pads cited as a nesting material by 19 programs. Mouse huts were the next most prevalent form of structural enrichment, with cardboard and plastic shelters used in 12 buildings. Novel foods were also given to some of the mice in 12 facilities, although most respondents reported that they were provided primarily for medical or breeding purposes. Chew toys, the only form of manipulanda used by more than one program, was provided to some mice by four programs (Figure 4). When asked to estimate the percentage of animals housed in various social settings, only one facility (5%) reported housing more than 30% of its mice in single enclosures, and the average reported percentage of singly housed mice was only 11%.

The survey results also revealed that rats at the 15 responding facilities largely receive treatment similar to that of mice. All respondents reported providing structural enrichment to at least some rats, and 11 (73%) included it for all. Nesting material was the frequently cited type of structural enrichment, as reported in 13 surveys, 12 of which referred specifically to compressed cotton pads. In contrast to the results for mice, an equal number of rat facilities also reported using tubes or prefabricated shelters, the most common of which were paper tubes, cited by eight facilities. Novel foods, primarily in the form of fruit, seeds, and manufactured treats, were provided to at least some of the animals by seven programs. Manipulanda were reportedly used for some rats in six facilities, with plastic chew toys the most common example (Figure 4). The social housing situation for rats was similar to that of mice, with only one respondent (7%) estimating that more than 25% of their animals were singly housed. The average reported rate of single housing was 12%.

Figure 4 Enrichment is widely used and varies per species. A survey evaluating enrichment use at 22 separate animal facilities at the National Institutes of Health was performed as described in the text. All respondents used enrichment for all three species in most instances (see text). Although facilities generally used nesting material for rats and mice, guinea pigs were more likely to be provided with structural enrichment.

EE strategies for guinea pigs emphasized social grouping rather than nesting material. Structural enrichment was given to some or all of the animals in six of the seven reporting facilities, and in each of these cases, polyvinyl chloride tubes were cited as the primary device in use. Novel foods, mostly fruits and vegetables, were provided to at least some guinea pigs in four programs. Manipulanda were only reported in use at one facility (Figure 4). When asked to estimate social housing rates, respondents of one facility estimated that 50% of its animals were singly caged, but all others reported at least 90% social caging.

The implications of the study described above are as follows:

Social housing of the most commonly used research rodents is the norm. Analysis of innate behaviors and preliminary studies suggests that this practice serves to improve the well-being of these animals.
Facilities implementing materials were more likely to use materials that promote species-specific behaviors. For example, nesting material already appears to be a standard enrichment practice in mouse housing. In contrast, the thigmotaxic tendencies of guinea pigs are facilitated by providing structures that increase the useful wall space of an enclosure.
A surprisingly uniform picture of rodent enrichment practices emerged over a wide variety of independently managed facilities. This consistency alleviates some concerns that large variability in enrichment practices from facility to facility could skew data interpretation and interfere with experimental reproducibility.

Instituting an EE Program for Rodents

In devising an enrichment program for rodents, animal care unit staff are forced to consider a variety of practical matters, including oversight, strategy selection, investigator notification, personnel education, and follow-up evaluation. This section provides an overview of pitfalls encountered when a rodent enrichment program was initiated at one of our facilities (Colorado State University [CSU¹]), as well as potential solutions for consideration by programs in the development phases of EE implementation in rodents.

Some large facilities have hired enrichment specialists to provide oversight for their enrichment programs, but this mechanism is not practical for most facilities dealing strictly with rodents. At CSU, implementation of a rodent EE program was spearheaded by interested animal care technicians who worked with a liaison on the veterinary staff. Establishing an EE coordinator from among the animal care staff was helpful in providing continuity between plans and practice. After several years of establishment at CSU, enrichment provision became a standard component of animal care and no longer required a specific employee for oversight.

When selecting appropriate enrichment strategies for their rodents, facility staff should consider a number of factors. As described above, species natural histories and existing literature should be used to determine which items may actually be desirable and beneficial for the animals. Literature and data presented in this article suggest avoiding superenrichment, at least in mouse caging. The type of research being conducted is also important because some studies may have specific needs, whereas others would be unable to use certain types of enrichment (i.e., a toxicology study may not permit use of manipulanda with glues or dyes). Although individual rodent enrichment devices are usually inexpensive, the large numbers of animals housed in many facilities makes the cost of these items a crucial factor. In addition to the obvious cost of materials, labor costs for purchasing, processing, sanitizing, and providing enrichment will also be incurred. Facilities may consider establishing an "enrichment budget" so that related costs can be tracked. At CSU, specific animal care technicians or managers became responsible for inventory, purchase, and materials processing to avoid shortages. In an effort to solve the problem of storage for enrichment items, a dedicated space was established for holding processed enrichment materials, which greatly increased utilization by caretakers.

It is vital that investigators receive notification of EE practices before implementation. Specific studies (e.g., behavioral studies) may often have absolute contraindication to enrichment, whereas others may have a more limited scope of permissible enrichment (e.g., food treats are not acceptable for nutrition studies). We addressed this issue by providing investigators with a questionnaire in which we asked them to provide permission for each of the categories of enrichment likely to be offered. Investigator requests to withhold enrichment were then prominently displayed and communicated to caretakers.

Animal care technicians must be instructed fully in the use of enrichment to ensure the standardization of practices throughout a facility. Standard operating procedures should be developed to aid in training and record keeping. We found that when new employee training included an introduction to rodent enrichment, consistency of use and record keeping improved. Record-keeping requirements vary with institution and enrichment scheme. Facilities offering novel enrichment items on a rotating basis will likely need a record-keeping log to avoid repetition. Technicians may find it useful to have enrichment instructions posted in each room. As the effects of enrichment on experimental outcome are more completely recognized, investigators may require access to enrichment records to report them in research materials and methods.

Ongoing evaluations of enrichment should be made by the animal care technician, veterinary staff, and investigator. This process is most critical when new enrichment is being introduced; however, continued evaluation of the literature is recommended in this rapidly evolving branch of laboratory animal science.

Summary

Although use of EE for rodents has many unanswered questions, we believe that several conclusive themes emerge from the literature. The following three themes are prominent:

EE for rodents is apparently widely practiced, despite the lack of scientifically based guidelines. Fortunately, it appears that use of enrichment tends to follow the logical paradigm of providing items that promote species-typical behavior and minimize the effects on research. For example, the widely reported use of nesting material in mouse cages is supported by the natural histories of these species. In addition, an increasing number of studies document that it has minimal effects on common experimental variables. Likewise, appropriate social contact for all rodents and structural enrichment for rats and guinea pigs are practices that closely mimic natural conditions for the nondomesticated counterparts of these species, and that have so far not been associated with detrimental physiological effects. Laboratory animal professionals should take pride that intuitive and educated observation about the behaviors of the species under their care have led to what appear to be the most useful rodent enrichment practices so far defined.
It is clear that certain enrichment schemes, especially the superenrichment condition, not only cause a number of potentially confounding changes in common behavioral and physiological variables, but also are detrimental to the rodents themselves. Introduction of novel items and structural complexities may increase stress and aggression via induction of "territorial" behaviors, which are deleterious to research, facility management, and animal health. It is apparent from the NIH survey findings and regulations under consideration that persons involved in biomedical research, from animal care staff to investigators, favor the notion that augmenting animal well-being is desirable and worthy of added expenses. However, in implementing enrichment strategies for rodents, we must remain focused on actual benefit to the animals and cautious of imposing our own perceptions on creatures more than three orders of magnitude smaller than ourselves.
EE is an experimental variable. Although this conclusion is obvious, the implications have been widely overlooked. The use of enrichment may have the following effect on research studies:
- Potentially greater variation among data, requiring more animals per group for statistical significance;
- Potential effect on experimental reproducibility; and
- Inappropriate conclusions from data skewed by environmental conditions.
For this reason and because recommendations for EE for rodents are still being debated, it would seem prudent to report types of enrichment provided during any study involving rodents, just as one would report other housing conditions and animal data. A quick perusal of the literature confirms that this information is rarely included in research article despite its obvious impact (whether positive or negative) on animal physiology.

Conclusions

Based on the foregoing information and identifiable themes, we conclude that continued studies are necessary in this area. Such results are especially relevant given that the actions of regulatory agencies and facilities implementing such practices suggest that EE is the next frontier of animal care that will be developed and scrutinized. Studies are needed that evaluate single-enrichment schemes that are practical, affordable, and likely to provide benefits to animals with minimal alterations to research data. Variation between preferences of different species, strains, age groups, and sex requires further evaluation. It is incumbent on investigators to link preference testing with studies evaluating animal well-being to prove that preferred environments result in positive changes in well-being. Although this area of study would not generally elicit great enthusiasm by funding agencies or by investigators, it behooves the scientific community to determine more completely how enrichment can be used to benefit both animal welfare and research objectives.

We hope that this review presents a balanced, scientific assessment of this important topic. We trust that the assessment will be helpful both in crafting rodent enrichment programs and in further outlining the studies necessary to devise schemes that have positive benefits for laboratory animals and experimental outcomes.

Acknowledgments

We thank Sara Gilbert and Michael Rollin for performing a substantial portion of the mouse enrichment study described in this article. These studies were funded by the Stanford University Undergraduate Research Program, Merck-Merial Summer Fellowship for Veterinary Students, and the ACLAM Foundation. We thank Katie Smith, Dr. Jim Weed, and the Trans-NIH Enrichment Advisory Panel, for their roles in gathering and compiling the enrichment survey data, and Dr. Phillip Chapman, for his discussions on biostatistical power analysis.

¹Abbreviations used in this article: AWA, Animal Welfare Act; CCAC, Canadian Council on Animal Care; CSU, Colorado State University; EE, environmental enrichment; ETS, European Treaty Series; Guide, Guide for the Care and Use of Laboratory Animals; IFNγ, gamma-interferon; IgG, immunoglobulin G; IL, interleukin; NIH, National Institutes of Health.

References

Abbot A. 2004. Laboratory animals: The Renaissance rat. Nature 428:464-466.

Ago A, Gonda T, Takechi M, Takeuchi T, Kawakami K. 2002. Preferences for paper bedding material of the laboratory mice. Exp Anim 51:157-161.

Alternative Research and Development Foundation v. Veneman. 2001. 262 F.3d 406. DC Cir 2001.

Augustsson H, van de Weerd HA, Kruitwagen CL, Baumans V. 2003. Effect of enrichment on variation and results in the light/dark test. Lab Anim 37:328-340.

Barnard CJ, Behnke JM, Sewell J. 1996. Environmental enrichment, immunocompetence, and resistance to Babesia microti in male mice. Physiol Behav 60:1223-1231.

Bartoletti A, Medini P, Berardi N, Maffei L. 2004. Environmental enrichment prevents effects of dark-rearing in the rat visual cortex. Nat Neurosci 7:215-216.

Bazille PG, Walden SD, Koniar BL, Gunther R. 2001. Commercial cotton nesting material as a predisposing factor for conjunctivitis in athymic nude mice. Lab Anim 30:40-42.

Berge, E. 2003. Potential impact of proposed changes to ETS 123. AAALAC Connection Newsltr Spring 1-4.

Boice R. 1977. Burrows of wild and albino rats: Effects of domestication, outdoor raising, age, experience, and maternal state. J Comp Physiol Psychol 91:649-661.

Bond TL, Neumann PE, Mathieson WB, Brown RE. 2002. Nest building in nulligravid, primigravid and primiparous C57BL/6J and DBA/2J mice (Mus musculus). Physiol Behav 75:551-555.

CCAC [Canadian Council on Animal Care]. 2003. Experimental Animal User Training: Core Topics Modules. Ottawa: CCAC. Available online (http://www.ccac.ca/english/educat/edframe.htm).

Chmiel DJ, Noonan M. 1996. Preference of laboratory rats for potentially enriching stimulus objects. Lab Anim 30:97-101.

Connor JL. 1975. Genetic mechanisms controlling the domestication of a wild house mouse population (Mus musculus L.). J Comp Physiol Psychol 89:118-130.

Coviello-Mclaughlin GM, Starr SJ. 1997. Rodent enrichment devices--Evaluation of preference and efficacy. Contemp Top Lab Anim Sci 36:66-68.

Crawley JN, Belknap JK, Collins A, Crabbe JC, Frankel W, Henderson N, Hitzemann RJ, Maxson SC, Miner LL, Silva AJ, Wehner JM, Wynshaw-Boris A, Paylor R. 1997. Behavioral phenotypes of inbred mouse strains: Implications and recommendations for molecular studies. Psychopharmacology (Berl) 132:107-124.

Dahlqvist P, Ronnback A, Bergstrom SA, Soderstrom I, Olsson T. 2004. Environmental enrichment reverses learning impairment in the Morris water maze after focal cerebral ischemia in rats. Eur J Neurosci 19:2288-2298.

Dudek BC, Adams N, Boice R, Abbott ME. 1983. Genetic influences on digging behaviors in mice (Mus musculus) in laboratory and seminatural settings. J Comp Psychol 97:249-259.

Ehninger D, Kempermann G. 2003. Regional effects of wheel running and environmental enrichment on cell genesis and microglia proliferation in the adult murine neocortex. Cereb Cortex 13:845-851.

Eskola S, Kaliste-Korhonen E. 1999. Aspen wood-wool is preferred as a resting place, but does not affect intracage fighting of male BALB/c and C57BL/6J mice. Lab Anim 33:108-121.

Espey MG, Basile AS. 1999. Glutamate augments retrovirus-induced immunodeficiency through chronic stimulation of the hypothalamic-pituitary-adrenal axis. J Immunol 162:4998-5002.

Faherty CJ, Kerley D, Smeyne RJ. 2003. A Golgi-Cox morphological analysis of neuronal changes induced by environmental enrichment. Brain Res Dev Brain Res 141:55-61.

Farmer J, Zhao X, van Praag H, Wodtke K, Gage FH, Christie BR. 2004. Effects of voluntary exercise on synaptic plasticity and gene expression in the dentate gyrus of adult male Sprague-Dawley rats in vivo. Neuroscience 124:71-79.

[FELASA] Federation of European Laboratory Animal Science Associations. 2003. Terms of references. London: FELASA. Available online (http://www.felasa.org/Documents/Workinggroups/TermsofReference/stenr-t.html).

Fernandez V, Fernandez B, Adaro L. 2003. Early polysensorial enrichment: Specific experience-induced structural changes in the parieto-occipital cortex of the rat. Growth Dev Aging 67:3-10.

Field KJ, Sibold AL. 1999. The Laboratory Hamster and Gerbil. Boca Raton: CRC Press.

Frank M, ed. 2004. Public affairs: European welfare standards may affect US research. The Physiologist 47:1-3.

Frick KM, Fernandez SM. 2003. Enrichment enhances spatial memory and increases synaptophysin levels in aged female mice. Neurobiology 24:615-626.

Green TA, Cain ME, Thompson M, Bardo MT. 2003. Environmental enrichment decreases nicotine-induced hyperactivity in rats. Psychopharmacology (Berl) 170:235-241.

Guilarte TR, Toscano CD, McGlothan JL, Weaver SA. 2003. Environmental enrichment reverses cognitive and molecular deficits induced by developmental lead exposure. Ann Neurol 53:50-56.

Hobbs BA, Kozubal W, Nebiar FF. 1997. Evaluation of objects for environmental enrichment of mice. Contemp Top Lab Anim Sci 36:69-71.

Jankowsky JL, Xu G Fromholt D, Gonzalez V, Borchelt D. 2003. Environmental enrichment exacerbates amyloid plaque formation in a transgenic mouse model of Alzheimer disease. J Neuropathol Exp Neurol 62:1220-1227.

Jennings M, Batchelor GR, Brain PF, Dick A, Elliott H, Francis RJ, Hubrecht RC, Hurst JL, Morton DB, Peters AG, Raymond R, Sales GD, Sherwin CM, West C. 1998. Refining rodent husbandry: The mouse. Report of the Rodent Refinement Working Party. Lab Anim 32:233-259.

Knight J, Abbot A. 2002. Full house. Nature 417:785-786.

Kunzl C, Kaiser S, Meier E, Sachser N. 2003. Is a wild mammal kept and reared in captivity still a wild animal? Horm Behav 43:187-196.

Kunzl C, Sachser N. 1999. The behavioral endocrinology of domestication: A comparison between the domestic guinea pig (Cavia aperea f. porcellus) and its wild ancestor, the cavy (Cavia aperea). Horm Behav 35:28-37.

Manser CE, Broom DM, Overend P, Morris TH. 1998a. Investigations into the preferences of laboratory rats for nest-boxes and nesting materials. Lab Anim 32:23-35.

Manser CE, Broom DM, Overend P, Morris TH. 1998b. Operant studies to determine the strength of preference in laboratory rats for nest-boxes and nesting materials. Lab Anim 32:36-41.

Marashi V, Barnekow A, Ossendorf E, Sachser N. 2003. Effects of different forms of environmental enrichment on behavioral, endocrinological, and immunological parameters in male mice. Horm Behav 43:281-292.

Morley-Fletcher S, Rea M, Maccari S, Laviola G. 2003. Environmental enrichment during adolescence reverses the effects of prenatal stress on play behaviour and HPA axis reactivity in rats. Eur J Neurosci 18:3367-3374.

NRC [National Research Council]. 1996. Guide for the Care and Use of Laboratory Animals. 7th ed. Washington DC: National Academy Press.

Olfert ED, Cross BM, McWilliam AA, eds. 1993. CCAC Guide to the Care and Use of Experimental Animals. Vol 1, 2nd ed. Ottawa: CCAC.

Olsson IAS, Dahlborn K. 2002. Improving housing conditions for laboratory mice: A review of "environmental enrichment." Lab Anim 36:243-270.

Powell SB, Newman HA, McDonald TA, Bugenhagen P, Lewis MH. 2000. Development of spontaneous stereotyped behavior in deer mice: Effects of early and late exposure to a more complex environment. Dev Psychobiol 37:100-108.

Pratt BL, Goldman BD. 1986. Environmental influences upon circadian periodicity of Syrian hamsters. Physiol Behav 36:91-95.

Rao PV. 1998. Statistical Research Methods in the Life Sciences. Pacific Grove CA: Duxbury.

Reebs SG, Maillet D. 2003. Effect of cage enrichment on the daily use of running wheels by Syrian hamsters. Chronobiol Int 20:9-20.

Schridde U, van Luijtelaar G. 2004. The influence of strain and housing on two types of spike-wave discharges in rats. Genes Brain Behav 3:1-7.

Sharp PE, La Regina MC. 1998. The Laboratory Rat. Boca Raton: CRC Press.

Sherwin CM. 1997. Observations on the prevalence of nest-building in non-breeding TO strain mice and their use of two nesting materials. Lab Anim 31:125-132.

Sherwin CM. 1998a. The use and perceived importance of three resources which provide caged laboratory mice the opportunity for extended locomotion. Appl Anim Behav Sci 55:353-367.

Sherwin CM. 1998b. Voluntary wheel running: A review and novel interpretation. Anim Behav 56:11-27.

Smith MA, Bryant PA, McClean JM. 2003. Social and environmental enrichment enhances sensitivity to the effects of kappa opioids: Studies on antinociception, diuresis and conditioned place preference. Pharmacol Biochem Behav 76:93-101.

Steiner B, Kronenberg G, Jessberger S, Brandt MD, Reuter K, Kempermann G. 2004 Apr. Differential regulation of gliogenesis in the context of adult hippocampal neurogenesis in mice. Glia 46:41-52.

Suckow MA, Danneman P, Brayton C. 2001. The Laboratory Mouse. Boca Raton: CRC Press.

Terril LA, Clemons DJ. 1998. The Laboratory Guinea Pig. Boca Raton: CRC Press.

Tsai PP, Oppermann D, Stelzer HD, Mahler M, Hackbarth H. 2003a. The effects of different rack systems on the breeding performance of DBA/2 mice. Lab Anim 37:44-53.

Tsai PP, Pachowsky U, Stelzer HD, Hackbarth H. 2002. Impact of environmental enrichment in mice. 1: Effect of housing conditions on body weight, organ weights and haematology in different strains. Lab Anim 36:411-419.

Tsai PP, Stelzer HD, Hedrich HJ, Hackbarth H. 2003b. Are the effects of different enrichment designs on the physiology and behaviour of DBA/2 mice consistent? Lab Anim 37:314-327.

Van de Weerd HA, Van Loo PLP, Van Zutphen LFM, Koolhaas JM, Baumans V. 1997a. Preferences for nesting material as environmental enrichment for laboratory mice. Lab Anim 31:133-143.

Van de Weerd HA, Van Loo PLP, Van Zutphen LFM, Koolhaas JM, Baumans V. 1997b. Nesting material as environmental enrichment has no adverse effects on behavior and physiology of laboratory mice. Physiol Behav 62:1019-1028.

VandeWoude S, Rollin M, Avery A. 2002. The effects of environmental enrichment on murine immune responses. Contemp Top Lab Anim Sci 41:69.

VandeWoude S, Rollin M, Avery A. 2003. The effects of environmental enrichment on murine immune responses. In: Proceedings of the ACLAM Forum held May 14-17, 2003, Sanibel Harbor, Fort Myers, Florida. Chester NH: ACLAM.

Van Loo PLP, Baumans V. 2004. The importance of learning young: The use of nesting material in laboratory rats. Lab Anim 38:17-24.

Van Loo PLP, Van der Meer E, Kruitwagen CLJJ, Koolhaas JM, Van Zutphen LFM, Baumans V. 2004a Apr. Long-term effects of husbandry procedures on stress-related parameters in male mice of two strains. Lab Anim 38:169-177.

Van Loo PLP, Van de Weerd HA, Van Zutphen LFM. 2004b. Preference for social contact versus environmental enrichment in male laboratory mice. Lab Anim 38:178-188.

Van Loo PLP, Van Zutphen LFM, Baumans V. 2003. Male management: Coping with aggression problems in male laboratory mice. Lab Anim 37:300-313.

van Praag H, Kempermann G, Gage FH. 2000. Neural consequences of environmental enrichment. Nat Rev Neurosci 1:191-198.

Watson DSB. 1993. Evaluation of inanimate objects on commonly monitored variables in preclinical safety studies for mice and rats. Lab Anim Sci 43:378-380.

Webster JI, Tonelli L, Sternberb EM. 2002. Neuroendocrine regulation of immunity. Annu Rev Immunol 20:125-163.

White WJ, Balk MW, Lang CM. 1989. Use of cage space by guinea pigs. Lab Anim 23:208-214.

Xerri C, Zennou-Azougui Y, Coq J-O. 2003. Neuroprotective effects on somatotopic maps resulting from piracetam treatment and environmental enrichment after focal cortical injury. ILAR J 44:110-124.

Zhu J, Green T, Bardo MT, Dwoskin LP. 2004. Environmental enrichment enhances sensitization to GBR 12935-induced activity and decreases dopamine transporter function in the medial prefrontal cortex. Behav Brain Res 148:107-117.