Review

An Update on Memory
Reconsolidation Updating
Jonathan L.C. Lee,1,* Karim Nader,2,* and Daniela Schiller3,*
The reactivation of a stored memory in the brain can make the memory tran-
Trends
siently labile. During the time it takes for the memory to restabilize (reconsoli-
Memory reactivation can lead to its
date) the memory can either be reduced by an amnesic agent or enhanced by destabilization, necessitating the resta-
memory enhancers. The change in memory expression is related to changes in bilization of the memory trace through a
process termed ‘reconsolidation’.
the brain correlates of long-term memory. Many have suggested that such
retrieval-induced plasticity is ideally placed to enable memories to be updated Interfering with reconsolidation by
with new information. This hypothesis has been tested experimentally, with a pharmacological means may lead to
blockade or strengthening of subse-
translational perspective, by attempts to update maladaptive memories to quent expression of the memory.
reduce their problematic impact. We review here progress on reconsolidation
This bidirectional modulation of synaptic
updating studies, highlighting their translational exploitation and addressing
plasticity during reconsolidation hints at
recent challenges to the reconsolidation field. its functional role: memory updating by
incorporating new information.
Reconsolidation and the Dynamic Nature of Memory
Evidence for memory updating is accu-
Despite our strong tendency to view memory as an accurate depiction of past events, scientific mulating across memory systems,
analysis suggests that memories are not fixed entities but are instead a dynamic process for developmental stages, and clinical
updating memories. The retrieval (see Glossary) and expression of an existing memory places populations, pointing to possible non-
invasive techniques for permanent
it into a labile state. What this means is that the reactivated memory becomes vulnerable to
modification of maladaptive memories.
interference. In real life, new experience can interweave with the memory it triggers and modify
future recollections. Psychological and physiological states could enrich or deplete neural However, replication failures and alter-
resources needed for the memory re-storage, thereby strengthening or weakening the mem- native explanations challenge clinical
translation. These obstacles can be
ory. In the laboratory, these interventions have typically been pharmacological in nature, but in
viewed constructively by highlighting
recent years the scope of intervention has broadened into non-pharmacological (i.e., behav- the need to demarcate the source of
ioral) treatments. The ability of these interventions to impair subsequent memory expression, failure: lack of destabilization or insuffi-
and the phenomenological similarity to the effects of post-learning treatment, have led to their cient updating at restabilization.

interpretation within a memory reconsolidation framework.

Three lines of evidence support the existence of a stabilization period on the order of hours after
the acquisition of new memories. First, performance can be impaired if amnesic treatments
such as electroconvulsive shock [1] or protein synthesis inhibitors [2] are given after learning. 1
School of Psychology, University of
Second, performance can be impaired if new competing learning occurs after the initial learning Birmingham, Edgbaston, Birmingham
[3]. Third, retention can be enhanced by administering various compounds after the initial B15 2TT, UK
2
learning, such as strychnine [4]. Crucially, all three manipulations are effective only when given Department of Psychology, McGill
University, Department of
soon after new learning and not when given after a delay. These findings gave rise to theories of Psychology,1205 Dr Penfield Avenue,
synaptic consolidation [5,6]. Montreal, QC H3A 1B1, Canada
3
Departments of Psychiatry and
Neuroscience, Icahn School of
The same three lines of evidence point toward the existence of a post-retrieval restabilization Medicine at Mount Sinai and
period. First, performance can be impaired if a range of amnesic treatments are given after Friedman Brain Institute, New York,
memory reactivation [7–11]. Second, performance can be impaired if new competing learning NY 10029, USA

occurs after the reactivation [12,13]. Third, retention can be enhanced by administration of *Correspondence:
various compounds after reactivation [14,15]. Crucially, again all three manipulations are [email protected] (J.L.C. Lee),
effective only when given soon after reactivation but not when given after a delay. Ultimately, [email protected] (K. Nader),
[email protected] (D. Schiller).

Trends in Cognitive Sciences, July 2017, Vol. 21, No. 7 http://dx.doi.org/10.1016/j.tics.2017.04.006 531
© 2017 Elsevier Ltd. All rights reserved.
these findings led to the concept of a post-retrieval reconsolidation process [16–18]. That is, Glossary
memory retrieval can lead to the destabilization of the memory, thereby necessitating a Conditioned stimulus: a previously-
reconsolidation process to restabilize it. neutral stimulus that has been
learned to predict an outcome;
presentation of the stimulus evokes
The existence of a reconsolidation process begs the question of what, if any, function it serves. the memory of the prior learning.
Many have suggested that the retrieval-induced plasticity is ideally placed to enable memories Counterconditioning: the learning
to be updated with new information [19–23]. This hypothesis has been tested experimentally, of an opposite outcome to that
experienced previously (e.g., a
with a translational perspective, by attempts to update potentially maladaptive memories to
stimulus now predicts reward after
reduce their problematic impact. We review here the progress on such reconsolidation previously being linked with an
updating studies, highlighting their translational exploitation and addressing recent challenges aversive outcome).
to the reconsolidation field. Destabilization: the active transfer
of a retrieved memory into an
unstable state upon the presentation
Reconsolidation Updates Memories of a reminder, such that the memory
The observation that pharmacological treatment at, or shortly after, the reactivation of a will decay if it is not actively
memory leads to subsequent long-lasting amnesia is held to be a defining feature of memory restabilized.
Extinction: the presentation of a
reconsolidation [23,24]. That is, memory reactivation can, but does not always [22,25], result in
conditioned/learned stimulus now in
the destabilization of the existing memory, thereby necessitating a protein synthesis-dependent the absence of the previously
phase of reconsolidation [26]. The bidirectional nature of synaptic plasticity modulation allows associated outcome; this results in
reactivated memories to be potentiated by pharmacological enhancement of neurochemical or the temporary decline of subsequent
memory expression.
cellular processes [15,27]. Given the slim chance of encountering any of these laboratory Reactivation: re-exposure to
amnesic agents in real life, these protocols indicate the existence of reconsolidation but not its memory reminders, which may result
functional role. If reconsolidation were to serve an adaptive function, there should be naturally in destabilization of the previously
occurring reconsolidation interference. A prominent working hypothesis is that reconsolidation learned neural representation of
memory.
enables the update of memories to maintain their relevance in the face of changing circum- Reactivation–extinction (retrieval–
stances [22–24,28–31]. By extension, non-pharmacological interventions should have the extinction): the combination of
ability to modify memory by influencing the reconsolidation process. memory reactivation (usually via a
reminder that results in memory
retrieval) and, after a brief interval,
Behavioral interference protocols have the same structure as pharmacological protocols: long- subsequent extinction.
term memory is reactivated, and behavioral interference (instead of administering a pharma- Reconsolidation: the active process
cological agent) ensues during the reconsolidation window. In the same way pharmacological that is necessary to restabilize a
reactivated/destabilized memory;
agents block or potentiate reconsolidation, behavioral interventions may similarly cause amne-
disruption of reconsolidation results
sia or strengthen the memory. Behavioral interventions, however, are uniquely poised to in memory impairment, while new
capitalize on reconsolidation in a constructive manner, allowing the incorporation of new information is incorporated during
information into an existing memory (Figure 1). The result of this update could be manifested reconsolidation into an updated
memory.
in different forms depending on the target memory system (Table 1). In procedural memories,
Retrieval: a reminder results in
for example, when a finger-tapping sequence memory was reactivated and followed by a new recollection of the previously learned
sequence, the accuracy of the initial memory diminished [13]. Declarative memories are also memory; the term encompasses the
vulnerable to new or conflicting information that follows reactivation, affecting the amount of multiple processes from reactivation
of the neural memory representation
information retrieved from the original memory or enhancing intrusions of the new material into it to behavioral expression of the
[32]. For example, learning a new list of objects following the reactivation of a previously learned memory.
list caused items from the second learning to infiltrate the memory of the first list [33]. By the State-dependent learning: the
same token, when targeting emotional memories, extinction learning following reactivation of embedding of a learned memory
within the physiological state present
a threat conditioned stimulus may lead to long-term reduction in conditioned defensive at the time, such that retrieval of the
responses [34,35]. In all these cases, the initial memory is not ‘erased' but instead incorporates memory is most successful if the
new information, consistent with the view of reconsolidation updating. physiological state is reinstated.
Unconditioned stimulus: a
motivating outcome (aversive or
The bidirectional nature of reconsolidation updating is especially relevant for emotion-related rewarding) which promotes learning
psychiatric disorders because it allows the modification of emotional memories in different about a conditioned stimulus that
directions. Neutral context memory can acquire threatening properties when reconsolidation predicts the outcome.
mediates the incorporation of new threat-learning experience [36], contextual threat memories
can strengthen with additional learning after retrieval in a process depending selectively on
reconsolidation mechanisms [37], and appetitive associative memories can lose their rewarding

532 Trends in Cognitive Sciences, July 2017, Vol. 21, No. 7

 Learning Reminder Updang Retrieval
Exncon

Countercondioning

Interference

Figure 1. The Stages of Reconsolidation Updating. Experiencing a significant event may result in the formation of a
long-term memory (in this case, an aversive emotional memory of a threatening dog). Encountering cues associated with
the event (such as the red collar) may serve as a reminder that triggers the memory and destabilizes it. Re-stabilization
(reconsolidation) of the memory ensues until the memory returns to a stable inactive state. During this time-window, update
may occur in several possible ways, for example: extinction, counterconditioning, or interference. These processes may
provide new information that is then incorporated into the memory (extinction, counterconditioning), or compete for
resources and interfere with re-storage of the memory, thereby hindering subsequent retrievals. The result is an updated
memory, in this case devoid of the negative emotional response. In the case of drug-related memories, such updating may
result in reduced drug craving. Reconsolidation updating in other memory systems might induce other alterations such as
modified memory content. For example, new learning during reconsolidation of episodic memories can result in reduced
correct retrievals of the episodic items; new learning during reconsolidation of motor memories can result in altered speed
or accuracy of motor performance.

properties using the reactivation–extinction procedure [38,39]. Another updating process

that has effectively reduced reward-related responding involves reactivation followed by
counterconditioning [40–42], where the appetitive conditioned stimulus is paired with an
aversive outcome during reconsolidation.

Beyond integrating new information, behavioral intervention may also affect reconsolidation by
engaging a cognitive process that competes for the same neural resources that reconsolidation
relies upon. A recent study [43] found that playing the computer game Tetris following
reactivation of a trauma film memory reduced intrusive memories. Memory reactivation alone,
or playing Tetris without reactivation, did not affect the number of intrusions. Because Tetris is a
visuospatial task, playing Tetris possibly competes with reconsolidation for memory-related
neural resources, and therefore interferes with the re-storage of the trauma film memory. Unlike
the effects of pharmacological amnesic agents, however, this behavioral manipulation did not
abolish subsequent retrievals (as in ‘erasure’), but affected only involuntary retrievals (the
‘intrusions’). Better understanding Tetris interference will need additional research assessing
non-affective tasks that engage other neural resources, thus gauging the degree of resource
competition and hedonic impact.

These studies suggest that reconsolidation updating may be a viable pathway for non-
invasively modifying maladaptive memories that are at the core of some psychiatric conditions.
In the next section, we describe the current state of psychiatric translation and possible paths
for moving forward.

Translation of Memory Reconsolidation Updating

To date, very little research in clinical populations has attempted testing and translating the
reconsolidation updating phenomenon for clinical treatment. The first study to harness the
reactivation–extinction procedure was carried out in heroin-addicted individuals [39]. One
group of participants was reminded of the drug memory using a short video clip of heroin
cues, and then underwent extinction training (repeated exposures to heroin-related cues);

Trends in Cognitive Sciences, July 2017, Vol. 21, No. 7 533

Table 1. Some Paradigms in Which Behavioral Reconsolidation Updating Has Been Observeda
Refs

Experimental paradigm Pavlovian threat conditioning [34,35,69,121–127,138]

Pavlovian threat conditioning with fear-relevant stimuli [85,88]

Context threat conditioning [36,38,57,128]

Pavlovian reward conditioning [38,42]

Instrumental reward conditioning [39,67,129,130]

Motor sequence learning [13]

Episodic memory [32,33,131]

Subliminal instrumental conditioning [132]

Spatial memory [133]

Updating treatment Extinction [34,35,38,69,75,121–127]

Vicarious extinction [88]

Imaginal extinction [134]

Counterconditioning [40–42]

Extinction following unconditioned stimulus reminder [57,85]

Tetris interference [43]

Interference with new learning [13,32,33,131,133]

Contextual threat reactivation concomitant [135]

with appetitive stimuli

Species Mice [118,128,136,137]

Rats [34,38,39,42,57,138]

Humans [32,35,57,69,85,88,121–127]

Developmental stages Adolescents, humans [126]

Juvenile/adolescent rats [139]

Pharmacological enabling Epigenetic priming of behavioral memory updating [136]

Clinical populations Heroin addicts [39]

Hazardous alcoholic drinkers [40,49]

Tobacco smokers [44]

Spider phobic [53–55]

a
Examples of various experimental paradigms, updating treatments, species, developmental stage, and clinical popula-
tions, from studies reporting evidence for reconsolidation updating.

another group went through a similar protocol but had a 6 h break after the reminder, and the
third group had extinction training without the reminder. The results show that only the first
group showed significant attenuation in craving that persisted for at least 6 months. These
findings point to the potential efficacy of reconsolidation updating mechanisms in drug addic-
tion interventions and prevention of relapse.

Another attempt at manipulating maladaptive addiction-related memories using the reactivation–

extinction procedure examined active cigarette smokers [44]. The participants viewed a short
video showing people smoking cigarettes to reactivate the relevant smoking-related memories.
Ten minutes later, the participants underwent extinction using extended exposure to smoking
cues. Compared to a control group who had not undergone reactivation, the participants in the
retrieval–extinction group showed rapid and lasting reduction in smoking frequency, as well as
larger and enduring decrease in cue-induced craving that generalized to novel smoking cues.

534 Trends in Cognitive Sciences, July 2017, Vol. 21, No. 7

The relatively short follow-up (1 month) and possibly insufficient statistical power to observe a full
range of smoking outcomes (e.g., cotinine level, days abstinent, and relapse milestones) are
limitations to the nevertheless encouraging demonstration of retrieval–extinction in another clinical
population.

Taking a potentially more robust approach, a recent study used reactivation followed by coun-
terconditioning in hazardous alcoholic drinkers [40]. The study induced memory reactivation using
an actual alcoholic drink as the reminder, which the participants were almost allowed to drink but
were stopped. In two other control conditions the participants were given either a non-alcoholic
beverage or an alcoholic beverage that they were allowed to consume. The premise of this
manipulation is that prediction error at the time of retrieval (in this case violating expectation of
alcohol consumption) is effective and even essential for successful memory destabilization [45–
48]. Counterconditioning ensued 10 min following reactivation, and included pairing of alcohol-
related cues and disgusting images. The authors found retrieval-induced reduction of attentional
bias, cue-induced craving, and liking ratings. A different manipulation for reconsolidation updating
in a similar population of hazardous drinkers was less effective [49]. In this case the study used an
emotion-regulation technique of cognitive reappraisal following reactivation, but only verbal
fluency for positive alcohol-related words diminished when reappraisal followed memory reacti-
vation. This effect was observed only with the reminder that consisted of omission-prediction error
(preventing from drinking) but not value-prediction error (drink is unexpectedly bitter), pointing to
specificity in the type of violation required for memory destabilization.

In anxiety and threat-related disorders there is some promising but very preliminary support in
implementing reconsolidation update mechanisms. Existing therapeutic interventions appear
to engage reconsolidation mechanisms, but direct scientific evidence is lacking. For example, in
a pilot study [50], individuals diagnosed with post-traumatic stress disorder (PTSD) underwent
a procedure the reconsolidation of traumatic memories (RTM) protocol a visual–kinesthetic
protocol also known as the rewind technique [51]. In brief, the procedure consists of a trauma
reminder evoked by the patient retelling the trauma narrative, the patient is then reoriented to
the present and imagines a movie theatre in which the trauma memory is presented in black and
white, from various vantage points, and so forth. The study reports a relatively strong effect
where a majority of treated patients no longer met the diagnostic criteria of PTSD. To directly
link the crucial features of such protocols with reconsolidation, however, a fully controlled and
randomized experimental design ought to be employed.

Three different studies targeted spider-phobia using the retrieval-extinction protocol. One study
found no superiority of memory reactivation compared to no-reactivation, as both groups
showed significant reduction in phobic symptoms, leaving open the possibility that retrieval-
induced updating may augment treatment efficacy [52]. Another study found significant
improvement following reactivation–extinction compared to the reversed (extinction–reactiva-
tion) procedure, but also found the unexpected immediate effect of rapid threat attenuation
during exposure itself [53], providing some proof of concept. More conclusively, a study in a
similar spider-phobic population [54], compared groups that saw a spider image either 10 min
or 6 h (within or outside the reconsolidation window, respectively) before undergoing an
extinction session. The study found retrieval-induced enhancement in approach behavior
toward the spiders in the 10 min compared to the 6 h group, an effect that lasted 6 months
[55]. Together, these studies indicate that post-retrieval exposure to spider stimuli may
effectively reduce spider phobia via reconsolidation updating but protocol optimization is
required.

Indirect evidence comes from a recent retroactive study [56] supporting the ecological validity
of the reactivation–extinction paradigm. The study reviewed the course of traumatic memories

Trends in Cognitive Sciences, July 2017, Vol. 21, No. 7 535

in youth from New Orleans who survived both hurricane Katrina in 2005 and hurricane Gustav in
2008. Participants who had a milder exposure to hurricane Gustav, which was evidently less
devastating than hurricane Katrina, recalled fewer negative memories of hurricane Katrina 1
month after hurricane Gustav, and showed lower levels of PTSD symptoms induced by
hurricane Katrina [56]. There may be parallels between these observations and the reactivation-
–extinction paradigm. The milder exposure to hurricane Gustav may have acted as a reminder
for Katrina memories. This form of reactivation resembles the use of an unconditioned
stimulus as a reminder, such as a milder shock reminder of threat conditioning [57], a nicotine
reminder in nicotine-seeking rats and humans [58], or a cocaine reminder in cocaine-seeking
rats [59]. This form of reactivation may suffer fewer limitations compared to the stimulus-
specificity of reconsolidation impairments with cue exposure. It may also more effectively
destabilize the memory owing to its potency or the unexpected presentation of the outcome
without the prior predictive stimulus [60]. By the same token, the experience of a milder storm
may both generate sufficient prediction error to trigger destabilization and provide the updating
experience to diminish the memory of hurricane Katrina. Thus, forms of treatment that resemble
re-exposure to the trauma, albeit in a milder form (such as guided imagery), may capitalize on a
reconsolidation updating mechanism. These few studies provide very preliminary evidence for
the potential therapeutic efficacy of reconsolidation updating. It is possible that existing
therapies or even naturally occurring sequence of events may already implement reconsoli-
dation updating processes. A direct link between treatment protocols and their underlying
mechanisms ought to be scientifically validated.

The memory-updating role of reconsolidation is consistent with the manifestation of abnormally

strong memories related to anxiety, trauma, and addiction, which take over behavior and
cripple adaptive function. On the one hand, recurrent retrieval of memories imbued with
exceptionally strong emotion may effectively destabilize, strengthen, and update these memo-
ries with added emotional impact. On the other hand, it is possible that, owing to conditions at
the time of formation, these memories might lack destabilization mechanisms, thereby imped-
ing reconsolidation and becoming update-resistant. Note that these two conflicting scenarios
the ‘snowball' effect of continuous strengthening or a ‘snapshot' of the original event result
in the same phenotypic expression of an exceptionally strong emotional memory, but require
opposite treatment. An overly reconsolidated memory would require reconsolidation interfer-
ence, and destabilization-resistant memory would be insensitive to reconsolidation blockers
and require destabilization promoters. These opposing scenarios are not mutually exclusive
within the psychiatric realm because each may explain a different cluster of symptoms or
individual variability within symptom domain. This may also explain why some treatments work
while others are ineffective. Failure to dissociate these two aspects of memory updating may be
at the heart of two main hurdles to translation: alternative explanations and replication failure. It
is therefore imperative to develop positive markers of memory destabilization and restabilization
rather than indirectly assume that these processes took place based on purely behavioral
indices. In the next section, we identify and justify specific research questions addressing the
obstacles that may have the greatest impact upon clinical translation.

The Theoretical Challenges of Reactivation-Dependent Lability

The appeal of the reconsolidation-based interpretation of reactivation-dependent amnesia is
that it implies robust and long-lasting disruption/updating of the memory, thereby providing the
theoretical basis for a persistently beneficial therapeutic strategy for memory-based psychiatric
disorders. However, memory reconsolidation is by no means a universally-accepted process,
and there have been a series of challenges to reconsolidation theory (Box 1). Recently,
alternative interpretations of reactivation-dependent amnesia have been (re)proposed. For
example, amnesia might result from a state-dependent learning process [61,62] or may
simply be a form of uninhibited extinction [63]. State-dependent amnesia suggests that the

536 Trends in Cognitive Sciences, July 2017, Vol. 21, No. 7

 Box 1. Alternative Interpretations of Retrieval-Dependent Amnesia
Alternative interpretations such as state-dependent learning and unleashed extinction follow a series of challenges to
reconsolidation theory (nonspecific drug effects such as lesions [89], state-dependent learning [61], new learning [90],
facilitated extinction [91], and retrieval impairment [92]). All these issues have been explicitly addressed in a previous
review and have been refuted on the basis of their inability to account for the richness of data supporting the
reconsolidation interpretation [23]. Briefly, the observations of recovery from amnesia do not necessitate a retrieval-
–impairment view of amnesia as is implicit within the alterative interpretations. Recovery from amnesia is by no means a
novel observation and has limited interpretative value [93] because recovery procedures can easily supplement an
incompletely disrupted memory [70,94,95]. There is also a computational model showing how a partially impaired
memory can be strengthened by reminder cues, bringing the amnesic group to parity with the controls [96]. Second,
many studies have explicitly tested for, and failed to show, spontaneous recovery, reinstatement, and state-dependent
learning [97]. Moreover, non-reconsolidation accounts fail to explain examples where drug treatment can potentiate,
rather than impair, the reactivated memory [15,27,98]. Such memory enhancements are easily interpreted as potentia-
tion of reconsolidation, but the drug treatment should create a similarly altered internal physiological state to that which
is hypothesized to account for memory impairments. While memory enhancement could result from an impairment of
extinction, the fact that the same treatment can cause bidirectional effects on memory depending upon the parameters
of cue exposure at memory reactivation strongly suggests that competing processes (i.e., reconsolidation vs extinction)
must exist [22,23,25,27,99,100]. Finally, given that extinction is unique to associative memories, unleashed extinction
cannot apply to the wider reconsolidation field.

amnestic drug treatment creates an altered internal physiological state which becomes crucial
for future memory retrieval. Thus, reinstatement of the internal physiological state can be
sufficient to recover the impaired memory, as has been demonstrated for the effects of
cycloheximide and lithium chloride on inhibitory avoidance and conditioned taste aversion
memories [62]. The idea of reactivation-dependent amnesia resulting from uninhibited extinc-
tion similarly takes, as a fundamental starting point, the observation of recovery from amnesia
[63]. In this case a weak footshock reminder delivered during contextual fear-memory testing
recovered subsequent contextual fear. This external reinstatement of the memory is perhaps
less likely to be mediated by reinstatement of an internal physiological state, leading to the
suggestion that reactivation-dependent amnesia is qualitatively similar to extinction, the latter
being well-established to result in easily recoverable memory.

Focusing on reconsolidation updating interventions, it is not clear if and how non-reconsoli-

dation accounts can provide explanations for the induced amnesia. It may be possible that
behavioral treatment can alter neural activity in a manner that creates a state-dependent effect.
Similarly, behavioral intervention may somehow reduce the endogenous inhibition of extinction
for associative memories. However, neither of these alternative accounts can explain the variety
of interventions that have been demonstrated to show beneficial effects: retrieval–extinction,
updating with counterconditioning, and competition for neural resources. Perhaps the only
viable alternative to a reconsolidation-based explanation is specifically related to the original
retrieval–extinction demonstration [34,35], in that it is difficult to disambiguate reconsolidation
update from enhancement of extinction. Pharmacological potentiation of extinction has been
shown previously not only to result in a quantitative enhancement in the reduction of fear but
also a qualitative reduction in the propensity of fear to recover [64–66], thereby mirroring the
persistent memory impairments characteristic of reconsolidation impairments and updates.
Therefore, it is possible that the combination of retrieval with extinction training shortly after
reduces fear not by updating the memory through the retrieval episode but by the retrieval
somehow priming or enhancing the subsequent extinction training. Such alternative interpre-
tations of behaviorally induced memory reduction are also pertinent to the observation that a
reversal of the order of retrieval and extinction (i.e., extinction followed shortly afterwards by
brief retrieval) can reduce alcohol-seeking to a quantitatively similar level as retrieval–extinction
[67]. While the reversed order is less consistent with a reconsolidation-based updating process,
we would argue that it is not a logical extension to conclude that retrieval–extinction similarly
does not exploit the reconsolidation process behaviorally.

Trends in Cognitive Sciences, July 2017, Vol. 21, No. 7 537

Returning to the concept of potentiated extinction, although such an account of reconsolidation
updating is plausible based upon the behavioral data alone, it may be less consistent with the
outcomes of further exploration of the retrieval–extinction phenomenon. First, a parametric
study in rats showed that retrieval–extinction only occurs under retrieval conditions that
promote memory destabilization [68]. More convincingly, retrieval–extinction does not result
in an enhancement of the neural mechanisms of extinction but instead appears to diminish
extinction-related prefrontal cortical activity [69]. Therefore, the updating of a destabilized
memory trace is the most parsimonious interpretation of retrieval–extinction and other behav-
ioral interventions.

While non-reconsolidation accounts of amnesia are unable to refute the existence of recon-
solidation as a memory process, or to explain all of the hundreds of studies demonstrating
reconsolidation, it remains possible that some observations of amnesia are attributable to non-
reconsolidation mechanisms. Moreover, the theoretical basis of reactivation-dependent amne-
sia is unimportant from a clinical perspective. Whether or not the suppression of maladaptive
memory expression is a result of impaired or updated memory reconsolidation, unleashed
extinction or state-dependent inhibition of memory retrieval is immaterial provided that the
suppression is robust and long-lasting. Therefore, the central observations of memory recovery
(i.e., only short-lasting memory impairment) that triggered the alternative explanations remain
important and may be instead viewed constructively for future translation. In particular,
translational studies will need to test explicitly for resilience to recovery procedures using
pharmacological and behavioral treatments. This is because it may be difficult to determine a
priori whether any beneficial effects are truly a result of reconsolidation impairments/updates,
and hence are likely to be persistent, or whether the treatment instead reduces maladaptive
memory expression by non-reconsolidation means, with the possibility of recovery. At the
clinical translational level, follow-up tests for psychiatric symptoms already implicitly test for
resilience against everyday triggers of relapse. However, rarely do they explicitly quantify
exposure to relapse-provoking situations or directly compare the consequences of different
interventions over long time-periods. Moreover, there is a lack of emphasis on understanding
the causes of the individual differences in risk to relapse. Certainly, no studies show 100%
persistent remission, and those patients who either fail to respond to the treatment in the first
place, or show relapse at follow-up, may be indicative of a fundamental propensity to relapse.
Alternatively, they may reflect inter-individual differences in boundary conditions on reconso-
lidation (see below), thereby suggesting that there is likely to be persistent benefit in those
patients who do respond. Moreover, it is not out of the question that the same intervention
might induce memory impairment via distinct mechanisms, or a combination of different
mechanisms, depending on the particular experimental or clinical setting [70].

The Reliability of Reconsolidation Effects

The previous discussion concerns the interpretation of reactivation-dependent amnesia. How-
ever, a more fundamental challenge to the reconsolidation literature is emerging namely
whether the fundamental preclinical findings are sufficiently reliable and replicable to warrant
translational application. If memory reactivation combined with pharmacological or non-phar-
macological treatment does not reliably induce any memory impairment, then the rationale for
clinical exploitation may be flawed. As a recent example, the original demonstration that human
motor memory could be disrupted by combining memory reactivation with interference [13]
was not replicated in a series of direct and conceptual replication attempts within a single
extensive study [71]. While perhaps it remains unclear whether human motor memories
undergo reconsolidation, there are very many studies showing reconsolidation effects in other
settings [72], particularly for fear memories. Many of these studies may indeed be suboptimally
designed to allow observation of memory recovery ([73] and see above). Nevertheless, to
suggest that memory reconsolidation, as a wider phenomenon, is invalidated by a single failed

538 Trends in Cognitive Sciences, July 2017, Vol. 21, No. 7

replication (or even several) suffers from the same risk of overinterpretation as does any
assumption that retrieval-dependent memory deficits necessarily reflect reconsolidation impair-
ments (e.g., [74]).

Given that retrieval–extinction is the most mature of the reconsolidation-update literature, it is

possible to draw some inferences about the replicability of the phenomenon. Some conceptual
replication attempts of the original findings were seemingly unsuccessful [75–78]. The resultant
uncertainty in the literature might suggest that attempts to combine memory reactivation with
extinction training (and perhaps, by extension, other reconsolidation updating procedures) are
not likely to be viable clinically. However, as reviewed earlier, there is a growing literature
supporting the likely translational efficacy of reconsolidation updating. Within this framework,
the observations of failures to show memory impairments in individual studies should, we
argue, not lead to obsessive discussion as to whether reconsolidation updating exists, but
rather should be viewed constructively in terms of increasing our understanding of reconso-
lidation updating and its potential for translational exploitation [79]. The failure of a treatment,
pharmacological or behavioral, to impact upon a reactivated memory may result from a lack of
efficacy on one of two fronts. First, for reconsolidation impairments to be effective the memory
must be successfully destabilized. Second, the reconsolidation process must be disrupted or
hijacked effectively. Therefore, a negative finding may result either from a failure to destabilize
the memory or because the treatment does not reliably impair/modify its reconsolidation.

Given the link between reconsolidation and memory updating, only those reactivation sessions
that induce memory updating are likely to destabilize the existing memory, rendering it
vulnerable to pharmacological or behavioral intervention. Many studies have determined
parametric conditions under which memories do and do not destabilize (e.g.,
[25,45,48,80–82]). While most of these boundary conditions have been identified in rodent
studies using a variety of amnestic treatments, there is emerging understanding of the
boundary conditions on human fear-memory destabilization. First, it has been suggested that
memory retrieval in itself is insufficient to trigger fear-memory destabilization, as assessed by
the impact of propranolol [83]. It appears that there is indeed a requirement for human fear
memory updating, conceptualized as an error in prediction about the forthcoming experience
during memory reactivation, so as to destabilize the underlying memory [48]. Armed with this
understanding, clinical interventions can be designed such that pharmacological or behavioral
updating treatments are only implemented if the memory reactivation procedure has been
deemed to have successfully evoked the problematic memory and induced some level of
prediction error [84]. Such a prediction error might also be elicited by unexpected presentation
of the aversive outcome (as opposed to its unexpected omission following stimulus exposure),
as evidenced by the capacity of outcome presentation to destabilize human fear memories
within a retrieval–extinction setting [85]. Moreover, personality traits may provide a modulatory
impact upon the success of reconsolidation treatments. High trait anxiety attenuated the
beneficial reduction in fear following reactivation-related propranolol, and this is likely due to
the lack of efficacy of the reactivation procedure in destabilizing the fear memory [86]. These
insights into modulatory effects on memory destabilization have emerged from studies in which
there has been a successful demonstration of reconsolidation impairment. Replication failures
might provide further valuable information concerning the reliability of memory destabilization
procedures. The challenge, however, is to disambiguate whether replication failures result from
a lack of efficacy in memory destabilization or in impairing memory reconsolidation. Advances in
our understanding of the neural and cellular mechanisms of memory destabilization may allow
some disambiguation, given the nascent ability to enhance destabilization pharmacologically
and thereby facilitate reconsolidation impairments (Box 2).

Trends in Cognitive Sciences, July 2017, Vol. 21, No. 7 539

 Box 2. Neural Mechanisms of Memory Destabilization and Updating Outstanding Questions
While reactivation procedures can be designed to maximize the chances of inducing prediction errors, and thereby How can we assess whether memory
produce successful memory destabilization, there remains the challenge of achieving this effectively on an individual has been destabilized? Behavioral trig-
basis. This raises the potential utility of pharmacologically enhancing memory destabilization. While there is relatively little gers, especially those carrying predic-
understanding of the neurochemical mechanisms of memory destabilization, several processes have been identified tion error, may return memory to a
[101]. At the intracellular level, there appears to be a requirement for protein degradation at the proteasome [102,103], labile state, but to reliably achieve
protein phosphatase activity [104], CaMKII [105], and nitric oxide [106,107]. At the cell surface, there is functional destabilization, specific neural markers
involvement of cholinergic [108] and dopaminergic receptors [109], at least for non-fear memory destabilization. Of should be demarcated.
particular relevance to fear memories is the necessity for cannabinoid CB1 receptor and calcium channel activity in the
dorsal hippocampus [110], and for NMDA receptor activity (specifically NR2B-containing NMDA receptors) in the
Is there an optimal process of recon-
basolateral amygdala [111–114]. We and others have exploited some of these mechanisms to stimulate fear-memory
solidation updating? Given individual
destabilization to render effective post-reactivation treatment even under conditions that do not normally trigger
variability in the key boundary param-
reconsolidation [115–117]. Use of pharmacological (partial) agonists to activate CB1 or NMDA receptors during
eters, the most effective updating
memory retrieval enabled post-retrieval drug treatment to impair the reconsolidation of fear memories. Therefore,
approach should probably be tailored
even within the realm of behavioral reconsolidation updating, there may be cause to incorporate pharmacological
to the individual.
treatment to maximize the likelihood of successfully destabilizing the maladaptive memory. Our assumption would be
that reconsolidation updating should share the same mechanisms of memory destabilization with pharmacological
reconsolidation impairments, and this has been shown to be true in the case of the requirement for AMPA receptor How can clinicians exploit reconsolida-
trafficking in the amygdala for fear-memory destabilization [118,119]. tion mechanisms? Novel treatment
protocols could be structured to
include behavioral triggers, behavioral
In the human brain, during post-retrieval extinction, the involvement of the ventromedial prefrontal cortex (vmPFC) and
and/or pharmacological destabiliza-
its functional connectivity with the amygdala diminish compared to standard extinction [69]. Following updating, when
tion and updating, followed by long-
reconsolidation is complete, subsequent encounters with the conditioned cue elicit lower physiological arousal, and
term recovery assays.
reduced amygdala activity and connectivity with vmPFC. Specifically, while amygdala threat-memory trace was
recovered in a group that underwent standard extinction, the group that underwent extinction during reconsolidation
showed no trace recovery in the amygdala even 18 months later [120,121]. During the phase immediately following the
reminder cue, and before updating, resting-state functional connectivity between amygdala and vmPFC is high,
compared to the no-reminder group, and the degree of connectivity predicts the ensuing update-induced reduction
in conditioned responses [122]. Together, these findings point to possible markers of reconsolidation updating in the
human brain: resting-state functional connectivity as a candidate for memory destabilization, dissociable neural patterns
indicating the update process (such as altered amygdala–vmPFC circuitry), and the resulting altered memory repre-
sentation (such as modified amygdala or hippocampal representation).

Assuming that a memory is successfully destabilized, a failure to observe a reconsolidation

effect may result from a lack of efficacy of the behavioral updating treatment. This may not be a
fundamental lack of efficacy but instead might reflect a ‘behavioral dosing' effect. In a similar
manner to the use of single standard drug doses, regardless of potential individual differences
that may impact upon drug response, inter-individual variability is highly likely to impact upon
the response to the invariant parameters of extinction (or other updating) training. That is, some
individuals (human and non-human) may require greater extinction training post-retrieval, or
various forms of it, to achieve reliable long-lasting reduction in memory expression. For
example, while previous studies failed to demonstrate retrieval–extinction effects on fear-
relevant stimuli (such as spider and snake images) using standard extinction [78,87], a recent
study using vicarious extinction (observing another person undergoing extinction) was more
efficacious [88]. Moreover, even at the individual level, the degree of behavioral updating
necessary to mitigate the memory will depend upon the strength of the memory in the first
place. An appreciation of the potential for such variability, and its consequences for clinical
impact, would guide the interrogation of current and future studies to identify causes and
mitigate/exploit them to improve the efficacy of the treatment. Perhaps indices of original
memory strength and the within-session success of update training would be predictive of
long-term efficacy.

Concluding Remarks
Over half a century ago the dominant memory paradigm has initially marginalized the recon-
solidation phenomenon, positing that consolidation takes place only once in the lifetime of a
memory during its initial formation. This paradigm shifted more than a decade ago with the
discovery of brain circuits mediating memory reconsolidation using well-defined behavioral

540 Trends in Cognitive Sciences, July 2017, Vol. 21, No. 7

procedures. Since then, reconsolidation has been described in fine detail spanning all levels of
analysis, from molecular, cellular, and physiological, to behavioral and large-scale neural
systems, over a spectrum of species, amnesic agents, and behavioral protocols. The fact
that reconsolidation interventions may lead to memory weakening, modification, or strength-
ening is consistent with an adaptive view of reconsolidation: a biological memory mechanism
enabling memory updating. Initial support for this idea came from a few key studies showing
that new learning during reconsolidation modifies the expression of motor, episodic, and
emotional memories, and that memory strengthening is mediated by reconsolidation mecha-
nisms. To date, the phenomenon of reconsolidation updating has greatly matured with
substantial evidence encompassing various memory systems (motor, episodic, emotional,
spatial), valences (negative, positive, and neural), species (mice, rats, and humans), develop-
mental stages (adolescence and adulthood), and clinical populations (addiction and anxiety).

Reconsolidation remains a topic of intense research not only for the basic understanding of
long-term memory but particularly for potential application to psychiatric conditions in which
persistent maladaptive memories are a core feature (see Outstanding Questions). The transla-
tion of reconsolidation theory to psychiatric practice utilizing pharmacological and behavioral
interventions is already underway, but significant fundamental issues remain to be addressed to
maximize (or at least fully evaluate) the therapeutic potential alternative explanations and
replication failure. Alternative explanations generally arise from datasets that appear, at first
sight, to be inconsistent with the assertion that reconsolidation impairments should be reacti-
vation-dependent and long-lasting. Usually, they center on observations that the memory can
be ‘recovered' via some manner of reminder even after reconsolidation impairment. Our
criticism of the alternative explanations is that they tend to overinterpret individual observations,
thereby extending an explanation of a single study to the entire literature. We argue that this
generalization of interpretation is both unnecessary and unwarranted. Instead, the central
observations that triggered the alternative explanation may be viewed constructively from a
clinical perspective. In particular, translational studies will need to test for resilience against
recovery more systematically because it may be difficult to determine a priori whether any
beneficial effects are truly a result of reconsolidation impairments, and hence are likely to be
persistent.

Replication failure encompasses observations that fail to replicate memory impairment as a

result of reactivation plus treatment. Several papers purport to fail to replicate or extend the
reconsolidation updating phenomenon. These replication failures could be held to provide
doubt concerning the phenomenon itself. However, we argue instead that the conflicting
literature provides a rich opportunity to understand more fully the ‘boundaries' and limitations
of reconsolidation updating. The sheer number of positive studies supports the existence of the
phenomenon, but the failures to replicate present a valuable insight into how reliably the
phenomenon might be exploited for translational benefit by highlighting the crucial factors
underlying reconsolidation updating. Thus, deeper analysis of replication failures, even at a
conceptual level, is essential. Especially because any optimization of translational application
will need to determine the source of failures. In particular, future research should focus on
whether the failures represent failure to destabilize the memory or a lack of behavioral updating
via the restabilization process. This may be possible by identifying specific markers of memory
destabilization and restabilization. Cellular and molecular indications could be utilized to
develop pharmacological agents that may aid memory destabilization and restabilization upon
need, and system-level markers in the human brain may signal effective behavioral
manipulations.

All in all, research on behavioral targeting of reconsolidation aligns with the futurist view that
non-invasive manipulations may one day make drug therapy obsolete. A second possibility is

Trends in Cognitive Sciences, July 2017, Vol. 21, No. 7 541

that there may be classes of psychopathologies that are responsive to non-pharmacological
and some to the pharmacological. The third possibility is some individuals may be more
responsive than others to the pharmacological versus non-pharmacological approaches for
some psychopathologies. Successful translation of the rich reconsolidation literature into
clinical applications may crucially depend on our ability to describe the processes of memory
destabilization and restabilization as separate and complementary targets for pharmacological
and behavioral interventions.

Acknowledgments
Funding was provided by Medical Research Council (MRC) grant MR/M017753/1 and Leverhulme Trust Research Grant
RPG-2015-006 to J.L.; Canadian Institutes of Health Research grants MOP-133444 and MOP-123430, and Natural
Science and Engineering Research Council grant RGPIN 249880-11 to K.N.; and National Institute of Mental Health
(NIMH) grant MH105515 and a Klingenstein-Simons Fellowship Award in the Neurosciences to D.S.

References
1. Duncan, C.P. (1949) The retroactive effect of electroshock on 21. Fernandez, R.S. et al. (2016) The fate of memory: reconsolida-
learning. J. Comp. Physiol. Psychol. 42, 32–44 tion and the case of Prediction Error. Neurosci. Biobehav. Rev.
2. Flexner, L.B. et al. (1965) Loss of memory as related to inhibition 68, 423–441
of cerebral protein synthesis. J. Neurochem. 12, 535–541 22. Lee, J.L.C. (2009) Reconsolidation: maintaining memory rele-
3. Gordon, W.C. and Spear, N.E. (1973) Effect of reactivation of a vance. Trends Neurosci. 32, 413–420
previously acquired memory on the interaction between mem- 23. Nader, K. and Hardt, O. (2009) A single standard for memory:
ories in the rat. J. Exp. Psychol. 99, 349–355 the case for reconsolidation. Nat. Rev. Neurosci. 10, 224–234
4. McGaugh, J.L. and Krivanek, J.A. (1970) Strychnine effects on 24. Dudai, Y. (2004) The neurobiology of consolidations, or, how
discrimination learning in mice: effects of dose and time of stable is the engram? Annu. Rev. Psychol. 55, 51–86
administration. Physiol. Behav. 5, 1437–1442 25. Suzuki, A. et al. (2004) Memory reconsolidation and extinction
5. McGaugh, J.L. (1966) Time-dependent processes in memory have distinct temporal and biochemical signatures. J. Neurosci.
storage. Science 153, 1351–1358 24, 4787–4795
6. Hebb, D.O. (1949) The Organisation of Behavior, Wiley 26. Nader, K. (2003) Memory traces unbound. Trends Neurosci. 26,
7. Misanin, J.R. et al. (1968) Retrograde amnesia produced by 65–72
electroconvulsive shock after reactivation of a consolidated 27. Lee, J.L.C. et al. (2006) Reconsolidation and extinction of con-
memory trace. Science 160, 554–555 ditioned fear: inhibition and potentiation. J. Neurosci. 26,
8. Schneider, A.M. and Sherman, W. (1968) Amnesia: a function of 10051–10056
the temporal relation of footshock to electroconvulsive shock. 28. Dudai, Y. (2006) Reconsolidation: the advantage of being refo-
Science 159, 219–222 cused. Curr. Opin. Neurobiol. 16, 174–178
9. Nader, K. et al. (2000) Fear memories require protein synthesis 29. Schiller, D. and Phelps, E.A. (2011) Does reconsolidation occur
in the amygdala for reconsolidation after retrieval. Nature 406, in humans? Front. Behav. Neurosci. 5, 24
722–726 30. Sara, S.J. (2000) Retrieval and reconsolidation: toward a neu-
10. Milekic, M.H. and Alberini, C.M. (2002) Temporally graded robiology of remembering. Learn. Mem. 7, 73–84
requirement for protein synthesis following memory reactivation. 31. Tronson, N.C. and Taylor, J.R. (2007) Molecular mechanisms of
Neuron 36, 521–525 memory reconsolidation. Nat. Rev. Neurosci. 8, 262–275
11. Litvin, O.O. and Anokhin, K.V. (2000) Mechanisms of memory 32. Scully, I.D. et al. (2016) Does reactivation trigger episodic mem-
reorganization during retrieval of acquired behavioral experience ory change? A meta-analysis. Neurobiol. Learn. Mem. Pub-
in chicks: the effects of protein synthesis inhibition in the brain. lished online December 23, 2016. http://dx.doi.org/10.1016/j.
Neurosci. Behav. Physiol. 30, 671–678 nlm.2016.12.012
12. Gordon, W.C. (1977) Similarities of recently acquired and rec- 33. Hupbach, A. et al. (2007) Reconsolidation of episodic memories:
tivated memories in interference. Am. J. Psychol. 90, 231–242 a subtle reminder triggers integration of new information.
13. Walker, M.P. et al. (2003) Dissociable stages of human memory Learn. Mem. 14, 47–53
consolidation and reconsolidation. Nature 425, 616–620 34. Monfils, M.H. et al. (2009) Extinction-reconsolidation bound-
14. Gordon, W.C. (1977) Susceptibility of a reactivated memory to aries: key to persistent attenuation of fear memories. Science
the effects of strychnine: a time-dependent phenomenon. 324, 951–955
Physiol. Behav. 18, 95–99 35. Schiller, D. et al. (2010) Preventing the return of fear in humans
15. Tronson, N.C. et al. (2006) Bidirectional behavioral plasticity of using reconsolidation update mechanisms. Nature 463, 49–53
memory reconsolidation depends on amygdalar protein kinase 36. Lee, J.L.C. (2010) Memory reconsolidation mediates the updat-
A. Nat. Neurosci. 9, 167–169 ing of hippocampal memory content. Front. Behav. Neurosci. 4,
16. Lewis, D.J. (1979) Psychobiology of active and inactive memory. 168
Psychol. Bull. 86, 1054–1083 37. Lee, J.L.C. (2008) Memory reconsolidation mediates the
17. Spear, N.E. (1984) Behaviours that indicate memory: levels of strengthening of memories by additional learning. Nat. Neurosci.
expression. Can. J. Psychol. 38, 348–367 11, 1264–1266
18. Miller, R.R. and Springer, A.D. (1973) Amnesia, consolidation, 38. Flavell, C.R. et al. (2011) Behavioural memory reconsolidation of
and retrieval. Psychol. Rev. 80, 69–79 food and fear memories. Nat. Commun. 2, 504
19. Agren, T. (2014) Human reconsolidation: a reactivation and 39. Xue, Y.X. et al. (2012) A memory retrieval–extinction procedure
update. Brain Res. Bull. 105, 70–82 to prevent drug craving and relapse. Science 336, 241–245
20. Exton-McGuinness, M.T. (2015) Updating memories the role 40. Das, R.K. et al. (2015) Rewriting the valuation and salience of
of prediction errors in memory reconsolidation. Behav. Brain alcohol-related stimuli via memory reconsolidation.
Res. 278, 375–384 Transl. Psychiatry 5, e645

542 Trends in Cognitive Sciences, July 2017, Vol. 21, No. 7

41. Goltseker, K. et al. (2017) Counterconditioning during reconso- 64. Ledgerwood, L. et al. (2003) Effects of D-cycloserine on extinc-
lidation prevents relapse of cocaine memories. tion of conditioned freezing. Behav. Neurosci. 117, 341–349
Neuropsychopharmacology 42, 716–726 65. Ledgerwood, L. et al. (2004) D-cycloserine and the facilitation of
42. Olshavsky, M.E. et al. (2013) Updating appetitive memory during extinction of conditioned fear: consequences for reinstatement.
reconsolidation window: critical role of cue-directed behavior Behav. Neurosci. 118, 505–513
and amygdala central nucleus. Front. Behav. Neurosci. 7, 186 66. Ledgerwood, L. et al. (2005) D-cycloserine facilitates extinction
43. James, E.L. et al. (2015) Computer game play reduces intrusive of learned fear: effects on reacquisition and generalized extinc-
memories of experimental trauma via reconsolidation-update tion. Biol. Psychiatry 57, 841–847
mechanisms. Psychol. Sci. 26, 1201–1215 67. Millan, E.Z. et al. (2013) Memory retrieval, extinction, and rein-
44. Germeroth, L.J. et al. (2017) Effect of a brief memory updating statement of alcohol seeking. Neurobiol. Learn. Mem. 101,
intervention on smoking behavior: a randomized clinical trial. 26–32
JAMA Psychiatry 74, 214–223 68. Pineyro, M.E. et al. (2013) Memory destabilization is critical for
45. Diaz-Mataix, L. et al. (2013) Detection of a temporal error trig- the success of the reactivation-extinction procedure. Learn.
gers reconsolidation of amygdala-dependent memories. Mem. 21, 46–54
Curr. Biol. 23, 467–472 69. Schiller, D. et al. (2013) Extinction during reconsolidation of
46. Pedreira, M.E. et al. (2004) Mismatch between what is expected threat memory diminishes prefrontal cortex involvement.
and what actually occurs triggers memory reconsolidation or Proc. Natl. Acad. Sci. U. S. A. 110, 20040–20045
extinction. Learn. Mem. 11, 579–585 70. Clem, R.L. and Schiller, D. (2016) New learning and unlearning:
47. Reichelt, A.C. et al. (2013) Ventral tegmental dopamine dysre- strangers or accomplices in threat memory attenuation? Trends
gulation prevents appetitive memory destabilisation. Neurosci. 39, 340–351
J. Neurosci. 33, 14205–14210 71. Hardwicke, T.E. et al. (2016) Postretrieval new learning does not
48. Sevenster, D. et al. (2013) Prediction error governs pharmaco- reliably induce human memory updating via reconsolidation.
logically induced amnesia for learned fear. Science 339, Proc. Natl. Acad. Sci. 113, 5206–5211
830–833 72. Walker, M.P. and Stickgold, R. (2016) Understanding the
49. Hon, T. et al. (2016) The effects of cognitive reappraisal following boundary conditions of memory reconsolidation.
retrieval-procedures designed to destabilize alcohol memories Proc. Natl. Acad. Sci. U. S. A. 113, E3991–E3992
in high-risk drinkers. Psychopharmacology 233, 851–861 73. Hardwicke, T.E. and Shanks, D.R. (2016) Reply to Walker and
50. Gray, R.M. and Bourke, F. (2015) Remediation of intrusive Stickgold: proposed boundary conditions on memory reconso-
symptoms of PTSD in fewer than five sessions: a 30-person lidation will require empirical verification.
pre-pilot study of the RTM Protocol Journal of Military. Vet. Proc. Natl. Acad. Sci. U. S. A. 113, E3993–E3994
Family Health 1, 13–20 74. Gershman, S.J. et al. (2013) Neural context reinstatement pre-
51. Gray, R.M. and Liotta, R.F. (2012) PTSD: extinction, reconso- dicts memory misattribution. J. Neurosci. 33, 8590–8595
lidation, and the visual-kinesthetic dissociation protocol. 75. Kredlow, M.A. et al. (2016) Harnessing reconsolidation to
Traumatology 18, 3–16 weaken fear and appetitive memories: A meta-analysis of
52. Shiban, Y. et al. (2015) Fear reactivation prior to exposure post-retrieval extinction effects. Psychol. Bull. 142, 314–336
therapy: does it facilitate the effects of VR exposure in a ran- 76. Ishii, D. et al. (2012) No erasure effect of retrieval–extinction trial
domized clinical sample? J. Behav. Ther. Exp. Psychiatry 46, on fear memory in the hippocampus-independent and depen-
133–140 dent paradigms. Neurosci. Lett. 523, 76–81
53. Telch, M.J. et al. (2017) Use of a brief fear memory reactivation 77. Hutton-Bedbrook, K. and McNally, G.P. (2013) The promises
procedure for enhancing exposure therapy. Clin. Psychol. Sci. 5, and pitfalls of retrieval-extinction procedures in preventing
367–378 relapse to drug seeking. Front. Psychiatry 4, 14
54. Bjorkstrand, J. et al. (2016) Disrupting reconsolidation attenu- 78. Kindt, M. and Soeter, M. (2013) Reconsolidation in a human fear
ates long-term fear memory in the human amygdala and facil- conditioning study: a test of extinction as updating mechanism.
itates approach behavior. Curr. Biol. 26, 2690–2695 Biol. Psychol. 92, 43–50
55. Bjorkstrand, J. et al. (2017) Think twice, it's all right: long lasting 79. Auber, A. et al. (2013) Post-retrieval extinction as reconsolida-
effects of disrupted reconsolidation on brain and behavior in tion interference: methodological issues or boundary condi-
human long-term fear. Behav. Brain. Res. 324, 125–129 tions? Psychopharmacology 226, 631–647
56. Weems, C.F. et al. (2014) Memories of traumatic events in 80. Merlo, E. et al. (2014) Reconsolidation and extinction are disso-
childhood fade after experiencing similar less stressful events: ciable and mutually exclusive processes: behavioral and molec-
results from two natural experiments. J. Exp. Psychol. Gen. 143, ular evidence. J. Neurosci. 34, 2422–2431
2046–2055
81. Flavell, C.R. and Lee, J.L.C. (2013) Reconsolidation and extinc-
57. Liu, J. et al. (2014) An unconditioned stimulus retrieval extinction tion of an appetitive pavlovian memory. Neurobiol. Learn. Mem.
procedure to prevent the return of fear memory. Biol. Psychiatry 104, 25–31
76, 895–901
82. Reichelt, A.C. and Lee, J.L. (2012) Appetitive Pavlovian goal-
58. Xue, Y.X. et al. (2017) Effect of selective inhibition of reactivated tracking memories reconsolidate only under specific conditions.
nicotine-associated memories with propranolol on nicotine crav- Learn. Mem. 20, 51–60
ing. JAMA Psychiatry 74, 224–232
83. Sevenster, D. et al. (2012) Retrieval per se is not sufficient to
59. Luo, Y.X. et al. (2015) A novel UCS memory retrieval–extinction trigger reconsolidation of human fear memory. Neurobiol. Learn.
procedure to inhibit relapse to drug seeking. Nat. Commun. 6, Mem. 97, 338–345
7675
84. Kindt, M. and van Emmerik, A. (2016) New avenues for treating
60. Dunbar, A.B. and Taylor, J.R. (2016) Reconsolidation and psy- emotional memory disorders: towards a reconsolidation inter-
chopathology: moving towards reconsolidation-based treat- vention for posttraumatic stress disorder.
ments. Neurobiol. Learn. Mem. Published online November 9, Ther. Adv. Psychopharmacol. 6, 283–295
2016. http://dx.doi.org/10.1016/j.nlm.2016.11.005
85. Thompson, A. and Lipp, O.V. (2017) Extinction during reconso-
61. Millin, P.M. et al. (2001) Interpretations of retrograde amnesia: lidation eliminates recovery of fear conditioned to fear-irrelevant
old problems redux. Nat. Rev. Neurosci. 2, 68–70 and fear-relevant stimuli. Behav. Res. Ther. 92, 1–10
62. Gisquet-Verrier, P. et al. (2015) Integration of new information 86. Soeter, M. and Kindt, M. (2013) High trait anxiety: a challenge for
with active memory accounts for retrograde amnesia: a chal- disrupting fear memory reconsolidation. PLoS ONE 8, e75239
lenge to the consolidation/reconsolidation hypothesis?
87. Golkar, A. et al. (2012) Are fear memories erasable? recon-
J. Neurosci. 35, 11623–11633
solidation of learned fear with fear-relevant and fear-irrelevant
63. Trent, S. et al. (2015) Rescue of long-term memory after recon- stimuli. Front. Behav. Neurosci. 6, 80
solidation blockade. Nat. Commun. 6, 7897

Trends in Cognitive Sciences, July 2017, Vol. 21, No. 7 543

88. Golkar, A. et al. (2017) Vicarious extinction learning during 113. Ferrer Monti, R.I. (2016) An appetitive experience after fear
reconsolidation neutralizes fear memory. Behav. Res. Ther. memory destabilization attenuates fear retention: involvement
92, 87–93 GluN2B-NMDA receptors in the basolateral amygdala complex.
89. Rudy, J.W. et al. (2006) Anisomycin and the reconsolidation Learn. Mem. 23, 465–478
hypothesis. Learn. Mem. 13, 1–3 114. Holehonnur, R. et al. (2016) Increasing the GluN2A/GluN2B ratio
90. Eisenhardt, D. and Menzel, R. (2007) Extinction learning, recon- in neurons of the mouse basal and lateral amygdala inhibits the
solidation and the internal reinforcement hypothesis. Neurobiol. modification of an existing fear memory trace. J. Neurosci. 36,
Learn. Mem. 87, 167–173 9490–9504

91. Myers, K.M. and Davis, M. (2002) Systems-level reconsolida- 115. Lee, J.L.C. and Flavell, C.R. (2014) Inhibition and enhancement
tion: reengagement of the hippocampus with memory reactiva- of contextual fear memory destabilization.
tion. Neuron 36, 340–343 Front. Behav. Neurosci. 8, 144

92. Lattal, K.M. and Abel, T. (2004) Behavioral impairments caused 116. Ortiz, V. et al. (2015) The effect of midazolam and propranolol on
by injections of the protein synthesis inhibitor anisomycin after fear memory reconsolidation in ethanol-withdrawn rats: influ-
contextual retrieval reverse with time. ence of D-cycloserine. Int. J. Neuropsychopharmacol. 18,
Proc. Natl. Acad. Sci. U. S. A. 101, 4667–4672 pyu082

93. Gold, P.E. and King, R.A. (1974) Retrograde amnesia: storage 117. Gazarini, L. et al. (2014) PTSD-like memory generated through
failure versus retrieval failure. Psychol. Rev. 81, 465–469 enhanced noradrenergic activity is mitigated by a dual step
pharmacological intervention targeting its reconsolidation.
94. Cherkin, A. (1972) Retrograde amnesia in the chick: resistance
Int. J. Neuropsychopharmacol. 18, pyu008
to the reminder effect. Physiol. Behav. 8, 949–955
118. Clem, R.L. and Huganir, R.L. (2010) Calcium-permeable AMPA
95. Gold, P.E. et al. (1973) Retrograde amnesia and the ‘reminder
receptor dynamics mediate fear memory erasure. Science 330,
effect’: an alternative interpretation. Science 180, 1199–1201
1108–1112
96. Amaral, O.B. et al. (2008) A synaptic reinforcement-based
119. Hong, I. et al. (2013) AMPA receptor exchange underlies tran-
model for transient amnesia following disruptions of memory
sient memory destabilization on retrieval.
consolidation and reconsolidation. Hippocampus 18, 584–601
Proc. Natl. Acad. Sci. U. S. A. 110, 8218–8223
97. Duvarci, S. and Nader, K. (2004) Characterization of fear mem-
120. Bjorkstrand, J. et al. (2015) Disruption of memory reconsolida-
ory reconsolidation. J. Neurosci. 24, 9269–9275
tion erases a fear memory trace in the human amygdala: an 18-
98. Schramm, M.J.W. et al. (2016) Bidirectional modulation of alco- month follow-up. PLoS ONE 10, e0129393
hol-associated memory reconsolidation through manipulation of
121. Agren, T. et al. (2012) Disruption of reconsolidation erases a fear
adrenergic signaling. Neuropsychopharmacology 41,
memory trace in the human amygdala. Science 337,
1103–1111
1550–1552
99. Eisenberg, M. et al. (2003) Stability of retrieved memory: Inverse
122. Feng, P. et al. (2016) Resting-state functional connectivity
correlation with trace dominance. Science 301, 1102–1104
between amygdala and the ventromedial prefrontal cortex fol-
100. Glickman, S.E. (1961) Perseverative neural processes and con- lowing fear reminder predicts fear extinction. Soc. Cogn. Affect.
solidation of the memory trace. Psychol. Bull. 58, 218–233 Neurosci. 11, 991–1001
101. Almeida-Correa, S. and Amaral, O.B. (2014) Memory labilization 123. Agren, T. et al. (2012) Human fear reconsolidation and allelic
in reconsolidation and extinction evidence for a common differences in serotonergic and dopaminergic genes.
plasticity system? J. Physiol. 108, 292–306 Transl. Psychiatry 2, e76
102. Lee, S.H. et al. (2008) Synaptic protein degradation underlies 124. Oyarzun, J.P. et al. (2012) Updating fearful memories with
destabilization of retrieved fear memory. Science 319, extinction training during reconsolidation: a human study using
1253–1256 auditory aversive stimuli. PLoS ONE 7, e38849
103. Lee, S.H. et al. (2012) A cellular model of memory reconsolida- 125. Steinfurth, E.C. (2014) Young and old Pavlovian fear memories
tion involves reactivation-induced destabilization and restabili- can be modified with extinction training during reconsolidation in
zation at the sensorimotor synapse in Aplysia. humans. Learn. Mem. 21, 338–341
Proc. Natl. Acad. Sci. U. S. A. 109, 14200–14205
126. Johnson, D.C. and Casey, B.J. (2015) Easy to remember, diffi-
104. Yu, Y.J. et al. (2016) Involvement of protein phosphatases in the cult to forget: the development of fear regulation. Dev. Cogn.
destabilization of methamphetamine-associated contextual Neurosci. 11, 42–55
memory. Learn. Mem. 23, 486–493
127. Asthana, M.K. et al. (2015) Preventing the return of fear using
105. Jarome, T.J. et al. (2016) CaMKII regulates proteasome phos- reconsolidation update mechanisms depends on the Met-allele
phorylation and activity and promotes memory destabilization of the brain derived neurotrophic factor Val66Met polymor-
following retrieval. Neurobiol. Learn. Mem. 128, 103–109 phism. Int. J. Neuropsychopharmacol. 19, pyv137
106. Balaban, P.M. et al. (2014) Nitric oxide is necessary for labiliza- 128. Rao-Ruiz, P. et al. (2011) Retrieval-specific endocytosis of
tion of a consolidated context memory during reconsolidation in GluA2-AMPARs underlies adaptive reconsolidation of contex-
terrestrial snails. Eur. J. Neurosci. 40, 2963–2970 tual fear. Nat. Neurosci. 14, 1302–1308
107. Bal, N.V. et al. (2017) Cued memory reconsolidation in rats 129. Ma, X. et al. (2012) Post-retrieval extinction training enhances or
requires nitric oxide. Eur. J. Neurosci. 45, 643–647 hinders the extinction of morphine-induced conditioned place
108. Stiver, M.L. et al. (2015) Cholinergic manipulations bidirectionally preference in rats dependent on the retrieval-extinction interval.
regulate object memory destabilization. Learn. Mem. 22, Psychopharmacology 221, 19–26
203–214 130. Sartor, G.C. and Aston-Jones, G. (2014) Post-retrieval extinc-
109. Merlo, E. et al. (2015) Amygdala dopamine receptors are tion attenuates cocaine memories. Neuropsychopharmacology
required for the destabilization of a reconsolidating appetitive 39, 1059–1065
memory. eNeuro 2, 0024-14 131. St Jacques, P.L. (2015) Modifying memory for a museum tour in
110. Suzuki, A. et al. (2008) Activation of LVGCCs and CB1 receptors older adults: Reactivation-related updating that enhances and
required for destabilization of reactivated contextual fear mem- distorts memory is reduced in ageing. Memory 23, 876–887
ories. Learn. Mem. 15, 426–433 132. Pine, A. et al. (2014) Unconscious learning of likes and dislikes is
111. Ben Mamou, C. et al. (2006) NMDA receptors are critical for persistent, resilient, and reconsolidates. Front. Psychol. 5, 1051
unleashing consolidated auditory fear memories. Nat. Neurosci. 133. Jones, B. et al. (2012) Remaking memories: reconsolidation
9, 1237–1239 updates positively motivated spatial memory in rats.
112. Milton, A.L. et al. (2013) Double dissociation of the requirement Learn. Mem. 19, 91–98
for GluN2B- and GluN2A-containing NMDA receptors in the 134. Agren, T. et al. (2017) Disruption of human fear reconsolidation
destabilization and restabilization of a reconsolidating memory. using imaginal and in vivo extinction. Behav. Brain Res. 319,
J. Neurosci. 33, 1109–1115 9–15

544 Trends in Cognitive Sciences, July 2017, Vol. 21, No. 7

135. Haubrich, J. et al. (2015) Reconsolidation allows fear memory to 138. Tedesco, V. et al. (2014) Extinction, applied after retrieval of
be updated to a less aversive level through the incorporation of auditory fear memory, selectively increases zinc-finger protein
appetitive information. Neuropsychopharmacology 40, 268 and phosphorylated ribosomal protein S6 expression in
315–326 prefrontal cortex and lateral amygdala. Neurobiol. Learn.
136. Graff, J. et al. (2014) Epigenetic priming of memory updating Mem. 115, 78–85
during reconsolidation to attenuate remote fear memories. Cell 139. Jones, C.E. and Monfils, M.H. (2016) Post-retrieval extinction in
156, 261–276 adolescence prevents return of juvenile fear. Learn. Mem. 23,
137. Escosteguy-Neto, J.C. (2016) Reconsolidation and update of 567–575
morphine-associated contextual memory in mice. Neurobiol.
Learn. Mem. 130, 194–201

Trends in Cognitive Sciences, July 2017, Vol. 21, No. 7 545