The effects of the stress response on memory and cognitive function have been widely studied with great emphasis on the peripheral effects of the hormones released, including noradrenaline (NA) and glucocorticoids. For example, recent findings have implicated hippocampal glucocorticoid receptor action in the recruitment of CaMKIIa-BDNF-CREB molecular pathways , pathways known to mediate long-term memory formation through induction of long-term potentiation in neurons. Due to the physiological response of stress procured from fearful memories, the use of fear conditioning has been widely used as a model for studying how stress hormones are implicated within the various phases of the modal model of memory.
Under the modal model of memory, memory is said to be liable, and thus susceptible to loss and disruption, at two points: after acquisition (i.e. before consolidation) and after retrieval (i.e. before reconsolidation). The implications of reconsolidation have been speculated to be adaptive by allowing the ability to update past memories and schemas with new information as well as maladaptive by potentially altering past memories causing misinformation or creation of false memories. Currently, the molecular mechanisms of reconsolidation of fear memories within the amygdala are being uncovered with recent research deeming the translational regulator mTOR  and the protein CREB as necessary in reconsolidation .
Taking into consideration the effects of stress hormone function in memory and the liable states of memory, especially during reconsolidation, the implications of systems and molecular research allow for the opportunity to perhaps disrupt and even selectively erase memories acquired during negative stressful events, specifically events which could lead to the onset of Post-Traumatic Stress Disorder (PTSD).
Table of Contents
1. Fear Memories
|Fig 1. Fear Conditioning Protocol and Circuitry|
|Top:Typical Fear Conditioning Protocol
Adapted from 
Bottom: Neuronal Circuitry Involved During Fear Conditioning
Important brain structures and nuclei include:
HIP = Hippocampus, LA = Lateral Amygdala, CEm = medial central nucleus of amygdala
LH = Lateral Hypothalamus, PVN = Paraventricular nucleus of hypothalamus
BNST = Bed nuclei of the stria terminalis, NAcc = Nucleus Accumbens
d/vPAG = dorsal/ventral periaqueductal grey
Adapted from 
Much of the information known about cognition during fearful and aversive conditions has been due to the behavioural paradigm of Pavlovian fear conditioning. In this associative learning task, a neutral conditioned stimulus (CS), e.g. an auditory tone, is paired with an emotionally harmful unconditioned stimulus (US) like an electric foot shock . In a normal fear conditioning protocol (Figure 1, top), an animal is first placed in a conditioning chamber on training day where it will learn to associate the context and tone played (CS) with the electric foot shock it will receive during the training trial. During test day, the context of the box is changed (e.g. through altering the smell and the interior of the box) in order for the CS to be presented by itself. In this task, measures of “freezing” behaviour seen in the animal during testing is used as an operational definition for learning, indicating to the researcher that the animal has learned to associate the US with the CS.
1.1 Fear Memory Circuitry
Through utilizing the paradigm outlined above, much work has been done in furthering the field of stressful learning to elucidate the neuronal circuits involved in the physiological and behavioural responses seen during testing. Generally speaking, the “fear circuit” (Figure 1, bottom) is a network of brain structures comprising of three main functions  :
1) areas which receive and relay sensory input
2) an integrating area which experiences plasticity and long-term potentiation during conditioned learning
3) output areas which control the responses generally seen after fear conditioning
Master regulation of this circuit is due to the action of the amygdala found within the medial temporal lobe. More specifically it is the action of the lateral amygdala (LA) which mediates and regulates most of the flow of information coming from the cortex and thalamus to output brain structures. Much research has been done to verify this claim with work on LA neurons showing that after fear conditioning they have increases in field potentials , gene expression  and protein phosphorylation all known to be required for the mechanism of long-term potentiation and synaptic plasticity in learning. To solidify this claim, recent work has also shown that the convergence of CS and US specifically occurs within the neurons of the LA through indirect analysis of the expression of immediate early gene Arc . Efferent targets of amygdalar neurons synapse onto many brain nuclei including the ventral periaqueductal grey (vPAG) and the paraventricular nucleus (PVN) of the hypothalamus both responsible for the freezing and stress responses seen, respectively, during fear conditioning.
|Fig 2. Molecular Mechanisms of Fear Memory Acquisition
and Consolidation within LA Neurons
|Blue lines indicate proposed pathways found to be involved in fear memory acquisition.
Black lines indicate molecular pathways shown to be required in memory consolidations and maintenance.
PKA = protein kinase A, CaMKII = Ca2+/calmodulin-dependent protein kinase
MAPK = Mitogen-activated protein kinase, βAR = Beta-adrenergic receptor
CREB = cAMP Responsive Element Binding Protein, CRE = cAMP Responsive Element
mTOR = Mammalian target of rapamycin
Adapted from 
1.2 Molecular Mechanisms of Fear Memory Acquisition
As expected with most forms of associative learning within the brain, lateral amygdalar neurons show many molecular properties of Hebbian synaptic plasticity and long-term potentiation (LTP)  . Recent work by Dalton and colleagues, reveal that the blockage of NMDA receptor (NMDAR) subunit GluN2A found on the post-synapse of lateral amygdalar neurons was enough to disrupt the acquisition of a fear memory . Electrophysiological recordings of these lateral amygdalar neurons indicated that blockage of GluN2A prevented LTP, further strengthening the claim that NMDAR and its GluN2A subunit is required for fear memory acquisition . Other molecules widely associated with LTP including AMPA receptor (AMPAR) surface expression and CaMKII autophosphorylation have also been implicated in the acquisition phase of fear memories. For example after a tone-shock associative pairing, increased AMPAR GluR1 subunit trafficking and surface expression was found post-synaptically on lateral amygdala neurons  along with increased populations of autophosphorylated CaMKII within lateral amygdala spines . CaMKII was also found to be necessary for the mechanism of fear memory acquisition with infusions of KN-62, a CaMKII selective inhibitor, within the amygdala causing impairment of acquisition but not expression of fear memories in rats .
Unique to the acquisition of fear memories within the lateral amygdala is the effect neuromodulators have on associative plasticity. Due to the close relationship between the fear circuit and the LA with the stress circuit (See Fig 1., bottom) it is no surprise that noradrenaline (NA) plays an important role in the acquisition of fear memories. Specifically wanting to determine the effects this stress hormone has within the LA, Bush and colleagues used microinfusions of propranolol, a Beta-adrenergic receptor (βAR) antagonist, within the LA to determine the effects NA blockage would have during the acquisition, consolidation and expression of a fear memory. Surprisingly, blockage of βAR activity pre-training disrupted acquisition impairing both consolidation of short-term and long-term memory while post-training intra-LA infusions did not . Taken together these results indicate the importance of βAR activity during acquisition of fear memories within the lateral amygdala.
1.3 Molecular Mechanisms of Fear Memory Consolidation
Important in the understanding of fear memories is their different states. Consolidation is the process by which fear memories are converted from a liable state after acquisition into stable long-term memories . Generally, the consolidation of acquired fear memories is said to involve the recruitment of secondary messengers and protein synthesis within the neurons of the LA . Extensive work done by Schafe and LeDoux over the years has shown that memory consolidation of fear memories within LA neurons require protein synthesis and the actions of PKA and ERK/MAP kinases  . Other pathways involved in fear memory consolidation are highlighted in black seen in Fig 2.
Surprisingly, not only is the LA crucial in the consolidation of acquired fear memories, work done by Gale and colleagues., also implicate the LA in the permanent storage of these memories. In an intricate and extensive study, mice with lesions within the basolateral amygdala (BLA) showed deficits in both remote (fear memory acquired sixteen months prior to lesions) and recent (fear memory acquired one day before lesions) fear memories, compared to mice with sham lesions . Results from this study not only implicate the LA in fear memory consolidation, but it also provides direct evidence of this structure as a potential permanent safe house for fear memories to reside in.
1.4 Involvement of Glucocorticoids in Fear Memories
1.4.1 Fear-Induced Release of Stress Hormones
As indicated above, one of the brain nuclei recruited by an evoked fear response is the paraventricular nucleus (PVN) of the hypothalamus (Fig. 1, bottom) resulting in the release of hallmark stress hormones noradrenaline and glucocorticoids. By activating the PVN, efferent amygdalar neurons stimulate the production of corticosterone releasing hormone (CRH) and vasopressin/antidiuretic hormone (ADH) . CRH and ADH then travel through the hypophyseal portal system transporting the tropic hormones from their site of production within the PVN to the anterior pituitary (AP). Here CRH binds to its receptors found on corticotropes within the AP to induce the synthesis and release of adrenocorticotropic hormone (ACTH) into systemic circulation. Once in the blood stream, ACTH primarily acts within the zona fasciculata of the adrenal cortex to stimulate production and release of glucocorticoids. The PVN also has CRH containing neurons which synapse onto cells found within the locus coeruleus . CRH binding to these cells, which synapse onto various subcortical structures i.e. the amygdala and hippocampus, stimulate the production of the peptide hormone noradrenaline causing increases in arousal .
1.4.2 Actions of Glucocorticoids within the Amygdala
Due to their lipophilicity, glucocorticoids have the ability to pass the blood-brain barrier and bind to either mineralcorticoid or glucocorticoid receptors (MR and GR respectively) eliciting both non-genomic and genomic responses within the limbic system . Due to pioneering work done by Donley and Roozendaal it has been widely accepted that GR action within the LA is required for memory consolidation. Together their work showed that the administering of GR antagonist RU486 into the LA and not the central amygdala (CEA) showed inhibition of memory consolidation   while GR agonist RU28362 was able to enhance consolidation of fear memories . Administration of corticosterone (CORT) has also shown to increase the intrinsic excitability of LA neurons in rats . Similar results in the LA neurons of mice have been shown through miniature excitatory post-synaptic currents (mEPSCs) seen initially via a non-genomic interaction of CORT with MR. Maintenance of these mEPSCs hours later required genomic interactions of CORT with GR to further induce protein synthesis and GR expression within these neurons .
Although studies implicating glucocorticoids and consolidation exist, work on MR and GR action in LTP and plasticity within the LA is limited. Current knowledge implicates the interaction of GR with βAR to induce the cAMP/PKA pathway during the consolidation of fear memories in LA neurons .
Reconsolidation is the time-dependent process by which previously consolidated memories are re-activated and enter a liable state . Said liable state is proposed to either be involved in the underlying mechanism of “lingering consolidation”, by which some memories take an extended time to consolidate, or to allow for the continual updating of past memories with new information . It is important to note however, that reconsolidation is not a universal property of all memories. Thus inherent to the premise of reconsolidation is the idea of boundary conditions of memories, which include the age of a memory, intensity of training, and reactivation predictability, to provide some insight as to why some memories do not destabilize after reactivation .
2.1 Mechanisms of Reconsolidation
|Fig 3. Molecular Mechanisms of Fear Memory
Reconsolidation within LA Neurons
|Adapted from |
Groundbreaking work by Nader and colleagues, provided initial reports on the necessity of de novo protein synthesis within LA neurons for the mechanism of reconsolidation . Infusion of anisomycin, a protein synthesis inhibitor, directly into the BLA of rats directly after exposure to just the CS of a fear memory produced impaired deficits in memory expression the next day when compared to direct BLA infusions without retrieval trials . Surprisingly after infusion of anisomycin, rats were able to relearn the fear conditioned memory indicating that LA function was not damaged. By showing the necessity of the retrieval trial in disrupting reconsolidation of a previously consolidated fear memory, Nader and colleagues provided conclusive evidence for the liable state of retrieved memories and the requirement of reconsolidation of said retrieved memories via time-dependent protein synthesis. Further work has reinforced this adaptive memory mechanism in a variety of other animals and memory types ( See Table 1 in ).
The liability of memory post-retrieval was translated into a tangible physiological mechanism through work done by Kim and colleagues. Since neurons that have undergone consolidation show mGluR1 depotentiation , Kim and colleagues went on to use this premise as a marker for consolidated neurons, to demonstrate that synapses enter a liable state post-retrieval through insensitivity to mGluR1 depotentiation of LA neurons . A recent paper by Li and colleagues (2013) has gone on to implicate mTOR activity in the strengthening of synapses during reconsolidation showing that inhibition of mTOR activity using rapamycin causes synaptic weakening via post-synaptic mechanisms.
Work on the molecular mechanism of reconsolidation within LA neurons is currently at an early stage. Interestingly, molecular mechanisms of reconsolidation implicate the same molecular pathways involved in memory acquisition and consolidation (see Fig. 3) including NMDAR receptor activation  and βAR activity . Other molecules found to be involved in reconsolidation of LA neurons include the transcription factor NF-κB  and the protein kinase cdk5 . As outlined in Fig. 3, a novel perspective of reconsolidation using epigenetics has also elucidated the roles of histone deacetylases (HDACs) and DNA methyltransferases (DNMTs) in fear memory reconsolidation. Inhibition of HDACs following fear memory retrieval allowed for the enhancement of reconsolidation while inhibition of DMNTs caused the reverse implicating a yin-yang regulatory role of both proteins within LA neurons during reconsolidation .
2.2 Intentional Disruption of Reconsolidation
|Vid 1. Dr. Elizabeth Phelps explaining
Reconsolidation and Extinction Trials
With knowledge that some memories can enter this liable state immediately post-retrieval, much of the work done on reconsolidation has looked at its disruption as a therapeutic method for PTSD treatments and for resolving emotionally laden fear memories . One method of such disruption is the use of extinction training and exposure therapy. By exposing an animal or a patient with a learned associative CS-US pairing to just the CS in repeated presentations, said association is often weakened in tested subjects . The main caveat of exposure therapy however is due to its temporary weakening of this CS-US association. Often after a prolonged period of extinction therapy, the associative fear memory can return either through spontaneous recovery, renewal or with mild re-exposure  .
|Fig 4. fMRI BOLD levels within the amygdala
of subjects during renewal trials
|Top: Subjects given exposure therapy 6 hours after reminder
Bottom: Subjects given exposure therapy 10 minutes after reminder
Adapted from 
However, with new knowledge of the specific timeframes established through reconsolidation studies, the blending of both reconsolidation and extinction therapy can perhaps weaken associative fear memories. A recent paper by Agren and colleagues was able to do this through the combining of temporal delays post-retrieval and extinction therapy to extinguish the fear memory association seen in human subjects . On Day 1 of the training paradigm, subjects were given fear conditioning of 16 mild shocks (US) paired to a visual cue (CS) with a reminder trial using the same visual cue 24 hours later (Day 2). After this reminder, subjects either underwent extinction trials immediately after exposure to the reminder trial (10 minutes after) or 6 hours after the reminder trial. On Day 3, subjects were exposed to renewal trials in a novel environment through electrode placements . fMRI images taken on this day indicated that amygdala activity was lower in subjects that received the extinction training immediately after the reminder trial (see Fig. 4) . The resultant decrease in fMRI BOLD levels within the amygdala indicated that by exposing patients to extinction trials immediately after a reminder one, researchers were able to disrupt reconsolidation of the induced fear memory. Reinstatement of the fear memory through unpaired shocks 48 hours later (Day 5) was also decreased in subjects who received extinction trials immediately after retrieval .
Outside of extinction trials, a lot of pharmacological agents exist to disrupt the reconsolidation of associative fear memories post-retrieval. Although rapamycin has been shown in various studies as an effective disruptor of reconsolidation, translational research requires the use of agents that would minimize harm when taken pharmacologically. Current candidate drugs show a mechanism of action which disrupt molecular processes that endogenous stress hormones would normally recruit. One widely studied drug considered in the treatment of PTSD is propranolol, a βAR antagonist, used as an anti-hypertensive to treat those with heart disease, anginas and certain types of tumours . Since βAR activity is fundamental in the acquisition, consolidation and reconsolidation of memories, blockage of this receptor has long been thought to be sufficient in preventing reconsolidation of such emotional memories. This hypothesis was recently validated in humans through a study done by Soeter and Kindt. In their study, acquired fear memories were shown to be selectively erased upon oral administration of propranol ninety minutes before reactivation of these memories . Subjects who did not undergo the reactivation trials also showed no observable decrease in fear response , providing evidence that both retrieval and propranolol were instrumental in the selective erasure of this fear memory.
Another pharmaceutical drug that has shown promise in the treatment of PTSD and dampening of emotional responses of fear memories is the glucocorticoid antagonist mifepristone or RU486. Since much is known about the pharmacological properties of glucocorticoids and mifepristone and because glucocorticoids are a hallmark of the stress response, preliminary research has aimed at addressing the use of this emergency contraceptive as a novel method in disrupting reconsolidation. Work using RU486 has also provided direct evidence of the importance of glucocorticoid receptors during the reconsolidation of fear memories . In line with these results is a study by Taubenfield and colleagues which showed that systemic administration of RU486 in mice was effective in forgetting and erasing fear memories post-retrieval . This forgetting seen via RU486 systemic injections could be due to the disruption of βAR activity normally caused by endogenous glucocorticoids . Since this pathway is one of the two major ones implicated in the reconsolidation process of memories (see Fig. 3), interferences in activity by a GR antagonist should also interfere with reconsolidation (see Fig. 3). Extensive post-testing was also done in the study to verify the absence of spontaneous recovery of the fear memory .