Effects of repeated social stress on risk assessment behaviors and response to diazepam in the elevated plus maze in adult male rats

Abstract

Anxiety is highly common and stress is a major trigger for anxiety. Anxiety includes heightened threat assessment and avoidance, but we do not fully understand which components are sensitive to stress. Rodents show a balance of exploration and avoidance that incorporates threat assessment prior to making the relatively risky decision to explore an open area. The purpose of this study was to determine if stress impacts risk assessment, and if this is tied to effects of stress on exploration. The present study used Elevated Plus Maze (EPM) to test the effects of repeated social defeat stress (RSDS) on risk assessment behaviors in adult male rats. We then tested the effects of diazepam, an anxiolytic that reduces the impact of stress on EPM exploration, to further clarify the relationship between risk assessment and risky behavior in the EPM. We found that RSDS decreased time in the open arm, similar to prior studies. We also found that RSDS increased the likelihood of the primary risk assessment behavior, stretch and attend (SAP), increased SAP prior to entering an open arm, and decreased the likelihood that a rat would enter an open arm after SAP. Diazepam ameliorated the effects of RSDS on both SAP and exploratory behavior, further linking risk assessment and subsequent exploratory behaviors. These results suggest that increased risk assessment, and reduced risky choices after risk assessment, is tied to effects of stress on exploration, and provide novel insight into how stress may increase avoidance by effects on risk assessment.

INTRODUCTION

The relationship between stress and anxiety is well documented in American adults across generations, and stress has only increased throughout recent years (American Psychological Association, 2020). In our social lives, repeated episodes of stress can predate the manifestation of many neuropsychological disorders, such as major depression, social anxiety, and post-traumatic stress disorder (Davis et al., 2017; Kessler et al., 2005; Ménard et al., 2016). Even when it does not trigger a psychiatric illness, stress can cause social withdrawal, when anxiety can outweigh the appeal of social engagement. This outcome is part of a wider balance that is seen across mammals, between the safety of avoidance and the potential risks of engagement and exploration.Stress factors that contribute to differences in risk aversion and risk taking are still unclear, but likely contribute to anxiety and avoidance behaviors. Thus, the concept of risk assessment, particularly in response to stress, has been emerging as a focal point in understanding anxiety disorders. Risk assessment involves determination of the relative safety or danger of exploration prior to a decision to approach or avoid a potential threat or reward (Bendahan et al., 2017; Blanchard et al., 1991; Stopper et al., 2014). Repeated social defeat stress (RSDS) is a common method for introducing social stress in rodent models (Jaisinghani & Rosenkranz, 2015; López-Moraga et al., 2022; Munshi et al., 2020). RSDS involves a resident-intruder paradigm in which an adult male rodent (intruder) is placed in the cage of an aggressive male rodent (resident) and allowed to freely interact. The resident initiates attacks on the intruder over several sessions, resulting in social stress that shifts social behavior and can develop into anxiety (Ferrara et al., 2020; Munshi et al., 2020; Ritger et al., 2023). The Elevated Plus Maze (EPM) has been used to evaluate exploratory behavior (Patki, Solanki, Atrooz, et al., 2014; Pellow et al., 1985). Most strains of laboratory rodents are nocturnal, and avoid brighter, open (unprotected) spaces and prefer darker, closed (protected) spaces (Godsil & Fanselow, 2004; Goff & Finger, 1966; Keller & Oberlin, 1942; Norton et al., 1975). This can apply to the EPM, where investigation into the open arms of the EPM is inherently riskier while the closed arm is safer (Rodgers & Dalvi, 1997). Substantial prior work has characterized exploration into the open arm as a reflection of lower anxiety-like behavior because many treatments that reduce anxiety in humans, such as benzodiazepines, increase open arm exploration, while treatments that increase anxiety decrease open arm exploration (Walf & Frye, 2007). Stress produces anxiety-like behavior in rodents often displayed as less EPM open arm exploration (Adamec et al., 2006; Belda et al., 2004; Calvo et al., 1998; Huynh et al., 2011; Rodgers & Cole, 1993; Zhang & Rosenkranz, 2012). When exploring novel areas, rodents must decide what areas seem safer and riskier, and use risk assessment behaviors to gather this information. Within the context of exploration, prior studies have defined risk assessment in three main categories: stretch and attend posturing (SAP), rearing, and head-dipping/over-edge looking behaviors (Barnett, 1958; Blanchard et al., 1991; Cole & Rodgers, 1994; Cruz et al., 1994; Loh et al., 2022; Rodgers et al., 1999; Setem et al., 1999). Although these behaviors have been evaluated in the EPM, little research has investigated the effects of stress on risk assessment behaviors in the EPM. The prior research has mostly used a single potent stressor or prolonged housing stressors, with mixed effects on risk assessment in the EPM and its relation to EPM exploration (Adamec et al., 2004; Adamec & Shallow, 1993; Gasparotto et al., 2005; Haller & Halász, 1999; Pohorecky, 2008; Singewald et al., 2009). It is still unclear if the effects of stress on EPM exploration are influenced by risk assessment. Thus, our first hypothesis is that social stress will alter risk assessing behaviors. We propose that RSDS will lead to more risk assessment, and bias exploration towards less risk taking behaviors (open arm exploration) after risk assessment. We further wanted to manipulate anxiety-like behavior to test the link between risk assessment and subsequent open arm exploration. Diazepam is a fast-acting benzodiazepine medication approved for management of anxiety disorders (Dhaliwal et al., 2024). Diazepam has been shown to increase open arm exploration in adult rodents, particularly in anxiogenic conditions and after stress (Hazim et al., 2014; Rico et al., 2019; Treit et al., 1993), and benzodiazepines have been reported to reduce risk assessment behaviors in the EPM (Albrechet-Souza et al., 2007; Cole & Rodgers, 1993; Molewijk et al., 1995). It is not clear if diazepam might accomplish this via a reduction of risk assessment and/or bias towards assessment of open arms as safe. Therefore, our second hypothesis is that diazepam will reverse the effects of RSDS and thus reduce risk assessing behavior, and mitigate the link between risk assessment and open arm exploration. We used a 7 day RSDS protocol to induce social stress in adult male Sprague Dawley rats. Rats were then administered vehicle or diazepam (1 mg/kg) prior to EPM assessment.

MATERIALS AND METHODS

Subjects

Subjects were male Sprague Dawley rats purchased from Envigo (n = 50; Indianapolis, IN, USA). Rats were group housed in two to three per cage in the Rosalind Franklin University Biological Resource Facility with ad libitum food and water and were maintained on a reverse light cycle (12/12 h light/dark) with lights on at 7:00 pm CDT. Adult rats arrived at postnatal day (PND) 64-69 and acclimated for at least 5 days prior to experimental procedures. Retired breeder male Long Evans rats (n = 25) were obtained from Envigo and single housed without enrichment for use as aggressor rats in the resident-intruder paradigm. All procedures were approved by the Institutional Animal Care and Use Committee and complied with the Guide for the Care and Use of Laboratory Animals (National Research Council (US) Committee for the Update of the Guide for theCare and Use of Laboratory Animals, 2011).

Repeated social defeat stress

Housing cages were randomly assigned to social defeat (stress) or control conditions. Socially defeated Sprague Dawley rats were exposed to the resident-intruder paradigm using a previously published protocol (Munshi et al., 2022). Experimental Sprague Dawley rats underwent the resident-intruder stress with a different, unfamiliar, and aggressive Long Evans rat for 7 sessions over 10 days. Social defeat sessions mostly occurred between 7 am and 12 pm. During each session, experimental Sprague Dawley rats (intruders) were brought to the stress procedure room in a transport cage and individually placed inside the aggressor Long-Evans (resident) rat home cage (53.3 × 29.2 × 20.3 cm (L x W x H)). Long-Evans rats were rotated so that residents were not used more than 3 times in a 24 h period. Stressed rats were exposed to up to 15 min of direct interaction (i.e. physical stress via attacks), followed by an additional 15 min of indirect exposure by using a smaller wire mesh cage (15.2 × 17.8 × 20.3 cm, 1.9 cm square mesh) inside the resident home cage to physically separate the intruder from the resident rat, while still permitting sensory stress via sight and smell. Rats were immediately separated prior to the end of 15 min of direct exposure according to protocol under the following conditions: 10 attacks, 5 min with no attacks, submission, or severe injury (>2 cm) and/or observation of blood. Submission was defined as the intruder laying immobile partially or wholly on its back or flank and its ventral surface of the abdomen exposed (Barnett, 1958; Blanchard & Blanchard, 1977; Ritger et al., 2023). Long-Evans that repeatedly did not attack were removed from rotation. Rats were humanely euthanized if injuries were severe (n = 1) and offending Long Evans were immediately removed from rotation. Injury was rare (8 incidences total out of 182 total sessions). Stressed rats were rarely separated due to reaching the attack limit (7 incidences total out of 182 total sessions). If resident rats attempted to attack intruders while they were physically separated with the wire mesh cage, this was noted. Matched control rats were placed into transport cages for an equivalent amount of time in their housing room. After the completion of each session, all experimental rats were returned to their home cages. Rats remained in their home cages for 3-4 days until behavioral testing (Figure 1A).

Figure 1.

Timeline of experimental protocol. Rats underwent resident-intruder stress or control handling. This was repeated seven times over 10 consecutive days. After 3-4 days, rats were given either diazepam (1.5 mg/kg) or saline vehicle and after 60 min were then placed in the EPM for 5 min exploration.

Drug administration

To elucidate the relationship between anxiety and risk assessment, diazepam was chosen as a positive control due to its known effect on EPM exploration and anxiolytic properties (Hazim et al., 2014; Kim et al., 2010; Leveleki et al., 2006; Mechan et al., 2002; Rico et al., 2019). Diazepam was administered prior to testing on the EPM. To avoid potential sedation seen in higher doses of diazepam (Hazim et al., 2014), rats were weighed and randomly assigned by a blinded experimenter to receive a single intraperitoneal injection of low dose diazepam in vehicle (1.5 mg/kg body weight) or vehicle (sterile saline, 0.9% NaCl) administered at 2 mL/kg at 90-95 PND. Upon injection, rats were placed back into their home cage. Injections were timed so that each rat was placed on the elevated plus maze (EPM) 60 min post injection (Figure 1B).

Elevated plus maze (EPM)

To test the effects of stress on risk assessment and anxiety-like behavior, we used the EPM test. The EPM (Scientific Designs, Pittsburgh, PA, USA; open arms width × length 5”×20” with 1.125” rails, closed arms width × length × wall height: 5”×20”×18”; square center width×length: 5”× 5”, 32” from the ground) was tested under dim white light conditions (25-30 lux). Prior work demonstrated that stress can strongly decrease exploration into open arms (Adamec et al., 2006; Belda et al., 2004; Calvo et al., 1998; Huynh et al., 2011; Rodgers & Cole, 1993; Zhang & Rosenkranz, 2012). Brighter light also decreases exploration (Garcia et al., 2005). Because we want to ensure that we will be able to detect an effect of stress, we chose conditions with dim light to avoid a floor effect. The EPM was cleaned with a 50% EtOH solution prior to testing each rat. Each rat was placed at the intersection of the open and closed arms (middle) with its nose pointing towards the open arm farthest from the experimenter. Rats were allowed to freely explore the EPM for 5 min. Video was acquired with ANY-maze (version 6.34, Stoelting Co, Wood Dale, IL) for automated analysis. Time in open arms, entries to the open arms, total number of closed and open arm entries, and total distance traveled were automatically quantified. Video was also manually scored by a trained observer blinded to the experimental conditions using Behavioral Observation Research Interactive Software (BORIS, version 8.20.3) (Friard & Gamba, 2016). The following behaviors were scored in BORIS: stretch and attend posturing (SAP), rearing, look over edge, go to open after stretch, and stay in closed after stretch. SAP was scored if the rat’s rear paws were stationary while the torso and head were extended beyond relaxed positioning and recorded in the open arm (oSAP) (Figure 2A). Following SAP in the closed arm (cSAP), the location of the rat was scored for staying in the closed arm or going to the open arm (Figure 2B). Rearing was scored if the rat's rear paws were stabilizing the rat's weight while the rat’s body was upright with front paws extended beyond the torso (against the closed arm wall or in the air when on the open arm) (Figure 2C-E). Looking over the edge was scored if the rat’s nose dipped below it’s torso while peering over either edge of the open arm or at the end of the open arm (Figure 2F).

Figure 2.

Risk assessment behaviors in the EPM. (A) Stretch and attend posture in the open arm (oSAP). (B) Stretch and attend posture from the closed arm (cSAP). (C) Rear in the middle/center of EPM. (D) Rear in the closed arm. (E) Rear in the open arm. (F) Look over the edge of the open arm.

Data analysis & statistics

The overall design of the experiments was 2 x 2 between groups, with all rats receiving either vehicle or drug, and either stress or control handling. The rats in the first section of the Results (Effects of stress) were administered vehicle (vehicle control compared to vehicle stress), and were the same compared between vehicle and diazepam groups [drug (vehicle, diazepam) x condition (control x stress)] in the later section of the Results (Effects of diazepam). Sample size was determined by power analysis based on prior studies (α = 0.05, power=0.8, G*Power version 3.1.9.7). Data are represented as mean ± SEM, difference between means ± SEM (DIFF), and 95% confidence interval (95% [X,X]). A value of p < 0.05 was considered as statistically significant. To measure effects of stress on weight gain, and to test if effects of stress were equivalent in groups prior to drug/vehicle treatment, two-tailed unpaired t-tests were used. Group analyses with more than one factor were compared by ordinary 2-way analysis of variance (ANOVA) or repeated measures (RM) ANOVA or 3-way ANOVA. Significant effects found in RM-ANOVAs or ordinary ANOVAs were followed by Šídák's multiple comparisons for groups with two factors or post hoc Tukey’s multiple comparisons test for groups with three or more levels for one factor.

To maintain organization, effects of stress are described first (main effects of stress and stress x drug interactions where there was a significant difference between stress and control in post hoc comparisons). Effects of diazepam are then described (main effects of drug and stress x drug interactions where there was a significant difference between vehicle and diazepam in post hoc comparisons). When there was a significant stress x drug interaction, to clearly display the effects of diazepam across groups, diazepam data was normalized to the vehicle group as a percentage [For non-stress groups: (control-diazepam value ÷ control-saline mean) x 100] and [For stress groups: (stress-diazepam value ÷ stress-saline mean) x 100] separately for control and stress groups. To verify that effects from ANOVA were still valid with the normalized data, diazepam drug groups were first analyzed individually with one sample t-tests to confirm difference (no difference = normalized value of 100) and then compared using two-tailed unpaired t-tests. Data in plots are displayed as individual data points and figure error bars expressed as the group means ± SEM.

Transparency and openness

We report how we determined our sample size, all data exclusions, all manipulations, and all measures in the study. Data were analyzed and graphed using PRISM 10.1.2 for Windows 64-bit (GraphPad, La Jolla, CA, USA). This study’s design and its analysis were not pre-registered. All data are available as a supplemental file and can also be obtained by emailing the corresponding author.

RESULTS

Effects of stress

Rats were subjected to 7 days of social defeat or control conditions (Figure 1A) and then administered vehicle or diazepam. Body weight was measured to ensure that RSDS was effective. Social defeat exposure greatly reduced weight gain (Figure 3A; t(48) = 5.133, p < 0.001, two-tailed unpaired t-test, Table 1). Analysis of social defeat was performed to quantify the magnitude of stress exposure, and to ensure that rats that eventually got vehicle or diazepam had similar previous stress exposure because quality of defeat sessions can modify the influence of stress on behavior (Golden et al., 2011). Social defeat resulted in a total average of approximately 20 total attacks with no statistical difference between drug groups (Table 1; t(24) = 0.3120, p = 0.7578, n = 13/group, two-tailed unpaired t-test). Attack latency (amount of time before the subject is first attacked by the resident per session), averaged across sessions was 84 s with no statistical difference between diazepam and saline groups (Figure 3B; t(24) = 1.645, p = 0.1131, n = 13/group, two-tailed unpaired t-test, Table 1), and a similar total stressor exposure of 19.25 min each day (direct physical exposure with resident + 15 min indirect exposure) for each subject over the span of 7 sessions (Figure 3C; t(24) = 0.3482, p = 0.7307, n = 13/group, two-tailed unpaired t-test, Table 1). Therefore, rats in diazepam and vehicle groups were similarly exposed to social stress prior to any pharmacological or behavioral testing.

Figure 3.

Effects of resident-intruder social defeat stress. (A) Rats gained weight over the timecourse of control handling. Rats exposed to repeated resident-intruder stress gained significantly less weight compared to control (p <0.001, two-tailed unpaired t-test, n = 24 control rat, 26 stress rats/group). (B) There was a steady latency until first attacks across the 7 resident-intruder stress sessions (n = 26 rats). (C) There was a stable degree of exposure to the resident prior to the physical separation phase across the 7 resident-intruder stress sessions (n = 26 rats). *** p < 0.001.

Table 1.

Group data for stress metrics and elevated plus maze activity. Control and Stress data are mean ± SEM, DIFF = difference between means ± SEM, 95% CI = 95% confidence interval.

MEASURE	DRUG	CONTROL	STRESS	DIFF	95% CI
Body weight gain	-	30.0 ± 9.5 g	16.2 ± 1.9 g	−13.8 ± 2.7 g	[−19.2, −8.4]
Total attacks	saline	-	20.5 ± 3.5	1.3 ± 4.2	[−10.0, 7.34]
	diazepam		19.2 ± 2.3
Attack latency	saline	-	95.2 ± 11.2 s	−22.9 ± 13.9	[−51.7, 5.8]
	diazepam		72.3 ± 8.3 s
Total stress exposure	saline	-	9.4 min ± 26.3 s	−15.1 ± 43.4	[−104.7, 74.7]
	diazepam		9.1 min ± 34.5 s
Total distance (m) in EPM	saline	3.69 ± 0.78 m	13.00 ± 0.77 m	−0.70 ± 1.10	[2.96, 1.57]
	diazepam	15.02 ± 1.37 m	15.59 ± 0.88 m	0.57 ± 1.58	[−2.71, 3.85]
Closed arm preference	saline	77.5 ± 12.1 s	126.4 ± 17.3 s	48.9 ± 21.2	[5.2, 92.6]
	diazepam	76.7 ± 24.4 s	38.3 ± 19.2 s	−38.4 ± 30.6	[−101.9, 25.1]

The distribution of exploration in the EPM is used as a standard measure of rodent anxiety-like behavior. We first focused on effects of stress in vehicle-treated rats. Stress decreased open arm exploration in the EPM compared to controls (Figure 4A; stress x drug interaction, F_(1,46) = 5.258, p = 0.0265, n = 12-13/group, 2-way ANOVA; saline control compared to saline stress, p < 0.05, Šídák's multiple comparisons test). Stress did not have significant effects on measures of exploration in the closed arm (Figure 4B; main effect of stress F_(1,46) = 0.02046, p = 0.8869; stress x drug interaction F_(1,46) = 1.371, P = 0.2476, n = 12-13/group, 2-way ANOVA). Stress had no impact on the total distance traveled on the EPM, suggesting no locomotor differences between groups (Figure 4C; main effect of stress F_(1,46) = 0.004331, p = 0.9478; stress x drug interaction F_{(1, 46)} = 0.4448, p = 0.5082, n = 12-13/group, 2-way ANOVA). While this demonstrates a clear shift in open arm exploration without an effect on general locomotion, the effect of stress on preference for closed arm can be directly measured. RSDS significantly increased preference for closed arm (Table 1; stress x drug interaction F_(1,46) = 6.032, p = 0.0179, n = 12-13/group, 2-way ANOVA; saline-control compared to saline-stress, p < 0.05, Šídák's multiple comparisons test). Although, regardless of stress, rats preferred the closed arm over the open arm. However, stress impacts arm preference, leading to less time spent exploring the open arm.

Figure 4.

General exploration in the EPM. (A) Stress reduced the time spent exploring the open arm (p < 0.05, Šídák's multiple comparisons test after 2-way ANOVA (stress x drug), n = 12-13 rats/group). (B) Closed arm exploration was not significantly different between control and stress groups (p > 0.05, 2-way ANOVA, n = 12-13 rats/group). (C) Total distance traveled was not significantly different between control and stress groups (p > 0.05, 2-way ANOVA, n = 12-13 rats/group). * p < 0.05.

Effects of stress on risk assessment

In the EPM, the occurrence of SAP was used as an indicator of risk assessment as an association of anxiety-related behavior (Rodgers et al., 1995; Setem et al., 1999). To determine if stress impacted the number of SAP events, comparisons were performed using a 3-way RM-ANOVA with parameters location (open, closed), stress (stress, control), and drug (saline, diazepam). A main effect of location was found (Figure 5A; F_(1,92) = 100.4, p <0.0001), but there were no effects of stress (main effect of stress F_(1,92) = 0.9813, p = 0.3245, stress x location interaction F_(1,92) = 0.06706, p = 0.7962; n = 12-13/group), with more SAP events when the body was in closed arms (Table 2. control-saline, events in open arms vs closed arms p <0.001; stress-saline, events in open arms vs closed arms p <0.001, post hoc Šídák's multiple comparisons test). Most SAP behaviors were performed from the closed arms, and stress did not significantly change this pattern. Similarly, the amount of time spent in SAP posture was greater from the closed arm (Figure 5B; 3-way RM-ANOVA with parameters location (open, closed), stress (stress, control), and drug (saline, diazepam); main effect of location F_(1,92) = 54.61, p < 0.0001; n = 12-13/group) and stress did not significantly alter this pattern (main effect of stress, F_(1,92) = 1.544, p = 0.2171; location x stress interaction, F_(1,92) = 0.1261, p = 0.7233; post hoc Šídák's multiple comparisons test significant difference between open and closed arm in stress (p = 0.0158) and control (p = 0.0035)). As described above, stress impacted the amount of time in the open arms, so we normalized values of stretching based on the individual time spent in the open arms or the number of open arm entries (SAP in the open arm, oSAP). There was no effect of stress on the number of oSAP events over the total time in the open arm (Table 2; main effect of stress F_(1,44) = 2.105, p = 0.15391; stress x drug interaction F_(1,44) = 3.020, p = 0.0892; 2 values omitted due to no open arm exploration: saline-stress n = 12, saline-control n = 13, diazepam-stress n = 11, diazepam control n = 13). There was also no effect of stress on the number of oSAP events over the number of open arm entries (Table 2; main effect of stress F_(1,44) = 2.124, p = 0.1521; stress x drug interaction F_(1,44) = 3.258, p = 0.0779; 2 values omitted due to no open arm exploration: saline-stress n = 12, saline-control n = 13, diazepam-stress n = 11, diazepam control n = 13). Stress strongly trended towards increasing the amount of time stretching in the open arm (oSAP) over the total time in the open arm (Figure 5C; stress x drug interaction F_(1,44) = 3.977, p = 0.0523; main effect of stress F_(1,44) = 3.358, p = 0.0736; 2 values omitted due to no open arm exploration: saline-stress n = 12, saline-control n = 13, diazepam-stress n = 11, diazepam control n = 13). This suggested heightened risk assessment when in an unprotected environment after stress.

Figure 5.

Stretch and attend (SAP) risk assessment behavior was sensitive to stress. (A) SAP occurs in the closed arm more frequently than the open arm in both control and stress groups (p <0.001, Sidak’s multiple comparisons test after 3-way ANOVA, n = 12-13 rats/group). (B) Time spent in SAP is higher in the closed arm is higher than the open arm (p < 0.01, Sidak’s multiple comparisons test after 3-way ANOVA, n = 12-13 rats/group). (C) SAP in the open arm (oSAP time) was normalized to the total open arm time. Stress strongly trended towards increasing the proportion of time performing oSAP when in the open arms (stress x drug interaction p = 0.052, 2-way ANOVA, n = 12-13 rats/group). (D) SAP from the closed arm (cSAP) is a risk assessment often performed prior to entering an open arm. The proportion of times a rat goes to the open arm or stays in the closed arm was measured. Stress significantly reduced the likelihood that a rat will enter the open arm after a cSAP risk assessment (p < 0.05, Sidak’s multiple comparisons test after 2-way ANOVA, n = 12-13 rats/group). (E) In the control group, if a rat enters an open arm, there is approximately 50% likelihood that a rat performed cSAP prior to entering. Stress increased the proportion of open arm entries preceded by cSAP (p < 0.05, Sidak’s multiple comparisons test after 2-way ANOVA, n = 12-13 rats/group). *** p < 0.001, ** p < 0.01, * p < 0.05.

Table 2.

Group data for stretch and attend postures. Control and Stress data are mean ± SEM, DIFF = difference between means ± SEM, 95% CI = 95% confidence interval. SAP = stretch and attend; oSAP = stretch and attend in open arm; cSAP = stretch and attend from closed arm.

MEASURE	DRUG	CONTROL	STRESS	DIFF	95% CI
SAP number: closed arm	saline	15.2 ± 1.7	17.6 ± 1.2	2.4 ± 2.1	[−1.96, 6.74]
	diazepam	15.8 ± 1.4	14.9 ± 1.6	−1.0 ± 2.1	[−5.38, 3.43]
SAP number: open arm	saline	4.4 ± 0.7	5.2 ± 1.1	0.8 ± 1.3	[−1.93, 3.47]
	diazepam	6.8 ± 1.1	8.5 ± 1.7	1.6 ± 2.1	[−2.72, 6.01]
SAP time: closed arm	saline	67.4 ± 8.2 s	75.1 ± 7.7 s	7.8 ± 11.2	[−15.38, 30.90]
	diazepam	63.3 ± 7.9	65.0 ± 7.5	1.7 ± 10.9	[−21.02, 24.33]
SAP time: open arm	saline	22.7 ± 4.8 s	29.2 ± 7.3 s	6.5 ± 8.8	[−11.61, 24.52]
	diazepam	25.8 ± 7.5 s	36.3 ± 8.4 s	10.5 ± 11.4	[−13.22, 34.21]
(oSAP events) / (open arm time)	saline	0.06 ± 0.01 events/s	0.24 ± 0.11 events/s	0.18 ± 0.11	[−0.04, 0.40]
	diazepam	0.10 ± 0.01 events/s	0.09 ± 0.01 events/s	−0.02 ± 0.01	[−0.04, 0.01]
(oSAP events) / (open arm events)	saline	1.07 ± 0.17 events/entries	1.93 ± 0.46 events/entries	0.87 ± 0.48	[−0.17, 1.85]
	diazepam	1.21 ± 0.10 events/entries	1.12 ± 0.15 events/entries	−0.09 ± 0.19	[−0.49, 0.31]
(oSAP time) / (open arm time)	saline	0.3 ± 0.06	0.5 ± 0.07	0.22 ± 0.09	[0.03, 0.41]
	diazepam	0.3 ± 0.04	0.03 ± 0.04	−0.01 ± 0.06	[−0.14, 0.12]
“go” after cSAP	saline	0.24 ± 0.04 events	0.12 ± 0.03 events	0.13 ± 0.05	[−0.22, −0.03]
	diazepam	0.25 ± 0.06 events	0.27 ± 0.05 events	0.02 ± 0.07	[−0.12, 0.17]
open arm preceded by cSAP	saline	0.43 ± 0.09	0.72 ± 0.07	−0.29 ± 0.12	[−0.53, −0.05]
	diazepam	0.57 ± 0.05	0.52 ± 0.08	−0.05 ± 0.10	[−0.26, 0.15]

Since SAP is a risk assessment behavior, and it is mostly performed from the closed arms (SAP in the closed arm, cSAP), we examined the outcome of this assessment. Choice to venture out of the protected space and into the unprotected space (“go”) was compared to the decision to stay in the protected area (“stay”) following risk assessment (cSAP). Overall, rats chose to “stay” more often than “go” (Figure 5D; 2-way RM-ANOVA with parameters of choice (stays, goes) and stress exposure (stress, control); n = 13/group; main effect for choice F(_1,24) = 71.02, p < 0.0001). Stress decreased the incidence of “go” after cSAP (stress x choice interaction F(_1,24) = 4.615, p = 0.0420; main effect of stress exposure F(_1,24) = 1.152, p = 0.9820; decreased open arm choice in stress compared to control, p = 0.0415 post hoc Šídák's multiple comparisons test; no effect on closed arm choice, p = 0.4023 (stress M = 0.5841 ± 0.04216 events, control M = 0.5195 ± 0.04183 events, post hoc Šídák's multiple comparisons test). SAP from the closed arm to the open arm presumably assesses risk of open arm entry. If stress increases risk assessment, it is expected that entries into the open arm would be preceded by risk assessment more frequently. Stress exposure significantly increased the proportion of open arm entries preceded by a cSAP and the total amount of open arm entries (Figure 5E; 2-way ANOVA, stress x drug interaction, F_(1,44) = 4.992, p = 0.03062; saline-control compared to saline-stress, p < 0.05 Šídák's multiple comparisons test 2 values omitted due to no open arm entry: saline-stress n = 12, saline-control n = 13, diazepam-stress n = 11, diazepam-control n = 13).

Rearing and over-edge looking can be either risk assessment or exploratory behaviors. We measured these to determine if they are impacted by stress. There was no significant effect of stress on the total time engaged in rearing (Figure 6A; main effect of stress F_(1,46) = 0.1740, p = 0.6785; stress x drug interaction F_(1,46) = 0.2038, p = 0.6538, two-way ANOVA). Further analysis comparing between EPM areas – closed, middle, and open arms – demonstrated that most of the rearing was performed in the closed arms (Figure 6B; 3-way RM-ANOVA with parameters location (closed, middle, open), stress exposure (stress, control), and drug (saline, diazepam) main effect of location, F_(2,92) = 141.8, p <0.0001, n = 12-13/group; significant differences between closed and open (p < 0.0001), open and middle (p < 0.0001) and between closed and middle (p < 0.0001; Tukey's multiple comparisons test). However, there was no significant effect of stress (main effect of stress, F_(1,46) = 0.1740, p = 0.6785; stress x location interaction, F_(2,92) = 0.001223, p = 0.9988, 3-way ANOVA). Similarly, rear frequency was different between EPM areas (Figure 6C; 3-way RM-ANOVA with parameters location (closed, middle, open), stress exposure (stress, control), and drug (saline, diazepam); main effect of location, F_(2,92) = 159.0, p <0.0001, n = 12-13/group; significant differences between closed and open (p < 0.0001), and between closed and middle (p < 0.0001), post hoc Tukey's multiple comparisons test). But there was not a significant effect of stress (main effect of stress, F_(1,46) = 0.1740, p = 0.6785, stress x location interaction, F_(2,92) = 0.001223, p = 0.9988). Additionally, no significant difference was observed in the total time spent looking over the side of the open arm (Figure 6D; main effect of stress F_{(1, 46)} = 0.04915, p = 0.8255; stress x drug interaction F_(1,46) = 0.5643, p = 0.4564, n = 12-13/group, 2-way ANOVA). Thus, rearing in the EPM and head dipping over the side may not be sensitive to the social stressor used here.

Figure 6.

Rearing and head-dipping in the EPM were not sensitive to stress. (A) Stress did not significantly impact the total time spent rearing (p > 0.05, 2-way ANOVA, n = 12-13 rats/group). (B) When rearing was separated based on location, most rearing time was observed in the closed arms compared to open arms and middle of the EPM (p < 0.001, Sidak’s multiple comparisons test after 3-way RM-ANOVA, n = 12-13 rats/group). But there was still no effect of stress even when analyzed by location (p > 0.05, Sidak’s multiple comparisons test). (C) Similarly most of the rearing events occurred in the closed arms compared to open arms and middle of the EPM (p < 0.001, Sidak’s multiple comparisons test after significant main effect of location in 3-way RM-ANOVA, n = 12-13 rats/group), with no significant effect of stress (p > 0.05, Sidak’s multiple comparisons test). (D) All head-dipping (looking over side) is performed in an open arm. There was no significant effect of stress on head-dipping (p > 0.05, 2-way ANOVA) **** p < 0.001.

Effects of diazepam

The data above demonstrate a link between risk assessment and exploration on the EPM that is sensitive to stress. Prior work has demonstrated that EPM exploration is sensitive to benzodiazepines. But it is not clear if risk assessment is sensitive to benzodiazepines. In addition, the suppression of exploration in the EPM after stress is thought to be produced by a process that is sensitive to benzodiazepines. Therefore, we used a benzodiazepine to test the link between risk assessment and exploration, and to test whether stress might exert effects on risk assessment via a process that is sensitive to benzodiazepines. We chose diazepam, a prototypical benzodiazepine, at a dose that modulates EPM exploration without causing sedation. We compared diazepam-treated rats run alongside vehicle-treated rats (groups from above). To verify absence of a sedative effect of diazepam, we first examined general locomotion. When comparing the total distance traveled in the EPM, there was a main effect of diazepam (2-way ANOVA with parameters drug (saline, diazepam) and stress exposure (stress, control), main effect of drug F_(1,46) = 4.273, p = 0.0444), with diazepam tending to increase total distance traveled (p > 0.05 between all groups, post hoc Šídák's multiple comparisons test). These results provide evidence that the dosage of diazepam used did not induce sedation. To focus on the effects of diazepam here, we report main effects and interactions of diazepam (effects of stress reported above). When there is a significant ANOVA main effect of diazepam or diazepam x stress interaction, to isolate interactions and facilitate visualization of the magnitude of these effects between control and stress, we normalized each diazepam value to its own vehicle control, separately for non-stress and stress groups, to obtain the magnitude of effect of diazepam. These were compared between control and stress groups. Diazepam selectively increased open arm time in the stress group (stress x drug interaction F_(1,46) = 5.258, p = 0.0265, 2-way ANOVA; p < 0.05 stress-saline compared to stress-diazepam, p > 0.05, control-saline compared to control-diazepam, Šídák's multiple comparisons test). To directly assay the amplitude of the effect, we examined the diazepam values normalized to their respective controls (diazepam normalized to saline; control p = 0.7852, t(10) = 0.2800, one sample t-test; stress group p = 0.0051, t(12) = 3.422, one sample t-test), with a significant difference in the magnitude of the effects of diazepam (Figure 7A; t(22) = 3.026, p = 0.0062, two-tailed unpaired t-test). This provides the opportunity to dissociate effects of diazepam on risk assessment by comparing between a condition that showed an effect of diazepam on anxiety-like behavior and a condition that did not.

Figure 7.

Effects of diazepam differ between control and stress groups. To test whether diazepam influenced risk assessment behaviors and if this was different between group, the effects of diazepam were normalized to saline vehicle to obtain relative effect of diazepam. This was compared between control and stress groups. (A) Diazepam increased the amount of time spent in the open arm in stressed rats (p < 0.05, one-sample t-test) but not control rats (p > 0.05, one-sample t-test), and there was a significant difference in the effect of diazepam between control and stress group (p < 0.01, two-tailed unpaired t-test, n = 11-13 rats/group). (B) Diazepam decreased the proportion of time performing oSAP in the open arm in stress rat (p < 0.05, one-sample t-test) but not control rats (p > 0.05, one-sample t-test), and there was a significant difference in the effect of diazepam between control and stress group (p < 0.01, two-tailed unpaired t-test, n = 11-13 rats/group). (C) Diazepam decreased the frequency of oSAP events in stress rats (number of events/time in open arm; (p < 0.05, one-sample t-test) rats but increased the frequency of oSAP in control (p < 0.05, one-sample t-test), and there was a significant difference in the effect of diazepam between control and stress groups (p < 0.001, two-tailed unpaired t-test, n = 11-13 rats/group). (D) Diazepam reduced the proportion of open arm trips wherein oSAP was displayed in stress rats (p < 0.05, one-sample t-test) but not in control rats (p > 0.05, one-sample t-test), and there was a significant difference in the effect of diazepam between control and stress groups (p < 0.001, two-tailed unpaired t-test, n = 11-13 rats/group). (E) Diazepam increased the probability of open arm entry following a cSAP in stress rats (p < 0.05, one-sample t-test) but not in control rats (p > 0.05, one-sample t-test), and there was a significant difference in the effect of diazepam between control and stress groups (p < 0.05, two-tailed unpaired t-test, n = 11-13 rats/group). (F) Diazepam reduced the proportion of open arm entries preceded by cSAP in stress rats (p < 0.05, one-sample t-test) but increased the proportion of open arm entries preceded by cSAP in control rats (p < 0.05, one-sample t-test), and there was a significant difference in the effect of diazepam between control and stress groups (p < 0.05, two-tailed unpaired t-test, n = 11-13 rats/group). Significant effect of diazepam (one-sample t-test) and the direction of effect is denoted with up/down black arrowheads. * p < 0.05, ** p < 0.01, *** p < 0.001 control compared to stress.

Next, we compared the effects of diazepam on risk assessment behaviors. Diazepam had an effect on oSAP number that was different across control and stress groups (Table 2; stress x drug interaction F_(1,44) = 0.00956, p = 0.9225; main effect of drug F_(1,44) = 6.106, p = 0.0174; p > 0.05 Šídák's multiple comparisons test). This effect of diazepam may be due to effects on open arm exploration in general. To test this more carefully, oSAP was normalized by open arm exploration. Diazepam significantly lowered the proportion of time spent performing oSAP (time of stretch in open normalized to time in open) in the stress group (p = 0.0009, t(12) = 4.349, one sample t-test) but not in the control group (p = 0.4157, t(10) = 0.8532, one sample t-test, 1 value omitted due to no open arm time), and the magnitude of the effects of diazepam were significantly different between control and stress (Figure 7B; stress n = 13, control n = 11, t(22) = 3.172, p = 0.0046, two-tailed unpaired t-test). Diazepam also significantly reduced the rate of oSAP events (number of oSAP events normalized to time in open) in stressed rats (Figure 7C; p < 0.0001, t(12) = 15.22; one sample t-test) and control rats (p = 0.0005, t(10) = 5.302; one sample t-test, 1 value omitted due to no open arm time), with different magnitude of diazepam effect between control and stress groups (Figure 7C; stress n = 13, control n = 11, t(22) = 10.63, p < 0.0001, two-tailed unpaired t-test). Similarly, there was a significant difference between between control and stress in the effects of diazepam on the incidence of oSAP (oSAP events normalized to open arm entry events; Figure 7D; stress n = 13, control n = 11, t(22) = 4.601, p = 0.0002, two-tailed unpaired t-test). This indicates that diazepam has a modulating effect on open arm risk assessment in stressed rats, and this is parallel to the effects of diazepam on open arm exploration.

We then examined choice behavior in rats that performed cSAP. The total number of cSAP were not significantly impacted by diazepam (Table 2; main effect of drug F_(1,46) = 0.5299, p = 0.470; drug x stress interaction F_(1,46) = 1.254, p = 0.269, 2-way ANOVA, n = 12-13/group), nor was time performing cSAP (Table 2; main effect of drug F_(1,46) = 0.8201, p = 0.370; drug x stress interaction F_(1,46) = 0.1511, p = 0.699, 2-way ANOVA, n = 12-13/group). However, in stressed rats diazepam significantly shifted behavior after cSAP. Diazepam impacted the likelihood of entering an open arm (“go”) after cSAP (Table 2; main effect of drug F_(1,46) = 5.156, p = 0.0279, 2-way ANOVA; p < 0.05, stress-saline compared to stress-diazepam, Šídák's multiple comparisons test). As above, to closely compare the magnitude of diazepam effects, control and stress diazepam groups were normalized to their respective saline, and diazepam increased the likelihood of “go” responses after cSAP in the stress group (Figure 7E; stress, n = 13, p = 0.0049, t(12) = 3.436), but not the control group (control, n = 11, p = 0.8865, t(10) = 0.1464), with a significant difference in the effects of diazepam between control and stress groups (t(22) = 2.772, p = 0.0111, two-tailed unpaired t-test). Similarly, diazepam impacted the proportion of open arm entries that were preceded by cSAP (Table 2; drug x stress interaction F_(1,44) = 4.992, p = 0.0306, 2-way ANOVA; p < 0.05, stress-saline compared to stress-diazepam, Šídák's multiple comparisons test), with diazepam significantly decreasing this proportion in stressed rats (Figure 7F; stress: n = 13, p = 0.0241, t(12) = 2.579, one sample t-test), and controls (1 sample omitted due to lack of open arm entries; control n = 10, p = 0.0139, t(9) = 3.047), with a significant difference between control and stress groups in the effect of diazepam (1 sample omitted due to lack of open arm entries; t(21) = 3.901, p = 0.0008, two-tailed unpaired t-test).

Diazepam did not significantly impact rearing when total rearing time (Table 3; main effect of drug F_(1,46) = 0.9835, p = 0.3265; drug x stress interaction F_(1,46) = 0.2038, p = 0.6538, 2-way ANOVA) or closed arm rearing was compared (main effect of drug F_(1,46) = 1.996, p = 0.164; drug x stress interaction F_(1,46) = 0.4318, p = 0.5144). Additionally, head-dipping over the side was not affected by diazepam (Table 3. main effect of drug F_(1,46) = 0.03931, p = 0.8437, drug x stress interaction F_(1,46) = 0.5643, p = 0.4564, 2-way ANOVA).

Table 3.

Group data for rearing and heap dip. Control and Stress data are mean ± SEM, DIFF = difference between means ± SEM, 95% CI = 95% confidence interval.

MEASURE	DRUG	CONTROL	STRESS	DIFF	95% CI
rearing time - closed	saline	20.2 ± 2.9 s,	22.6 ± 3.6 s	2.4 ± 4.6	[−7.1, 11.9]
	diazepam	18.0 ± 3.0	16.6 ± 1.9	−1.4 ± 3.4	[−8.5, 5.7]
rearing time - middle	saline	2.8 ± 1.4 s	2.8 ± 0.8 s	0.01 ± 1.6	[−3.2, 3.2]
	diazepam	3.0 ± 0.9 s	3.8 ± 1.2 s	0.8 ± 1.5	[−2.3, 3.9]
rearing time - open	saline	0.04 ± 0.04 s	0.5 ± 0.4 s	0.5 ± 0.4	[−0.4, 1.4]
	diazepam	0.27 ± 0.12 s	0.8 ± 0.2 s	0.5 ± 0.3	[0.012, 1.05]
rearing number - closed	saline	12.9 ± 1.7 events	14.9 ± 2.1 events	2.1 ± 2.7	[−3.5, 7.7]
	diazepam	12.1 ± 1.7 events	11.0 ± 1.4 events	−1.1 ± 2.1	[−5.5, 3.4]
rearing number - middle	saline	1.6 ± 0.7 events	2.3 ± 0.6 events	0.7 ± 0.9	[−1.1, 2.5]
	diazepam	2.4 ± 0.6 events	3.2 ± 0.9 events	0.9 ± 1.2	[−1.5, 3.3]
rearing number - open	saline	0.08 ± 0.08 events	0.5 ± 0.4 events	0.5 ± 0.4	[−0.4, 1.3]
	diazepam	0.5 ± 0.2 events	0.5 ± 0.3 events	0.08 ± 0.35	[−0.6, 0.8]
over edge head dip time	saline	12.7 ± 2.0 s	10.9 ± 2.4 s	−1.9 ± 3.1	[−8.3, 4.6]
	diazepam	10.7 ± 1.5 s	12.1 ± 1.4 s	1.3 ± 2.1	[−3.0, 5.7]

DISCUSSION

The present study investigated change in EPM risk assessing behaviors in rats subjected to RSDS. Consistent with our first hypothesis, we found that social stress indeed alters risk-assessing behaviors. Stress increased the relative number and amount of time engaged in SAP behaviors in the open arms. While stress did not impact the total amount of time engaged in SAP from a closed arm, it significantly altered the relationship between closed arm SAP and subsequent open arm exploration. After stress, rats were more likely to perform SAP from the closed arm prior to open arm exploration, and they were less likely to enter an open arm after SAP. This indicates that rats used a key risk assessment behavior to make decisions more frequently after stress. Because rats also performed this SAP behavior more frequently before entering an open arm, but were less likely to enter an open arm after SAP behavior, it implies that they also evaluated the open arm as more riskier. To further understand the link between SAP and open arm exploration, we treated rats with diazepam, based on its known ability to reverse effects of stress on open arm exploration in the EPM. Diazepam reversed the effects of stress on open arm exploration, and also mitigated the effects of stress on SAP behaviors, consistent with the second hypothesis. This further links SAP from the closed arm to subsequent open arm exploration. Overall, these results support the interpretation that SAP from the closed arm to the open arm is a risk assessment that is used prior to making a decision, and indicates that this risk assessment is sensitive to stress and contributes to the effects of stress on open arm exploration.

Prior work established that risk assessment behaviors in the EPM are dissociable factors from behaviors thought to reflect anxiety (Augustsson et al., 2005; Carobrez & Bertoglio, 2005; Doremus et al., 2006; Wall & Messier, 2001). Rodent SAP is believed to reflect an approach-avoidance conflict (Grant & Mackintosh, 1963; Kaesermann, 1986; Molewijk et al., 1995; Van Der Poel, 1979). While the link between SAP risk assessment and subsequent exploration is fairly robust in our results, it is unclear why stress modifies this link. One interpretation is that stress leads to assessment of the open arm as riskier. Alternatively, it is possible that rats evaluate the open arms as a similar level of risk whether or not they have undergone stress, but are more sensitive to risky options, such as the open arms, after stress. Prior work that examined effects of stress on risk assessment also observed increased SAP behaviors in the EPM after stress (Adamec et al., 2004; Reis et al., 2012; Wang et al., 2014), but most did not assess the relationship between SAPs and subsequent open arm behaviors. One study did find a negative correlation between SAP behaviors and open arm exploration, but that this relationship was disrupted by stress (Wang et al., 2014). Other studies on risky decision-making in reward tasks suggest that stress may push rodents to become more risk averse instead of shifting their assessment of level of risk (Chiavegatti & Floresco, 2024) although this may depend on trait anxiety (Zalachoras et al., 2022). Overall, this may suggest that stress increases risk assessment in the EPM, and leads to reduced open arm exploration, due to making rats more risk averse.

We found that diazepam can decrease risk assessment in the EPM after stress. Prior studies have also found that benzodiazepines can decrease risk assessment behaviors in the EPM (Albrechet-Souza et al., 2007; Cole & Rodgers, 1993; Weiss et al., 1998). Interestingly, the effects of diazepam for some measures were greater after stress compared to control. Indeed, in some instances diazepam had minimal effect in control animals, but a large effect after stress. We found that diazepam increased the likelihood to enter the open arm without prior risk assessment, and exploration in an unprotected space, in stressed animals but not control animals. Similarly, diazepam also greatly reduced the prospect of risk assessing SAPs in these unprotected spaces in stressed but not control animals. This further supports the link between SAPs and the decision to explore, aligning with the interpretation that stress makes rats more risk averse, resulting in decreased open arm exploration upon risk assessment. Moreover, our results suggest that diazepam’s anxiolytic effects, at the dose used, are greater under stress conditions, leading to a more pronounced reduction in risk assessment and an increase in exploratory behavior. This could imply an antagonistic effect between diazepam and the stress response, in which prior stress enhances the drug’s efficacy in reducing anxiety-like behaviors.

Throughout all the experiments, there was a dissociation between the effects of stress on SAPs and other potential risk assessment behaviors (head-dipping and rearing). These observations align with prior work that also finds differences in their sensitivity to anxiolytic drugs (Cole & Rodgers, 1994; Griebel et al., 1997; Weiss et al., 1998) and in their neural substrates (Deacon et al., 2002; Dean et al., 1980; Foreman et al., 1978; Harley & Martin, 1999; Loh et al., 2022; Marshall, 1978). This would align with the interpretation that head-dipping and rearing behaviors might not be utilized by the rat as risk assessment towards making a decision, but might reflect more broad exploratory aspects of behavior. The differential sensitivity of these behaviors to an anxiolytic drug underscores the complexity of anxiety-like behaviors and their underlying mechanisms. For instance, SAPs are closely linked to the decision-making process regarding the exploration of potentially threatening environments, making them a more direct measure of risk assessment. In contrast, head-dipping and rearing may reflect more general aspects of exploration rather than specific risk assessment behaviors sensitive to the EPM.

While our results demonstrate a sensitivity of SAP to repeated stress, and a link between this risk assessment and subsequent exploration, there are some limitations of this study. We chose a dose of diazepam that reversed the effects of stress, but did not substantially alter EPM exploration based on preliminary studies. This was beneficial here because it allowed more selective effects, and avoided known sedating effects of diazepam at higher doses (Hazim et al., 2014; Rex et al., 2004) that could potentially disrupt risk assessment. By avoiding these sedative effects, we ensured that the observed changes in behavior were due to the anxiolytic effects of diazepam rather than its potential to impair motor function or induce lethargy. Another significant consideration is the type of stressor used in the study. We utilized a social defeat stress, where a rat was physically attacked by a larger, more aggressive rat. The physical nature of this stressor provides a real external threat to these animals, making it a robust and ethologically relevant model for inducing stress-related behavior. However, it also raises the question of how different forms of stress that do not rely on an external threat might influence risk assessment and exploratory behavior, such as sickness or sleep deprivation. These types of stressors can also impact physiological and behavioral responses through different mechanisms, potentially leading to variations in how animals exhibit anxiety and risk assessment behaviors. Additionally, the use of social defeat stress poses sex-specific challenges when applying this model to female rats. The social dynamics and aggression patterns in female rats differ from those in males, making it difficult to establish a comparable model of social defeat stress for female subjects (Blume et al., 2017, 2019; Kim et al., 2010; Kuske & Trainor, 2021; Patki, Solanki, & Salim, 2014). This is a significant limitation in our study, as it restricts the generalizability of our findings to male rats. Understanding sex differences in stress responses is crucial. Future research should aim to address this gap by utilizing appropriate social stressors for female rats, thereby providing a more comprehensive understanding of the effects of stress on risk assessment across sexes.

CONCLUSIONS

Stress is a strong trigger for anxiety, significantly influencing expression of anxiety and the manifestation of defensive behaviors. These behaviors can arise from a confluence of factors that include individual differences in the anticipation or perception of a threat. A key intermediate component in this process is the assessment of threats and risks. Our results demonstrate that stress can increase risk assessment, and alter the immediate subsequent behaviors following risk assessment, biasing a rodent towards more cautious decisions. This indicates that stress modulates defensive decisions through effects on risk assessment. A better understanding of these effects of stress on risk assessment may lead to novel approaches to treat disorders triggered by stress.

Transparency and openness.

We report how we determined our sample size, all data exclusions, all manipulations, and all measures in the study. Data were analyzed and graphed using PRISM 10.1.2 for Windows 64-bit (GraphPad, La Jolla, CA, USA). This study’s design and its analysis were not pre-registered. All data are available as a supplemental file and can also be obtained by emailing the corresponding author.