Patients with stress-related psychiatric disorders experience debilitating emotional symptoms, including excessive fear that they are unable to control. Decades of research have shown that such disorders have opposite effects on two key structures in the brain. Normally, a region called the amygdala helps to form fear-related memories, while the prefrontal cortex helps control these memories and eliminate them through a process called extinction. But brain imaging on patients with stress-related psychiatric disorders reveals a smaller prefrontal cortex, and an overactive amygdala. Animal studies confirm that chronic stress shrinks brain cells in the prefrontal cortex, but grows them bigger in the amygdala.

These and other studies have led scientists to believe that stress causes people to both form stronger fear memories and then have difficulties getting rid of such memories. These include studies in which stressed animals were trained to fear a sound, and then followed while they learned to overcome that fear. One problem with many of these studies is that the animals were repeatedly stressed before the fear memory was formed. This made it hard to tease apart whether the strength of the fear memory itself or a problem extinguishing the fear memory were to blame for the animals’ difficulties overcoming their fear.

Now, Rahman et al. address this problem by exploring how the timing of chronic stress affects how well an animal can overcome fear memories. In the experiments, some rats learned to fear a sound before exposure to chronic stress, while a second group was exposed to chronic stress first then learned to fear the sound. When the animals were stressed after they learned to fear the sound, they could still eliminate their fear. But the animals stressed before exposure to the fear-inducing sound struggled to extinguish the fear. Recordings of the brain activity in the rats that were exposed to stress after learning to fear the sound showed the amygdala remained overly active even after these animals overcame their fear. However, the stress did not seem to disrupt the normal activity of the prefrontal cortex in these rats. This shows that memories formed before stress reflect normal activity in the prefrontal cortex and not the abnormally high activity in the amygdala.

Exposure therapy helps people overcome stress-disorder related fears. For the therapy to work, the prefrontal cortex must be able to extinguish fear. Rahman et al. show that this is possible when the fear-enhancing effects of prior stress are not in play. More studies exploring why prior stress makes fears stronger and harder to overcome may help scientists develop ways to make therapies for stress disorders more effective.

Accumulating evidence from animal models shows that stress elicits divergent patterns of plasticity across brain regions (Chattarji et al., 2015). For instance, repeated stress causes loss of dendrites and spines in the medial prefrontal cortex (mPFC) (Shansky and Morrison, 2009). In the basolateral amygdala (BLA), by contrast, chronic stress strengthens the structural basis of synaptic connectivity through dendritic growth and spinogenesis (Chattarji et al., 2015). Physiological and molecular measures of synaptic plasticity also exhibit these contrasting features. As useful as these studies have been in examining the effects of stress across biological scales, much of this evidence was gathered from postmortem analyses (Chattarji et al., 2015). Less is known about how stress affects neural activity in the intact brain of behaving animals. Further, in many of these studies, stress-induced plasticity was viewed as stand-alone effects intrinsic to individual brain areas, despite extensive interconnections between them. Indeed, interactions between these brain areas together give rise to behaviors that are affected by stress (Quirk and Mueller, 2008).

One such behavior involves the expression of fear memories, various facets of which depend on both the BLA and mPFC (Sierra-Mercado et al., 2011). Repeated stress has been shown to enhance fear memories, as well as impair their extinction (Miracle et al., 2006; Suvrathan et al., 2014). These studies, however, first exposed animals to stress, and then subjected them to fear conditioning followed by extinction (Izquierdo et al., 2006; Miracle et al., 2006). In other words, these studies analyzed the impact of stress on both the formation and extinction of fear memories, not the latter alone. This experimental design has led to two broad classes of findings. In a majority of these studies, stress did not necessarily lead to higher freezing levels during fear recall. Even though expression of fear was not elevated by prior exposure to stress, it was resistant to subsequent extinction in stressed animals (Noble et al., 2017; Zhang and Rosenkranz, 2013). In a few studies, however, the behavioral data suggest that prior exposure to stress may have also led to stronger fear memories manifested as higher levels of freezing in stressed animals at the beginning of extinction training (Chauveau et al., 2012; Hoffman et al., 2014; Miracle et al., 2006). This too could have contributed to the subsequent deficit in fear extinction recall. An alternative strategy to test if stress impairs the extinction of fear memories would be for animals to form fear memories before stress exposure, thereby dissociating the effects of stress on acquisition versus extinction of conditioned fear. This would offer an opportunity to examine how stress specifically affects expression of fear during and after extinction, without the confounds of stronger fear memories caused by prior exposure to stress. Hence, the present study combines simultaneous behavioral and in vivo electrophysiological analyses to address this question.

Rats were first subjected to auditory fear conditioning (Day 0, Figure 1) at the end of which they exhibited significantly higher freezing behavior in response to the tone conditioned stimulus (CS) (Figure 1b). These animals were then divided into two groups – one was subjected to 10 days of chronic immobilization stress (Days 1–10, Figure 1a) while the other served as unstressed control. 24 hr after the end of chronic stress there was no difference in CS-induced freezing behavior between the two groups (Day 11, Block1, Figure 1b). Thus, the recall of fear memory formed earlier was not affected by subsequent stress. This ensured that both stressed and unstressed animals were at the same levels of freezing when the extinction protocol was initiated after stress. Next, repeated tone presentations reduced freezing levels significantly such that both groups eventually underwent comparable extinction of fear, though the stressed rats were slower in achieving the same reduction in freezing (Day 11, Figure 1b). Notably, a day later stressed animals showed no difference in recall of extinction memory compared to unstressed animals (Day 12, Figure 1b). Thus, freezing levels during both fear and extinction recall were unaffected in stressed animals. This result differs from past reports of stress-induced deficits in fear extinction. However, as mentioned earlier, those earlier studies subjected animals to fear conditioning and extinction after exposure to stress (Izquierdo et al., 2006; Maren and Holmes, 2016; Miracle et al., 2006). Taken together, this suggests that the timing of stress may be a critical determinant of whether extinction recall is impaired. Thus, to examine this possibility we repeated the experimental design adopted in earlier studies by administering the same 10 day chronic immobilization stress protocol before (Figure 1c,e) the same fear conditioning paradigm depicted in Figure 1a. Consistent with earlier findings, prior exposure to chronic stress caused a significant impairment in the recall of fear extinction (Figure 1e). However, these animals, unlike those used in Figure 1b (i.e. conditioning before stress) were not implanted with electrodes for simultaneous in vivo recordings. Hence, we repeated the behavioral experiments described in Figure 1b without surgical interventions related to in vivo recordings. This too yielded the same results as seen in the implanted animals – extinction recall was intact in stressed animals (Figure 1d,f) when they were subjected to conditioning before chronic stress.

Figure 1

Download asset Open asset

Figure 1—source data 1

Data for animals across groups representing freezing response to CS and pretone during the different phases of behaviour (Figure 1b,e and f).

https://doi.org/10.7554/eLife.35450.004

Download elife-35450-fig1-data1-v2.xlsx

Next, we examined the neural basis of this result by recording local field potentials (LFPs) in these freely behaving rats (Karalis et al., 2016; Likhtik et al., 2014). While a role for potentiation of amygdalar neuronal responses to the tone CS in conditioned fear behavior is well established, in vivo recordings have also shown correlations of tone responses in the dorsal mPFC (dmPFC) with freezing behavior in fear conditioning and extinction. Taken together with earlier pharmacological inactivation studies, these findings identified an important role for the dmPFC in underlying conditioned fear responses and the expression of fear extinction (Burgos-Robles et al., 2009; Sierra-Mercado et al., 2011). Therefore, in addition to the BLA, we also monitored responses in the dmPFC (Figure 2a, Figure 2—figure supplement 1). We first analyzed CS-evoked LFPs in the BLA at three key behavioral time points described in Figure 1 – tone habituation before conditioning, fear recall and extinction recall (Figure 2b–d). We measured auditory evoked potential (AEP) amplitudes as the difference between first peak and first trough in the AEP (Figure 2—figure supplement 1). During fear recall, AEP amplitudes (Rogan et al., 1997) were enhanced in both stressed and unstressed animals (Figure 2c,e). However, while this increase was reversed in unstressed animals, it persisted in stressed rats even during extinction recall (Figure 2c,e). Previous work also identified increase in CS-evoked theta power as a neural correlate of conditioned fear (Likhtik et al., 2014). We found BLA theta power (measured as the power of auditory evoked responses in the 2–12 Hz frequency band) in unstressed animals also paralleled the increase, followed by decrease, in freezing during fear and extinction recall respectively (Figure 2d–e). By contrast, BLA theta power remained high in stressed animals (Figure 2d–e). Thus, despite stress-induced theta hyperactivity in the BLA, fear expression was not enhanced during extinction recall. To probe this further, we also analyzed the same LFP parameters in the dmPFC, which according to recent studies plays an important role in fear expression (Karalis et al., 2016; Likhtik et al., 2014). In the dmPFC of unstressed rats, AEP amplitude increased during fear recall, and this was reversed during extinction recall (Figure 2f). Stressed animals, however, did not exhibit any change in dmPFC AEP amplitudes during either fear or extinction recall. Further, fear conditioning enhanced dmPFC theta power in both stress and unstressed animals (Figure 2f). Interestingly, this was reversed in both groups during extinction recall. In other words, unlike the BLA, changes in theta power in the dmPFC, during fear and extinction recall, were not affected by stress. Moreover, in stressed animals, bidirectional modulation of theta power in the dmPFC, but not the BLA, accurately mirrored the changes in freezing, a behavioral expression of fear. Interestingly, the gradual decrease in dmPFC theta power paralleled the significant within-session reduction in freezing during the acquisition of extinction (Day 11, Figure 1b) in both control and stressed animals (Figure 2—figure supplement 2b,d). However, unlike the slower time course of extinction in the stressed rats, there was no difference in the time course of reduction in dmPFC theta power between stressed versus control animals. Similar analysis of BLA theta power during extinction learning (Day 12, Figure 1b) did not reveal any significant differences that paralleled the gradual within-session decrease in freezing levels (Figure 2—figure supplement 2b,c).

Figure 2 with 2 supplements see all

Download asset Open asset

Figure 2—source data 1

Data for animals representing freezing response to CS (Figure 2b) and AEP amplitude and theta power in dmPFC and BLA (Figure 2e,f)

https://doi.org/10.7554/eLife.35450.008

Download elife-35450-fig2-data1-v2.xlsx

Figure 2—source data 2

Data for animals representing freezing response to CS (Figure 2—figure supplement 2b) and AEP amplitude and theta power in dmPFC and BLA (Figure 2—figure supplement 2c,d)

https://doi.org/10.7554/eLife.35450.009

Download elife-35450-fig2-data2-v2.xlsx

Finally, there is growing appreciation of the importance of interactions between the mPFC and BLA, not just activity within these areas, in regulating fear behavior (Karalis et al., 2016; Lesting et al., 2011; Likhtik et al., 2014; Popa et al., 2010). This issue comes into sharp focus here because of the distinct effects of stress on the mPFC versus BLA. Thus, in light of recent reports that theta frequency oscillations synchronize dmPFC–BLA circuits during expression of fear behavior (Karalis et al., 2016; Likhtik et al., 2014), we investigated whether the tone-evoked increases in theta power (Figure 2) were accompanied by enhanced theta-frequency synchrony between the two areas, and if this was in anyway affected by stress. Hence, we quantified CS-evoked coherence to assess moment-by-moment synchrony across LFPs recorded from the dmPFC and BLA for all three time points (Figure 3a) (Likhtik et al., 2014). In unstressed rats, consistent with earlier reports, the CS elicited significantly higher theta coherence during fear recall (Likhtik et al., 2014), and this increase persisted during extinction recall as well (Figure 3b). Notably, in stressed animals, there was no change in BLA-dmPFC theta-frequency coherence (Figure 3b). Thus, stress appears to completely suppress the dynamic, behaviorally relevant enhancement in BLA-dmPFC coherence that is seen during both fear and extinction recall in unstressed animals. In light of strong reciprocal connections between the mPFC and BLA, increases in theta synchrony have led earlier studies to analyze the direction of information flow between the two areas (Karalis et al., 2016; Likhtik et al., 2014). Hence, we estimated the directionality of functional connectivity and leads between the dmPFC and BLA using a previously validated method of calculating cross-correlations of instantaneous amplitude of filtered LFPs (Adhikari et al., 2010). This reveals that theta activity in the dmPFC leads that in the BLA during recall of both fear and extinction memories in unstressed rats (Figure 3c). However, this dmPFC-to-BLA directional influence is absent in stressed animals. Together, these data suggest that chronic stress causes a decoupling of activity between the two brain areas, as evidenced by a complete disruption of the increase in dmPFC-BLA theta synchrony and dmPFC-to-BLA directional influence normally seen during the recall of fear and extinction memories. This decoupling is also evident across all trial blocks during the acquisition of extinction the previous day (Figure 3—figure supplement 1). This breakdown in mPFC-BLA theta-frequency synchrony and directional coupling could be one reason why enhanced BLA theta activity was not manifested as higher freezing in stressed animals.

Figure 3 with 1 supplement see all

Download asset Open asset

Figure 3—source data 1

Data for animals representing theta coherence between dmPFC and BLA (Figure 3b) and dmPFC-BLA phase difference (Figure 3c)

https://doi.org/10.7554/eLife.35450.012

Download elife-35450-fig3-data1-v2.xlsx

Figure 3—source data 2

Data for animals representing theta coherence between dmPFC and BLA (Figure 3—figure supplement 1b) and dmPFC-BLA phase difference (Figure 3—figure supplement 1c)

https://doi.org/10.7554/eLife.35450.013

Download elife-35450-fig3-data2-v2.xlsx

The present study, specifically designed to administer chronic stress after the formation of fear memory, reveals that when stressed animals started extinguishing fear memories from the same level of freezing as their unstressed counterparts, their ability to recall extinction memory remained intact. This is in contrast to past findings wherein stressed animals exhibited a deficit in extinction recall when faced with the challenge of extinguishing fear memories strengthened by prior exposure to stress. This is consistent with earlier findings that repeated immobilization stress (4 h/day for 14 days) administered after auditory fear conditioning did not affect the expression of previously acquired fear memories (Meyer et al., 2014). However, the same immobilization stress, when administered before fear conditioning, elicited a robust increase in fear recall. Together, these results suggest that the timing of chronic stress exposure, with respect to fear conditioning, is a key determinant of how subsequent extinction of fear is affected by stress. This dissociation of the effects of stress on acquisition versus extinction of fear conditioning suggests that stress acts primarily on acquisition, and the earlier findings from pre-conditioning stress experiments were not the result of a deficit in fear extinction per se. Age may also influence how stress affects fear extinction in rats. For instance, deficient fear extinction was only observed in previously stressed adolescent, but not adult, rats (Zhang and Rosenkranz, 2013).

Interestingly, despite no visible behavioral effect of stress on fear expression, our in vivo recordings reveal a robust impact of stress on amygdalar activity, as evidenced by enhanced theta activity in the BLA that failed to reverse even after the animals exhibited normal extinction recall (Figure 2e). This is consistent with earlier findings on physiological and structural strengthening of excitatory synaptic connectivity, as well as reduced inhibitory tone, in the BLA after stress (Suvrathan et al., 2014). Together, these changes point to possible synaptic mechanisms in the BLA for the stress-induced strengthening of subsequent encoding of fear memories.

The same chronic stress, on the other hand, did not disrupt normal bidirectional modulation of mPFC theta activity, which in turn was reflected in normal freezing behavior during recall of fear and extinction (Figure 2f). This is consistent with growing evidence for a pivotal role played by the mPFC in fear expression (Dejean et al., 2016; Sierra-Mercado et al., 2011). For instance, recent work has demonstrated strong correlations between mPFC theta-frequency oscillations and conditioning-induced freezing behavior (Likhtik et al., 2014). Furthermore, we find normal mPFC theta activity to be decoupled from theta hyperactivity in the BLA, possibly reducing the latter’s influence on fear expression. This is similar to a report that even a single episode of stress can weaken functional connectivity between the two areas measured by resting state fMRI (Liang et al., 2014). Indeed, such stress-induced disruptions in prefrontal-to-amygdala connectivity is also known to affect social interaction and anxiety-related behaviors in rodents (Adhikari et al., 2015; Hultman et al., 2016).

Our findings, taken together with earlier behavioral studies (Izquierdo et al., 2006; Meyer et al., 2014; Miracle et al., 2006; Rau and Fanselow, 2009; Sierra-Mercado et al., 2011; Suvrathan et al., 2014) suggest a model for why the effects of stress on the recall and extinction of fear memories depend on the timing of stress exposure with respect to when the fear memory is formed and extinguished (Figure 4). Specifically, these findings suggest that the regulation of remote fear memories by stress (i.e. pre-stress conditioning, Figure 4b) is different from that of more recent fear memories (i.e. post-stress conditioning, Figure 4c). As reported here, when fear memories were formed before exposure to the 10 day chronic stress, recall of fear extinction was not affected. Although stress caused a persistent increase in theta activity in the BLA, it did not affect bidirectional regulation of dmPFC theta activity (Figure 4b). While stress-induced disruption in mPFC-BLA directional coupling and theta-frequency synchrony (Figure 3) may explain why enhanced BLA theta activity did not result in stress-induced enhancement in freezing, another intriguing possibility arises from a recent report that neural circuits mediating the recall of fear memories shifts over time (Do-Monte et al., 2015). Specifically, this study showed that optogenetic silencing of BLA reduced freezing when it was performed 6 hr, but not 7 days, after auditory fear conditioning. Since our post-conditioning stress protocol was carried out for 10 days, it may have outlasted BLA involvement in expression of fear and could also explain why enhanced BLA excitability was no longer reflected in freezing at that remote time point (Day 11; Figure 4b). As a first step towards testing this possibility, we carried out in vivo infusions of muscimol directly into the BLA just before fear recall and extinction learning on Day 11 (Figure 4—figure supplement 1). Strikingly, this targeted inactivation of the BLA had no impact on fear expression at this time point in either stressed or control rats (Figure 4—figure supplement 1b). Further, freezing levels in the muscimol-infused control and stressed rats did not differ from the freezing levels seen in stressed/control animals (Figure 4—figure supplement 1c) that did not receive any infusions in earlier experiments (Figure 1f). Thus, we found that in vivo inactivation of the BLA no longer affects the recall of fear memories that were formed before the 10 day stress exposure. While more detailed analyses will be required to test all aspects of this model, these preliminary results suggest that fear memories formed 10 days ago no longer depend on BLA activity in either control or stressed animals. Consequently, the expression of fear after extinction reflects normal regulation of theta activity in the mPFC, not theta hyperactivity in the BLA (Figure 4b).

Figure 4 with 1 supplement see all

Download asset Open asset

Figure 4—source data 1

Data for animals across groups representing freezing response to CS during the different phases of behaviour (Figure 4—figure supplement 1c,d)

https://doi.org/10.7554/eLife.35450.016

Download elife-35450-fig4-data1-v2.xlsx

On the other hand, when fear memories were formed after exposure to the same chronic stress, the behavioural effects on freezing are very different, as has also been reported in earlier studies that employed a post-stress conditioning strategy (Maren and Holmes, 2016; Miracle et al., 2006; Rau and Fanselow, 2009). Animals that were fear conditioned after stress show enhanced freezing during both acquisition and recall of extinction memory. What could be the neural basis for this difference? Electrophysiological data from our post-conditioning stress experiments point to possible mechanisms that can be examined in future studies (Figure 4c). For instance, in contrast to the pre-stress conditioning situation, stressed animals will recall and extinguish fear memories that were formed only 1 day ago. As a result, stress-induced BLA hyperactivity will be manifested as stronger fear memories that are resistant to subsequent extinction (as evidenced by higher freezing during extinction recall in CIS rats, Figure 1e). Thus, now it is the sustained enhancement in BLA theta activity, not normal regulation of dmPFC activity, that is likely to drive higher freezing in stressed rats during extinction learning and extinction recall (Figure 4c). Further, the detrimental effects of stress will be compounded by the disruption of mPFC-BLA theta-frequency coherence and directional coupling, which is likely to impair the mPFC’s ability to regulate BLA hyperactivity. Interestingly, pre-conditioning stress has been shown to cause subsequent enhancement in fear learning that can last up to 90 days later (Rau and Fanselow, 2009). This naturally raises questions about the involvement of the BLA in the enhanced expression of conditioned fear at remote time points. Our preliminary findings (Figure 4—figure supplement 1), taken together with the study by Do-Monte et al. (2015), suggest that despite the effects of pre-conditioning stress on the encoding of fear memories, subsequent expression at later time points may no longer depend on BLA activity. This, in turn, raises the possibility that other brain areas that interact with the BLA in the regulation of fear, such as the dmPFC, may play a role in maintaining this expression at remote time points. Thus, a better understanding of the neural basis of earlier behavioral findings (Izquierdo et al., 2006; Meyer et al., 2014; Miracle et al., 2006; Noble et al., 2017; Rau and Fanselow, 2009) will require detailed in vivo electrophysiological analyses, coupled with targeted inactivation, to examine the role of other brain areas that are engaged in stress-induced changes of fear expression.

Finally, it is interesting to note that the chronic immobilization stress paradigm used in the present study was previously shown to strengthen functional connectivity from the amygdala to the hippocampus (Ghosh et al., 2013), which undergoes stress-induced deficits similar to the mPFC (Arnsten, 2015; Chattarji et al., 2015; McEwen and Morrison, 2013). In other words, although both the hippocampus and mPFC undergo similar forms of stress-induced deficits, the impact of stress on their individual interactions with the amygdala are strikingly different. A better understanding of these divergent features of aberrant interactions distributed across the amygdala-mPFC-hippocampal network, not just those confined within each area, may offer new insights into therapeutic interventions against the cognitive and emotional symptoms of stress-related psychiatric disorder.

Data generated or analysed during this study are included in the manuscript and supporting files. Source data files have been provided for all the Figures.

1. 1. Adhikari A
   2. Lerner TN
   3. Finkelstein J
   4. Pak S
   5. Jennings JH
   6. Davidson TJ
   7. Ferenczi E
   8. Gunaydin LA
   9. Mirzabekov JJ
   10. Ye L
   11. Kim SY
   12. Lei A
   13. Deisseroth K

   (2015) Basomedial amygdala mediates top-down control of anxiety and fear

   Nature 527:179–185.

   https://doi.org/10.1038/nature15698

   • PubMed
   • Google Scholar
2. 1. Adhikari A
   2. Sigurdsson T
   3. Topiwala MA
   4. Gordon JA

   (2010) Cross-correlation of instantaneous amplitudes of field potential oscillations: a straightforward method to estimate the directionality and lag between brain Areas

   Journal of Neuroscience Methods 191:191–200.

   https://doi.org/10.1016/j.jneumeth.2010.06.019

   • PubMed
   • Google Scholar
3. 1. Arnsten AFT

   (2015) Stress weakens prefrontal cortex network connections: molecular events affecting cognitive disorders

   Nature Neuroscience 18:1376–1385.

   https://doi.org/10.1038/nn.4087

   • Google Scholar
4. 1. Blanchard DC
   2. Blanchard RJ

   (1988) Ethoexperimental approaches to the biology of emotion

   Annual Review of Psychology 39:43–68.

   https://doi.org/10.1146/annurev.ps.39.020188.000355

   • PubMed
   • Google Scholar
5. 1. Burgos-Robles A
   2. Vidal-Gonzalez I
   3. Quirk GJ

   (2009) Sustained conditioned responses in prelimbic prefrontal neurons are correlated with fear expression and extinction failure

   Journal of Neuroscience 29:8474–8482.

   https://doi.org/10.1523/JNEUROSCI.0378-09.2009

   • PubMed
   • Google Scholar
6. 1. Chattarji S
   2. Tomar A
   3. Suvrathan A
   4. Ghosh S
   5. Rahman MM

   (2015) Neighborhood matters: divergent patterns of stress-induced plasticity across the brain

   Nature Neuroscience 18:1364–1375.

   https://doi.org/10.1038/nn.4115

   • PubMed
   • Google Scholar
7. 1. Chauveau F
   2. Lange MD
   3. Jüngling K
   4. Lesting J
   5. Seidenbecher T
   6. Pape HC

   (2012) Prevention of stress-impaired fear extinction through neuropeptide s action in the lateral amygdala

   Neuropsychopharmacology 37:1588–1599.

   https://doi.org/10.1038/npp.2012.3

   • PubMed
   • Google Scholar
8. 1. Dejean C
   2. Courtin J
   3. Karalis N
   4. Chaudun F
   5. Wurtz H
   6. Bienvenu TC
   7. Herry C

   (2016) Prefrontal neuronal assemblies temporally control fear behaviour

   Nature 535:420–424.

   https://doi.org/10.1038/nature18630

   • PubMed
   • Google Scholar
9. 1. Do-Monte FH
   2. Quiñones-Laracuente K
   3. Quirk GJ

   (2015) A temporal shift in the circuits mediating retrieval of fear memory

   Nature 519:460–463.

   https://doi.org/10.1038/nature14030

   • PubMed
   • Google Scholar
10. 1. Ghosh S
    2. Laxmi TR
    3. Chattarji S

    (2013) Functional connectivity from the amygdala to the Hippocampus grows stronger after stress

    Journal of Neuroscience 33:7234–7244.

    https://doi.org/10.1523/JNEUROSCI.0638-13.2013

    • PubMed
    • Google Scholar
11. 1. Hoffman AN
    2. Lorson NG
    3. Sanabria F
    4. Foster Olive M
    5. Conrad CD

    (2014) Chronic stress disrupts fear extinction and enhances amygdala and hippocampal Fos expression in an animal model of post-traumatic stress disorder

    Neurobiology of Learning and Memory 112:139–147.

    https://doi.org/10.1016/j.nlm.2014.01.018

    • PubMed
    • Google Scholar
12. 1. Hultman R
    2. Mague SD
    3. Li Q
    4. Katz BM
    5. Michel N
    6. Lin L
    7. Wang J
    8. David LK
    9. Blount C
    10. Chandy R
    11. Carlson D
    12. Ulrich K
    13. Carin L
    14. Dunson D
    15. Kumar S
    16. Deisseroth K
    17. Moore SD
    18. Dzirasa K

    (2016) Dysregulation of prefrontal cortex-mediated slow-evolving limbic dynamics drives stress-induced emotional pathology

    Neuron 91:439–452.

    https://doi.org/10.1016/j.neuron.2016.05.038

    • PubMed
    • Google Scholar
13. 1. Izquierdo A
    2. Wellman CL
    3. Holmes A

    (2006) Brief uncontrollable stress causes dendritic retraction in infralimbic cortex and resistance to fear extinction in mice

    Journal of Neuroscience 26:5733–5738.

    https://doi.org/10.1523/JNEUROSCI.0474-06.2006

    • PubMed
    • Google Scholar
14. 1. Karalis N
    2. Dejean C
    3. Chaudun F
    4. Khoder S
    5. Rozeske RR
    6. Wurtz H
    7. Bagur S
    8. Benchenane K
    9. Sirota A
    10. Courtin J
    11. Herry C

    (2016) 4-Hz oscillations synchronize prefrontal-amygdala circuits during fear behavior

    Nature Neuroscience 19:605–612.

    https://doi.org/10.1038/nn.4251

    • PubMed
    • Google Scholar
15. 1. Lesting J
    2. Narayanan RT
    3. Kluge C
    4. Sangha S
    5. Seidenbecher T
    6. Pape HC

    (2011) Patterns of coupled theta activity in amygdala-hippocampal-prefrontal cortical circuits during fear extinction

    PLoS One 6:e21714.

    https://doi.org/10.1371/journal.pone.0021714

    • PubMed
    • Google Scholar
16. 1. Liang Z
    2. King J
    3. Zhang N

    (2014) Neuroplasticity to a single-episode traumatic stress revealed by resting-state fMRI in awake rats

    NeuroImage 103:485–491.

    https://doi.org/10.1016/j.neuroimage.2014.08.050

    • PubMed
    • Google Scholar
17. 1. Likhtik E
    2. Stujenske JM
    3. Topiwala MA
    4. Harris AZ
    5. Gordon JA

    (2014) Prefrontal entrainment of amygdala activity signals safety in learned fear and innate anxiety

    Nature Neuroscience 17:106–113.

    https://doi.org/10.1038/nn.3582

    • PubMed
    • Google Scholar
18. 1. Maren S
    2. Holmes A

    (2016) Stress and fear extinction

    Neuropsychopharmacology 41:58–79.

    https://doi.org/10.1038/npp.2015.180

    • PubMed
    • Google Scholar
19. 1. McEwen BS
    2. Morrison JH

    (2013) The brain on stress: vulnerability and plasticity of the prefrontal cortex over the life course

    Neuron 79:16–29.

    https://doi.org/10.1016/j.neuron.2013.06.028

    • PubMed
    • Google Scholar
20. 1. Meyer RM
    2. Burgos-Robles A
    3. Liu E
    4. Correia SS
    5. Goosens KA

    (2014) A ghrelin-growth hormone axis drives stress-induced vulnerability to enhanced fear

    Molecular Psychiatry 19:1284–1294.

    https://doi.org/10.1038/mp.2013.135

    • PubMed
    • Google Scholar
21. 1. Miracle AD
    2. Brace MF
    3. Huyck KD
    4. Singler SA
    5. Wellman CL

    (2006) Chronic stress impairs recall of extinction of conditioned fear

    Neurobiology of Learning and Memory 85:213–218.

    https://doi.org/10.1016/j.nlm.2005.10.005

    • PubMed
    • Google Scholar
22. 1. Noble LJ
    2. Gonzalez IJ
    3. Meruva VB
    4. Callahan KA
    5. Belfort BD
    6. Ramanathan KR
    7. Meyers E
    8. Kilgard MP
    9. Rennaker RL
    10. McIntyre CK

    (2017) Effects of vagus nerve stimulation on extinction of conditioned fear and post-traumatic stress disorder symptoms in rats

    Translational Psychiatry 7:e1217.

    https://doi.org/10.1038/tp.2017.191

    • PubMed
    • Google Scholar
23. Book

    1. Paxinos G
    2. Watson C

    (2009)

    The Rat Brain in Stereotaxic Coordinates

    Elsevier/Academic Press.

    • Google Scholar
24. 1. Popa D
    2. Duvarci S
    3. Popescu AT
    4. Léna C
    5. Paré D

    (2010) Coherent amygdalocortical theta promotes fear memory consolidation during paradoxical sleep

    PNAS 107:6516–6519.

    https://doi.org/10.1073/pnas.0913016107

    • PubMed
    • Google Scholar
25. 1. Quirk GJ
    2. Mueller D

    (2008) Neural mechanisms of extinction learning and retrieval

    Neuropsychopharmacology 33:56–72.

    https://doi.org/10.1038/sj.npp.1301555

    • PubMed
    • Google Scholar
26. 1. Rau V
    2. Fanselow MS

    (2009) Exposure to a stressor produces a long lasting enhancement of fear learning in rats

    Stress 12:125–133.

    https://doi.org/10.1080/10253890802137320

    • PubMed
    • Google Scholar
27. 1. Rogan MT
    2. Stäubli UV
    3. LeDoux JE

    (1997) Fear conditioning induces associative long-term potentiation in the amygdala

    Nature 390:604–607.

    https://doi.org/10.1038/37601

    • PubMed
    • Google Scholar
28. 1. Shansky RM
    2. Morrison JH

    (2009) Stress-induced dendritic remodeling in the medial prefrontal cortex: effects of circuit, hormones and rest

    Brain Research 1293:108–113.

    https://doi.org/10.1016/j.brainres.2009.03.062

    • PubMed
    • Google Scholar
29. 1. Sierra-Mercado D
    2. Padilla-Coreano N
    3. Quirk GJ

    (2011) Dissociable roles of prelimbic and infralimbic cortices, ventral Hippocampus, and basolateral amygdala in the expression and extinction of conditioned fear

    Neuropsychopharmacology 36:529–538.

    https://doi.org/10.1038/npp.2010.184

    • PubMed
    • Google Scholar
30. 1. Suvrathan A
    2. Bennur S
    3. Ghosh S
    4. Tomar A
    5. Anilkumar S
    6. Chattarji S

    (2014) Stress enhances fear by forming new synapses with greater capacity for long-term potentiation in the amygdala

    Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences 369:151–159.

    https://doi.org/10.1098/rstb.2013.0151

    • Google Scholar
31. 1. Zhang W
    2. Rosenkranz JA

    (2013) Repeated restraint stress enhances cue-elicited conditioned freezing and impairs acquisition of extinction in an age-dependent manner

    Behavioural Brain Research 248:12–24.

    https://doi.org/10.1016/j.bbr.2013.03.028

    • PubMed
    • Google Scholar

Thank you for submitting your article "Extinction recall of fear memories formed before stress is not affected despite amygdalar hyperactivity" for consideration by eLife. Your article has been reviewed by three peer reviewers, including Jennifer L Raymond as the Reviewing Editor and Reviewer #1, and the evaluation has been overseen by Michael Frank as the Senior Editor. The following individual involved in review of your submission has agreed to reveal their identity: Gregory Quirk (Reviewer #2).

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

Summary:

This manuscript challenges the idea that stress impairs the extinction of fear memories. The authors replicate previous observations of impaired extinction when chronic stress occurs before fear conditioning, which had provided support for the idea that stress affects extinction. However, they report normal extinction and extinction recall in rats that underwent chronic restraint stress before extinction, but after fear conditioning. This dissociation of the effects of stress on acquisition vs. extinction of fear conditioning suggests that stress acts primarily on acquisition. Surprisingly, stress affected the CS (tone)-evoked field potential amplitude and theta power in the basolateral amygdala during extinction recall, although the behavior was normal. This raises the interesting question of what prevents the elevated BLA responses in the stressed animals from affecting behavior. Some clues are provided by recordings from the dorsal mPFC. In mPFC the theta power paralleled the behavior during fear recall and extinction recall in the stressed and unstressed animals, although the evoked potentials during fear recall were altered by stress. In addition, stress blocked the enhanced theta mPFC and the BLA that is observed in the unstressed rates during fear recall and extinction. The paper is very accessible and the figures are clear and convincing. Analysis of the effects of chronic stress on fear extinction, independent of stress's effects on fear conditioning, is long overdue, given the developing story that chronic stress supposedly impairs extinction. Impact on the field should be significant.

Essential revisions:

1) Substantial revision of the text is required to more carefully interpret the results, and more fully consider how previous work could affect the interpretation.

a) The behavioral hypothesis stated in the Introduction and Discussion is that earlier studies induced stress prior to conditioning, which caused rats to start extinction at higher levels of freezing, which would then take longer to extinguish and appear as an extinction deficit. However, this is not exactly what the authors show in their replication of the pre-conditioning stress experiment (Figure 1—figure supplement 1). The stressed rats seem to start extinction only slightly higher than controls, but then show absolutely no within-session extinction, and impaired recall the following day. This suggests that the effects of stress on BLA excitability and theta that they observed causes conditioning to occur in a way that resists subsequent extinction (even though expression of fear is not elevated that much). The authors may wish to speculate on this, which may be more interesting than simply starting extinction from a higher freezing level.

The following statement would especially benefit from more careful wording: "stressed animals had to extinguish fear memories that were invariably stronger than unstressed animals" The metric for "invariably stronger" is not clear, since such as Noble et al. (2017) show that this statement is not true if one's metric of "stronger" is% freezing during fear recall. If, on the other hand, one's metric of "stronger" is resistance to extinction, the authors may be correct, but probably don't have enough data yet to claim "inevitably."

b) Discuss the potential reasons for the different outcomes of pre- and post-conditioning stress on extinction. One possibility is that the encoding of fear is different in stressed animals so that extinction operates on this in a different manner. Another possibility is that the circuits supporting fear recall are different at different times after acquisition, and therefore undergo extinction differently – this could come into play in the current study, since the time between acquisition and extinction was different for the pre- and post-conditioning stress conditions (see point #2 below).

c) Previous studies showing that chronic post-conditioning stress doesn't have any impact on the recall of fear memories acquired prior to stress should be discussed; see Meyer et al. (2014).

d) Results, second paragraph. It is not clear why the authors are equating high levels of pip-evoked theta power with "stress-induced hyperactivity". High levels of pip-evoked theta do not necessarily reflect hyperactivity or hyperexcitability in neurons. In this regard, the title, which refers to hyperactivity rather than theta, is misleading.

e) The deficit in fear extinction in previously stressed rats in Zhang and Rosenkranz was only observed in adolescent rats. No effect of stress on extinction was observed in adult rats. It is this latter finding that is most relevant to the findings in the manuscript. This point should be highlighted in the current manuscript.

2) It is convenient that post-conditioning stress did not increase the expression of freezing, so that extinction could be objectively and cleanly assessed. However, this raises a question: if BLA is so important for expression of conditioned fear, why was expression of freezing not increased, especially with the observed hyperexcitability in BLA? One (unlikely?) possibility is that the circuits within BLA affected by stress were separate from the circuits that were previously conditioned. Another possibility is that BLA is not needed for expression of conditioned fear at that timepoint. This (somewhat heretical) idea comes from a 2015 study showing that optogenetic silencing of BLA reduced freezing when it was performed 6 hrs, but not 7 days, after auditory fear conditioning (Do-Monte et al., 2015, Extended Data Figure 8). Because post-conditioning stress was carried out for 10 days, this may have outlasted BLA involvement in expression of freezing, and could explain why excitability changes were not reflected in freezing at that timepoint. To test this hypothesis, the authors could inactivate BLA at that timepoint to determine the extent to which BLA is (or is not) necessary for expression of older fear memories. Alternatively, they could suggest this possibility in the Discussion.

3) Figure 1 shows a slower the time course of extinction in the stressed rats. Are there parallels in the physiology? More generally, it would be helpful to see a time course in Figures 2 and 3.

4) In the Figure 2 legends for panels E and F, and in Figure 3, it is not clear how the degrees of freedom were calculated. These should be integer values, not values with decimal places. Please clarify this.

[Editors' note: further revisions were requested prior to acceptance, as described below.]

Thank you for submitting your article "Extinction recall of fear memories formed before stress is not affected despite higher theta activity in the amygdala" for consideration by eLife. Your article has been reviewed by two peer reviewers, and the evaluation has been overseen by Jennifer Raymond as the Reviewing Editor and Michael Frank as the Senior Editor. The following individual involved in review of your submission has agreed to reveal their identity: Gregory Quirk (Reviewer #2).

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

Summary:

This manuscript provides an analysis of the effects of chronic stress on fear extinction, independent of the effects of stress on fear conditioning. Whereas the prevailing view in the field has been that stress impairs the extinction of fear memories, the authors report normal extinction and extinction recall in rats that underwent chronic restraint stress before extinction, but after fear conditioning. This finding should have significant scientific impact and may also have substantial clinical impact. The reviewers are enthusiastic about the manuscript, and agreed that the new experiments included in the revised submission have the potential to add considerably to the story. However, the reviewers also felt those experiments were difficult to interpret without a positive control for the effective pharmacological inhibition of BLA.

Essential revision:

The reviewers agree that the manuscript is enhanced by the new experiments showing that pharmacological inactivation of the BLA 12 days after conditioning had no effect on fear expression or extinction. This new finding explains a conundrum from the previous version, namely, why was fear expression normal in recently stressed rats even though BLA excitability was increased? The answer seems to be that BLA is not involved in fear expression at that timepoint. This increases our understanding and interpretation of chronic stress experiments: not only is the timing relative to conditioning important, but so is the duration of the stress period, as BLA's involvement is time-limited.

The reviewers were impressed that the authors provided new experiments in response to the request in the initial review to simply "Discuss the potential reasons for the different outcomes of pre- and post-conditioning stress on extinction." However, given the negative result (no effect of BLA inactivation on fear expression at the late time point), a positive control is needed to show that the pharmacological inactivation of BLA was successful, by showing that BLA inactivation in the same place/dose but at an earlier time point impairs memory expression whereas at a later time it does not (with histological placements of cannula tips shown to confirm location in BLA and the volume and concentration of muscimol provided). The reviewers did not feel comfortable with the inclusion of the new experiments in the manuscript without this essential control, and hope that the authors will choose to do these additional experiments.

Alternatively, since the authors had gone above-and-beyond the reviewers' request to simply "discuss" potential reasons, an acceptable alternative would be to remove the new experiments, and simply discuss potential reasons based on the published results in the literature.

Essential revisions:

1) Substantial revision of the text is required to more carefully interpret the results, and more fully consider how previous work could affect the interpretation.

a) The behavioral hypothesis stated in the Introduction and Discussion is that earlier studies induced stress prior to conditioning, which caused rats to start extinction at higher levels of freezing, which would then take longer to extinguish and appear as an extinction deficit. However, this is not exactly what the authors show in their replication of the pre-conditioning stress experiment (Figure 1—figure supplement 1). The stressed rats seem to start extinction only slightly higher than controls, but then show absolutely no within-session extinction, and impaired recall the following day. This suggests that the effects of stress on BLA excitability and theta that they observed causes conditioning to occur in a way that resists subsequent extinction (even though expression of fear is not elevated that much). The authors may wish to speculate on this, which may be more interesting than simply starting extinction from a higher freezing level.

The following statement would especially benefit from more careful wording: "stressed animals had to extinguish fear memories that were invariably stronger than unstressed animals" The metric for "invariably stronger" is not clear, since such as Noble et al. (2017) show that this statement is not true if one's metric of "stronger" is% freezing during fear recall. If, on the other hand, one's metric of "stronger" is resistance to extinction, the authors may be correct, but probably don't have enough data yet to claim "inevitably."

We thank the reviewer for raising this important point. The reviewer rightly points out that our original text implied that the metric for “stronger” fear memory is higher % freezing. However, in our pre-conditioning stress experiments, exposure to chronic stress caused conditioning to occur in a way that resists subsequent extinction even though expression of fear is not significantly higher to start with(data originally shown in Figure 1—figure supplement 1, now moved to main Figure 1C-F). Hence, we agree that resistance to extinction offers a metric that better reflects our data. Accordingly, we have revised the Introduction to incorporate this suggestion by the reviewer (last paragraph). We have also removed the term “invariably” from the Introduction. Both the Introduction (last paragraph) and Results and Discussion (fifth paragraph) have been modified to elaborate on this point. These sections have also been revised to mention the references to earlier studies relevant to these points, as advised by the reviewer.

b) Discuss the potential reasons for the different outcomes of pre- and post-conditioning stress on extinction. One possibility is that the encoding of fear is different in stressed animals so that extinction operates on this in a different manner. Another possibility is that the circuits supporting fear recall are different at different times after acquisition, and therefore undergo extinction differently – this could come into play in the current study, since the time between acquisition and extinction was different for the pre- and post-conditioning stress conditions (see point #2 below).

We thank the reviewer for this suggestion. We have now revised the Results and Discussion to elaborate on the potential reasons underlying the different outcomes of pre- versus post-conditioning stress on fear extinction (Results and Discussion, seventh paragraph). Specifically, the new data we have added from our analysis of the effects of muscimol inactivation of the BLA (new Figure 4) shed light on the different possibilities suggested by the reviewer (see response to point #2 below). In addition to these new results, we have also added a new figure (Figure 5) to summarize all of these points. And we have devoted a new paragraph in the Results and Discussion section to elaborate on the key points depicted in Figure 5.

c) Previous studies showing that chronic post-conditioning stress doesn't have any impact on the recall of fear memories acquired prior to stress should be discussed; see Meyer et al. (2014).

We thank the reviewer for drawing our attention to the study by Meyer et al. (2014). We apologize for this omission and have now modified the Results and Discussion section to include the relevant findings reported in the earlier study (Results and Discussion, fifth paragraph).

d) Results, second paragraph. It is not clear why the authors are equating high levels of pip-evoked theta power with "stress-induced hyperactivity". High levels of pip-evoked theta do not necessarily reflect hyperactivity or hyperexcitability in neurons. In this regard, the title, which refers to hyperactivity rather than theta, is misleading.

We have modified the title to in response to the concern expressed by the reviewer:

“Extinction recall of fear memories formed before stress is not affected despite amygdalar theta hyperactivity”.

We have revised the Abstract to address this concern.

Finally, we have modified the Results and Discussion section, wherever relevant, to clarify this point.

e) The deficit in fear extinction in previously stressed rats in Zhang and Rosenkranz was only observed in adolescent rats. No effect of stress on extinction was observed in adult rats. It is this latter finding that is most relevant to the findings in the manuscript. This point should be highlighted in the current manuscript.

We thank the reviewer for pointing this out. We have now mentioned this finding in the revised Results and Discussion section (fifth paragraph).

2) It is convenient that post-conditioning stress did not increase the expression of freezing, so that extinction could be objectively and cleanly assessed. However, this raises a question: if BLA is so important for expression of conditioned fear, why was expression of freezing not increased, especially with the observed hyperexcitability in BLA? One (unlikely?) possibility is that the circuits within BLA affected by stress were separate from the circuits that were previously conditioned. Another possibility is that BLA is not needed for expression of conditioned fear at that timepoint. This (somewhat heretical) idea comes from a 2015 study showing that optogenetic silencing of BLA reduced freezing when it was performed 6 hrs, but not 7 days, after auditory fear conditioning (Do-Monte et al., 2015, Extended Data Figure 8). Because post-conditioning stress was carried out for 10 days, this may have outlasted BLA involvement in expression of freezing, and could explain why excitability changes were not reflected in freezing at that timepoint. To test this hypothesis, the authors could inactivate BLA at that timepoint to determine the extent to which BLA is (or is not) necessary for expression of older fear memories. Alternatively, they could suggest this possibility in the Discussion.

This point raised by the reviewer is a very interesting one indeed. Some of the issues mentioned in point #1 also center on this question of why theta hyperactivity in the BLA, despite its pivotal role in the expression of conditioned fear, is not manifested as enhanced freezing in the stressed animals. The reviewer suggests two possible explanations for our findings. The more likely, and particularly intriguing (and “somewhat heretical”, according to the reviewer), possibility is based on a study by Do-Monte et al. (2015) that showed that the BLA is not needed for expression of conditioned fear 7 days after auditory fear conditioning. Since our chronic stress protocol lasts for 10 days, it is possible that BLA hyperactivity no longer affects the expression of the older fear memories. Hence, as suggested by the reviewer, we carried out new experiments to test this possibility (instead of just elaborating on this is in the Discussion) using in vivo infusion of muscimol directly into the BLA just before fear recall and extinction on Day 12 (Figure 4). Strikingly, these new experiments confirm that fear expression at this time point no longer depends on the BLA (in both stressed and control rats). These results are now described in a new figure (Figure 4). We added a new Figure 5 to discuss the implications of these new findings.

3) Figure 1 shows a slower the time course of extinction in the stressed rats. Are there parallels in the physiology? More generally, it would be helpful to see a time course in Figures 2 and 3.

As advised by the reviewer, we have carried out analyses of our electrophysiological recordings to examine if these in vivo measurements parallel the time course of gradual changes in freezing behavior. These additional analyses revealed that the gradual decrease in dmPFC theta power was a reliable electrophysiological correlate of the significant within-session reduction in freezing during the acquisition of extinction (Day 12, Figure 1B) in both control and stressed animals (new Figure 2—figure supplement 2). However, the slower time course of extinction in the stressed rats was not reflected as a difference in the time course of reduction in dmPFC theta power between stressed and control animals. Similar analysis of BLA theta power during extinction learning (Day 12, Figure 1B) did not reveal any significant differences that paralleled the gradual within-session decrease in freezing levels (new Figure 2—figure supplement 2). Though we detected a decreasing trend in BLA CS-evoked theta power by trial block 4 during extinction learning (Day 12), this was not statistically significant. In contrast, no such change was evident in stressed animals. Finally, the stress-induced disruption in both the dmPFC-BLA theta synchrony and dmPFC-to-BLA directional influence is evident across all trial blocks during the acquisition of extinction the previous day (Figure 3—figure supplement 1). Hence, the absence of any changes in these measures across trial blocks meant that they were not good indicators of the gradual and significant reduction in within-session freezing during extinction learning.

These results are now reported in Figure 2—figure supplement 2 and Figure 3—figure supplement 1, as well as in the main text (Results and Discussion, second and third paragraphs).

4) In the Figure 2 legends for panels E and F, and in Figure 3, it is not clear how the degrees of freedom were calculated. These should be integer values, not values with decimal places. Please clarify this.

We used repeated measures ANOVA to analyze the results. While this design reduces error in variance due to individuals between groups, it also introduces a correlation of errors in individuals between groups. This happens because the same individual is tested across groups. At times this causes a violation of assumptions of sphericity of data. Since sphericity of data is one of the assumptions for ANOVA, its violation needs to be corrected using Greenhouse-Geisser correction (Abdi, Encyclopedia of research design 1, 2010). Assuming the data set might violate the assumptions of sphericity, we applied the Greenhouse-Geisser correction. However, we agree that with the sample size we have, it is very difficult to conclude if the assumptions of sphericity are violated. Our earlier analysis adopted a cautious approach. Hence, we have now modified our analysis. Hence, we no longer assume violation of assumptions of sphericity and not use the Greenhouse-Geisser correction. Therefore, the degrees of freedom are integers and not values with decimal places now. This is consistent with the suggestion made by the reviewer. We have rectified this in Figure 3 and the corresponding legends. Importantly, this change in analysis does not affect the interpretation of any of our results.

[Editors' note: further revisions were requested prior to acceptance, as described below.]

Essential revision:

The reviewers agree that the manuscript is enhanced by the new experiments showing that pharmacological inactivation of the BLA 12 days after conditioning had no effect on fear expression or extinction. This new finding explains a conundrum from the previous version, namely, why was fear expression normal in recently stressed rats even though BLA excitability was increased? The answer seems to be that BLA is not involved in fear expression at that timepoint. This increases our understanding and interpretation of chronic stress experiments: not only is the timing relative to conditioning important, but so is the duration of the stress period, as BLA's involvement is time-limited.

The reviewers were impressed that the authors provided new experiments in response to the request in the initial review to simply "Discuss the potential reasons for the different outcomes of pre- and post-conditioning stress on extinction." However, given the negative result (no effect of BLA inactivation on fear expression at the late time point), a positive control is needed to show that the pharmacological inactivation of BLA was successful, by showing that BLA inactivation in the same place/dose but at an earlier time point impairs memory expression whereas at a later time it does not (with histological placements of cannula tips shown to confirm location in BLA and the volume and concentration of muscimol provided). The reviewers did not feel comfortable with the inclusion of the new experiments in the manuscript without this essential control, and hope that the authors will choose to do these additional experiments.

Alternatively, since the authors had gone above-and-beyond the reviewers' request to simply "discuss" potential reasons, an acceptable alternative would be to remove the new experiments, and simply discuss potential reasons based on the published results in the literature.

We are grateful to the reviewers for their flexibility in how this issue can be addressed in our revised manuscript. We agree with the reviewers and appreciate the value of having a positive control confirming that pharmacological inactivation of BLA at an earlier time point is successful in impairing fear memory expression. Indeed, it is the importance of directly testing why BLA hyperactivity in stressed animals did not affect fear expression that prompted us to do the new inactivation experiments. This, in turn, led to the new question about a positive control raised by the reviewers.

Unfortunately, Dr. Rahman, the first author who carried out all of the in vivo experiments for this study, has already left the lab to start his post-doctoral work abroad. As a result, we do not have anybody else in the lab at this time who can complete these additional control experiments in a timely fashion. Further, the elegant study by Do-Monte et al. (Nature, 2015), showing optogenetic silencing of BLA reduced freezing when it was performed 6 hours, but not 7 days, after auditory fear conditioning, addresses the same issue. So, we hope this will partly make up for the absence of a positive control confirming the same finding using a pharmacological inactivation in the present study. Thus, as much as we would like to do these control experiments, we are unable to do so due to constraints that are beyond our control. Please accept our sincere apologies for this.

Hence, we have removed the findings on the BLA inactivation experiments (Figure 4 in the earlier version) from the Results and Discussion section. Instead, we have opted for the “acceptable alternative” by addressing these points in the revised Results and Discussion section. Further, since the reviewers have expressed positive views on the overall utility of the final summary figure (new Figure 4, which was Figure 5 in the earlier version), we have covered these points in the form of preliminary results in Figure 4—figure supplement 1.

Author details

1. Mohammed Mostafizur Rahman

   National Centre for Biological Sciences, Bangalore, India

   Present address

   Department of Molecular and Cellular Biology, Center for Brain Science, Harvard University, Cambridge, United States

   Contribution

   Conceptualization, Data curation, Formal analysis, Validation, Investigation, Visualization, Methodology, Writing—original draft, Writing—review and editing

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-9355-4867

2. Ashutosh Shukla

   National Centre for Biological Sciences, Bangalore, India

   Contribution

   Formal analysis, Validation, Investigation

   Competing interests

   No competing interests declared

3. Sumantra Chattarji

   1. National Centre for Biological Sciences, Bangalore, India
   2. Centre for Brain Development and Repair, Institute for Stem Cell Biology and Regenerative Medicine, Bangalore, India
   3. Centre for Integrative Physiology, Deanery of Biomedical Sciences, University of Edinburgh, Edinburgh, United Kingdom

   Contribution

   Conceptualization, Resources, Data curation, Supervision, Funding acquisition, Methodology, Writing—original draft, Project administration, Writing—review and editing

   For correspondence

   shona@ncbs.res.in

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0001-9962-3635

• 2,980

  Page views

• 328

  Downloads

• 9

  Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.