Understanding the neural mechanisms of complex human behaviors and the behavioral anomalies accompanying neurobiological disorders, including substance use disorder (SUD; or substance addiction), represents one of the most formidable challenges in behavioral and cognitive neuroscience research at present (Lynch et al., 2010; Gulinello et al., 2019; Kuhn et al., 2019). The past decade has seen a resurgence of studies using laboratory rodents, coupled with an impressive array of genetic modifications, providing unprecedented opportunities to generate suitable rodent models to research human pathologies. In contrast, behavioral assays, and their application to large numbers of animals, trailed behind in terms of throughput if compared with genetic advances for rodent models. Notwithstanding, there is a growing need for reliable and robust high-throughput behavioral assessment platforms to elucidate the cognitive and behavioral performances in both wild-type and transgenic rodent strains.

To address this issue, we focus on the IntelliCage system, a home-cage-based rodent behavioral assessment platform, and specifically, its utility to investigate the neurobehavioral underpinnings of SUD in rodent models. This fully automated live-in environment approach helps eliminate the confounding effects and considerable stress from environmental and experimental variables that may obscure the behavioral measures. Owed to its low-dependency on human interference, the IntelliCage system also enables investigators to monitor the multi-dimensional processes in group-housed mice regulated over longer time scales in a straightforward, time-, and cost-effective manner (Lipp, 2005; Lipp et al., 2005; Spruijt and DeVisser, 2006; Wolfer et al., 2012; Kiryk et al., 2020). Such an approach may not only improve throughput, but also provide new insights into the regulation of rodent behaviors that is not as practical with conventional behavioral assays.

In this narrative review, we provide an overview on several commonly employed animal models of SUD, describe the IntelliCage system apparatus and its application in modeling human neurological disorders, and provide in-depth reviews of the related SUD studies utilizing the IntelliCage in the assessment of multi-symptomatic animal physiology, behaviors, and cognitive functions.

Substance use disorder is a chronically relapsing disorder characterized by compulsive and uncontrollable substance-seeking and use, which persists even in the face of negative consequences (Koob and Volkow, 2010, 2016; Uhl et al., 2019). Animal models of SUD are recognized as indispensable tools in defining our current knowledge of the neurobiology and pathophysiology of addiction, and the neuropharmacological aspects of substances of abuse (Koob, 2014; Venniro et al., 2016; Wingo et al., 2016; Müller, 2018; Kuhn et al., 2019). Although animal models may not fully emulate and reproduce the complex human experience, they nevertheless provide means for the researchers to conduct addiction research under highly controlled conditions that may not be possible or ethical to replicate in humans. Earlier animal models of SUD emphasized on the use of operant paradigms in non-human primates and the mechanisms of acute reward. However, recently, these paradigms have been extrapolated and utilized in small rodents (namely, laboratory mouse and rat). Current research has also shifted to include consequent neuroadaptations in long-term or chronic substance abuse paradigms. The use of rodent models, together with the recent advancement, has provided significant new knowledge and understanding in the neurobiology of SUD.

The IntelliCage system (Figure 1) is a social-group environment developed by Hans-Peter Lipp and colleagues of the University of Zurich, Switzerland primarily for the use of Neural Plasticity & Repair, National Centre for Competence in Research (NCCR) research groups (Lipp, 2005; Lipp et al., 2005; Kiryk et al., 2020). IntelliCage is the first fully automated cage system designed for the assessment of spontaneous activity, spatial learning, memory, and cognitive abilities of rodents living in social groups. It allows the individual recording of the long-term and multi-dimensional behavioral patterns of up to 16 animals simultaneously. Various experimental paradigms and protocols can be freely programmed and executed with this system, thus, allowing maximum flexibility in the experimental design. Data are recorded while the animals are housed in the IntelliCage system, which provides considerably more information for analysis compared to any conventional method. The IntelliCage system was designed to circumvent practical issues often encountered with the standard behavioral test paradigms (as summarized in Table 1). The automated generation and collection of data by standardized procedures allow for high data comparability and reproducibility between labs thereafter, permitting a reduced number of animal replications needed to obtain reliable findings. This system also minimizes the need for human or experimenter’s handling, thus, reducing external artifacts that interfere with the animals’ activities throughout the desired period of monitoring (Lipp, 2005; Lipp et al., 2005; Kiryk et al., 2020).

The IntelliCage system is a standard polycarbonate cage (55 cm width × 38 cm depth × 21 cm height) equipped with four triangular operant test chambers (15 × 15 × 21) fitted at each corner (Figure 1D). Animals are identified by the individual subcutaneously injected radio-frequency identification (RFID) tags (known as microtransponder; size: 12 mm × 2 mm) before being released into the IntelliCage system. Entry into each operant chamber is via the ring antenna which detects the animal’s unique RFID tags and records their visits. The round apertures on the walls of each chamber provide free access to water bottles. To date, mice have been reported to be supplied with tap water, sweetened water as a natural reward (i.e., sucrose or saccharin) (Radwanska and Kaczmarek, 2012; Iman et al., 2017; Heinla et al., 2018), aversive liquid (quinine solution) (Knapska et al., 2013; Smutek et al., 2014), or diluted liquid drug rewards (Radwanska and Kaczmarek, 2012; Marut et al., 2017; Skupio et al., 2017; Ajonijebu et al., 2018, 2019; Heinla et al., 2018). Small motorized doors at the aperture can be programmed too close to limit water access according to mice identification, time constraint, and conditioned action. Mice can be trained to perform a fixed or progressive ratio of nose-pokes at the door to allow access to water. The amount of liquid consumed is precisely measured by a lick-o-meter, while nose-pokes are measured by dedicated sensors. Three colored LEDs above the door in each corner provide visual cues. Aversive stimuli, or aversive reinforcement, employ the effective use of bitter tasting solution in one corner or brief air-puffs directed to the head of the mouse, therefore, eliminating the need for a more aggressive, painful, and fear-inducing stimulus (i.e., foot shock, vibration, loud noises). Four small triangular-shaped shelters are placed at the middle of the cage as a form of enrichment on which the mice could climb to reach for food (ad libitum). Shelters are red and transparent, but mice see red color as black so they are willing to hide inside which allows for their observation. The IntelliCage system also provides a continuous recording of the ambient variables (such as temperature and illumination) (Galsworthy et al., 2005; Lipp, 2005; Lipp et al., 2005).

Substantial evidence has documented and recognized the practicality and effectiveness of the IntelliCage system for the short-term and/or long-term cognitive assessment of group-housed rodents. Earlier experiments with the IntelliCage system demonstrated its value for measuring spontaneous and simple conditioned behaviors. Ensuing studies developed and tested numerous parameters/protocols for the assessment of rodent social behaviors and cognitive functions, including spontaneous behavior and spatial navigation, learning and memory-related tasks, circadian activities, and place/drug preference or avoidance tasks. See Kiryk et al. (2020) for a more comprehensive description of the IntelliCage system parameters and protocols developed by approximately 80 research groups on a wide spectrum of rodent behaviors, to date.

Systematic phenotyping of rodent models in automated home-cage systems is presently receiving considerable attention as an effective means of monitoring general and complex activity parameters, as well as detecting perturbations in the neural circuitry function. These complex tasks are achieved while eliminating the tedious and error prone bias of human assessment over extended periods, allowing researchers to address and recognize larger arrays of behavioral outputs than those traditionally assayed. Indeed, the automated home-cage monitoring is a promising frontier for improving translational neurobehavioral research in rodents (Jhuang et al., 2010; Mingrone et al., 2020; Voikar and Gaburro, 2020), including in SUD models. In addition to the IntelliCage system, there are several other automated home-cage monitoring systems available at present (as summarized in Table 3).

Thus far, a growing body of literature has attested the practical utilities and importance of the IntelliCage system in addiction-related mouse models, primarily in alcoholism research. Each IntelliCage corner chamber permits a voluntary oral consumption of liquid reward (via nose-poke), which is useful for self-administration paradigms, as well as the application of operant and Pavlovian conditioning tasks for studying the rewarding properties of various substances of abuse.

The IntelliCage system allows for mimicking different aspects of human behavior to meet the addiction criteria defined in the Diagnostic and Statistical Manual V (DSM-V) of the American Psychiatric Association (Hasin et al., 2013). The DSM-V recognizes that individuals are not all equally vulnerable to developing SUD, and that SUD is a pattern of symptoms. DSM-V sets a diagnostic threshold of 2 or more out of 11 criteria to be met. The IntelliCage allows for examining the following criteria: withdrawal, tolerance, craving, amount of consumption, and time spent on seeking. By analyzing these measures, it is possible to differentiate animals in terms of level of compulsive drinking into low and high drinkers (Radwanska and Kaczmarek, 2012; Stefaniuk et al., 2017; Beroun et al., 2018; Skóra et al., 2020). Radwanska and Kaczmarek (2012) designed the first longitudinal study of animal models of addiction using extensive IntelliCage system parameters in BALB/cJ and C57BL/6 male mice. The study successfully elucidated the behavioral traits associated with alcohol addiction, such as: (i) novelty-seeking (number of corner visits in the novel IntelliCage system); (ii) impulsivity (inability to withhold nose-pokes at rewarding corners); (iii) anxiety (suppression of reward consumption at air-puff associated corners); (iv) motivation and persistence for natural reward (i.e., 10% sucrose); (v) withdrawal; and (vi) relapse in mice for a span of 128 days. This study corroborated the IntelliCage system as a reliable tool for an efficient, high-throughput screening of mice addiction-prone behavioral traits. The data suggested that high levels of anxiety-related traits (i.e., low novelty-seeking, low resistance to punishment, increased compulsivity and impulsivity) predicted addiction-like alcohol drinking in mice (Radwanska and Kaczmarek, 2012). Parkitna et al. (2013) later adapted this alcohol abuse model for a 3-month assessment of ethanol self-administration, abstinence, circadian pattern of chronic ethanol consumption, and cue-induced alcohol relapse. Ensuing studies utilized the IntelliCage system paradigms to develop more complex learning and memory procedures in alcohol addiction models, including intermittent-access schedule (Smutek et al., 2014; Koskela et al., 2018), delay-discounting impulsivity (Szumiec and Parkitna, 2016), motivation for alcohol-seeking behaviors (Stefaniuk et al., 2017), alcohol-deprivation-induced effects (Thomsen et al., 2017), and cue-induced conditioning procedures (Koskela et al., 2018).

Additionally, the IntelliCage system has been used for oral morphine self-administration (0.1–0.5 mg/ml) in a progressive ratio nose-pokes, with the co-administration of dexamethasone [a selective glucocorticoid receptor (GR) agonist], and CPP paradigms to evaluate the GR effects on the rewarding properties of morphine in mice. This model represents a novel approach for investigating the behavioral and molecular mechanisms underlying opioid addiction (Marut et al., 2017). In a follow-up study, Skupio et al. (2017) evaluated mice compulsive morphine self-administration features, including progressive ratio nose-pokes, intermittent-access schedule, enhanced resistance to punishment, withdrawal, and reinstatement of morphine-seeking behaviors, for over 100 days. More recently, to induce and assess symptoms of compulsive cocaine intake, similar paradigms were adapted to the IntelliCage by Ajonijebu et al. (2018). Compared with the control animals, cocaine-addicted C57BL/6J female mice exhibited a higher preference for natural reward and failure to discriminate rewarded from non-rewarded corners, suggestive of significant learning deficits with a prolonged cocaine exposure (Ajonijebu et al., 2018, 2019).

Overall, considering the flexible task design and longitudinal monitoring in a social cage environment, the IntelliCage system indicates invaluable and promising abilities to be a novel model for short-term and long-term SUD studies for other substances of abuse. Therefore, based on this knowledge, our laboratory had successfully designed a new protocol (Iman et al., 2017), which was an adaptation from Radwanska and Kaczmarek’s mice alcohol addiction model (Radwanska and Kaczmarek, 2012), for the study of extended behavioral and cognitive effects of socially interacting Swiss albino mice chronically exposed to the widely abused substances, i.e., morphine, Δ-9-tetrahydrocannabinol (THC), and mitragynine, a major alkaloid of Thai medicinal plant, kratom or Mitragyna speciosa Korth leaves, with psychostimulant and opioid-like properties (Suwanlert, 1975; Ahmad and Aziz, 2012; Hassan et al., 2013; Saingam et al., 2013; Iman et al., 2017). In brief, data collected from our IntelliCage sensitization model (Iman et al., 2017) effectively presented the behavioral and cognitive impairment evoked by the chronic administration of morphine, THC, and mitragynine, which are consistent with the reports from previous studies using conventional animal addiction assays (Justinova et al., 2009; Lu et al., 2010; O’Brien et al., 2013; Harvey-Lewis et al., 2015; Yusoff et al., 2016; Vanderschuren et al., 2017; Hassan et al., 2019).

In addition, the water bottles and programmable conditioning corners in the IntelliCage system would allow researchers to study hedonic behavior and multitudes of spatial learning and memory function in drug-addicted rodents, as well as compulsive behaviors after being punished with an air-puff. Different colored LEDs feature can be utilized to study cue-induced drug memory. The system also differs from conventional tests of SUD in that it examines behavior over an extended period of time. Nevertheless, it is important to note that the main drawbacks of the IntelliCage system include its upfront setting-up costs, regular maintenance, absence of visual tracking of in-cage behavior and social interaction, as well as male aggression issue in socially grouped rodents. However, extracting together the current data and the analysis from previous addiction-related IntelliCage studies, we can assert that the IntelliCage system provides an effective and reliable platform to detect and characterize addiction-related behavioral phenotypes of rodent, chiefly mice, in a social dimension.

In summary, current findings from the SUD mouse model characterize the IntelliCage system as a biologically valid, sensitive, and efficient system in the phenotypic detection of drug effects concurrently across multiple behavioral measures. Moreover, the IntelliCage system permits the assessments of behavior in a controlled environment for socially grouped rodents that minimizes human investigators’ interference. Concurrently, a fully valid model of substance addiction in the IntelliCage system can be further refined with more complex conditioning tasks and parameters, to complement other conventional behavioral assays. Thus, this platform can be beneficial in eliminating the bottleneck in rodent behavioral addiction studies. More importantly, it expands the opportunities to design better preclinical models of SUD to further elucidate the neurobiological mechanisms that contribute to addiction-related behaviors, as well as discovering related treatment options.

II, NY, and MMu conceptualized, drafted, and revised the manuscript. UT and NA contributed to the animal models of the SUD section. AN, JK, MMe, NJ, SM, MS, LK, and MMu critically reviewed the manuscript. All authors made substantial contribution to the review and approved the final manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.