Sensorimotor developmental factors influencing the performance of laboratory rodents on learning and memory

Abstract

Behavioral studies in animal models have advanced our knowledge of brain function and the neural mechanisms of human diseases. Commonly used laboratory rodents, such as mice and rats, provide a useful tool for studying the behaviors and mechanisms associated with learning and memory processes which are cooperatively regulated by multiple underlying factors, including sensory and motor performance and emotional/defense innate components. Each of these factors shows unique ontogeny and governs the sustainment of behavioral performance in learning tasks, and thus, understanding the integrative processes of behavioral development are crucial in the accurate interpretation of the functional meaning of learning and memory behaviors expressed in commonly employed behavioral test paradigms. In this review, we will summarize the major findings in the developmental processes of rodent behavior on the basis of the emergence of fundamental components for sustaining learning and memory behaviors. Briefly, most sensory modalities (except for vision) and motor abilities are functional at the juvenile stage, in which several defensive components, including active and passive defensive strategies and risk assessment behavior, emerge. Sex differences are detectable from the juvenile stage through adulthood and are considerable factors that influence behavioral tests. The test paradigms addressed in this review include associative learning (with an emphasis on fear conditioning), spatial learning, and recognition. This basic background information will aid in accurately performing behavioral studies in laboratory rodents and will therefore contribute to reducing inappropriate interpretations of behavioral data and further advance research on learning and memory in rodent models.

1. Introduction

1.1. Rationale

Laboratory rodents have been utilized as animal models in a wide range of biomedical and translational studies (Denayer et al., 2014). The behavior of laboratory rodents, as measured in various experimental settings, including learning and memory tasks, provides valuable information regarding brain function, as well as the pathophysiology of human cognitive diseases (Peters et al., 2015). In particular, the use of rodent models enables us to delve into precise mechanistic depth in highly social species using exhaustive transgenic techniques. To this aim, the behaviors of laboratory rodents must be understood on the basis of the species-typical, age-specific social demands and individual ontogenetic backgrounds, which may interact to mediate behaviors expressed in an experimental setting. This state will facilitate the rigorous assessment and accurate interpretation of performance in standard behavioral tests.

The purpose of this review is to provide a practical set of descriptions regarding the behavioral development of several fundamental components that cooperatively regulate processes of learning and memory, including sensory modalities, motor behaviors, and emotional and defensive behaviors in laboratory rodents, with an emphasis on mice. Specifically, we will discuss the developmental demands that determine the emergence and expression of learning and memory behaviors, particularly in fear conditioning, spatial learning, and recognition tests, to provide valuable guidance that is required to achieve comprehensive learning and memory studies in neuroscience.

1.2. Background

Laboratory mice and rats possess adaptively tuned sensory and cognitive systems (Berry, 1970; Grant & Mackintosh, 1979) that affect their behavioral performance during tests. Several sensorimotor functions that underlie the emergence and expression of learning and memory behaviors differ based on age (Calhorn, 1963); thus, the behavioral characteristics that represent specific cognitive abilities in these animal models are substantially altered throughout development (Crowcroft, 1973). Compared to humans, laboratory mice and rats rapidly develop during infancy and become sexually (physiological to behavioral) mature at approximately 6–8 weeks of age, with a life-span of approximately 2.5 years (1:35 difference compared to that of humans) (Dutta & Sengupta, 2016). Timescales and developmental stages also differ regionally between the central nerve system and peripheral body parts, as well as between simple biochemical processes (e.g., enzyme kinetics and tRNA turnover) (Giuffrida, 1983; Agoston, 2017) and complex behavioral processes.

Mouse (and rat) infants are born before the eyelids (Tkatchenko et al., 2010) and ear canals (Crowley & Hepp-Raymond, 1966) open, and these sensory modalities are functionally mature by the 3^rd postnatal (P) week, the end of neonatal period, which is referred to as the weaning period (Cramer et al., 1990). Weaning is considered to be the period during which mother-infant interactions decline and suckling is replaced by solid food intake. In nature, rodent pups gradually migrate from their natal nest to a new habitat during the juvenile through pubertal stages (Berry, 1970; Crowcroft, 1973; Galef, 1981). Thus, the juvenile stage of rodents is defined as the period from weaning (approximately P-21) to prepuberty (P-35) (Laviola et al., 2003). The pubertal period appears to be the prominent biological transition from the juvenile to adult system (Spear, 2000), during which gonadal hormones organize the neural circuits and mature reproductive function (Sisk et al., 2003; Sisk & Zehr, 2005). Although the onset of sexual maturation in rodents can be influenced by environmental and social factors (vom Saal, 1989), pubertal onset occurs physiologically at the 4^th week of life [ ] along with a gonadal hormonal surge at approximately P-40–43 in female rodents (Khan et al., 2008), which contributes to the development of sexual characteristics and interest (Adriani et al., 1998; Laviola et al., 1999). On the other hand, the age of male pubertal onset is unclear due to a lack of biomarkers, although morphological changes (i.e., preputial separation along with an abrupt increase in testosterone occurs at approximately P-40 (Korenbrot et al., 1977; Kauffman et al., 2009) and ejaculation are observed at approximately P-50 to 55 (Clegg, 1960). Behavioral maturation including onset of sexual receptivity, copulatory behavior, and adult social behaviors occurs later than physiological modifications, between P-45 and 80 (Arakawa et al., 2008; Meisel & Sachs, 1994; Hull & Dominquez, 2007). During this period, sex differences in particular behaviors along with physiological mechanisms are developed, which are also considerable factors that influence behavioral measurements in most learning and memory tasks.

2. Sensorimotor development sustaining behavioral performance

The most commonly used laboratory rodent species, such as mice and rats, are nocturnal (Crowcroft, 1973); thus, they use sensory modalities in a species-specific manner that differs from humans. The priority of sensory modalities is determined based on the social adaptive demands and ethological background of the species; this factor has become increasingly recognized in laboratory rodent studies on their social, sexual, and cognitive behaviors (Arakawa & Iguchi, 2018). Studies on sensory development have primarily been conducted using rats in the late twentieth century. For example, the comprehensive review of sensory development in rats (Alberts, 1984) has provided principle evidence that sensory function begins in a specific order: tactile (thermal)/chemical → vestibular → auditory → visual during the neonatal period, and these sensory modalities are functionally mature by the 3^rd postnatal week (the weaning period) (Cramer et al., 1990). Duo to the limited available mouse studies on ontogeny and the substantial overlap with rat studies, we will summarize rat studies on sensory development that are relevant to performance in learning and memory behaviors (Fig. 1).

Figure 1.

Neonatal sensorimotor development in laboratory rodents. The pictures of C57BL/6 mice represent physical appearance development from P-0 (at birth) to P-15. Motor control indicates the development of motor control from partial, primitive movement to whole-body integrated movement. Emergence of behaviors observed through the neonatal period includes; huddling (temperature control) and suckling (milk), righting reflex (posture control), ultrasound vocalization (distress call), walking, startle response (defense), behavioral inhibition (defense), running, and jumping. Sensory development includes visual, auditory, tactile, and olfactory functions.

2.1. Olfaction (chemoreception)

Olfaction is a major modality through which rodents are able to detect and identify other animals and environmental features (Brown & MacDonald, 1985). They use olfaction, along with auditory and tactile senses in the major modes of social and nonsocial behavior from birth to adulthood (Brown & MacDonald, 1985; Eisenberg & Kleiman, 1972). Odorant chemicals as social signals can provide the rodent recipients with unambiguous information regarding the individual background, health and social status of the odor-donor animal (Hurst & Beynon, 2004; Arakawa et al., 2008).

Much of this information is based on genetic factors of the odor-donor. Indeed, individuals from highly inbred strains are unable to discriminate between each other via volatile urinary odors (Nevison et al., 2000; Arakawa et al., 2008). While olfactory signals typically present a delay between the signal emission and reception (Eisenberg & Kleiman, 1972), odors may remain for longer durations and thus provide information to a wider range of recipients concerning animals that have been but are no longer present (Brown & MacDonald, 1985; Hurst & Beynon, 2004).

Olfaction represents a fundamental key to sustaining neonatal development in rodents from immediately after birth (Bouslama et al., 2005; Armstrong et al., 2006). Odorants are primarily sensed by the main and accessory olfactory subsystems, which consist of several sensory substructures in the nasal cavity, including the main olfactory epithelium (MOE) and vomeronasal organ (VNO), respectively (Stowers & Logan, 2010). Within these chemosensing systems, distinct populations of sensory neurons project to separate regions of the main or accessory olfactory bulb (Stowers & Logan, 2010). In rat pups (P-4 to 10), damage to the MOE and VNO severely impairs discriminative responses to lactating dams, while selective lesions to the VNO disrupt only the recognition of mammary cues (Singh et al., 1976; Teicher et al., 1984). Neonatal rodents, particularly those born in litters with multiple offspring, utilize olfaction in various activities with siblings and adult caregivers (Alberts, 2007). Newborn pups use olfactory cues to direct and sustain suckling (Teicher & Blass, 1976; Singh & Hofer, 1978), and when the nipples of lactating rats are washed with an organic solvent, the pups no longer exhibit grasping of nipples (Teicher & Blass, 1976). At P-2, rat pups can form a conditioned response associated with the ‘natural’ odor cues of the dam, conspecifics, or home nest (Campbell & Alberts, 1979; Rosenblatt, 1983; McLean & Harley, 2004) as well as show odor (e.g., lemon extract)-associated aversion (Rudy & Cheatle, 1977; Bollen et al., 2012), which indicates that both the MOE and VNO are activated from the neonatal period. In addition, a newly found olfactory organ, the Grueneberg ganglion, which is distinctly located at the anterior end of the nasal cavity (Liberles & Buck, 2006), is functional from P-1 and senses maternal conditions, such as the ambient temperature, carbon dioxide, and lactating scent, to maintain the safety of the pups (Brechbuhl et al., 2008; Schmid et al., 2010).

2.2. Auditory

Auditory signals are particularly important for group-living animals (Knutson, 1998; Arriaga & Jarvis, 2013), and rodents utilize several ranges of ultrasound vocalizations (USVs) for intraspecies communication (Arriaga & Jarvis, 2013). The range of the mouse’s hearing is approximately 1 kHz to 80 kHz (Ehret, 1974) (rats: 200Hz to 80kHz (Kelly & Masterson, 1977)), and there are strain differences in commonly used inbred mice (Zheng et al., 1999). Some degree of genetically determined, progressive hearing loss is also observed in the most frequently used strains, including C57BL/6, BALB/c, and A/J (from 2 months over 1 year), DBA/2 (from 1 month), and 129 (from 3 months)(Turner et al., 2005).

The rodent’s auditory system begins to function later in development relative to its olfactory and taste systems (Ehret, 1976; Rudy & Hyson, 1982; Song et al., 2006). Although rat pups first detect sound at approximately P-4, the external auditory meatus and ear canal, which sustain the functional impact of sound stimuli (Huangfu & Saunders, 1983), do not open until the rat is at least P-12 (Sonntag et al., 2009; Crins et al., 2011). By assessing startle-like reactions to discrete pure-tone stimuli, rat pups show a response at low-frequency ranges (e.g., 1–4 kHz) at P-1 to 4 (Brunjes & Alberts, 1981). Cochlear function in the range of rat ultrasounds (40 kHz) emerges at approximately the time of the ear-opening at P-12 (Crowley & Hepp-Raymond, 1966). Accordingly, the startle response to ultrasonic tones begins approximately 1 day after ear-opening at P-13 (Brunjes & Alberts, 1981).

Rodent pups produce USVs as a distress call (Benton & Nastiti, 1988) or for thermoregulation (Blumberg & Sokofoff, 2001) when separated from their mother and siblings (for comprehensive review see Ehret, 2005). The neonatal cries of the pups alert the lactating female rats and lead to engagement in searching behaviors (Allin & Banks, 1972). USV production in rat pups by dam separation is low at birth, reaches a maximum at P-6 to 8, and then progressively declines to low levels by P-18 (Allin & Banks, 1971; Hahn et al., 1998; Thornton et al., 2005). The dam separation is unlikely necessary to produce USVs in P-18 pups because at this age, they are able to independently maintain their body temperature and switch from milk suckling to solid food chewing (Crowcroft, 1973; Alberts, 2007). Thereafter, rodents produce USVs in several social situations with different frequencies of range (for review see Panksepp et al., 2007; Wohr & Schwarting, 2013), such as play (Knutson et al., 1998) or sexual behavior (Barfield and Geyer, 1972), as well as defensive calls to an external threat (Blanchard et al., 1991; Litvin et al., 2007).

2.3. Tactile (somatosensation)

Rapid-on and offset auditory signals involve alarm or pleasure information during social interactions, which are coordinately used with tactile, whisking contact (Rao et al., 2014). The sensory communication of nocturnal rodents in close range heavily relies on the tactile sense (Welker, 1964). Rodents use their whiskers as a specialized sensory organ to scan the environment and navigate in the dark, as well as communicate with conspecifics (Mitchinson et al., 2007; Hartmann, 2011). The tactile sense via whiskers provides rich 3-D shape and texture information about the surrounding world (Grant et al., 2011). Tactile senses, including skin contact and pain, also provide a biological detection system against external potential threats (Damasio & Carvalho, 2013).

The tactile sense via whiskers and bodily fur/skin plays a significant role in the survival of neonatal rodents, such as thermal regulation and suckling (Sullivan et al., 2003; Alberts, 2007). At an early age, pups are not able to orient their heads or whiskers toward a stimulus source; however, they are capable of passively sensing via whiskers, which is demonstrated in tactile conditional learning in P-3 to 5 rats (Sullivan et al., 2003; Landers & Sullivan, 1999) and electrical recording of neural signals on whiskers even before birth (Maklad et al., 2010). Neonate (<P-10) rodent pups have short, rudimentary whiskers on the snout that sense but are unmovable (Welker, 1964). Neonatal rat pups display a twitch response to passive whisker contacts as early as P-3 (Tiriac et al., 2012), which is a reflex of whisker skeletal muscles, distinguishable from whisking. Whisking is active sensing in which repetitive rapid protraction and retraction of whiskers at high frequencies (>10 Hz) scan the environment (Grant et al., 2011). At the onset of whisking, at approximately P-11, mouse pups are able to walk slowly; however, an integrative motor performance including body coordinative, locomotor movement, is not yet established (Arakawa & Erzurumlu, 2015). At P-13 to 14, a context-dependent active whisking (asymmetrical wiggling whiskers toward scanning objects) with adjustable head and body movement emerges, which is a prerequisite for active scanning of the environment and distinguishing the surface of interest (Grant et al., 2011). At this stage of development, integrative motor coordination is established, accompanied by the eyes opening at approximately P-15 (Sachdev et al., 2003). In parallel with behavioral development, a discrete neural structure, referred to as the barrel system, which represents whisker-sensorimotor control, forms an integrative refined network from the whisker-thalamus-cortex (Erzurumlu & Gaspar, 2012). The tactile sense via whiskers also appears to be important for handling and consuming solid food when the transition from milk suckling occurs at approximately P-15 to 17 (Arakawa & Erzurumlu, 2015).

2.4. Vision

As nocturnal rodents, mice and rats are often considered to not depend on their visual system. Mouse visual acuity is poor, and mice are estimated to be legally blind in terms of human vision (Baker, 2013). While there are clear limits to the visual capabilities of mice, numerous behavioral studies have demonstrated that mice (and rats) display visually guided behaviors, including using visual cues in conditioning tasks (Fanselow & Rudy, 1998) and landmark cues in spatial memory tests (Douglas et al., 2005), as well as evaluating visual abilities in several assays, such as the visual cliff (Crawley, 1999), visual discrimination task (Prusky et al., 2001), or optomotor test (Prusky and Douglas, 2008). Recent research has identified a species-specific retinal photoreceptive system in mice (Peirson, 2018), including the classical rod and cone visual pathway. Compared with humans, mice show an increased sensitivity to ultraviolet (UV) light, while they are less sensitive to longer wavelength light (Sun et al., 1997). UV vision may have an advantage in social communication with conspecifics, as scent marks, which are social signal molecules deposited on the ground (Arakawa et al., 2008), are visible with UV vision, along with smell via odorants. Moreover, visual sensitivity to longer wavelength (red) light is considerable even in albino strains, as mice are certainly capable of responding to red light (>600 nm) with a very low threshold (Peirson, 2018). Regardless of a lack of red-sensible cones in the mouse (rat) retinal system, they are exquisitely sensitive to the intensities of light (Peirson, 2018). Considering that nocturnal rodents use their vision mainly during the dark phase of the daily light/dark cycle for detecting intensity differences in light, it is noteworthy that using red light to black out rodents’ vision may not be effective.

Some popular inbred strains of mice (e.g., Balb/c, A/J, AKR, 129/SvJ, and albino B6) and rats (e.g., Wistar, Sprague Dawley, and Fisher-344, cf. 117 strains of albino rats (Kuramoto et al., 2012) are albino, a condition that results in a deficiency of melanin in the retinal pigment epithelium and thus gives rise to a pink eyes (in addition to neuro-retinal abnormalities (Jeffery, 1998)) and white (discolored) body fur (Jeffery, 1997). Visual ability, as measured by several visual tests, is significantly worse in albino mice than in strains with normal vision (Wong & Brown, 2006; Brown & Wong, 2007). Although there are strain differences regard to the visual performance of albino mice (e.g., better in AKR/J)(Wong & Brown, 2006), visual ability is an unigonorable factor that influences behavioral performance, particularly in visuo-spatial tasks (Yeritsyan et al., 2012).

The visual sense is not important for neonatal rodents, as their eyelids are closed until approximately P-15, when pups start moving around and searching outside of the nest (Sachdev et al., 2003). Their eyes remain in the process of growth to adjust the quality of the image received by the retina, and the growth (expansion) is decelerated after P-40 (Prusky & Douglas, 2003; Schmucker & Schaeffel, 2004). Juvenile mice (C57BL/6) are highly myopic, and beginning at P-40, mice become progressively hyperopic (Tkatchenko et al., 2010) and stable throughout aging (>20-month old)(Lehamann et al., 2012). Using a specific visual based-task, a forced-choice visual discrimination task (Prusky et al., 2001), adult mice are capable of learning the task quickly in which the visual cues are presented in 17 inch monitors with a 70–140 cm distance, which suggests that adult mice are able to use their vision functionally in most of the regular test chambers.

2.5. Motor development

Neonatal motor development, a fundamental component for sustaining normal behavioral expression, including particular forms and postures, has been comprehensively reviewed in landmark papers regarding rats (Altman & Sudarshan, 1975) and mice (Fox, 1965). Protocols for measurements of sensorimotor (Roubertoux et al., 2018) and motor development (Hill et al., 2008; Feather-Schussier & Ferguson, 2016) in mice as well as mouse strain differences (McFadyen et al., 2003; Bothe et al., 2005) are available. After birth, the pups of laboratory rodents immediately begin to spontaneously breathe and search for nipples to suckle (Alberts, 2007). They are able to move their bodies to their mother’s nipples and adjust their body positions alongside their siblings in the nest (Brunjes & Alberts, 1979)(Fig. 1). During the first few days after birth, the laboratory rodent dams maintain contact, or huddle, with the pups, and the littermates remain together in a nest (Grota & Ader, 1969).

Neonatal pups are hairless, and the ear canals and eyes are shut. They stay huddled together and double over each other’s bodies to regulate their body temperatures from birth (Alberts & May, 1984). Rat pups drag and pivot their heads by P-2, despite being incapable of supporting their body with their limbs (Lelard et al., 2006). At P-5, pups develop light, fuzzy dorsal fur, and the head can be raised and angled downwards (Alberts, 2005). From P-7 to 10, the back and belly of the pups is covered in fur, which provides decent body temperature control. At approximately P-10, a stable quadruped posture is attained, and the posterior body can be supported off the ground (Lelard et al., 2006), after which the pups start to walk (Alberts, 2005). By P-11, the teeth begin to erupt. Standing on four limbs to sustain the body emerges at approximately P-12 (Palanza, Parmigiani, et al., 2001; Grant et al., 2011), and fully locomotive behavior is observed shortly afterward (Arakawa & Erzulumlu, 2015). At this point, however, the auditory, visual, and tactile systems are not fully functional, and motor patterns may require complementary development with a maturation of sensory modalities throughout the weaning period, when pups are fully capable of feeding solid foods, exploring outside of the nest, and thermal controlling alone.

3. Stabilization of behavioral performance through development

In most behavioral testing, the subject animals are required to move around in the test chamber or apparatus and explore the stimuli that they encounter. We measure certain interactive movement/activity of animals toward a stimulus as a regular index for learning and memory behavior. The behavior of animals expressed during the tests would indicate certain aspects of their learning and memory abilities, as well as involve their innate components of emotional and defensive reactions. Here, we will discuss how locomotor performance that reflects emotional state and defensive reactions impacts learning performance and address several key factors, including age and sex, that require attention for proper interpretation of behavioral responses in learning and memory tests (Fig. 2).

Figure 2.

Development of sensorimotor, defense, and social behaviors and specific learning tasks including the fear conditioning (associative learning), spatial learning, and recognition, in laboratory rodents; mice and rats. Sensorimotor. Most sensory modalities are functional accompanied by the emergence of motor control before the weaning period (around postnatal day 21). Defense. Behavioral inhibition is observed from around P- 12, when rodent pups begin walking. After the weaning period, several defensive behaviors including exploratory behavior, stretch-attend postures, and defensive burying emerge and reach the highest in the late juvenile stage. Social. During the neonatal period, the primary relationship includes mother-offspring interaction is associated with separation-induced distress call by pups. After the weaning period, rodent pups engage playful interaction with siblings and decrease it afterward. Following the pubertal stage, sexually-matured rodents develop territorial behavior that is relevant to dominance-subordinate relationships and sexual behaviors. Associative learning using a fear conditioning paradigm is observable in the neonatal stage, the emergence of which alters with sensory cues utilized, and from P-17 with artificial auditory cues in the regular conditioning chamber and from P-23 with spatial, contextual cues. Conditioned fear response including tolerance of extinction is enhanced during the late-juvenile to pre-pubertal stage. Spatial learning that is represented by Morris water maze performance emerges around P-17 when intramaze cues are available, and P-23 when exteriormaze landmarks are presented. Impairment in reversal learning in the Morris water maze was demonstrated in late-juvenile mice. Recognition memory performed in a novel object recognition test emerges during P-17 when the interval was very short (5 min), and during before P-22 when the interval was short (e.g., <1 h), and then long-term recognition memory (>24 h) is documented in the mid-juvenile stage.

3.1. Locomotor activity sustaining behavioral performance

Locomotor activity and movement are prerequisites for measuring animals’ performance in learning and memory tests as well as emotional tests, such as the open field, elevated plus maze, and light-dark box (Walf & Frye, 2007). For example, in learning and memory tests, such as novel object recognition or maze tests, animals remaining in the corner of the test chamber for long periods of time with little exploratory movement should not be considered to have learning deficits, but rather suppressed locomotive activity from heightened anxiety. In this vein, several types of behavioral performance in learning and memory tasks also largely depend on locomotor activity in test animals. Most spatial learning tasks, such as the T-maze, Y-maze, and Barnes maze (Vorhees & Williams, 2014), and most exploration-based tasks, such as the object recognition (Ennanceur & Delacaur, 1988) and stimulus discrimination tasks in the Pavlovian paradigm (Fanselow & Rudy, 1998), require test animals to explore the stimuli/environment and exhibit approach/avoidance responses toward the exposed stimuli. If the animal does not move around in the test chamber (particularly observable in aged animals), it is difficult to accurately evaluate their learning and memory abilities. Locomotor activity, along with emotional state, is a considerable factor that influences behavioral performance in rodent models, and these factors are strongly influenced by the developmental stage of the test rodents. Here, we provide a discussion regarding the impacts of locomotor activities in three different age stages.

3.1.1. Juvenile stage

The juvenile stage of rodents is the period in which pups are weaned from their dam and prepare for adult-like social adaptive life and behavior (Laviola et al., 2003). It is noteworthy in the use of mouse strains that this weaning period is determined based on pups’ physical abilities including solid food intake and body thermal control, and thus varies depending on mouse strains. C57BL/6 mice are the standard strain used as a genetic background and in behavioral studies (Blanchard et al., 2013), and they have a shorter period of pregnancy (fetal stage)(18.5 days) than other major strains (e.g., 21 days), which results in preterm birth, leading to the requirement for longer neonatal care by their dam (Murray et al., 2010). Thus, the recommended weaning period for the C57BL/6 strain is 4 weeks rather than 3 weeks as other major strains according to the Jackson laboratories colony management guideline.

Sensory systems and motor abilities, including climbing and swimming are functionally developed at approximately the weaning period; thus, juvenile rodents express a strong tendency to explore an unfamiliar environment (Laviola et al., 2003), as is the nature of the juvenile development (Smith et al., 1996; Pellis et al., 1999). Exploratory behavior typically emerges in males during the juvenile stage, peaks during prepuberty at approximately P-35, and subsequently decreases to adult levels (Douglas et al., 2003; Laviola et al., 2003). Female exploratory behavior emerges during the juvenile stage and is relatively stable over pubertal development (Laviola et al., 2003; Arakawa, 2007). Duo to this ethological nature, the behavior of mice and rats during this stage is characterized as hustle-bustle locomotive movement, linking to the emergence of neural control in activity-based behavior. Several typical innate behaviors associated with exploratory behavior circuits emerge during this period, including risk assessment behavior (Hubbard et al., 2004) and play behavior with siblings (Pellis & Pellis, 1998). These behaviors are a significant background factor influencing specific behavioral performance (e.g., object recognition and social recognition).

Although we will discuss risk assessment behavior in a later chapter (Ch. 3.2.3), playful interactions with siblings and cagemates are particularly exhibited during this age; thus, they show a high motivational drive to express play behavior during the testing and/or after returning to their home-cages. Rats typically exhibit rough-and-tumble play with their siblings including specific postures and patterns (Pellis & Pellis, 1991), while mice are known to exhibit strain-specific patterns of play (Panksepp & Lahvis, 2007). Mouse play behaviors include frisky hops (ventral jumping hops) and fighting and chasing with siblings (Curley et al., 2010). There are sex differences in play behavior in which juvenile male rats engage in playful interaction with siblings more frequently than female rats (Takahashi & Lore, 1983). Male play behavior decreases after puberty and is accompanied by an emerging agonistic interaction that is associated with territoriality (Pellis & Pellis, 1991; Lore & Flannelly, 1977), while female play behavior appears in the juvenile period and moderately gradually diminishes after puberty (Meaney & Stewart, 1981; Palanza, Merley-Fletcher et al., 2001). The function of play behavior remains under discussion (Vanderschuren et al., 1997; Bell et al., 2009); however, it is potentially associated with the integration of the neural circuitry relevant to coordinative behavioral controls and repertories through puberty (van Kerkhof et al., 2013).

3.1.2. Pubertal stage

The pubertal period, which involves a physiological transition (P-30–50) and behavioral formation (P-45 to 80), appears to be the most prominent biological transition from the juvenile to the adult system. Therefore, Spear (2000) defines the broad array of biological transitions as adolescence. Adult-like social behaviors, including pheromonal communication regarding social and health status are also established later (>P-60) in this stage (Arakawa et al., 2011). Several social learning tests such as the social recognition test (Thor et al., 1982) and social odor learning (Engelmann, Hadicke, & Noack, 2011) are predominantly regulated by olfactory cues. A scent communication study in male mice demonstrated that the urinary odor of P-90 mice is distinguishable from that of P-60 mice and induces an adult behavioral reaction (e.g., counter-scent marking) in adult male mice (>P-90)(Arakawa et al., 2007), which indicates that the maturation of pheromonal signals is established later (>P-90) than those of behavioral reaction (>P-60) (Arakawa et al., 2008).

Pubertal mice and rats typically exhibit more anxiety than adults (>P-65) when assessed in various anxiety-related behavior tests, such as the open field, elevated plus maze, and light-dark box (Hefner & Holmes, 2007; Lynn & Brown, 2010; Moore et al., 2011). This fluctuation is shown to be associated with elevated physiological reactivity to emotional cues, such as stress-related hormonal responses (e.g., corticosterone)(Schulz et al., 2009; Spear, 2009) and alternations in limbic circuitry (Andersen, 2003; Schulz et al., 2009). This limbic circuit consists of interconnected brain areas, including the medial prefrontal cortex (mPFC), amygdala, thalamus, and periaqueductal gray, which participate in sensory processing and the generation of social and defensive behavior (LeDoux, 2000; Rosen, 2004). In particular, the interconnection between the mPFC and amygdala plays a crucial role in the mediation of the cognitive processes in emotional/defense behaviors, as well as regulating locomotor activity (Robbins & Amsten, 2009; Marek & Sah, 2018). While amygdala to mPFC projections emerge earlier (Cunningham et al., 2002), mPFC to amygdala projections continue to develop through the pubertal stage (Cressman et al., 2010; Gee et al., 2013). The integration of neural circuits, including dopaminergic neurons interconnected with GABA neurons in the mPFC network occurs coincidently with pubertal fluctuations of locomotor activity and inhibitory behavioral control (Andersen, 2003; Kim et al., 2009; Watt et al., 2009). Several studies on neural processes/circuits illustrate that the inhibitory control of expressing behaviors, such as the extinction process of fear conditioning (Vidal-Gonzalez et al., 2006; Kim et al., 2009), reversal learning (Willing et al., 2016), and attentional shift (Bissonette et al., 2014), is involved in the pubertal integration of mPFC behavioral circuits. Due to these systemic and physiological transitions, behavioral performance in activity-related learning and memory tasks are not consistent; thus, a large variance in behavioral performance is expected during this period.

3.1.3. Sex difference

Following puberty, the rodent behavior is matured and typically mediated by sex-dependent physiological mechanisms (Sisk & Zehr, 2005). It is noteworthy that sex differences of behaviors are observed throughout development, including during the juvenile stage, in which male’s predominance is demonstrated in exploratory activity, playful behavior, and some forms of defense behavior (cf. chapter 3.1.1.). Sex differences in adult behaviors have been comprehensively described in several contexts, one of which is activity-relevant emotional behaviors (Palanza, 2001). Adult females show a less-anxious-like profile compared to males, which is associated with lower defensiveness (e.g., enter more in open arms or arena) in the elevated plus maze and the open field (Johnston & File, 1991; Lucion et al., 1996). Females tend to display greater locomotive activity than males in a novel environment (Elliott & Grunberg, 2005). A higher anxious/defensive property in males is relevant to male competitiveness in social structure and strategy after puberty (Simpson & Kelly, 2012). Males tend to establish social dominant-subordinate relationships and develop territorial behaviors following puberty (Crawcroft, 1973; Grant & Mackintosh, 1963). Dominance and subsequent social defeat in male groups have a significant impact on neuroendocrinology and behavior (Koolhaas et al., 1999). Although there are controversial reports (Martinez et al., 1998), defeated or subordinate animals show a motor inhibition and increased anxiety (Blanchard et al., 1995; Raab et al., 1986). An inconsistency in the effect of social dominance may be partly due to strain differences in the social process (Rzazzoli et al., 2011; Martinez et al., 1998) and uncertain consequences of social interaction between adult males (Blanchard et al., 1995). For example, C57BL/6 mice are able to form social dominance relationships and exhibit discriminable behavior, which are relatively unstable and thus quickly disappear within a few days (Arakawa et al., 2009).

Despite the significance of emotional behaviors, the male advantage in body size, accompanied by muscle strength, is indicated as a sex difference in behavioral performance in several test paradigms. For example, males exhibit higher amplitudes of the startle response than females in the acoustic startle response test (Plappert et al., 2005), while males tend to show higher error steps than females in the ladder rung walking test (Jadavji & Metz, 2008). These differences are largely ascribable to the male advantage in muscle strength and body size, which is also a considerable factor that influences behavioral performance and the sex differences in several learning and memory tasks. In addition, one potential factor that creates a variable in female behavior regarding locomotor (activity) relevant tasks is the estrous cycle. Rodents possess a reproductive cycle of related hormonal levels that is dependent on the function of the hypothalamus-pituitary-gonads axis and lasts approximately 4–5 days in rats and 4–7 days in mice (Champlin et al., 1973). These hormonal fluctuations, including estrogen and progesterone, produce significant behavioral changes throughout the cycle (Sisk & Zehr, 2005). Although an increasing amount of evidence has indicated that the behavioral (quantitative) variability in female rodents stems from the estrous cycle does not vary with those individual variability in male rodents (e.g., Prendergast et al., 2014), qualitative differences according to sex must be considered to be significant biological factors when measuring behavioral performance.

3.2. Defensive behaviors mediating behavioral characteristics during testing

Animals exhibit several types of defensive behaviors that should match the situation with which they are confronted (Blanchard et al., 1995) in a behavioral test setting. Defensive behaviors are vital for inter- and intraspecies interactions and are evoked as innate, inflexible response patterns (Bolles, 1970). However, the response repertoire of rodents varies with age and develops from a simple reaction to more complex, multimodal behaviors (Alberts & May, 1984; Curio, 1993; Fanselow & Rudy, 1998). In the standard learning and memory tasks, the repertoire of defensive behaviors that animals can express is restricted to those exhibited in an enclosed, inescapable situation in which they are unable to flee or hide. In response to such threatening situations, animals may be expected to take several defensive strategies that lead to the most appropriate, adaptive response (Kavaliers & Choleris, 2001; Blanchard et al., 2001). Accordingly, mice and rats tend to exhibit passive defensive behaviors when they are in an inescapable situation (e.g., fear conditioning chamber); active defensive behaviors when they are in a manageable situation (e.g., locomotor-based learning setting); and risk assessment behaviors when the confronted situation is ambiguous.

3.2.1. Passive defensive behaviors

Successful defense against an array of ambiguous situations or the initial phase of a threatening situation typically requires rodents to inhibit their ongoing behavior (e.g., locomotor inhibition) and subsequently search around with a specific patterned posture, interposed with immobility to detect an ambiguous threat and evaluate the risk (Blanchard et al., 2011). Thus, the expression of defensive behavior requires that the first-round reaction is the inhibition of ongoing behavior, typically referred to as behavioral inhibition (Takahashi, 1992). This typical response emerges in neonatal pups at the presentation of threat stimuli, such as an unfamiliar adult male or electric footshock (Canteras et al., 1997; Lichtman & Fanselow, 1990). Unrelated adult males may kill pups to bring the dam back into estrus sooner (McCarthy & vom Saal, 1986), which is an innate threat for pups in a specific age range. Rats aged P-14 inhibit their emission of USVs as a distress call when exposed to the unrelated adult male odor (Allin & Banks, 1972; Noirot, 1972). The behavioral inhibition is observed at approximately P-12 to 14 (Takahashi & Rubin, 1993; Sullivan et al., 2000; Moriceau et al., 2004), whereas at P-7, the pups are quiescent and immobile in a huddle almost all of the time, and they do not change this behavior in response to a threat (Wiedenmayer & Barr, 2001; Wiedenmayer, 2009).

Behavior measured as behavioral inhibition is closely linked with specific reactions in the startle response, which is elicited by the presentation of an intense, unexpected stimulus, such as sound, light, and touch (Davis, 2000). The startle response or reflex is characterized by an exaggerated flinching response to an unexpectedly strong stimulus (Koch, 1999). Rats display a basic startle response from as early as P-12 (Parisi & Ison, 1979), which is likely to appear along with the emergence of behavioral inhibition (P-12 to 14) (Takahashi & Rubin, 1993; Takahashi, 1992). This consistency of behavioral emergences implies that both the startle response and behavioral inhibition are regulated by a similar underlying mechanism, in parallel with a similarity in behavioral features. At P-21, rats display decreased immobility in response to unfamiliar adult males, while they maintain immobility in response to natural predator odors, such as cat odors (Wiedenmayer & Barr, 2001). Predation is an innate threat for rodents across development, while the threat of unfamiliar males is age-specific (Wiedenmayer, 2009). The immobile response to predatory stimuli is observed throughout development, including P-14 (preweaning), P-26 (juvenile), P-35 to 45 (pubertal), and adult rats (Kabitzke & Wiedenmayer, 2011).

Fear potentiation of the startle response, a classical conditioning component of the startle stimulus, requires two processes in the neural circuit, including the startle reaction and fear-conditioned learning, which can be observed in P-18 rats but not in P-14 rats (Weber et al., 2003). Similarly, juvenile P-23 rats show behavioral inhibition as expressed by locomotor suppression, followed by risk assessment behavior (Hubbart et al., 2004) or analgesia (Weidenmayer & Barr, 1998, 2001) in response to a predatory odor, while preweaning P-18 rats do not exhibit this response (Hubbart et al., 2004). This integrative formation of complicated behaviors appears in a set of specific defensive responses with a sequence of learning and memory behavior during the juvenile stage.

The freezing response has been found in several contexts of learning and memory tests and is one of the major reactions that animals exhibit toward an unfamiliar stimulus that is presented in a closed chamber. In the fear-conditioning paradigm, the freezing response that occurs following the pairing of an aversive stimulus (e.g., footshock) with a conditioned stimulus is the most common measurement that represents fear-related learning and memory (Fanselow, 1994). Freezing in the fear conditioning paradigm is defined as the absence of all movement aside from that required for respiration, without regard to posture (Bolles & Riley, 1973; Fanselow, 1980). If the animal stops moving in the test chamber for a while, it will be simply counted as a freezing, which is problematic. For example, a high dose of serotonergic agonists induce tonic immobility in several species (Henning et al., 1986; Cryan & Mombereau, 2004), which would not be due to enhanced memory performance or involved in any innate defensive means.

In the ethological context, the freezing response is a natural defensive behavior in which an animal is inhibited from continuing a behavior and becomes motionless in a specific body position with a suspended arch-back body and straightened tail away from the ground (Blanchard & Blanchard, 1972; Blanchard et al., 1986). Behavioral inhibition and the immobile response are involved in a process of freezing behavior and discriminable from the matured freezing response with a specific posture. A neural circuitry study illustrated that the mPFC, a limbic structure, is involved in the regulation of conditioned freezing and the innate freezing response in adult rodents (Vidal-Gonzalez et al., 2006; Corcoran & Quirk, 2007). The inactivation of the mPFC by a GABA receptor agonist attenuates freezing in response to predator odor in prepubertal (P38 to 42) or adult rats, whereas it has no effect on the “freezing-like” response in P-14 and P-26 rats (Chan et al., 2011). These findings may imply that the mPFC is not involved in the regulation of the “freezing-like” response in younger rats in which the neural structure of the mPFC appears to be reconstructed (Benes et al., 2000). To this extent, one questionable concern is whether preweaning rats are capable of exhibiting a freezing response with a specific posture or are simply displaying startle-like behavioral inhibition. There is a technical difficulty that limits the ability to answer this question, as the video-based automated “freezing” detection method that was widely used in current research cannot distinguish behavioral inhibition from a matured set of freezing responses (i.e., crouching (Blanchard and Blanchard, 1969; Arakawa, 2019)). Given the hypothesis that the freezing response contains the integration of several elements of defensive behaviors, including behavioral inhibition and the immobile response, interposing with risk assessment behavior, it is likely that a matured freezing response, which is regulated by the integrated neural circuit including the mPFC, may emerge around prepuberty, by which other elements of defensive behaviors involved in the freezing response are developed. In this context, the maturation of freezing behavior that contains the sequence of behavioral patterns and postures significantly influences the performance in several learning and memory tests.

3.2.2. Active defensive behavior

The patterning of defensive behaviors contains elements of functional behaviors that develop from minimal passive reactions that inhibit ongoing behavior or movement to integrated multimodal behaviors that require certain action to the environment (Blanchard et al., 2011). For example, biting and attacking experimental stimuli (e.g., biting a pellet feeder and kicking a novel object) are observed in adult, particularly male, mice and rats, which may be associated with aggressiveness (Wagner, Beuving, & Hutchinson, 1980). These proactive styles of defense are characterized by active control of the environment, including active approaches to threat stimuli and directly reacting to threatening stimuli (Koolhaas et al., 1999; de Boer & Koolhaas, 2003).

Development of proactive defense behavior has been reported as agonistic behavior (Terranova et al., 1998) or marble burying behavior (Boivin et al., 2017). Defensive burying behavior is a proactive style of defense and is demonstrated in the shock-probe burying test (de Boer & Koolhaas, 2003). The test, which was developed by Pinel and Treit (1978), involves confronting subjects with a probe connected to a shock source that is presented through a hole in the chamber wall. When contact with the electrified probe is made as a risk assessment behavior, the animals displace their bedding material with typical alternating forward-pushing movements of their forepaws directed at the probe. Defensive burying behavior in male rats appears infrequently in the weaning period (approximately P-21) and gradually increases until the prepubertal stage (P-35 to 40), decreasing steadily thereafter to the adult level (Lopez-Rubalcava et al., 1996; Arakawa, 2007a). Female rats showed no changes in burying behavior from the juvenile stage (P-26) to adulthood (P-80)(Arakawa, 2007a), and no sex difference in defensive burying is observed in adult rats (Falconer & Galea, 2003) and mice (Sluyter et al., 1999).

A higher tendency in the expression of active defensive behavior may lead to offsetting effects on other types of defensive behaviors, such as freezing or immobility (De Boer & Koolhaas, 2003). For example, if animals find an object or environment that would enable the expression of active defensive behaviors, such as the bedding materials (e.g., burying) or levers (e.g., biting) in the conditioning chamber, the presentation of electrical shocks may induce active defensive behaviors rather than freezing responses, as a behavioral (coping) strategy (Carpenter & Summers, 2009). High aggressiveness is a considerable factor that influences behavioral performance in learning and memory tests, as biting and attacking a presented stimulus disrupts the accurate sampling of the measurable reactions during tests. Moreover, active defense is associated with higher escape tendencies and induces mouse jumping on and/or to the outside of experimental chambers or escaping from experimenter handling. Furthermore, the defense strategies that animals innately possess or acquire through learning sessions are important factors, especially when animals experience more than two test situations. For example, when animals are exposed to a fear-conditioning test, in which animals learn a passive defense strategy (i.e., freezing) as an effective behavioral pattern, this experience will influence the behavioral strategies the animal will take in subsequent test settings. Spatial memory tests, such as the Morris water maze or T-maze, require animals to adopt an active locomotive (swimming) strategy for better performance. Therefore, animals’ previous experience in the fear-conditioning test, in which they develop a passive strategy, would interrupt their behavioral performance in subsequent behavioral tests. These innate or learned strategies individual animals acquire are the major factors that need to be estimated to measure learning and memory behaviors (e.g., Iguchi et al., 2015).

3.2.3. Risk assessment behavior

Ethological studies of semi-natural habitats have illustrated that the juvenile stage of rodents is the period of migration from their natal habitat (Calhoun, 1963; Crowcroft, 1973), which requires postulating the emergence of exploratory behavior and adaptive defensive responses. Therefore, fundamental components of defense behaviors emerge and are integrated during the juvenile stage. When pups begin to exhibit robust investigatory activity outside of their nest, the first vital reaction to an ambiguous threat or a potential sign of danger that is encountered is to assess a risk of the potential threat in the surrounding environment (Blanchard et al., 1990).

Risk assessment behavior is observed as the first component of defensive behavior (Blanchard et al., 2011). In this mode, animals confronted with a potential threat display a series of particular postured motions, including a back-and-forth movement that utilizes a low back and stretched form, referred to as a stretch-attend posture. Stretch-attend postures are displayed for searching a potential threat stimulus or danger sign using all relevant sensory modalities (Blanchard et al., 1994), interspersed with periods of immobility that reduce the likelihood of being detected by the potential danger (Blanchard et al., 1990). Stretch-attend postures can be observed in rats and mice during the juvenile period (at approximately P-23) (Hubbard et al., 2004) and show no change in level throughout male adulthood (Macri et al., 2002; Arakawa, 2005; Marques et al., 2008). A sex difference in the risk assessment behavior when confronted with an ambiguous situation has been reported in mice in a strain dependent manner (e.g., Holmes et al., 2003; Marques et al., 2008; O’Leary et al., 2013). Stretch-attend postures are typically observed accompanied by exploratory behavior, when animals are initially confronted with an unfamiliar stimulus/situation in the test chamber, such as the novel object recognition test.

4. Optimization of developmental factors for measuring learning behavior

The behavioral performance of rodents during learning and memory tests is governed by a multitude of factors, including sensorimotor ability that sustains behavioral performance and innate behavioral patterns, and strategies that mediate behavioral expression during the tests. These factors underlying learning behaviors emerge and mature throughout the development of rodents, which must be taken into account when measuring learning behavior. Here, we briefly summarize the developmental factors of sensorimotor abilities and behavioral strategies underlying learning and memory behaviors in several popular test paradigms, including the fear-conditioning test (associative learning), spatial memory tests, and recognition tests.

4.1. Associative learning

The fear-conditioning paradigm contains key elements of associative learning that enable an organism to anticipate events (LeDoux, 2000). Rapid and robust learning of the association with an unconditioned (fear) stimulus is made with a proximal (assemble) cue that is presented as a conditioned stimulus (CS) and a contextual, spatial cue to which the animal is exposed in contextual learning. Associative learning can be observed in neonatal animals when a specific innate reaction or reflex in the natal cage environment is chosen as a conditioned response, such as suckling milk or nipple (Campbell & Alberts, 1979) or nestling and huddling with ambient temperature (Bollen et al., 2012). However, the performance of associative learning in the standard chamber (e.g., fear conditioning chamber) requires expressing a certain level of locomotor activity outside of the natal environment accompanied by sensory and cognitive abilities in recognizing exposed stimuli and contexts (Fig. 3). Neonatal mouse pups are capable of expressing fully locomotive movement at approximately P-14–15, in which most sensory modalities are also available (e.g., active whisking for tactile exploration and eye lids and ear canals are open). Therefore, performance of the cue-dependent learning in the standard chamber that does not require the spatial ability emerges at approximately P-17 (Rudy, 1993; Rudy & Morledge, 1994). During the juvenile stage, a repertoire of behavior increases when the neural circuitry undergoes fine-tuning and multimodal integration (Benes et al., 2000; Willing & Juraska, 2015). This is documented in learning tasks that require spatial abilities, which are strongly linked to exploratory locomotor behavior and typically emerge in the juvenile stage (Smith et al., 1996). It has been reported that before weaning, pups exhibit behavioral inhibition as a type of freezing response, while they are incapable of forming an association with contextual spatial cues (Revillo et al., 2015). The performance of spatial-dependent variants, such as contextual fear conditioning has been reported to emerge at approximately P-23 (Rudy, 1993; Rudy & Morledge, 1994; Burman et al., 2009). It is noteworthy that (i.e., male) mice at weaning age exhibit enhanced, nonspecific freezing responses in both conditioned and non-conditioned context (Arakawa, 2019). Rodents in this age may show a heightened freezing response in the conditioned context; however, we must evaluate whether the test rodents exhibit clear differences in the freezing response between conditioned context and non-conditioned context (e.g., missing in Park et al., 2017).

Figure 3.

Considerable factors for examining the fear conditioning in mice. Functional sensory modalities are required for detecting conditioned stimulus (CS). Tactile (pain) sense is relevant to reception of unconditioned stimulus (US). Development of behavioral patterns including freezing (i.e., passive defense) and extinction has been reported.

The pubertal transition appears in the performance of the fear-conditioning paradigm (Pattwell et al., 2011, 2013; Hunt et al., 2016). Juvenile (P-28 to 35) mice exhibit enhanced acquisition of a cued fear response compared to pubertal (P-42 to 49) and adult (P-56 to 63) mice, as expressed by a higher freezing ratio of juvenile mice toward conditioned stimulus exposure (Hefner & Holmes, 2007; also Ito et al., 2009). Juvenile (P-28 to 35) mice and rats also showed sustained fear responses during extinction compared to other age groups (Hefner & Holmes, 2007; Kim et al., 2009; McCallum et al., 2010), which indicates that a sustained fear response is diminished following the late-juvenile period. A study using pharmacological inactivation (GABAA agonist) suggests that the mPFC is not involved in the extinction of conditioned fear in P-17–18 rats, while it becomes responsive in P-24–25 rats (Kim et al., 2007, 2009) and significantly involved in P38–42 rats (Chan et al., 2011), as in adult rats (Herry & Garcia, 2003). A similar finding reported that NMDA-receptor antagonism in the mPFC in P-26 rats blocked the acquisition of reversal learning in the T-maze discrimination task (Watson & Stanton, 2009). These studies support the hypothesis that an integration of the neural network, including the excitatory/inhibitory balance in the mPFC, appears to be activated from the juvenile stage and is accomplished during puberty (Pattwell et al., 2013; Willing & Juraska, 2015; Hurt et al., 2016).

Several reports have indicated that males exhibit a higher level of freezing response than adult females in the fear conditioning context (Maren et al., 1994; Pryce et al., 1999; Kosten et al., 2005). This finding is ascribable not to an impaired learning ability but is rather likely due to decreased defensiveness in females or a sex difference in the defensive strategy expressed in a threatening situation (Arakawa, 2019), although no clear mechanism underlying this sex difference is provided. In addition, a practical problem for using mouse strains in the fear conditioning paradigm is the progressive hearing loss that is observed in several commonly used strains (Turner et al., 2005). A recent genome-wide mouse phenotyping project reported that 67 candidate hearing loss genes, of which 52 genes had not previously been linked with hearing loss, were identified from 3,006 strains of knockout mice (Bowl et al., 2017), which is a potential issue in the use of knockout mice for the fear conditioning test. Typically, fear conditioning employs a tone auditory stimulus as a neutral CS, in conjunction with electrical footshock as an unconditioned fear stimulus (US)(LeDoux, 2000). In some inbred strains developing hearing loss, middle and low frequency tones (4–8 kHz) remain salient compared to high frequency (>12 kHz) hearing loss (Willott et al., 1998). A potential way to avoid the hearing loss issue is likely to use white noise that contains a wide range of frequencies in equal amounts as a CS for the fear conditioning. However, a high-intensity white noise (particularly >87.5 dB) spontaneously develops a freezing response in C57BL/6 mice (Mollenauer et al., 1992), which indicates that a presentation of white noise as the CS can induce an unconditioned response. These impacts of noise and relevant environmental factors on the performance of laboratory rats are substantially reviewed in Castelhano-Carlos and Baumans (2009).

4.2. Spatial learning

Some spatial memory tasks, such as maze tests, require a physical maturation of rodents that is associated with locomotor performance. The Morris water maze is a widely used method for assessing spatial learning and memory (D’Hooge & De Deyn, 2001), in which animals have to find a submerged platform in a pool based on spatial location cues surrounding the pool (Fig. 4). Performance in the water maze requires swimming abilities, including complex motor-coordination, and spatial recognition, including cue- and context (spatial)-related memory formation. In rats, swimming and proximal cue (intermaze landmarks) learning can be observed at P-17, while water maze learning with distal cue (extramaze landmarks) utilization develops immediately after weaning (P-20) (Rudy et al., 1987). In mice, a report demonstrated abundant swimming abilities and spatial learning performance in P-22 mice, similar to those of P-65 adult mice (Chapillon & Roullet, 1996), regardless of the ethological view that the mouse is not a semiaquatic species with limited swimming ability in contrast to rats (Peters et al., 2015). In this context, a recent study demonstrated that mice exhibit spatial learning in the Morris water maze when training began at P-35 (Salmaso et al., 2012); however, it is noted that the escape latencies reported for juvenile mice in the water maze do not decrease as rapidly as reported for weanling rats (Jett et al., 1997; Akers et al., 2011) or adult mice (Varvel & Lichtman, 2002; Malleret et al., 1999, Logue et al., 1997).

Figure 4.

Considerable factors for examining spatial learning in mice. Visual detection of intra- and extra-landmarks surrounding the apparatus is required. Functional tactile sense is prerequisite for sensing a (visually) hidden platform in the Morris water maze. Intact locomotive (swimming) abilities are required. Developmental fluctuation in exploratory behavior and response to the reversal learning has been reported.

One potential factor that remains to be discussed is visual abilities in rodents. Spatial learning tasks such as the Morris water maze require animals to find and use spatial landmarks “surrounding” the pool as exterior distal cues. The distance of the (swimming) mice to these landmarks is calculated to be approximately 40 to 160 cm (*the regular mouse pool diameter is approximately 90–120 cm). Mouse vision is poor with particularly high levels of myopia in juvenile mice (Tkatchenko et al., 2010), which indicates that juvenile mice may not be capable of using the surrounding landmark cues in the Morris water maze, which is relevant to unfavorable performance in juvenile mice. Furthermore, some inbred strains of mice suffer from age-related blindness due to gene mutations (e.g., DBA/2J, C3H, FVB/NJ, and SJL/J)(Chang et al., 2002), which is a significantly factor when measuring behavioral performance in visuo-spatial navigation tasks (Charman & Mactutus, 2002).

There is no pubertal difference in the acquisition of spatial memory in the water maze; however, reversal learning (changing the location of the goal platform) following the acquisition in the water maze was impaired in pubertal (P-40 to 43) rats but was present in postpubertal (P-47) or young adult (P-60) rats (Willing et al., 2016). The performance of reversal learning, as well as extinction following fear acquisition, requires exerting inhibitory control over the acquired response (Kim et al., 2009). The reversal learning performance in the water maze task is mediated by mPFC associated function (de Bruin et al., 1994; Lacroix et al., 2002), which suggests a close linkage between the immature mPFC network and age-typical fluctuation in mPFC-mediated inhibitory control. Thus, the inhibitory control of behavior particularly regulated by mPFC function is established during the pubertal period, which is relevant to fine-tuning the process of integrated behavior (Andersen et al., 2000; Lacroix et al., 2002).

Sex differences in learning and memory in laboratory mice have been less consistent than those observed in laboratory rats (Bettis & Jacobs, 2009). Studies using rats have shown a consistent male advantage in spatial learning tasks (Kelly et al., 1988; Brandeis et al., 1989; Markowska, 1999; D’Hooge & De Deyn, 2001), in contrast to mouse studies (Barnhart et al., 2015). Although one study reported a significant female advantage only at older ages (Frick et al., 2000), most studies using mice have reported no advantage for either sex in learning or memory tasks (Berger-Sweeney et al., 1995; Lukoyanov et al., 1999; Bucci et al., 1995; Lamberty & Gower, 1988). A meta-analysis of sex differences in spatial memory performance based on extracted data from a large amount of articles demonstrated large, reliable advantages for male rats in the water maze data, while no consistent sex differences were apparent in mouse models (Jonasson, 2005). The behavioral strategies that are shown in spatial navigation tasks are potential key factors in understanding sex differences in behavioral expression. Male rodents (rats) are better at some spatial learning tasks, while female rodents more rigidly learn cue-dependent memory tasks (Perrot-Sinal et al., 1996; Baldan Ramsey & Rittenger, 2010). The behavioral strategies that individual rodents employ in a specific experimental setting differ based on how they recognize the confronted situation, including relationships with intra- and extra-apparatus landmarks, cues, and sizes.

The estrous cycle of female rodents would be a considerable factor that influences the performance regarding (spatial) learning and memory. There are fluctuations in both the anatomy and physiology of the hippocampus across the estrous cycle of female rats (Foy, 2011; Koss & Frick, 2017). While some performances of female mice have substantially varied depending on the phase of the estrous cycle, such as the open field behavior, tail flick pain response, and tail suspension depressive behavior (Meziane et al., 2007), memory performance in the water maze or radial maze showed inconsistent cyclic variations (Berry et al., 1997; Frye, 1995; Healy et al., 1999; Frick & Berger-Sweeney, 2001). Currently, it is assumed that cyclic variations may be relevant to the learning strategy (Baldan Ramsey & Rittenger, 2010; Bettis & Jacobs, 2009; Ter Horst et al., 2013). For example, females in proestrus performed better than those in estrus in the cue-dependent (intramaze landmarks) water maze task (Savva & Markus, 2005), while the performance of females was least efficient in proestrus in the spatial version (extramaze landmarks) of the task (Warrens & Juraska, 1997). Moreover, studies using physiological doses of gonadal hormones have consistently demonstrated that estradiol injection enhanced memory for platform location in the water maze (Harburger et al., 2007) as well as memory performance in other (non-spatial) types of tests, such as novel object recognition (Luine, 2014). There are two types of estrogen receptors that have been identified (ERα and ERβ), and the expression patterns and site-specific function of these receptors may play a critical role in estrogen-mediation in learning and memory behavior in female rodents (Ervin et al., 2013).

4.3. Recognition and behavioral strategies

Recognition memory tasks have been introduced relatively recently to provide a method of testing episodic (nonspatial) memory in rodents (Ennaceur & Delacour, 1988). These tasks are based on the spontaneous tendency of rodents to spend more time exploring a novel (unfamiliar) object than a familiar object. The choice to explore the novel object reflects the use of learning and recognition memory. This concept of recognition testing is prevalent; thus, several other versions of recognition tests have been developed. The habituation/dishabituation test has been developed to assess the social recognition abilities (Thor et al., 1982), and it consists of two phases. In the habituation phase, test animals are confronted with a social stimulus that is initially unfamiliar over several trials. The test animals will investigate with sniffing and contacting a social stimulus, as an innate novelty assessment, and decrease the investigation time over trials, i.e., habituation. Following habituation, in which the investigation time reaches a plateau, a stimulus is changed to a novel social stimulus. If the test animal detects a difference between the previous and present social stimuli, the investigation time recovers to the initial level i.e., dishabituation. The other modified version of recognition testing has been applied to assess spatial memory in rodents with changing the location of the objects (Ennaceur, Neave, & Aggleton, 1997). The protocol of recognition tests has also been used to assess whisker-tactile discrimination ability in mice, with changing only the texture of the objects (Arakawa et al., 2014). To this extent, the recognition tests have been theoretically considered nonspatial memory tasks; however, studies with brain lesions and pharmacological inactivation have reported that brain regions, including the hippocampus and perirhinal and entorhinal cortices, are involved in the process of recognition (Hammond, Tull, & Stackman, 2004; Wilson et al., 2013), which suggests that the spatial components of memory are also processed during the performance of the recognition task.

Novel object recognition memory, which requires both cue- and context-dependent recognition, is developmentally regulated. Object recognition in animals is measured by the difference in the exploration time of novel vs. familiar (previously exposed) objects (Antunes & Biala, 2012); thus, the memory retention (e.g., difficulty in tasks) can vary with the duration of the interval between the first exposure and second-choice trials (Ennanceur & Delacour, 1988). In this context, P-17 rats are able to show novel object recognition over a very short (5 min) interval (Kruger et al., 2012; Westbrook et al., 2014), which, in fact, indicates that performance in novel object recognition requires a certain level of locomotive (exploratory) activity as well as memory retention ability. Weaning (approximately P-21) rats exhibited novel object recognition only in short-term intervals (<1 h), while juvenile (P-29 to 40) and adult (P >50) rats demonstrated remembrance of the novel object over 24 h (Reger et al., 2009; Jablonski et al., 2013). This finding indicates that weaning rats are incapable of forming long-term memory (cf. >24hr), which is consistent with the finding in contextual fear conditioning in rats (Rudy, 1993). P-18 rats show a certain amount of conditioned freezing when tested immediately after conditioning; however, they show less conditioned freezing when the retention interval is 24 h compared to P-24 rats (Rudy & Moriedge, 1994). In addition to the lack of long-term memory representation in weanling rats, the object exploration time during the novel object recognition test (novel vs. familiar) was higher in P-26 rats than in P-21 rats (Jablonski et al., 2013), which indicates that performance is also likely affected by the emergence of exploratory behavior because a longer exploratory behavior is associated with the potentially decreasing detection threshold of differences between the investigation times of familiar and novel objects.

Although a distinct influence of pubertal development on learning is demonstrated in other types of learning tasks, memory performance in the novel object recognition tests does not differ during the pubertal stage (Reger et al., 2009). Sex differences in the performance of the novel object recognition test using mice are highly inconsistent (see review in Koss & Frick, 2017), even where data are obtained from the same mouse strain (C57BL/6) (Frick & Gresack, 2003; Benic et al., 2006; Bettis & Jacobs, 2009). The somewhat different procedure (e.g., definition of behavior measured) and apparatus (including objects) used may contribute to these inconsistencies. A lower defensiveness in females may influence exploratory performance in certain object recognition tasks. One study demonstrated a female advantage only when very similar shaped objects for both novel and familiar objects were used (Bettis & Jacobs, 2012). Interestingly, sex differences appear in the social version of recognition memory tests in which the reduction in investigation time over repeated trials was substantially smaller in female mice than in male mice (Karlsson et al., 2015), a finding that has also been shown in rats (Bluthe & Dantzer, 1993; Choleris et al., 2009; Markham & Juraska, 2007; Reeb & Tang, 2005). Performance in the social recognition test relies on an innate drive of novelty investigation; thus, a young female or juvenile male is typically used as a stimulus animal to avoid aggressive interaction between the test and stimulus animals (Young, 2002). Therefore, a sustained approach in females would be relevant to a heightened drive of social contact, cf. amicability, in females with their (non-aggressive) stimulus mice. To this extent, the sex difference in social investigation in the recognition memory test would be ascribable to sociality rather than memory performance.

5. Conclusion

Behavioral studies using animal models are important for unraveling the neural mechanisms of behavior and human diseases. One significant advantage in the use of laboratory rodents, particularly mice, is the growing use of exhaustive transgenic techniques. These studies using mice as a genetic tool presuppose a firm understanding of the standard control condition of mouse behavior for applying systematic comparisons. However, most strains of mice are inbred in which all individuals in the same strain are genetically identical. This is a completely unnatural situation for laboratory rodents that were originally wild-captured and possess substantial variety in their gene pool, and it may cause a residual effect on behavioral formation through developmental processes with maternal or sibling interactions. These specificities in the use of laboratory mice as an animal model require consideration of the ontogenetic factors that influence ‘standard’ behavioral performances during the tests, which must take into account how age-related physical and sensory abilities and social demands interact with the emergence and functioning of behaviors. The behavior of rodents exhibited as a performance in learning and memory tests contains several elements of microbehaviors that represent each neural process underlying behavior, including pattern detection via sensory modalities, assessment with innate exploratory (defensive) strategies, and particular reactions governed by age- and sex-specific processes. These factors do not receive sufficient attention in the vast majority of research using rodent models and are thus crucial in evaluating the precise meanings of behavior expressed in an experimental setting, which makes a firm commitment to the accurate interpretation of behavioral performance in learning and memory tests in rodent models.

Figure 5.

Considerable factors for investigating recognition memory in mice. Detection and discrimination of the objects (or social or unsocial cues) via available sensory modalities (e.g., visual, tactile, or olfactory) are necessary. Some developmental fluctuation of investigatory (locomotor) behavior is known.

Highlights:

• Summary of sensorimotor development sustains behavioral performance in rodents

• Age-related change in locomotor activity stabilizes learning and memory behavior

• Discussion on defensive strategies of rodents that is involved in learning behavior

• Optimizing age-related factors for performing cognitive tests in rodents