Dissociable role of the basolateral complex of the amygdala in the acquisition and extinction of conditioned fear following reproductive experience in female rats

In female rats and humans, reproductive experience (i.e., pregnancy) alters the behavioral, hormonal and molecular substrates of fear extinction. Here, we assessed whether the role of a central neural substrate of fear extinction, the basolateral amygdala (BLA), also changes following reproductive experience. Nulliparous (virgin) and primiparous (one prior pregnancy) female rats received infusions of the GABA A agonist, muscimol, to temporarily inactivate the BLA prior to fear conditioning or extinction training. In follow up experiments, the BLA was also inactivated immediately after extinction training. BLA inactivation impaired the acquisition and expression of conditioned fear in both nulliparous and primiparous rats. In nulliparous rats, BLA inactivation prior to or immediately after extinction training impaired extinction retention. In contrast, in primiparous rats, BLA inactivation prior to or immediately after extinction training did not impair extinction retention, despite suppressing freezing during extinction training. These results suggest that, consistent with past findings in males, the BLA is a central component of the neural circuitry of fear acquisition and its extinction in virgin female rats. However, after pregnancy, female rats no longer depend on the BLA to extinguish fear, despite requiring the BLA to acquire conditioned fear. Given that fear extinction forms the basis of exposure therapy for anxiety disorders in humans, the present findings may have clinical implications. To improve the efficacy of exposure therapy for anxiety disorders, we may need to target different mechanisms in females dependent on their reproductive history.


Introduction
Globally, anxiety disorders are the sixth largest contributor to nonfatal health loss and, with depression, cost US$ 1 trillion per year (The Lancet Global Health, 2020).Relative to men, anxiety disorders are twofold more prevalent in women, who also experience greater symptom severity, burden of illness, comorbidity, and relapse (GBD, 2019Mental Disorders Collaborators,2022;Li & Graham, 2017;McLean et al., 2011).Given the disproportionate number of women diagnosed with anxiety, identifying the female-unique factors underpinning anxiety may help to reduce its overall burden.Yet, two thirds of animal research on fear and anxiety in 2021 exclusively studied males (Kaluve et al., 2022), raising questions regarding the degree to which knowledge developed from preclinical models of anxiety accurately represents females.
Extinction training is procedurally and mechanistically analogous to exposure therapy, the first-line psychological treatment for anxiety disorders (Craske et al., 2014); thus, fear extinction is often studied to develop knowledge on how to improve exposure therapy (Graham & Milad, 2011).The behavioral and neurobiological mechanisms of fear extinction have been extensively mapped in rodents (Bouton et al., 2020;Orsini & Maren, 2012;Tovote et al., 2015), but the routine omission of females from fear extinction research is a shortcoming that has created a substantial knowledge gap in how females regulate fear.To close this gap, we must not only include females in fear extinction research, but also, consider variables of unique relevance to females.For example, female rodents and humans experience changes in ovarian hormones over their reproductive cycle (estrous and menstrual, respectively).Research has consistently shown that female fear extinction is facilitated during periods of high ovarian hormones, and impaired during periods of low ovarian hormones (Glover et al., 2012;Graham & Scott, 2018a;Li & Graham, 2016;Milad et al., 2009Milad et al., , 2010;;Rey et al., 2014;Wegerer et al., 2014), which may be in part driven by the impact of estradiol on the mechanisms of extinction learning, including the N-methyl-d-aspartate receptor (NMDAr) (Graham & Scott, 2018b).
Another female-unique factor is motherhood (encompassing pregnancy, parturition, and early maternal experience), which leads to long-term modifications in fear-relevant brain regions in mammals (Carmona et al., 2019;de Lange et al., 2019;Duarte-Guterman et al., 2019;Hoekzema et al., 2017;Pawluski et al., 2022;Pestana et al., 2023).Our investigations comparing fear extinction in nulliparous (virgin) and primiparous (one prior pregnancy) rats have revealed that pregnancy changes fear extinction in ways that challenge the current model of fear extinction.For example, a key tenet of this model is that fear extinction involves context-dependent inhibitory learning that depends on NMDAr activation and leaves the original feared association intact (Bouton et al., 2020;Craske et al., 2014;Orsini & Maren, 2012;Tovote et al., 2015).As such, conditioned responses re-emerge under certain conditions, e.g., when the extinguished CS is presented in a different context (renewal), or following un-signaled stress (reinstatement).Yet, we have shown that whereas fear extinction in nulliparous female rats is impaired by a systemic injection of an NMDAr antagonist, and susceptible to renewal and reinstatement, fear extinction in primiparous female rats is NMDAr-independent (Tang & Graham, 2019b), and resistant to renewal and reinstatement (Milligan-Saville & Graham, 2016).Moreover, we have demonstrated that whereas fear extinction is modulated by ovarian hormones in nulliparous female rats and humans, ovarian modulation of fear extinction is abolished after pregnancy in female rats and humans (Milligan-Saville & Graham, 2016;Pestana et al., 2021;Tang & Graham, 2020).
Given the differences in the behavioral, hormonal, and molecular mechanisms of fear extinction in primiparous versus nulliparous rats and humans, it is possible that reproductive experience fundamentally alters the neural circuitry of fear extinction.Based on studies predominantly conducted using males, it is widely agreed that activitydependent plasticity in the basolateral amygdala (BLA) is essential for the acquisition, expression, and extinction of conditioned fear (Bouton et al., 2020;Orsini & Maren, 2012;Tovote et al., 2015).Although amygdala activity has also been associated with fear extinction in female rats and humans (Zeidan et al., 2011), to the best of our knowledge, no research has assessed whether the BLA is necessary to fear extinction in females, irrespective of reproductive status.Therefore, in this study we sought to investigate the role of the BLA in fear extinction in nulliparous female rats, and to determine whether this role is altered in primiparous female rats.We infused muscimol, a GABA A receptor agonist, to selectively inhibit BLA activity prior to fear conditioning or extinction training in nulliparous and primiparous female rats.Follow-up experiments assessed the impact of post-extinction training BLA inactivation on extinction retention in primiparous and nulliparous rats.The outcomes demonstrated that whereas the BLA is necessary for fear acquisition and expression in female rats irrespective of reproductive status, the involvement of the BLA in the acquisition and consolidation of fear extinction differs as a function of reproductive status.

Animals
Age-matched, naturally cycling nulliparous and primiparous female Sprague-Dawley rats (17-24 weeks) from the Animal Resources Centre (ARC), WA, Australia were used in all experiments.See Figure captions for the number of animals in each experiment.Primiparous rats were bred, gave birth, and remained with their litter until weaning at the ARC, prior to their arrival at UNSW, as per Pestana et al. (2021).Upon arrival, rats were housed in groups of 5-8 in plastic boxes (67 cm long × 30 cm wide × 22 cm high) filled with corncob bedding, and covered with a wire lid.The boxes were kept in a colony room maintained at 20-22 • C on a 12 h light/dark cycle (lights on at 7:00 A.M.).Food and water were available ad libitum and rats remained under these conditions for an acclimatization period of 2 weeks prior to the commencement of any procedures.Behavioral testing occurred at ~ 6 months of age, during the light phase, in line with previous studies (Milligan-Saville & Graham, 2016).All rats were treated according to The Australian Code of Practice for the Care and Use of Animals for Scientific Purposes (8th edition, 2013), and all procedures were approved by the Animal Care and Ethics Committee at UNSW.

Vaginal cytology and experimental design
Vaginal smears were conducted daily to determine estrous phase (Graham & Daher, 2016).Primiparous rats were fear conditioned at least 2 weeks post-weaning, when estrous cycling had recommenced (Milligan-Saville & Graham, 2016).Rats that did not exhibit a regular 4-5 d estrous cycle were excluded.In Experiment 1, half of the rats were conditioned during diestrus and tested for conditioning retention during proestrus (heightened ovarian hormones), while the other half were conditioned during estrus and tested during metestrus (lower ovarian hormones).This was to control for potential differences in fear conditioning due to estrous stage.As there was no significant effect of estrous phase on levels of baseline freezing and on CS-elicited freezing during conditioning and retention test, data were collapsed across estrous phase in the reported analyses.In Experiments 2-4, rats were conditioned during diestrus, extinguished during proestrus, and tested for extinction retention during estrous.We hypothesized that the BLA infusion of muscimol prior to extinction training would impair extinction retention in nulliparous rats extinguished during proestrus.By contrast, salinetreated nulliparous rats extinguished during metestrus should show complete fear relapse at extinction retention (Chang et al., 2009;Graham & Daher, 2016;Graham & Milad, 2013;Milad et al., 2009), and therefore, any potential impairing effect of muscimol in metestrus rats would not be detected.We confirmed this hypothesis in an initial experiment (Fig S1).Although estrous effects on fear extinction are typically abolished in primiparous rats (Milligan-Saville & Graham, 2016;Pestana et al., 2021;Tang & Graham, 2020), primiparous rats were also extinguished during proestrus in Experiments 2-4 for consistency with nulliparous rats.Previous research in nulliparous rats has shown that differences between metestrus and proestrus groups (based on estrous phase during extinction training) were evident at extinction retention irrespective of whether estrous phase during fear conditioning and extinction retention was held constant or free to vary (Graham & Daher, 2016;Milad et al., 2009).This suggests that ovarian hormone levels during extinction training have a greater impact on extinction retention than those during conditioning or retention.See Figure S3 for a schematic mapping the training phases against the estrous phases for each experiment.

Surgery
Rats were anesthetised with isoflurane (5 % in 1 L/min of oxygen, adjusted to 3 % in 0.6 L/min of oxygen) and mounted onto the stereotaxic apparatus (David Kopf Instruments).Rats received an injection of 0.1 mL of 0.5 % Bupivacaine (Merck) to the incision site under the scalp.Twenty-six-gauge guide cannulas (Bioscientific) were implanted bilaterally in the BLA via drilled holes in the skull (relative to bregma: anteroposterior, − 2.5 mm; mediolateral, ±4.9 mm; dorsoventral, − 8.4 mm).Coordinates and placements were selected based on past studies that have shown a role for the BLA in fear extinction in male rats (Laurent & Westbrook, 2008;Sierra-Mercado et al., 2011).Guide cannulas were affixed in position with dental cement supported by four jeweler's screws in the skull.Thirty-three-gauge dummy cannula were kept in each guide to maintain patency until infusions commenced.At completion of surgery, rats received subcutaneous injections of 1 mL/kg of the non-steroidal anti-inflammatory, Carprofen, (Rymadil; Cenvet Australia) 0.15 mL/kg Duplocillin (Cenvet, Australia), and 1 mL/100 g 0.9 % saline.Animals received post-operative monitoring for 7 d during which they were handled, weighed and swabbed daily.

Histology
Upon completion of behavioral testing, rats were euthanized via carbon dioxide, decapitated, and brains were extracted.Coronal sections were cut 40 µm thick, mounted on glass slides and stained with cresyl violet.Slides were observed under a microscope to verify cannula placements using Paxinos and Watson's atlas (Paxinos & Watson, 2006).A total of 18 rats were excluded across Experiments 1-4 due to incorrect cannula placement.

Drugs
The GABA A agonist, muscimol (Tocris, In Vitro Technologies, Australia), was administered 30 min prior to conditioning (Experiment 1) or extinction training (Experiment 2), or immediately after extinction training (Experiments 3 and 4) to inactivate the BLA.Dummy cannulas were removed and 33-gauge infusion cannulas extending 1 mm below the guide cannula were inserted.Muscimol, dissolved in 0.9 % saline to obtain a final concentration of 2.63 nmol/3µL, or saline vehicle, was infused at a rate of 0.1 µL/min/per hemisphere for a duration of 3 min (dose based on previous investigations (Laurent et al., 2008;Laurent & Westbrook, 2008;Sierra-Mercado et al., 2011).One day before infusions, the dummy cannulas were removed and the infusion pump was turned on for 3 min in order to acclimate the rats to the procedure.Two rats were excluded in Experiments 1 and 4 due to blocked cannula.

Behavioral apparatus
Training and testing took place in Med Associates chambers as previously described (Graham & Daher, 2016).Two distinct contexts (A and B) were used; the CS was a 10 s white noise (62-dB) delivered via highfrequency speakers embedded into the right wall of each chamber and the US was a 0.4-mA foot shock.Fear conditioning involved 5 x CS-US pairings in context A. Fear extinction involved 30 x 10 s presentations of the CS in context B. Fear conditioning and extinction retention tests involved 15 x 10 s CS presentations in context B. A computer running Med Associates Med-PC V controlled presentations of the CS and the US.

Context pre-exposure
Prior to commencement of behavioral training procedures, rats were individually pre-exposed to context A for 10 min for two consecutive days.The purpose of context pre-exposure was to reduce the amount of fear accrued to the context during conditioning when the CS and US are presented.

Conditioning
Rats were placed in context A on day 1.Following a 2 min adaptation period, the CS was presented for 10 s and co-terminated with the foot shock US.Rats received 5 x CS-US pairings with an intertrial interval (ITI) of 85-135 s (average = 110 s).

Fear conditioning retention
In Experiment 1 only, rats were tested for fear conditioning retention 24 h post-conditioning in context B (day 2).Following an adaptation period of 2 min, rats received 15 x 10-s presentations of the CS with an ITI of 10 s.

Extinction training
In Experiments 2-4, rats underwent extinction training 24 h postconditioning in context B (day 2).Following an adaptation period of 2 min, rats received 30 x 10-s presentations of the CS with an ITI of 10-s, except in Experiment 2, in which half of the rats did not receive the CS presentations (group 'no extinction').

Extinction retention
In Experiments 2-4, rats were returned to context B 24 h postextinction training (day 3).The extinction retention test was identical to the fear conditioning retention test.

Scoring and statistical analysis
Freezing, defined as the absence of all movements except those required for respiration (Blanchard & Blanchard, 1969) was manually scored as a measure of conditioned fear.Rats were scored as either freezing or not freezing every 3 s by an observer blind to group condition.Measurements of freezing were obtained during the pre-CS adaptation period, and then during each white-noise CS presentation, and were not obtained during the intertrial intervals.A percentage of observed freezing was calculated.Trials during extinction training, fear conditioning retention, and extinction retention were averaged into blocks consisting of five trials each.
Data were analyzed via SPSS using ANOVAs (Experiments 1-2), with the between subject factors of reproductive status (nulliparous versus primiparous), drug (saline versus muscimol), and extinction group (extinction versus no extinction, Experiment 2 only), and t-tests (Experiments 3-4).Repeated measures data was analyzed using mixedmodel ANOVA with the within-subjects factor of trial or block.A significance value of p < 0.05 was applied throughout.When sphericity was violated, nominal degrees of freedom were reported but the Greenhouse-Geisser correction was applied.Two statistical outliers (>2 STDEV away from the mean) were removed from Experiments 2 and 4.

Experiment 1
To evaluate the role of the BLA in fear acquisition in nulliparous and primiparous rats, saline or muscimol was infused into the BLA 30 min prior to fear conditioning on day 1.Rats were tested for conditioning retention, drug free, on day 2. Fig. 1 depicts freezing throughout each experimental phase.Pre-CS freezing did not differ between groups in any experimental phase (no main effects of drug condition or reproductive experience, or interaction between the two factors; largest F = 3.7, P = 0.06).During fear conditioning (Fig. 1B and F), CS-elicited freezing increased in saline-treated, but not muscimol-treated rats, irrespective of reproductive status, indicating that muscimol treatment prevented fear acquisition in both nulliparous and primiparous rats.This is supported by a main effect of trial (F (4,108) = 24.03,P < 0.001, η p 2 = 0.47), a significant trial-by-drug interaction (F (4,108) = 13.13,P < 0.001, η p 2 = 0.33), but no trial-by-reproductive status interaction, or trial-bydrug-by reproductive status interaction (largest F = 2.08, P = 0.12).On average, muscimol-treated rats froze significantly lower during fear conditioning relative to saline-treated rats, irrespective of reproductive status (main effect of drug; F (1,27) = 59.77,P < 0.001, η p 2 = 0.69; no main effect of reproductive status or drug-by-reproductive status interaction; Fs < 1).

Experiment 2
To evaluate the role of the BLA in fear extinction in nulliparous and primiparous rats, rats received fear conditioning on day 1, and on day 2, intra-BLA infusions of saline or muscimol 30 min prior to extinction training or no extinction (equivalent exposure to context B without CS presentations).On day 3, rats were tested for extinction retention, drug free.
Fig. 2 depicts freezing throughout each experimental phase.Pre-CS freezing prior to fear conditioning or extinction training did not differ between groups (no main effect of drug, reproductive status, extinction group, or interaction between any factor; largest F = 2.46, p = 0.12).During fear conditioning (Fig. 2B and 2G), CS-elicited freezing increased across conditioning trials (significant effect of trial; F (4,248) = 199.78,P < 0.001, η p 2 = 0.76).On average, there was no main effect of drug, reproductive status, extinction group, or interaction between any factor (largest F = 2.35, P = 0.13).However, there was a significant trial-bydrug-by-extinction group interaction (F (4,248) = 4.65, P = 0.003, η p 2 = 0.07).The interaction was due to 'no extinction' rats assigned to the saline groups exhibiting higher freezing on trial 4 relative to 'no extinction' rats assigned to the muscimol groups.In contrast, 'extinction' rats assigned to the saline groups exhibited lower freezing on trial 4 relative to 'extinction' rats assigned to the muscimol groups.Importantly, there were no differences on any other conditioning trial (largest F = 3.21, P = 0.08; trial 2).
To identify the cause of these interactions, extinction retention in nulliparous and primiparous rats was examined separately.In nulliparous rats, averaged across extinction retention blocks, there was no main effect of drug, a main effect of extinction group (F (1,28) = 23.1,P < 0.001, η p 2 = 0.45), and a significant drug-by-extinction group interaction (F (1,28) = 19.72,P < 0.001, η p 2 = 0.41).This was due to saline-treated rats exhibiting lower freezing than muscimol-treated rats in groups that received extinction training (P < 0.001); whereas saline-treated rats exhibited non-significantly higher freezing than muscimol-treated rats in groups that did not receive extinction training (P = 0.06).Thus, BLAinfusion of muscimol prior to extinction training impaired extinction retention in nulliparous rats, without impacting freezing during extinction retention in non-extinguished rats.In primiparous rats, averaged across extinction retention blocks, there was a main effect of extinction group (F (1,33) = 19.35,P < 0.001, η p 2 = 0.37), due to nonextinguished rats exhibiting higher freezing relative to extinguished rats, but there was no main effect of drug, and crucially, no drug-byextinction group interaction (Fs < 1).Thus, BLA-infusion of muscimol prior to extinction training did not impair extinction retention in primiparous rats, and had no impact on freezing during extinction retention in non-extinguished rats.Demonstrating the robustness and reliability of these findings, we replicated the dissociable role of the BLA in fear extinction in nulliparous and primiparous rats using different fear conditioning and extinction parameters in a separate cohort of rats (Fig S1 and S2).

Experiment 3
Experiments 1-2 found that both nulliparous and primiparous female rats require BLA activity to acquire and express conditioned fear.Furthermore, nulliparous rats require BLA activity during extinction training to retain fear extinction learning.In contrast, primiparous rats did not require activity in the BLA in order to retain fear extinction.
This raises the possibility that primiparous rats do not recruit the BLA to extinguish fear.An alternative possibility, however, is that primiparous rats may use the BLA to extinguish fear under normal conditions, but when neuronal transmission is suppressed in the BLA prior to extinction training via muscimol, primiparous rats are uniquely able to compensate by recruiting a different neural region to extinguish fear.Conceptually, this idea is consistent with the theory of dynamic memory systems, which suggests that when the primary neural circuit that mediates development of a particular type of memory under typical circumstances is damaged or compromised, alternate neural circuits can be recruited to develop and express that memory (Poulos et al., 2010).If it is the case that primiparous rats use the BLA to extinguish fear under typical conditions, but can recruit a compensatory brain circuit when the BLA is inactivated, we reasoned that primiparous rats should exhibit impaired extinction if BLA activity was suppressed immediately after extinction training, once the BLA circuit had already been engaged during extinction training.That is, it seems unlikely that compensatory mechanisms would be effective in the post-extinction (memory consolidation) phase, if extinction-induced plasticity has already commenced in the BLA during extinction acquisition.Therefore, in Experiment 3, primiparous rats received intra-BLA infusions of saline or muscimol immediately after extinction training, to test whether shifting the timing of BLA inactivation to post extinction training would unmask a role for the BLA in extinction in primiparous rats.
During extinction retention (blocks depicted in Fig. 3D, individual data points averaged across block depicted in Fig. 3E), both groups showed comparable reductions in CS-elicited freezing (significant effect of retention block; F (2,30) = 7.343, P = 0.003, η p 2 = 0.33).However, there was no significant main effect of drug, or block-by-drug interaction (Fs < 1).Thus, blocking BLA activity immediately after extinction training had no impact on extinction retention in primiparous rats.

Experiment 4
Experiment 3 demonstrated that primiparous rats do not require BLA activity during extinction consolidation in order to exhibit long-term extinction retention, providing further evidence that primiparous rats do not depend on the BLA to extinguish fear, even when their BLA is not compromised during extinction acquisition.However, past studies in male rats have produced mixed findings on the effects of post-extinction BLA inactivation on fear extinction, with some reporting no effects (Sierra-Mercado et al., 2011), whereas others have reported impairing effects (Laurent et al., 2008).To our knowledge, Experiment 3 was the first to examine the impact of post-extinction training BLA inactivation in female rats; thus, it is unclear whether the observed null effect is unique to primiparous females, or reflects a more general absence of a role for the BLA in the consolidation of extinction in females.Therefore, in Experiment 4, we infused saline or muscimol into the BLA of nulliparous rats immediately after extinction training to verify the role of the BLA in the consolidation of fear extinction in non-reproductively experienced females.Note that because Experiments 3 and 4 were conducted at different times and with different personnel, we did not directly compare the outcomes between the two studies.

Discussion
The current study found that in both nulliparous and primiparous female rats, BLA inactivation prior to fear conditioning suppressed CS-  elicited freezing during fear acquisition, and prevented the development of a long-term conditioned fear memory.BLA inactivation prior to extinction training also suppressed CS-elicited freezing during extinction training in both groups.Thus, consistent with past findings in males (Bouton et al., 2020;Orsini & Maren, 2012;Sierra-Mercado et al., 2011;Tovote et al., 2015), the BLA is necessary for both the acquisition and expression of conditioned fear in female rats, irrespective of reproductive status.The current study also demonstrated that the BLA is necessary for the extinction of conditioned fear in nulliparous female rats.As has been reported in males (Sierra-Mercado et al., 2011), BLA inactivation prior to extinction training led to high CS-elicited freezing (comparable to that of non-extinguished rats) during the extinction retention test, after BLA activity was restored.However, in contrast to nulliparous females, and previous observations in males (Sierra-Mercado et al., 2011), BLA inactivation prior to extinction training produced no detrimental effects on extinction retention in primiparous rats.Muscimol-treated primiparous rats exhibited low CS-elicited freezing during extinction retention, comparable to that of saline-treated rats, and significantly lower than non-extinguished rats treated with either saline or muscimol.Post-extinction training BLA-inactivation similarly failed to impair extinction retention in primiparous rats, yet led to impaired extinction retention in nulliparous rats, as has previously been reported in male rats in some studies (Laurent et al., 2008, but not in Sierra-Mercado et al., 2011).The dissociable effects of BLA inactivation on extinction retention in nulliparous versus primiparous rats were replicated across different conditions and extinction parameters.As muscimol impaired both fear expression and fear acquisition in primiparous rats, the absence of an impairing effect of muscimol on fear extinction cannot be explained by primiparous rats potentially requiring a different dose or timing of muscimol infusion to nulliparous rats.Combined, these findings indicate that while the BLA is involved in the acquisition and expression of conditioned fear in female rats, it has a dissociable role in the acquisition and consolidation of fear extinction dependent on reproductive status.
There are several reasons why pre-extinction training BLA inactivation led to high CS-elicited freezing during test for extinction retention in nulliparous rats.The first is that BLA inactivation disrupted the retrieval of the original CS-US association, consistent with the finding that CS-elicited freezing was reduced in muscimol-treated rats at the start of extinction training in Experiment 2. A failure to retrieve the CS-US association would prevent the acquisition of extinction and lead to high CS-elicited freezing when tested during extinction retention, drugfree.However, we think this explanation is unlikely for several reasons.First, we also observed extinction retention deficits in muscimol-treated nulliparous rats in Experiment S1, in the absence of any effect of muscimol on CS-elicited freezing during extinction training.The reason for the discrepancy between the two studies with regards to the acute effects of muscimol on freezing is unclear, but could be due to the different conditioning parameters (weaker parameters were used in Experiment S1, resulting in lower overall freezing).Irrespective, it is clear that preextinction training infusions of muscimol do not need to suppress conditioned fear expression or retrieval in order to lead to deficits in extinction retention.Second, despite exhibiting lower CS-elicited freezing at the start of extinction training, muscimol-treated nulliparous rats did exhibit reductions in CS-elicited freezing across extinction training in Experiment 2, suggestive of some level of extinction acquisition.Finally, post-extinction training infusions of muscimol also impaired extinction retention in nulliparous rats (Experiment 4), which demonstrates a role for the BLA in extinction consolidation.Of course, while this does not rule out the possibility that pre-extinction training BLA inactivation had additional effects that led to an apparent deficit in extinction retention, it does strongly suggest that at least part of the deficit is driven by an impact on consolidation processes.
There are several reasons why fear extinction may be BLAindependent in primiparous rats.The first is that the decrement in fear exhibited by primiparous rats over the course of extinction training and during extinction retention is due to primiparous rats forgetting the CS-US association, rather than an active process like extinction learning.However, this is not the case because non-extinguished primiparous rats exhibited high levels of CS-elicited freezing when presented with the CS two days after conditioning, during extinction retention.Thus, primiparous rats require non-reinforced presentations of the CS (i.e., extinction training) in order to exhibit decrements in conditioned fear responses.We also tested the possibility that primiparous rats could use the BLA under typical circumstances, but compensate when the BLA is not available during extinction acquisition by recruiting a different brain structure.If this is occurring, we reasoned that inhibiting BLA activity immediately after extinction training, once activity-dependent plasticity in the BLA has already been initiated, should disrupt the consolidation of these changes and in turn disrupt extinction retention in primiparous rats.Yet no impairment in extinction retention following post-extinction training BLA inactivation was observed in primiparous rats, despite post-extinction training BLA inactivation causing impairments in extinction retention in nulliparous rats.This suggests that whereas nulliparous rats also require the BLA for the consolidation of extinction learning, primiparous rats do notand this finding speaks against the speculation that primiparous rats were recruiting a compensatory structure in Experiment 2.
Another possibility for the present results is that fear extinction in primiparous rats depends on a different, BLA-independent neural circuitry.This possibility may seem surprising given the robust evidence for BLA involvement in male fear extinction, and our present findings in nulliparous rats.Nonetheless, this explanation is consistent with our recent evidence of other dissociations in behavioral, hormonal, and neurobiological fear extinction mechanisms in female rats as a function of reproductive experience (Milligan-Saville & Graham, 2016;Tang & Graham, 2019b, 2019a).In combination with the present study, these findings suggest that the mechanisms of fear extinction undergo a fundamental shift from pre-to post-pregnancy.For example, extinction in primiparous rats may involve unlearning or erasure, which would account for the lack of relapse phenomena, NMDAr involvement, and ovarian hormone modulation (which is in part mediated by NMDAr; Graham & Scott, 2018b).However, if it is the case that unlearning is the primary mechanism of extinction in primiparous rats, it nonetheless seems plausible that they would still require BLA activity in order to engage this process, assuming that the CS-US memory is stored in the BLA.Indeed, a puzzling aspect of our findings is that the neurobiology of fear conditioning itself seems relatively unchanged in primiparous rats.That is, in the present study, the BLA was needed for both the acquisition and expression of conditioned fear.Likewise, we have previously shown that fear conditioning is dependent on NMDAr activity in both nulliparous and primiparous rats (Tang & Graham, 2019b).One possibility is that the location of the CS-US memory, although initially formed in the BLA, rapidly shifts from the BLA to a different region/s that also coordinate fear responses (e.g., the bed nucleus of the stria terminalis; BST, or the central nucleus of the amygdala; CeA), and it is this/these regions that are engaged during fear extinction (whether this involves new learning, erasure, or otherwise).The BST and CeA are good candidate regions for the regulation of fear extinction in primiparous ratsboth undergo extensive modification during pregnancy (Duarte-Guterman et al., 2019), and both receive input from cortical regions that regulate fear extinction, e.g., the prefrontal cortex (Shackman & Fox, 2016).If it is the case that fear extinction in primiparous rats is regulated by alternative structures such as these, the reduction in fear expression at the beginning of extinction training observed following BLA-infusion of muscimol could merely reflect the BLA's feedforward inhibition of the CeA and BST, or non-specific locomotor effects of the drug.Clearly, more work is required to map out the specific structures and pathways that mediate the storage and extinction of conditioned fear memories in female rats of differing reproductive statuses.This includes determining whether there are dissociable roles for more specific sub-regions of the BLA, like the lateral v basal amygdala.Moreover, further insights could be gained by extending our measures of fear to other behavioral outputs using finer grained analyses (e.g., Wiltschko et al., 2015).
These findings are not the first to show that the fear extinction circuitry can deviate from what is typically reported in adult male rodents under certain circumstances.For example, fear extinction mechanisms shift over development, e.g., preweaning male rat pups also exhibit fear extinction that is both relapse-resistant and NMDAr-independent, yet dependent on BLA activity (Kim & Richardson, 2007, 2008, 2010).This is also not the first example of extinction that can occur in the absence of BLA-involvement.In adult male rats, re-extinction (in which the extinguished CS is re-paired with the US, before being extinguished for a second time) is unaffected by BLA inactivation (Laurent et al., 2008), which the authors suggested was due to re-extinction involving retrieval of an already established extinction memory, which involves different mechanisms than those underlying initial extinction.Although this explanation does not easily transfer to the current study, in which primiparous rats were undergoing extinction training for the first time, it provides an example that underscores the lability of extinction mechanisms (including the involvement of the BLA) under different circumstances.
The mechanisms that cause the shift in fear extinction circuitry in primiparous rats are likely related to pregnancy, rather than maternal experience.In our recent study (Pestana et al., 2021), estrous effects on fear extinction were mitigated even in primiparous rats that had their pups permanently removed on the first day of life.In addition, we have replicated the mitigation of ovarian hormone modulation of fear extinction in parous human women (Milligan-Saville & Graham, 2016), in whom the hormonal changes associated with pregnancy are more similar to rats; relative to the divergent post-partum experiences of rats and humans.Thus, it seems likely that evolutionary conserved features of pregnancy may account for post-partum shifts in fear extinction mechanisms across species.Importantly, these shifts appear to be longlastingwe have demonstrated that altered fear extinction in primiparous rats persists at least three months post-weaning (Milligan-Saville & Graham, 2016); the changes observed in the present study were evident at least two weeks post-weaning, and our past observation in parous women were observed amongst those up to four years postpartum (Milligan-Saville & Graham, 2016).These long-lasting changes fit with recent observations of persistent changes in gray matter in human mothers (Carmona et al., 2019;de Lange et al., 2019;Hoekzema et al., 2017).Future work is required to delineate the precise hormonal changes that promote the long-term alterations in fear extinction in primiparous rats.Additionally, the present findings indicate that it would be worthwhile to collect data regarding pregnancy history, and assess pregnancy history as a moderating factor, in human fMRI studies on fear extinction, which, to our knowledge, has never been assessed.This is particularly important given recent meta-analyses in humans demonstrate inconsistent evidence for a role of the BLA in human fear extinction (Fullana et al., 2018).
In summary, the present findings add to a growing body of work illustrating that the mechanisms of fear extinction in males do not always transfer neatly to females, with the gap becoming wider when translating to primiparous female rats.Given that 85 % of women are mothers by the age of 44 (Martinez et al., 2011), it is possible that our current conceptualization of how fear extinction operates is far from representative for a substantial proportion of the human population.Moreover, one in five women who give birth experience perinatal anxiety (Fairbrother et al., 2016), in addition to anxiety disorders being twice as prevalent in women relative to men across adolescence and adulthood (Li & Graham, 2017;McLean et al., 2011).Thus, a failure to include females in our preclinical work on anxiety, and the resultant lack of systematic investigations of female-unique physiological factors like pregnancy, could have detrimental consequences for the development of appropriate treatments for anxiety.Although we have previously reported stark differences in other mechanisms of fear extinction between nulliparous and primiparous female rats, the present findings of the apparent lack of involvement of the BLA in fear extinction in primiparous rats are particularly notable because the BLA is widely conceptualized as an integral component of the fear extinction circuitry.These findings therefore underscore the need to conduct additional work to not only validate the neurobiological model of fear and its extinction in females, but also, to systematically investigate the impact of femaleunique physiological factors.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.