Characterizing behavioral effects of early-life stress in an animal model of auditory processing

Abstract

Animal models provide significant insight into the development of typical and disordered sensory processing. Such models have been established to take advantage of physical and behavioral characteristics of specific species. For example, the Mongolian gerbil is a well-established model for auditory processing, with a hearing range similar in frequency to that of humans and an easily accessible cochlea. Recently, early-life stress (ELS) has been shown to affect sensory processing in auditory, visual, and somatosensory neural regions. To understand the functional impact of ELS, it is necessary to evaluate the susceptibility of sensory perceptual abilities to this early perturbation. Yet measuring sensory perception – e.g., using operant conditioning – often concurrently involves animal behavioral elements such as attention, memory, learning, and emotion. All of these elements are well-known to be impacted by ELS, and may affect behavioral measurements in ways that could be misconstrued as sensory deficits. Thus, it is critical to characterize which behavioral elements are affected by ELS in any sensory model. Here we induced ELS during a developmental time window for maturation of the auditory cortex in Mongolian gerbils. We conducted behavioral measures in juveniles, a developmental age when ELS is known to impair the auditory pathway. ELS had no effect on overall activity but reduced anxiety-related behavior, impaired recognition memory, and improved spatial memory, with some sex-specific effects. These effects may influence the ability of gerbils to learn and retain operant training, particularly if anxiety-provoking reinforcement is used.

Introduction

Early life experiences shape neural circuitry, which in turn influences behavior [1–3]. Stress during this postnatal time period has widespread effects on physiology and behavior related to executive functions, memory, learning, and anxiety, which are all well-studied in murine species [4–6]. Evidence is growing that sensory processing is also susceptible to the effects of early-life stress (ELS) [7–12]. One challenge to assessing the effects of ELS on sensory systems is that behavioral measures of sensation, whether evoked by natural stimuli or measured in trained animals, may be influenced by top-down cognitive and emotional factors which are directly altered by ELS. Thus, to best interpret the effects of ELS on sensory systems, it is important to characterize these higher-order factors in any species new to the field. While rat and mouse models have many benefits, multiple other species are well-established model systems for specific types of sensory processing. These other species offer an opportunity to evaluate the effects of stress on sensory processing by establishing them as additional ELS models.

Our lab has recently developed such an ELS model to study auditory processing in the Mongolian gerbil (Meriones unguiculatus), demonstrating that ELS impairs perception and neural encoding of vocalizations and sound elements intrinsic to speech [12–14]. The gerbil is a well-established model for auditory processing, with over 1,200 extant publications on the topic [15–17]. The gerbil audiogram, unlike that of mice or rats, is comparable to that of humans, and gerbils have a longer time window of physiological maturation than mice or rats [18,19]. These characteristics produce a valuable model system for studying developmental perturbations such as hearing loss or ELS on the auditory system [20–22]. This approach requires reliable measures of sensory perception.

Operant conditioning is the gold standard method for measuring the sensory perceptual abilities of animals. Performance on operant tasks involves attention, learning, memory, motivation, decision-making, and emotional state. These elements and their neural substrates (e.g., amygdala and prefrontal cortex (PFC)) are known to be affected by stress, and to impact sensory cortices via top-down activation. These stress-induced effects could confound the interpretation of behavioral measures of perception. For example, anxiety and attention activate amygdala and prefrontal cortex, which shift auditory cortical responses and behavioral acuity to sound [23–27]. Indeed, acute stress itself enhances both attention and ACx responses to auditory stimuli, resulting in more sensitive and selective hearing [28–31]. Thus, to best study ELS in animal models of sensory processing, it is critical to establish a baseline understanding of how ELS impacts higher-level functions.

Characterization of ELS in a new model is essential, because stress induces variable effects. ELS can dysregulate the feedback loop of the hypothalamic-pituitary-adrenal (HPA) axis, altering the release of glucocorticoids and inducing either hyper- or hypo-reactivity [32–34]. This can bias individuals towards enhanced or suppressed behavioral responses. Importantly, a chronically active HPA axis can lead to differing neural and behavioral outcomes depending on factors such as the stage of development during the stressor, manner of stress, sex, neural region involved, and the age at which measurements are taken [35–37]. In relation to timing, ELS can be particularly disruptive when induced during or prior to sensitive periods of development for specific brain regions [1,38]. As an example of region specificity, ELS associated with threat accelerates maturation in frontal cortex, amygdala, and hippocampus [39–41]. In contrast, ELS associated with deprivation can delay maturation in sensory regions [9,42]. Our model of ELS is designed to target auditory processing, thus induces stress after ear opening during the critical period for auditory cortex (ACx) maturation. While this seems to delay auditory maturation [43,44], it may also shift or accelerate maturation of regions involved in recognition memory, spatial memory, and anxiety, which have an extended period of development [45,46].

Here we examined anxiety-related behavior and memory, both of which are typically altered by ELS, and are likely to influence behavioral measures of perception. Following 10 days of ELS in gerbils of both sexes, anxiety-related behavior and memory were characterized in juveniles, at an age where ELS impairs both auditory perception and neural responses [12–14]. Compared with Control juveniles, ELS gerbils exhibited reduced anxiety-like behavior, impaired recognition memory, and improved spatial memory, suggesting shifts in maturation of regions underlying these behaviors. Sex also affected some of these behaviors, despite the pre-reproductive age of the animals. Additionally, our Control animals provide one of the first characterizations of anxiety-related behavior and memory in juvenile gerbils.

Materials and methods

Subjects

All procedures relating to the maintenance and use of animals in this study were approved by the Institutional Animal Care and Use Committee at Northeast Ohio Medical University (NEOMED). Mongolian gerbils (Meriones unguiculatus) from multiple litters were raised by internally bred pairs. Lineage of the breeding pairs originated from Charles River Laboratories Inc. (Wilmington, MA, US). Animals were randomly assigned to one of two Treatment groups: early-life stress animals (ELS) and age-matched, non-early-life stress animals (Control). Treatments were by cage, where each litter was either ELS or Control without cross-fostering, to avoid any stress on Control pups or parents induced by removing siblings for ELS treatment. This creates a potential cohort-source level of variance. A total of 208 gerbils were used, with breakdowns by sex and treatment indicated under each test in the Results. Gerbils were weaned at postnatal day 25 (P25) and were housed as two to five same-sex littermates in a 12 h light/dark cycle (07:00–19:00h) with food and water ad libitum. The variable number of animals per cage depended upon the numbers of each sex per litter, and did not differ between control and ELS treatments (Х2 = 2.37, p = 0.499). Most cages contained 3 or 4 animals (28 of 31 cages for controls, 24 of 31 for ELS), and only one cage per treatment contained two animals. All animals were handled by an experimenter for two minutes one to two days before the start of the experiment to assist with habituation. Animals were brought to the testing room 30 minutes before each experiment to habituate to the novel environment.

Stress induction

Early-life stress was induced by unpredictable (i.e., at varying times during the light cycle), intermittent parental separation with isolation and restraint for 2 h on ten nonconsecutive days between P9–24, an age at which there is high neural plasticity in the auditory cortex [47]. Unpredictability of stressors is a crucial contributor to stress-related behavioral changes [48]. Pups were removed from their home enclosure and transferred to a separate room before being placed in a plastic, cylindrical restraining device. Each restraining device was then placed in a small sound-attenuating box such that no pup could see, hear, or smell any other pups. During this time, Control animals remained undisturbed in their home enclosure. This model of stress induction in gerbils significantly shifts two measures indicative of an altered HPA axis: blood corticosterone levels and responses to a startling stimulus [12]. The 2 h isolation period induced slight hypothermia, where pups lost 1.17 ± 0.35 °F (n=8). While gerbil pup bodyweight is susceptible to reduction via a 24 hour period of parental separation at P7 [21], our ELS model did not reduce body weight. This reduces the likelihood of differential results between treatment groups arising from the results of poor nutrition and failure to thrive.

Behavioral testing overview

All training (if applicable) and testing occurred between 09:00–14:00 h. To avoid possible stress, order, or habituation effects, each test was performed on a separate cohort of juvenile animals. For each test type, all animals were tested at a specific postnatal day ±1d: Elevated Plus Maze at P30, Novel Object Recognition and Object Place Recognition at P35, and Light-Dark Box at P37. Unless otherwise specified (i.e., for the elevated plus maze), all tests took place under typical laboratory lighting, within a sound-proof booth (dimensions (cm) 195 H × 150 D × 122 W). All apparatuses were cleaned with 70% ethyl alcohol and allowed to dry between animals. Behavior from all tests were scored offline by an experienced, blinded investigator. The Elevated Plus Maze (EPM) was recorded with a digital-single-lens reflex (DSLR) camera (EOS Rebel Ti3, Canon, Inc., Tokyo, Japan) and location was tracked automatically in EthoVision XT 16 (Noldus Information Technology, Wageningen, the Netherlands). For the remaining tests, location and other behaviors were manually scored offline using the SMPlayer application (open source, v23.12), from videos recorded in the ManyCam application (ManyCam ULC) with a Logitech camera (270c, Logitech International S.A., Lausanne, Switzerland). Transitions between locations were determined when all four paws were in the new location.

Elevated plus maze

The EPM test was used to measure proxies for anxiety, exploration, and risk assessment [49,50]. We based our EPM parameters on gerbil studies by Starkey [49], using a maze with a smooth surface under low lighting. Testing took place under very low red light in an open room. The apparatus (Fig. 1 A) made of black acrylonitrile butadiene styrene (ABS) was in the shape of a plus sign where two opposing arms had 30cm walls and two opposing arms had no walls. The EPM apparatus was raised 50 cm from the ground. The arms were each 10cm wide by 50cm long and the intersecting central platform was 10 × 10cm. Each session was video recorded from an overhead camera for offline location tracking with EthoVision. The test began by placing the gerbil in the center, facing an open arm. The gerbil was allowed to explore for 5 minutes [51,52]. Transitions between locations were determined when all four paws were in the new location. On rare occasions where an animal walked off the edge of the EPM, it was placed back in the same location from where it fell. One animal was excluded for continually walking off the edge. Animals were scored for amount of time in each location (open arms, closed arms, center), number of visits to each location, and first latency to enter open and closed arms. Gerbils were tested at P30 (± 1d).

Fig. 1. Behavioral Tests.

(A) The EPM measures anxiety-related behavior by comparing time spent in two closed arms with opaque walls (translucent only for sake of diagram here) vs. two open arms without walls. (B) The LDB test measures anxiety-related behavior via the latency to transition from the light to the dark side, and the amount of time in and number of transitions to the dark side. (C) The NOR test measures recognition memory for a familiar (“Objects”, left) vs. novel object (“Objects”, right). (D) The OPR test measures spatial memory for object location (“Object”). Both the NOR and OPR assume a preference for novelty.

Light-Dark Box

As prey animals, rodents tend to avoid brightly lit areas because they are more likely to be seen by a predator. Thus, the Light Dark Box (LDB) test measures proxies for anxiety, exploration, and risk assessment [49,53]. The apparatus (19.5cm H, 40.5cm W, 19cm D) was an acrylic box (Panlab LE918, Harvard Apparatus, Holliston, MA) divided into light (well-lit) and dark (dim) halves, separated by an open door 5 × 5cm (Fig. 1 B). The box was designed for mice and was thus smaller than that used for adult gerbils [54], but was sized comfortably for juvenile gerbils to freely explore. The box width matched that used by Starkey [54], allowing similar traveling distance from one end of the box to the other. The opening between the compartments was smaller, but comfortably sized for juvenile gerbils, though this may have contributed to the reduced number of transitions in our study compared with Starkey [54]. To avoid gerbil escape, our box was close-topped rather high open-topped as in Starkey [54]. The black walls, ceiling and floor of the dark compartment blocked all light except that which entered through the doorway from the light side. The light compartment was made of three white walls, one translucent wall, and a black floor. Except for the necessary overhead camera placement, the ceiling of the light side was open to let in typical laboratory lighting. This lighting was enhanced by an additional 60 watt white light directly overhead to increase the aversive nature of the light side.

At P37 (± 1d), each gerbil was placed in the light side and allowed to roam for five minutes. As with most rodents, gerbils tend to cross quickly from the light to dark compartment. A shorter crossing latency and increased time in the dark side serve as proxies for anxiety-like behaviors. Additional behaviors include exploration, measured by number of transitions (moving from one compartment to another), and risk assessment, measured by number of headpokes into the light (i.e., an animal’s body remains in the dark compartment while its head crosses the doorway into the light, akin to stretch-attends [55,56]).

Novel object recognition

Recognition memory and novelty preference were tested by determining whether animals spent more time with a novel object than a familiar object during a Novel Object Recognition (NOR) test in an open arena. This paradigm relies on an innate and assumed preference for novelty, which can vary based on both age and stress [57]. The arena was a lid-less box made from high-density polyethylene (HDPE) (Fig. 1 C left, 19cm H, 46 cm W, 19 cm D). The walls of this clear arena were covered with various shapes and colors in acrylic paint to provide fixed contextual cues for gerbils’ spatial orientation.

The NOR task requires two sets of identical objects of approximately the same dimensions. We used one pair of stacked Legos alternating in color horizontally (Fig. 1 C right, dimensions in cm: 7.9 H, 1.6 D, 3.2 W) and a pair of dark glass vials with safety orange vertical stripes (6cm height, 3cm diameter). Magnets held these objects in place at the bottom of the arena. The assignment of object type to novel or familiar differed across gerbils in a counterbalanced fashion. Preliminary tests verified that juvenile gerbils showed no inherent differences in preference for either of these objects.

The NOR task included one or more 5-minute training trials (see below) with identical objects, followed by a single 5-minute test trial where one object was replaced by a novel object in the same location. At the start of each training trial, the animal was placed in the lower right corner of the arena with two symmetrically located identical objects and allowed to explore for five minutes. During the test trial, one of the original objects was replaced by a novel object while the other original object remained in place as a familiar object. Again, the animal was placed in the lower right corner of the arena with two symmetrically located identical objects and allowed to explore for five minutes. For offline scoring, the start and end of each object investigation were marked based on the gerbil’s orientation toward, and nose within, 2cm of the object.

A set of control animals were tested in a NOR-Single Trial (NOR-ST) test with one 5-minute training session. Gerbils experienced the training session approximately 30 min after a 5-min habituation to the empty arena. The testing session occurred 24h later. Behaviors were scored from this testing session. To enhance recognition of the original objects for juvenile gerbils, we developed a multi-trial NOR (NOR-MT) with four 5-minute training sessions: the first training session on day 1 (30 min after a 5-min habituation), the second 24h later on day 2, and the third and fourth another 24h later, back-to-back (0-min inter-session interval) on day 3. The memory retention testing session took place directly after the fourth training session. The arena and objects were cleaned immediately before each training or testing session. Five gerbils were excluded from the NOR-MT because they did not meet our attention criteria during testing: (1) investigate each object for >1sec and (2) investigate both objects for >3 sec [58]. The first training day was P34 ±1d for the NOR-ST and P33 ±1d for the NOR-MT; thus all animals were tested at P35 ±1d.

From these investigation times, we quantified five behavioral measures for memory: Order (which object was approached first), Object Investigation Time (IT = N+F, where N = time with the novel object and F = time with the familiar object), Difference Time (DT = N-F, absolute discrimination), and a Recognition Index for investigation time (RI = DT/IT, relative discrimination). We used the same formula to create a Recognition Index for number of visits (RI_Visits = (# visits to novel object - # visits to familiar object) / total # visits to objects). DT is an absolute measure of discrimination where the magnitude indicates the time difference between investigation of the two objects (DT > 0 indicates more time spent with the novel object). RI is a relative measure of discrimination that ranges from −1 to 1. RI = 0 indicates no object discrimination, RI = 1 indicates that the animal only visited the novel object, and RI = −1 indicates that it only visited the familiar object.

Object place recognition

We employed an Object-Place Recognition multi-trial task (OPR-MT) to assess spatial memory and novelty preference using the same principles, training, and measurements as the NOR-MT. Instead of introducing a Novel object in the testing session, one of the identical familiar objects (small Lego tower) was moved to a novel location within the arena (Fig. 1 D). Five animals were excluded from the OPR-MT because they did not meet our attention criteria (same as those used in the NOR-MT). Behavioral measures were the same as those used for Novel Object Recognition. The first training day was P33 ±1d, with testing at P35 ±1d.

Statistical analysis

Results were analyzed using SPSS Statistics, version 30.0 (SPSS Inc., Chicago, IL., USA). Levene’s test was used to evaluate homogeneity of variance, an assumption necessary for parametric tests. In cases without homogeneity of variance (as indicated in the Results), data were optimally normalized via power transformation with the Box-Cox method (for positive values) or the Yeo-Johnson method (for data sets with negative values) [59]. Along with Treatment, Sex was treated as an independent variable for all tests except the EPM (where the n was too low to separate by sex). Statistical significance was determined as p < 0.05. Post-hoc t-tests were conducted when there were significant Treatment by Sex interactions in two-way ANOVAs; these were conducted separately on males and females, to identify any sex-specific effects of Treatment. All t-tests were corrected for multiple comparisons using the Benjamini-Hochberg method (critical value = 0.1). In violin plots, box edges are 25^th and 75^th percentiles, the median is a white circle, and each dot represents data from one animal. Data are presented as mean ± standard error of the mean (SEM).

Results

Early-life stress does not affect overall activity levels in the EPM

Because the measures in this study are dependent on mobility, we first compared treatment groups for two measures of activity within the Elevated Plus Maze: rate of travel and frequency of crossing through the center space. P30 (± 1d) gerbils (CTR n = 14, ELS n = 14) were tested in the EPM. Due to unequal numbers of animals collected in each group for the EPM (CTR 5F 9M, ELS 10F 4M), we collapsed our analyses across sex. Animals traveled at similar rates regardless of treatment (t₍₂₆₎=1.204, p=0.239; CTR 5.19±0.28, ELS 4.80±0.16 cm/min). The groups also did not differ in frequency of crossing through the center area (t₍₂₆₎=0.509, p=0.615; CTR 56.14±3.22, ELS 54.29±1.72 visits/min). Together these findings suggest that ELS does not affect overall activity within the EPM.

Elevated plus maze: ELS does not affect anxiety-like behaviors related to thigmotaxis under low lighting

The EPM measures thigmotaxic (wall-hugging) behaviors under low lighting as a proxy for anxiety, where animals exhibiting greater anxiety-related phenotypes spend relatively more time inside the enclosed arms and less time in the open arms. At P30 (± 1d), gerbils were tested under low lighting in the EPM (CTR n = 14 [5F 9M], ELS n = 14 [10F 4M]). ELS did not affect the amount of time juvenile gerbils spent in either set of arms or the center location (Fig. 2 A: open arms, t₍₂₆₎=−0.335, p=0.740; closed arms, t₍₂₆₎=−0.509, p=0.615; center area, t₍₂₇₎=1.590, p=0.124). Nor did ELS affect the number of visits to each area (Fig. 2 B: open arms, t₍₂₆₎=1.092, p=0.285; closed arms, t₍₂₆₎=−0.030, p=0.977; center area, t₍₂₆₎=0.509, p=0.615; t-test on transformed values for open arms to achieve homogeneity of variance). The first latency to enter an open arm was not affected by ELS, (Fig. 2 C: t₍₂₆₎=0.731, p=0.471). The first latency to enter a closed arm trended towards indicating less anxiety for ELS animals but did not reach significance after correction for multiple comparisons (Fig. 2 D: t₍₂₆₎=−2.106, p=0.045 [>BH Critical value 0.025]; t-test on transformed values). These findings suggest that ELS and Control animals display similar amounts of anxiety-like behaviors such as risk assessment and safety seeking when exploring an exposed area under low lighting.

Fig. 2. Early-life stress in juvenile gerbils did not affect anxiety-like behavior in the dimly-lit elevated plus maze.

(A) Control and ELS animals spent similar amounts of time in each location. There were no group differences in the measures of anxiety-like behavior including (B) number of open-arm visits, (C) latency to enter the open arms, and (D) latency to enter the closed arms.

Light-Dark Box: ELS and sex each affect behaviors related to risk-taking

While the EPM measures proxies for anxiety-related behaviors under low lighting and high elevation conditions, the LDB measures proxies for anxiety-related behaviors pertaining to brightly lit ground level locations. Reduced risk-taking (higher anxiety-related behavior) in the LDB is indicated by low initial latency to enter the dark compartment and increased time spent within the dark compartment. At P37 (± 1d), gerbils were tested in the LDB (CTR n = 22 [11F 11M], ELS n = 23 [13F 10M]). Unlike in the EPM, ELS treatment did affect this proxy for anxiety: the proportion of time spent in the dark area (Fig. 3 A: two-way ANOVA, main effect of treatment: F_(1,55)=10.241, p=0.002, ηp²=0.157). ELS animals spent less time in the dark than Controls over a 300s time window, suggesting a diminished anxiety-like response to bright areas. There was also a sex effect, with male juvenile gerbils showing more risk-taking with less time in the dark than females (main effect of sex: F_(1,55)=9.484, p=0.003, ηp²=0.147). There was no interaction between treatment and sex, indicating similar effects of ELS on each sex (F_(1,55)=0.731, p=0.396). Overall, gerbils did not show a strong light-dark preference, spending equivalent amounts of time in the light and dark sides (48% in dark, 52% in light).

Fig. 3. Early-life stress and sex in juvenile gerbils affect anxiety-like behavior in the light-dark box.

(A) Control animals spent more time in the dark area than ELS animals (** at top), and females spent more in the dark area than males (** at bottom), suggesting higher anxiety-related behavior for Controls and females. (B) There was a Treatment by Sex interaction for latency to enter the dark area, but no Treatment effects within each Sex. (C) Females had more headpokes into the light area than males (** at bottom), suggesting higher risk assessment. (D) ELS animals had more transitions between the light and dark sides than Controls (* at top), indicating increased environmental exploration. ** p ≤ 0.001, ** p < 0.01, * p < 0.05.

There was an interaction between ELS treatment and sex on the initial latency to cross to the dark compartment, a measure of safety-seeking. This indicates an effect of ELS that differs by sex, where ELS decreased male but increased female anxiety-related behavior (Fig. 3 B: F_(1,54)=5.396, p=0.024, ηp²=0.091; ANOVA on transformed values). There was no main effect of ELS treatment (F_(1,54)=0.003, p=0.959) or sex (F_(1,54)=0.262, p=0.611), and post hoc t-tests found no effect of treatment within either sex (CTR vs ELS females: t₍₂₅₎=1.641, p=0.113; CTR vs ELS males: t₍₂₈₎=−1.685 p=0.103).

Light-Dark Box: Sex, but not ELS affect behaviors related to risk-assessment

The number of headpokes into the light side was affected by sex (Fig. 3 C: F_(1,55)=11.494, p=0.001, ηp²=0.173), but not treatment (F_(1,55)=0.678, p=0.414) or an interaction between the variables (F_(1,55)=0.076, p=0.784). Juvenile females headpoked more often than their male peers. Since females spent more time in the dark side, they had more opportunities to headpoke into the light side. However, the sex difference remained when controlling for the amount of time spent in the dark side (ANCOVA, F_(1,55)=10.922, p=0.002, ηp²=0.168), which suggests that females engaged in more risk-assessment behaviors. Together, females had higher risk-assessment and spent more time in the dark, which indicates higher anxiety-related behaviors relative to the males.

Light-Dark Box: ELS increases environmental exploration

ELS increased exploration in the LDB as measured by the number of times the animal fully crossed into the other side (Fig. 3 D, transitions: main effect of treatment, F_(1,55)=4.945, p=0.030, ηp²=0.083). Because animals from both treatment groups traveled at similar rates in the EPM (Fig. 2 A), the increase in exploration of the ELS animals in the LDB may be independent from differences in overall activity. Due to camera placement angle it was not possible to confirm this by measuring locomotor activity in the LDB, which could differ from that in the EPM. There was no main effect of sex (F_(1,55)=0.019, p=0.890) and no interaction (F_(1,55)=0.702, p=0.406).

NOR-ST: control juvenile gerbils do not strongly discriminate objects in a single-training trial NOR task

The NOR-ST is a commonly used test of memory in rodents which assumes novel object preference [57], and often has a single training trial with a 24-h interval between training and testing sessions [60]. However, adult gerbils typically require multiple days of training on this task [61,62]. To determine whether juvenile gerbils would spend more time with novel objects following a single training trial, Control gerbils were tested on a NOR-ST (P35 ± 1d on the test day, n=24, 13F, 11M). A one-sample t-test versus chance was run on the relative discrimination (Recognition Index (RI)), which is the difference in time spent with each object normalized to the time spent investigating both objects (ranges from −1 to 1). Over the 5-min testing trial, these Control juveniles spent equal time with each novel and familiar object (Fig. 4 A left: one-sample t₍₂₃₎=1.579, p=0.128). This suggests that either juvenile gerbils do not have a novel object preference, the memory of juvenile gerbils is poor, or that a single training trial is not sufficient.

Fig. 4. Effects of ELS and sex in the NOR test: ELS reduced recognition memory and/or novelty preference.

(A) Juvenile control gerbils did not show robust memory recognition in the single-training trial NOR-ST. Left: RI Time over 5 min was no different from chance, suggesting a lack of recognition memory using the traditional RI metric. Right: Visit Order: 75% of animals visited the familiar object first, indicating discrimination and a familiarity preference rather than a novelty preference. (B) Juvenile Control gerbils displayed strong memory recognition in the NOR-MT. Gerbils scored significantly better than chance for RI Time (left), and Difference Time (middle). Visit Order: 65.2% of animals visited the Novel object before the Familiar, indicating a novelty preference (right). In the NOR-MT, Treatment and Sex affected preference scores: (C) Respectively, Control and male animals spent relatively more time with the novel object than ELS and female animals (* at top and bottom indicating main effects of Treatment and Sex, respectively). (D) Control animals made relatively more visits to the novel object than ELS animals (* at top, main effect of Treatment), where ELS reduced visits particularly in males but not females (* at bottom, posthoc effect of ELS on males but not females). Gray dotted lines indicate the level of equal preference for the two objects. * p < 0.05, ** p < 0.01, *** p < 0.001.

Another measure of discrimination with less top-down influence on behavior is the Visit Order. Seventy-five percent of Control juveniles visited the Familiar object first (Fig. 4 A right). This frequency is significantly greater than chance (one sample binomial test, 75% < 50%, z₍₂₂₎=−2.449, p=0.007), which suggests that juvenile gerbils are able to discriminate between the novel and familiar objects in the single-training trial NOR test if a non-traditional variable is used. When measured conventionally with RI, the appeal of familiarity over novelty appears to wash out for juvenile gerbils over the five-minute trial (Fig. 4 A left). The Visit Order indicated that juvenile gerbils remembered the familiar object but displayed a familiar rather than a novel preference after a single training trial. We wanted to employ the RI measure commonly used in other studies, and to determine whether juvenile gerbils could acquire and display a novel object preference. Thus as in previous studies with adult gerbils, we adapted the training method to achieve an RI-based novelty-familiarity discrimination.

NOR-MT: control juvenile gerbils strongly discriminate objects in a multi-training trial NOR task

Novel object recognition in adult gerbils often requires multiple training trials [61,62]. By piloting different training schedules in juveniles (data not shown), we identified a schedule that elicits an RI different from chance. The multi-trial NOR (NOR-MT) used a total of four training trials (identical objects) and a single testing trial (one familiar and one novel object). Control juvenile gerbils (P35 ±1d on the test day, 11F, 11M) showed better discrimination than chance, measured by both relative (RI time) and absolute discrimination (difference time, DT) (Relative: Fig. 4 B left, one-sample t-test t₍₂₂₎=6.978, p<0.001. Absolute: Fig. 4 B middle, t₍₂₂₎=4.549, p<0.001). This indicates that juvenile gerbils display adequate recognition memory in the adapted NOR-MT regardless of treatment and sex. Additionally, 65.2% of control animals visited the novel object before the familiar object, which is significantly higher than chance (Fig. 4 B right, binomial test, z₍₂₂₎=7.715, p<0.001). Novel object preference held true separately for both Control and ELS animals (y-axis of Fig. 6 B, based on DT: CTR: t₍₂₂₎=4.523, p<0.001; ELS: t₍₂₃₎=5.089, p<0.001). We thus used the NOR-MT to examine the effects of ELS and sex on novel object recognition.

Fig. 6. ELS decreased investigation time in the spatial memory test, but not the non-spatial.

(A) In the NOR-MT, ELS treatment did not affect Investigation Time. (B) In the NOR-MT for both ELS and Control, there was a positive correlation between Investigation Time (IT) and Difference Time (DT). (C) In the OPR-MT, ELS decreased IT (* at top, main effect of Treatment). (D) While ELS animals showed a positive correlation between IT and DT in OPR-MT, Control animals showed no correlation. Black and red dashed lines show correlations between DT and IT for Control and ELS gerbils, respectively. Gray dotted lines indicate the level of equal preference for the two objects or locations.* p < 0.05, ** p = 0.001, *** p < 0.001.

NOR-MT: object discrimination ability and/or novelty preference is affected by ELS and sex

Control and ELS juvenile gerbils were tested at P35 (±1d on the test day, CTR n = 22 [11F 11M], ELS n = 23 [13F 10M]). A two-way ANOVA on transformed values on the RI for time spent with the objects revealed a significant effect of treatment (Fig. 4 C, F_(1,46)=5.002 p=0.031; ηp²=0.104) and sex (F_(1,46)=16.899, p<0.001, ηp²=0.282), with no interaction between the variables (F_(1,46)=0.325, p=0.571). Thus, based on RI Time, both Control and male gerbils preferred the novel object more than their respective counterparts. Similarly, Control animals had a significantly higher RI for number of visits to the objects than ELS animals (Fig. 4 D, main effect of treatment, F_(1,46)=4.364, p=0.043, ηp²=0.094, two-way ANOVA on transformed values). Sex alone didn’t affect RI Visits (F_(1,46)=1.779, p=0.189). Instead, there was a significant interaction between treatment and sex (F_(1,46)=5.258, p=0.027, ηp²=0.111). Matching RI time, post hoc tests found that ELS males had significantly lower RI Visits than Control males (t₍₁₉₎=3.89, p=0.001). There was not a significant effect of treatment within the female gerbils (t₍₂₃₎=−0.131, p=0.897). Distinct from the normalized RI time, there was no effect of ELS on absolute difference time with the two objects (y-axis of Fig. 6 B, based on DT: F_(1,47)=0.882, p=0.353). One possible explanation for effects of sex and treatment may be that training was differentially effective across groups.

OPR-MT: control juvenile gerbils weakly discriminate in a multi-training trial OPR task

Because spatial and non-spatial memories involve different neural regions, they may be differentially affected by ELS or sex. We thus employed an OPR task to assess spatial memory based on novel location preference. As the experimental schedule of NOR-MT was successful in assessing juvenile gerbil recognition memory, the same approach was used to test spatial memory.

In the OPR-MT, animals were tested on their ability to discriminate between objects in a novel vs familiar location (P35 ±1d on the test day, CTR n = 33 [20F 13M], ELS n = 33 [18F 15M]). As with the NOR, we examined whether Control juvenile gerbils have sufficient novel location preference and spatial memory to discriminate between locations. Using two-sided t-tests, Control juveniles showed marginally better relative (RI) time and better RI visits discrimination than chance, albeit without a significant discrimination based on absolute difference time (Fig. 5 A left, one-sample two-sided t-test, RI Time: t₍₃₄₎=2.016, p=0.052; Fig. 5 A middle, RI Visits: t₍₃₄₎=3.125, p=0.004; Fig. 5 A right, DT: t₍₃₄₎=1.359, p=0.183). Furthermore, the proportion of animals which visited the object in a novel location before one in a familiar location was marginally higher than chance for Control animals (Fig. 5 B, binomial test, CTR: 65.7%; z₍₃₄₎=1.859, p=0.063). Compared with the NOR, weaker discrimination in the Controls may be due to immaturity in spatial processing and/or spatial novelty preference at the age tested here (see Discussion).

Fig. 5. Effects of ELS in the OPR test: ELS improved spatial memory and/or novelty preference.

(A) Juvenile Control gerbils displayed weak but clear spatial memory in the OPR-MT. Juvenile Controls spent marginally more relative RI Time (left) with significantly higher RI Visits (middle) in the novel location, albeit without a preference based on absolute Difference Time (right). (B) Visit Order: 66% of animals visited the Novel location before the Familiar, indicating a novelty preference. In the OPR-MT, stress treatment affected preference scores: In contrast to the NOR-MT, ELS animals showed an increased preference for novelty compared with Controls based on (C) RI Time (* at top, main effect of Treatment), but not (D) RI Visits. Gray dotted lines indicate the level of equal preference for the two locations. * p < 0.05, ** p < 0.003, *** p < 0.001, ◊ p < 0.065.

OPR-MT: ELS improves location discrimination and/or increases novelty preference

Like the NOR-MT findings, there was a significant effect of treatment on relative discrimination, RI Time (Fig. 5 C, two-way ANOVA, F_(1,69)=4.303, p=0.042, ηp²=0.061). Yet unlike the NOR-MT (Fig. 4 C), the ELS animals spent more time at the novel location than the Control animals. Another dissimilarity with the non-spatial version of this test is that there was no effect of sex on RI Time (Fig. 5 C, F_(1,69)=0.008, p=0.931). There were also no interaction effects (treatment x sex F_(1,69)=1.261, p=0.266). Neither treatment nor sex affected the number of RI Visits, nor was there an interaction (Fig. 5 D, Treatment F_(1,69)=1.123, p=0.293; Sex F_(1,69)=0.624, p=0.432; Treatment x Sex (F_(1,69)=0.172, p=0.679). The advantage for ELS animals measured as relative discrimination data (RI Time) shifted to just a trend when measured as the absolute discrimination time (F_(1,69)=2.288, p=0.135). There were no significant effects of sex or any interaction on DT (Sex F_(1,69)=0.223, p=0.638, Treatment x Sex F_(1,69)=0.229, p=0.634). The proportion of animals which visited the object in a novel location before one in a familiar location was significantly higher than chance for ELS animals (71.4%, z₍₃₄₎=2.535, p=0.011), compared with a marginal significance for Controls (Fig. 5 B). In addition, only ELS animals showed a significant absolute discrimination above chance for the novel object location (Fig. 5 A right and 6D, DT on y-axis, CTR: t₍₃₄₎=1.359, p=0.091; ELS: t₍₃₄₎=3.762, p<0.001).

ELS induces early maturation of investigation behavior in the OPR-MT but not in the NOR-MT

The presence of a treatment effect on relative but not absolute discrimination time in both the NOR-MT and OPR-MT suggests a difference in how ELS and Control animals interact with the novel and familiar items. Because RI is calculated by normalizing each individual’s DT to their IT, a possible cause for this disparity could be a difference in total exploration of the objects. Investigation in the NOR-MT, measured as the total time spent exploring both objects, was not affected by treatment, sex, or an interaction (Fig. 6 A, two-way ANOVA, Treatment: F_(1,46)=0.005, p=0.946; Sex: F_(1,46)=2.204, p=0.145; Treatment x Sex: F_(1,46)=0.558, p=0.459). We also measured the relationship between NOR-MT DT and IT. Within animals, these values were positively correlated in both Control and ELS animals (Fig. 6 B, Pearson’s: CTR r₍₂₁₎=0.657, p=0.001, ELS: r₍₂₂₎=0.720, p<0.001). Animals that spent more overall time investigating the objects were preferentially engaging with the novel object. Note that clear novel object preference by both Control and ELS animals based on DT is visible in Fig. 6 B (values > 0 indicate novel object preference; CTR: t₍₂₂₎=4.523, p<0.001; ELS: t₍₂₃₎=5.089, p<0.001).

In the OPR, ELS animals spent less time investigating both locations than Controls (Fig. 6 C, Treatment: F_(1,69)=21.745, p<0.001, ηp²=0.248). There was no effect of sex and no treatment x sex interaction (Sex: F_(1,69)=0.279, p=0.599; Treatment x Sex: F_(1,69)=1.806, p=0.184). Within-animal comparisons indicate that the relationship between DT and IT is treatment-specific in the OPR. As in the NOR-MT, the ELS animals that investigated for the longest amount of time in the OPR-MT, investigated the novel location more than the familiar (Fig. 6 D, Pearson’s r _(1,34)=0.384, p=0.023). However, unlike NOR-MT, there was no correlation between DT and IT in Control animals, which means the Control animals that spent more time investigating did not preferentially engage with objects in either location (Pearson’s r_(1,34)=−0.219, p=0.207), indicating a potential increased interest combined with lack of novelty preference. As described above, only ELS animals exhibited novel location preference based on DT, while Control animals showed a marginal preference (Fig. 6 D).

Discussion

Recent studies reveal that ELS, a known risk factor for many cognitive and emotional issues, also has detrimental effects on bottom-up neural representations of sensory signals, affecting auditory, visual, and somatosensory pathways [8,9,11,12]. This raises the possibility of perceptual deficits arising from poor sensory encoding, which we and others have described in the Mongolian gerbil, an established model for auditory processing [12,13,21]. Such deficits would lead to increased listening effort (i.e., cognitive load), which could contribute to the learning and attentional difficulties faced by ELS humans, such as listening in noisy environments and perception of emotional content in conversations [63–65]. Measuring sensory deficits in ELS animals can improve our understanding and treatment of communication disorders in humans.

Here we used a battery of behavioral tests to assess activity levels, exploration, and behavioral correlates of anxiety and memory. Tests were conducted on juvenile Mongolian gerbils that had been exposed to stress during a critical window for ACx maturation (P9–24). We first showed that ELS had no effect on overall activity within the context of the EPM. This reduces the likelihood of activity as a confounding factor in interpreting the behavioral results, though we did not measure locomotor activity in the other behavioral contexts, so can not fully rule out effects of locomotion in those contexts. Our cognitive tests demonstrated that ELS reduced recognition memory, improved spatial memory, and altered anxiety-related behavior in a sex-specific manner. Our results provide insight into the maturation of basic behaviors in gerbils, allowing comparison with other species of developing rodents and with adult gerbils. These findings are necessary to accurately interpret auditory measures in future studies with this novel ELS model.

Effects of ELS on anxiety-like behavior

In the context of the LDB but not the EPM (aside from a trend based on closed arm latency), ELS decreased anxiety-related behavior in juvenile gerbils. In the EPM, ELS had no effects on juvenile gerbils when thigmotaxis (wall-hugging behavior) was used as a proxy for anxiety in the context of this dimly lit, elevated platform (Fig. 2). Thigmotaxis is a defensive response from aerial predators seen in many rodents and appears in the EPM as spending more time in the closed than open arms [66]. However, our juvenile gerbils spent similar time in the open and closed arms (Fig. 2 A), suggesting that juvenile gerbils do not display thigmotaxis. The literature is mixed for whether adult gerbils display thigmotaxis, with some studies even reporting a preference for being far from the wall in an open field arena [67–69]. This low thigmotactic behavior may arise in part due to gerbils’ crepuscular nature, with bouts of activity throughout the circadian cycle [70–72], unlike mice and rats which are nocturnal. While this suggests that thigmotaxis may not be an effective measure of anxiety-related behavior in gerbils, adult gerbil thigmotaxis can be induced by adult stress in males and by neonatal stress in females [49,73]. Effects from ELS on thigmotaxis in the EPM might emerge if ELS gerbils were tested after puberty. This idea is supported by research that found no effect of ELS on mice when tested as juveniles, but that ELS increased thigmotaxis when tested as adults [74,75].

In contrast, ELS animals showed less anxiety-like behavior than Controls when risk assessment was measured in a small, brightly-lit enclosure (LDB), indicating that the context of the anxiety-related test (e.g., lighting level, elevation) influences the outcome measures. In the LDB, there were effects of ELS, sex, and their interaction using risk assessment, risk-taking, and safety-seeking as proxies of anxiety (Fig. 3). Relative to control animals, ELS animals showed a diminished anxiety-like response based on reduced time in the dark (increased risk-taking, Fig. 3 A) and increased exploration (Fig. 3 D). A differential effect of sex on ELS emerged based on initial latency to cross into the dark, where ELS increased anxiety-like behavior in females but reduced it in males (Fig. 3 B). The same pattern occurred in another study of ELS gerbils [73]. However, there is a wide range of sex effects on anxiety-related behaviors [49,54,67,76], and multiple studies show differential male and female responses to developmental stress [77,78]. In our study, juvenile males regardless of treatment showed less anxiety-like behavior than females (Fig. 3 C).

Unlike mice and rats, our juvenile gerbils spent a similar amount of time in light and dark sides. This is also seen in adult gerbils who spend an equal amount or more time in the light side [49,67]. Compared with the EPM, the difference in ambient lighting (bright in LDB, dim red in EPM) may account for seeing ELS effects in only the LDB context. Young adult female gerbils show significantly less behavior indicative of anxiety in the EPM under low lighting than under typical laboratory lighting [50]. This suggests our EPM was not overly anxiety-producing, and that an effect of ELS could emerge in the EPM if gerbils were tested under brighter light.

Effects of ELS on memory

For juvenile gerbils to achieve discrimination via the standard RI, it was necessary to add additional training trials to the typical single-training trial NOR. Such a lack of discrimination in the NOR-ST (Fig. 4 A left) suggests either a lack of novelty preference or insufficient learning and memory, likely driven by gradual maturational trajectories in attention, arousal, or learning abilities [22,79–81]. However, most gerbils approached the familiar object first during NOR-ST, demonstrating a familiarity preference which requires memory, although this effect was not sufficiently robust to endure over the 5 minutes of exploration. The additional training of NOR-MT enabled discrimination via the RI and, as desired, shifted the initial preference from familiar to novel (Fig. 4 B,C). This shift may have arisen due to increased familiarity through repeated training, or to the slightly older age at testing, as novelty preference is known to shift over development in other rodent species [58,82,83]. Alternatively, adult gerbils are often given multiple training trials [61,62], suggesting that gerbils may require additional familiarization time to attend to, remember, and establish preference for objects. Indeed, gerbils differ from rats and mice on a variety of behavioral tests, exhibiting greater exploratory activity [69,84,85].

In the multi-training trial NOR-MT, ELS animals and females spent less time with the novel object (compared with Controls or males; Fig. 4 C), indicating either poorer recognition memory or reduced novelty preference. This result is in concordance with ELS effects in rats, which impaired recognition memory in the NOR [86]. Further, our male gerbils were more affected by ELS than females, making fewer visits to the novel object than Control males (Fig. 4 D). This sex-specific susceptibility of recognition memory to ELS was also found in male rats, and correlated with reduced functional connectivity in PFC – hippocampal – perirhinal cortical (PRC) networks [87]. Recognition memory during the NOR relies primarily on an intact perirhinal cortex [88,89]; and the hippocampus and PFC are implicated in attention and working memory [90], both of which are required to form and retrieve new memories in recognition tasks. The NOR involves cholinergic release in both the PRC and the PFC [91–93]. Thus, stress effects in these regions, potentially involving cholinergic signaling [94–96], may be involved in the recognition memory deficits seen in ELS animals. However, attention and working memory in our ELS gerbil model were unaffected by ELS when measured as spontaneous alternations in a Y-maze [13]. Additionally, we cannot exclude the possibility that Control animals learned to recognize the objects better than ELS animals over the training sessions. Alternatively, the reduced recognition memory in ELS gerbils may be due to reduced novelty preference or emergent neophobia.

A challenge in interpreting both the NOR and OPR in juvenile animals is that tests of recognition and spatial memory are not exclusively memory tests – they rely on a preference for novelty, which can vary across development [57]. While a decreased time spent with novelty is generally interpreted as a memory impairment, a decrease in novelty preference would produce an identical effect. In spatial tests, juvenile rats exhibit a preference for familiar locations, early adolescent rats show no preference, and mid-adolescent and adult rats prefer novel locations [82]. Importantly, the youngest rats tested exhibited spatial memory by performing well on the OPR but with a familiar preference. Further, object recognition memory matures earlier than spatial location memory [97,98]. This would explain the clear preference seen here for novel objects but not novel locations in Control juvenile gerbils (Fig. 6 B, D): at this age, novel preference may be mature for objects but not yet for locations.

Spatial and recognition memory rely on distinct neural circuits that can be differentially affected by stress. In the present study, ELS surprisingly improved spatial memory (Fig. 5 B) which relies primarily on the hippocampus [89,99,100], while worsening recognition memory. One explanation of this dichotomy is that spatial memory and/or novel location preference is immature at the age tested here. ELS is known to accelerate maturation of the hippocampus and amygdala, along with behaviors related to those structures [39,41]. Thus, ELS may have hastened maturation in our juvenile gerbils via effects on the hippocampus. In contrast, at the same developmental stage, object recognition memory was mature and thus susceptible to the detrimental effects of ELS. This idea is supported by the differential investigation behavior in the two tasks in Control animals. In the NOR, object recognition was correlated with investigation time in both treatment groups (Fig. 6 B). But in the OPR, spatial recognition was only correlated with investigation in ELS animals, who spent more time investigating the novel object. While Control animals did not exhibit a preference for the novel location, they investigated objects in the OPR for more time than ELS animals but did not preferentially engage with either location (Fig. 6 D). This increased investigation suggests that Control animals noticed a spatial shift despite having no location preference. Such behavior indicates that spatial memory may be intact at this age despite immature novelty preference. Future OPR studies of developing gerbils across a range of ages are needed to test this interpretation.

Implications for sensory processing

A major goal of sensory neuroscience is to determine how neural responses to sensory input enable accurate detection of that input. This involves measuring perception, often in conjunction with neural recordings. The gold standard for assessing animal’s perceptual abilities is operant conditioning, in which animals learn to associate a sensory signal with an aversive or appetitive stimulus. With appropriate training, this method provides very sensitive detection or discrimination thresholds. Operant conditioning relies on the assumption that animals have similar attention, motivation, anxiety, memory, expectation, associative learning, and perceptual learning abilities. Differences in these factors across treatment groups can affect performance, and thus potentially be misconstrued as sensory deficits. This concern can at least partly be addressed by directly comparing learning and proxies for attention and motivation during operant performance [101].

An alternative technique for measuring sensory acuity is prepulse inhibition of the acoustic startle, where a reflexive response to a startling stimulus is reduced upon detection of a sensory signal [102–104]. This more automatic method of measurement does not require training and can be minimally influenced by attention with an appropriate interval between signal and startle [105]. Thus, this measure is less susceptible to top-down influences than operant conditioning.

Elements examined here that may affect operant assessment of sensory abilities in our gerbil model of ELS include attention, learning/memory, and anxiety-related behavior. Though we did not explicitly test it here, attention is required to form and retrieve new recognition and spatial memories, and could have contributed to the ELS effects on memory in our model. Another possible contributor would be differential learning of object or location. Our memory tests involved four training sessions, and performance differences upon testing could arise from differential object and spatial learning abilities. Future studies measuring performance during training would indicate whether learning is affected by ELS in this model, which could affect operant measures of sensory processing because they require training. Finally, our gerbil model of ELS reduces anxiety-like behavior and alters memory performance. This may affect the ability to learn and retain operant training, particularly if aversive, anxiety-provoking reinforcement such as a shock is used. Additionally, the interaction of sex and stress on anxiety-related behavior and recognition memory in ELS gerbils indicates the importance of including sex as a factor in future studies.