- Email: [email protected]
Emotional memory expression is misleading: delineating transitions between memory processes
Available online at www.sciencedirect.com
ScienceDirect Emotional memory expression is misleading: delineating transitions between memory processes Leonidas Faliagkas1, Priyanka Rao-Ruiz1 and Merel Kindt2 The hypothesis that fear memory is not necessarily permanent but can change when retrieved opens avenues to develop revolutionary treatments for emotional memory disorders. Memory reconsolidation is however only one of several mnemonic processes that may be triggered by memory reactivation and subtle environmental differences may cause a transition from a malleable to a stable state. This poses a major challenge to translating the reconsolidation intervention to clinical practice. Here we review recent advances in understanding the transitions between memory processes in animals and humans, and discuss how the cognitive expression (i.e. threat expectancies) of fear memory in humans may serve as read-out to delineate the underlying processes necessary for memory reconsolidation, independent from the emotional expression of fear memory. Addresses 1 Department of Molecular and Cellular Neurobiology, Center for Neurogenomics and Cognitive Research, Amsterdam Neuroscience, VU University Amsterdam, Amsterdam, The Netherlands 2 Department of Clinical Psychology, University of Amsterdam, Amsterdam, The Netherlands Corresponding author: Faliagkas, Leonidas ([email protected]), Kindt, Merel ([email protected])
Current Opinion in Behavioral Sciences 2018, 19:116–122 This review comes from a themed issue on Emotion–cognition interactions Edited by Mara Mather and Michael Fanselow
https://doi.org/10.1016/j.cobeha.2017.12.018 2352-1546/ã 2017 Published by Elsevier Ltd.
Introduction For more than a century a widely accepted view in cognitive science was that memories are only initially labile and sensitive to disruption, after which they become imprinted or consolidated into the physical architecture of the brain. At the turn of this century, a major breakthrough in neuroscience was achieved with the (re) discovery that fear memory is not inevitably permanent, but can change when retrieved [1,2]. Nader and colleagues [1] demonstrated in rats that the administration of a protein synthesis blocker (i.e. anisomycin) upon memory Current Opinion in Behavioral Sciences 2018, 19:116–122
reactivation eliminated the expression (i.e. freezing) of a previously formed fear memory. Further observations of post-reactivation amnesia have drastically changed the neuroscience literature on memory by generating a novel and influential conceptual framework — usually referred to as the memory reconsolidation hypothesis. The reconsolidation hypothesis states that memory is an intrinsically dynamic process allowing modification of an established memory trace, should conditions require such adaptation [3]. Upon recall, the memory trace may transfer to a transient destabilized state, requiring time-dependent restabilization to persist further (Box 1). Gene transcription and protein synthesis are necessary for reconsolidation and offer a window of opportunity to fundamentally change the memory trace [1,4]. The fear-conditioning paradigm is widely used across different species to study the process of memory reconsolidation [5,6]. In the lab, fear memories can be established through Pavlovian fear conditioning, which involves the repeated pairing of an initially neutral cue (e.g. a tone; conditioned stimulus, CS) with an inherently aversive stimulus (e.g. an electrocutaneous shock; unconditioned stimulus, US). As a result, the representation of the CS and US will become connected in the brain, such that a later presentation of the CS will retrieve the US representation and elicit a conditioned fear response. This so-called associative fear memory represents both cognitive and emotional aspects of fear learning: it involves contingency learning between the originally neutral or ambiguous stimulus (CS) and the reinforcer (US), while the CS becomes also imbued with the affective properties of the reinforcers (US) they predict. Upon a reminder trial of the CS, a conditioned fear response (CR) (i.e. freezing) is usually taken as evidence that a CS-US association has been formed, but this behavioral read-out does not distinguish between the cognitive and emotional aspects of fear memory in animals. Yet, this distinction can easily be made in human fear-conditioning studies where the cognitive expression of fear memory is typically assessed by threat expectancies and the emotional expression by the fear-potentiated startle reflex (e.g. [7,8]). In order for reconsolidation to occur, memory has to be reactivated (Box 1) and destabilized during a generally brief reminder session. Once the memory is rendered labile, reconsolidation can take place and may be experimentally modified by pharmacological [1,9,10] or behavioral manipulations [11–13]. In particular, it has been shown that fear memories can be enhanced or weakened, www.sciencedirect.com
Delineating transitions between memory processes Faliagkas, Rao-Ruiz and Kindt 117
Box 1 Definitions of core processes and concepts of fear memory.
Memory reactivation: The return of memory from a neurobiologically inactive to an active state in the presence of salient reminder cues. Reactivation of memory may take place without any behavioral output. Memory retrieval: The complex process by which previously stored information is recalled, which leads to behavioral output. Match/mismatch: An estimation of how consistent the environment is with previous experience. Prediction error: For fear conditioning, the magnitude of the outcome or the outcome itself is not being fully predicted. Memory destabilization: The transient labile state that a memory may return to upon its reactivation. Memory reconsolidation: The proces of restabilization of memory, once it has been destabilized after memory reactivation. Current Opinion in Behavioral Sciences
depending on the manipulation used [6,14,15]. Since maladaptive memory processing lies at the core of emotional memory disorders (i.e. anxiety disorders, post-traumatic stress disorder, addiction) [8], targeting the process of memory reconsolidation opens avenues to develop a revolutionary treatment. Even though the findings obtained from basic neuroscience are promising, the critical conditions to target complex and pervasive emotional memories typically encountered in clinical practice are still unknown. Very often basic findings from animal literature are translated to clinical trials without a full understanding of the mechanisms of change, leading to disappointing and confusing results. A thorough understanding of the necessary and boundary conditions to trigger and observe memory reconsolidation is essential before we can witness a paradigm shift in clinical practice [6,8,10,16,17]. The reactivation of fear memory is often considered to be synonymous with its behavioral expression, but fear memory may also be reactivated without an observable mnemonic read-out. Hence, a reminder stimulus may return memory to a malleable state necessitating reconsolidation, even when it is not behaviourally expressed [18,19,20,21]. Furthermore, there is substantial evidence demonstrating that a memory is not labilized every time it is retrieved [7,20,22]. If retrieval (Box 1) is not sufficient to trigger memory reconsolidation, then the critical question arises: How do we know whether a particular fear memory actually requires an adaptation? Memory reactivation is thought to trigger memory reconsolidation when the reminder stimuli are similar but not identical (match/mismatch, Box 1) to the original learning environment [5,23]. Yet, a reminder session that is too different www.sciencedirect.com
from the original learning procedure might not trigger memory reconsolidation, but instead initiate other memory processes such as an intermediate pharmacologically insensitive state of limbo [24,25,26] or the formation of a new memory, such as in extinction learning [27,28]. Without an independent index of the necessary and sufficient conditions to trigger memory reconsolidation, other than the memory enhancing or amnesic effects of the manipulations themselves, determining the degree of similarity (or dissimilarity) between the original learning and reminder session presents a challenge to empirical falsifiability. In this review we will discuss different post-reactivation memory processes and illustrate how the transition between these processes depend on subtle changes in the reactivation procedures in interaction with the learning history. Given that there is no single method of memory reactivation that always triggers memory reconsolidation, we claim that an independent read-out is imperative to delineate the underlying processes necessary for memory reconsolidation. In particular, we will briefly review a selection of animal and human studies and describe how memory expression is neither necessary nor sufficient to trigger reconsolidation, but that assessment of a match/mismatch experience (Box 1) may indicate whether a certain reminder trial triggers reconsolidation [6,16,29]. In fear-conditioning studies the match/ mismatch between the expected and actual outcome during memory reactivation is operationalized as prediction error, meaning that the magnitude of the aversive outcome or the outcome itself (i.e. US) is not fully predicted (PE, Box 1) [6,16,29]. Furthermore, we will argue that PE is also not a sufficient condition for Current Opinion in Behavioral Sciences 2018, 19:116–122
118 Emotion–cognition interactions
reconsolidation, but under specific circumstances, it may in fact be a boundary condition. In animal studies the occurrence of PE can be defined by the experimental parameters of the reactivation trial. Even though molecular indicators of memory destabilization have also been identified, the assessment of these markers requires either the firstly, post hoc analysis of brain tissue immediately after the reactivation session, or secondly, pharmacological manipulation of specific proteins/pathways involved in the destabilization of memory [30–34]. While the former method of analysis cannot be used for longitudinal measurements of memory in the same animals, pharmacological interventions may be used to successfully block/enhance [35] the destabilization of memory. The molecular markers of memory destabilization may indicate the effect of PE on downstream memory process, but they do not evaluate the degree of PE itself. In humans, however, the cognitive expression (i.e. threat expectancies) of fear memory may serve as read-out to indicate the degree of PE itself. In a series of studies on fear memory reconsolidation in which the emotional (i.e. fear potentiated startle reflex) and cognitive (i.e. threat expectancies) expression of fear memory were measured in parallel, we have demonstrated a double dissociation between these different response systems [7,24,36]. By capitalizing upon this double dissociation, we revealed that the cognitive expression of fear memory might serve as read-out to demarcate the underlying processes necessary for memory reconsolidation, independent from the emotional expression of fear memory.
Memory expression is neither sufficient, nor necessary to trigger memory reconsolidation In animal models, the reactivation of fear memories is typically assessed by behavioral output to salient cues, while there is convincing evidence that memory can be reactivated without behavioral responding [18,19,20,21]. However, when a reminder stimulus effectively elicits behavioral memory expression (i.e. retrieval, Box 1), this is not sufficient to destabilize memory and trigger subsequent reconsolidation [7,36,37,38]. In humans, we [7] demonstrated that the mere reactivation of a previously formed fear memory did not lead to reconsolidation. We addressed this issue by eliminating the very possibility of reinforcement of the reminder cue. A previously established associative fear memory (pictureshock pairings) was reactivated with shock-electrodes either attached or unattached. Reconsolidation was not triggered when nothing could be learned during the reminder trial (i.e. shocks unattached). Only when the outcome of the reactivation was not fully predictable (i.e. uncertainty about receiving a shock), memory reconsolidation was engaged. We concluded that memory reconsolidation does not take place when there is nothing to be learned during memory reactivation that would be relevant to the previously formed memory. It is also striking that memory reactivation, either with or without Current Opinion in Behavioral Sciences 2018, 19:116–122
anticipation of the threat (US), did not affect the emotional expression of the fear memory (startle response). Memory can also be reactivated and destabilized at a neurobiological level without being expressed at all [18,19,20,21]. In an elegant set of experiments, utilizing a context-signal memory model in crabs, Caffaro and colleagues [18] found that scopolamine-induced amnesia resulted in severe memory expression deficits. Interestingly, even under these amnestic conditions this nonexpressible memory could be reactivated, and the deficits were subsequently reversed by reconsolidation-specific memory enhancement. Thus, under defined circumstances memory can be destabilized at a neurobiological level and trigger reconsolidation without being behaviourally expressed. This dissociation has also been observed in other memory paradigms [20,39], and further evidenced by studies showing differences in glutamate receptor activity for memory expression and reconsolidation in the amygdala, with AMPAR antagonists specifically affecting the former and NMDAR antagonists the latter process [21,40] (but see [19]). Taken together, these data indicate that the expression of fear memory is not synonymous with its reactivation and destabilization. While reactivation is needed for reconsolidation, behavioral output is not sufficient or necessary to drive this adaptive process (but see [41]). This inevitably leads to the subsequent question: if memory expression during reactivation is not a necessary or sufficient condition for destabilization and reconsolidation, what triggers these processes? The studies discussed above [18,20,39] induced a discrepancy between the training (reinforced CS) and the reminder session (unreinforced CS presentation) to trigger reconsolidation. Once reactivated, memory might follow several paths based on the detection of a match/mismatch experience.
Prediction error is necessary, but not sufficient to trigger memory reconsolidation Drawing upon the idea that memory reconsolidation serves as a means of incorporating new information into memories to adapt to ever changing environments [2,3], it may be postulated that the reminder session should not simply convey new information, but it should also be strongly related to past experience. For associative learning paradigms, a discrepancy between expectation and actual outcome (match/mismatch) during reactivation can signal a PE, which in turn may serve to destabilize memory necessitating restabilization in order to persist further [29,42,43]. Pedreira and colleagues [37] provide one of the first reports of PE as a necessary condition for fear memory destabilization and subsequent reconsolidation. Since then, PE has been used in several experimental settings, across different species, to trigger destabilization and reconsolidation [7,18,19,24,29,36,38,44– 46,47,48]. For fear-conditioning paradigms, PE signals www.sciencedirect.com
Delineating transitions between memory processes Faliagkas, Rao-Ruiz and Kindt 119
may be experimentally generated during a reminder session by a non-reinforced CS exposure (no US), a reinforced CS exposure (positive PE) [36], uncertainty [49] or a shift in CS-US temporal association [45]. This gives rise to unmatched expectation that triggers memory destabilization and subsequent reconsolidation. Destabilization can be measured most reliably at a molecular level [30–34], while reconsolidation can be quantified at a molecular, cellular and behavioral level (blocking/ enhancing reconsolidation) [29,36,45]. In animal fearconditioning studies PE itself cannot be reliably measured and can only be pre-defined by the specific experimental parameters introduced during the memory reactivation trial. The analysis of molecular markers of PEdriven memory destabilization [25,32,33] or modified memory expression [25,29,36,37] at a later retention test may be used to indirectly test if a match/mismatch occurred, but these indicators require post hoc analysis and cannot be used to quantify the extent of mismatch generated during the reminder trial. Translational approaches from animal to humans offer a unique advantage in observing and quantifying the match/mismatch experience by threat expectancy ratings. Whereas freezing and startle potentiation basically reflect an expression of fear memory, threat expectancy is not a mere reflection of what has previously been learned: It also incorporates information from the current situation. As such, the threat expectancy ratings can be used to estimate how consistent the environment is with previous experience. The potential of expectancy ratings in unveiling the necessary conditions for memory reconsolidation is demonstrated in a series of human fear-conditioning studies [7,36,49]. An amnesic agent (b-adrenergic receptor blocker, propranolol) presented before or after memory reactivation specifically affected the emotional expression of fear memory (i.e. fearpotentiated startle response), while the cognitive level (i.e. threat expectancies) remained unaffected. This striking dissociation between the emotional and cognitive expression of fear memory in humans created an opportunity to develop a novel measure of PE-driven memory reconsolidation that does not rely on manipulation induced changes in the conditioned fear response. In particular, we demonstrated in a fear-conditioning study that reconsolidation was induced by reactivation conditions that led to a subsequent change in ratings of threat expectancy upon CS presentation (i.e. match/mismatch). Reconsolidation occurred in both unreinforced and reinforced reminder trials where the aversive stimulus (US) was not fully predicted by the CS, even though no reinforcement resulted in a decrease in threat expectancy from learning to trial while reinforcement resulted in an increase [36]. In contrast, reminder trials that left reported US expectancies unaffected did not induce fear memory reconsolidation. Notwithstanding these new insights on PE as a necessary condition to trigger memory reconsolidation, a critical www.sciencedirect.com
question is whether PE is a sufficient condition for reconsolidation, since PE can also give rise to new learning (e.g. extinction learning) and we know that extinction training puts a constraint on reconsolidation [27,50]. Pedreira et al. [51] (also, see [27]) demonstrated that the duration of reactivation may act as a switch to trigger either reconsolidation of the original fear memory or the formation of an extinction memory. A short unreinforced reminder session has been repeatedly found to induce reconsolidation, whereas longer exposure leads to the formation of an extinction memory that will be subsequently expressed [22,51–53]. Although reconsolidation and extinction are both vulnerable to amnestic agents, distinctive features of the reminder session in interaction with the original learning determines [24,36] which memory trace is predominant and subsequently more sensitive to disruption [51]. The studies of Sevenster et al. [24] and Merlo et al. [25] investigated the transition from reconsolidation to extinction by utilizing multiple match/mismatches (CS-no US) during the reminder session. Repeated or prolonged unreinforced exposure generated multiple PEs, which eventually reduced both threat expectancy and fear responding resulting in extinction learning at the behavioral [24,25] and molecular level [25]. Intriguingly, a CS-exposure of intermediate duration has been observed to lead to an intermediate phase (limbo state), insensitive to amnestic agents [24,25], presumably because neither of the two opposite processes of memory reconsolidation or extinction is engaged. In addition to factors such as the age and strength of memory [22,52], Alfei and colleagues [47] found that the specific temporal conditions under which initial learning took place are also critical in determining the duration of CS exposure needed to trigger one of four mnemonic processes: retrieval, reconsolidation, an intermediate state of limbo, or extinction. The existence of a pharmacologically-insensitive limbo state, as well as the indefinite transitional boundaries between these distinctive memory processes is a challenge to unequivocally trigger reconsolidation in both pre-clinical and clinical settings. In sum, behavioral studies in animals and humans have shown that the degree of similarity between encoding and reactivation demarcates the transition from retrieval, to reconsolidation, to limbo, to extinction. If reminder conditions fully match with encoding conditions, the memory is just retrieved, without further updating the memory. If conditions are quite similar, yet hold an element of novelty or surprise, memory destabilization and subsequent reconsolidation is triggered and the memory will be updated. If, however, conditions are too dissimilar, the reminder episode will transform into either a limbo state, or new learning by creating a new memory that coexists with the old, unaltered memory. The assumption made here is that minor changes in threat expectancy ratings in humans may serve to delineate the transitions between these memory processes (Figure 1). Current Opinion in Behavioral Sciences 2018, 19:116–122
120 Emotion–cognition interactions
Figure 1
humans/animals
humans/animals
humans
humans/animals
post reactivation memory processes
match/mismatch
change in expectancy during reactivation
change in fear expression during reactivation
Retrieval
no
no
no
Reconsolidation
yes
no
no
Limbo state
yes
yes
no
Extinction
yes
yes
yes Current Opinion in Behavioral Sciences
Overall boundary conditions that affect the post reactivation fate of memory: Behavioral expression of fear during memory reactivation is not necessary or sufficient to trigger post-reactivation mnemonic processes. However, a match/mismatch based on prior learning, introduced during a reactivation trial, results in prediction error that is necessary for destabilization of memory. Depending on the number of errors generated during memory reactivation, as well as the nature of the original learning trial, memories may enter one of several states, namely; sole retrieval, reconsolidation, an intermediate limbo state or extinction. Delineating the transition between these different states may be possible in humans using a read-out of threat expectancy, but is more difficult in animals where the only quantifiable read-out is the expression of the conditioned fear response, which only changes when extinction learning takes place. The quantifiable changes mentioned in the figure may be bi-directional in the case of reconsolidation and limbo, but always result in a decrease for extinction.
Conclusion and future directions Translational research across species is critical in fostering our understanding of the necessary and sufficient conditions to trigger memory reconsolidation in clinical practice. A selection of research demonstrates that behavioral output during a reminder trial is not necessary or sufficient to induce memory reconsolidation. A match/mismatch experience, generally referred to as PE, is emerging as a necessary condition to return memory to a labile state. However, multiple PE’s during memory reactivation can lead to other mnemonic states like limbo or extinction, which are clear boundary conditions on the induction of memory reconsolidation. The subtle and undefined transitions between the distinctive memory processes pose a huge challenge to research on memory reconsolidation. Although fear-conditioning research in humans offer a unique advantage in quantifying the match/mismatch experience by threat expectancy ratings, animal research does not differentiate between the affective and cognitive properties of fear memory. So far, the occurrence of PE in animals is typically defined by specific experimental parameters of the memory reactivation trial and inferred from the effect that a match/mismatch experience has on downstream memory processes like destabilization and reconsolidation. However, this form of analysis does not provide an assessment of the degree of match/mismatch Current Opinion in Behavioral Sciences 2018, 19:116–122
generated during reactivation, which can profoundly affect the subsequent fate of memory. To solve this problem of empirical falsifiability in animal research, future studies on PE might dissociate the cognitive and emotional representations of fear memory by combining a declarative memory task with a fear-conditioning procedure. Alternatively, dissecting the intricacies of a match/mismatch experience at a neurobiological level could also advance our understanding of the dynamic balance between the stability and malleability of fear memory. Identifying brain regions that signal match/mismatch experiences in combination with in vivo techniques like optogenetics would enable us to directly manipulate PE signals with an unprecedented level of precision.
Conflict of interest statement Nothing declared.
References and recommended reading Papers of particular interest, published within the period of review, have been highlighted as: of special interest of outstanding interest 1.
Nader K, Schafe GE, Le Doux JE: Fear memories require protein synthesis in the amygdala for reconsolidation after retrieval. Nature 2000, 406:722-726. www.sciencedirect.com
Delineating transitions between memory processes Faliagkas, Rao-Ruiz and Kindt 121
A seminal fear-conditioning study in rodents launching the reconsolidation hypothesis by showing a post-retrieval protein-synthesis dependent restabilization of memory. 2.
Sara SJ: Retrieval and reconsolidation: toward a neurobiology of remembering. Learn Mem 2000, 7:73-84.
3.
Dudai Y: Predicting not to predict too much: how the cellular machinery of memory anticipates the uncertain future. Philos Trans R Soc B 2009, 364:1255-1262.
4.
Mamiya N, Fukushima H, Suzuki A, Matsuyama Z, Homma S, Frankland PW, Kida S: Brain region-specific gene expression activation required for reconsolidation and extinction of contextual fear memory. J Neurosci 2009, 29:402.
5.
Finnie PSB, Nader K: The role of metaplasticity mechanisms in regulating memory destabilization and reconsolidation. Neurosci Biobehav Rev 2012, 36:1667-1707.
6.
Elsey JWB, Kindt M: Tackling maladaptive memories through reconsolidation: from neural to clinical science. Neurobiol Learn Mem 2017, 142(Part A):108-117.
7.
Sevenster D, Beckers T, Kindt M: Instructed extinction differentially affects the emotional and cognitive expression of associative fear memory. Psychophysiology 2012, 49:14261435.
8.
Kindt M: A behavioural neuroscience perspective on the aetiology and treatment of anxiety disorders. Behav Res Ther 2014, 62:24-36.
9.
De?biec J, Ledoux JE: Disruption of reconsolidation but not consolidation of auditory fear conditioning by noradrenergic blockade in the amygdala. Neuroscience 2004, 129:267-272.
10. Kindt M, Soeter M, Vervliet B: Beyond extinction: erasing human fear responses and preventing the return of fear. Nat Neurosci 2009, 12:256-258. 11. Schiller D, Monfils M-H, Raio CM, Johnson DC, LeDoux JE, Phelps EA: Preventing the return of fear in humans using reconsolidation update mechanisms. Nature 2010, 463:49-53. 12. Rao-Ruiz P, Rotaru DC, van der Loo RJ, Mansvelder HD, Stiedl O, Smit AB, Spijker S: Retrieval-specific endocytosis of GluA2AMPARs underlies adaptive reconsolidation of contextual fear. Nat Neurosci 2011, 14:1302-1308. 13. Monfils M-H, Cowansage KK, Klann E, LeDoux JE: Extinctionreconsolidation boundaries: key to persistent attenuation of fear memories. Science 2009, 324:951-955. 14. Tronson NC, Wiseman SL, Olausson P, Taylor JR: Bidirectional behavioral plasticity of memory reconsolidation depends on amygdalar protein kinase A. Nat Neurosci 2006, 9:167-169. 15. Lee JLC: Memory reconsolidation mediates the strengthening of memories by additional learning. Nat Neurosci 2008, 11:1264-1266. 16. Elsey JWB, Kindt M: Breaking boundaries: optimizing reconsolidation-based interventions for strong and old memories. Learn Mem 2017, 24:472-479. 17. Kindt M: The surprising subtleties of changing fear memory: a challenge for translational science. Philos Trans R Soc B 2017 http://dx.doi.org/10.1098/rstb.2017.0033. (in press). 18. Caffaro PA, Sua´rez LD, Blake MG, Delorenzi A: Dissociation between memory reactivation and its behavioral expression: scopolamine interferes with memory expression without disrupting long-term storage. Neurobiol Learn Mem 2012, 98:235-245. A context-signal memory study in crabs showing that non-expressible memory could still be reactivated without a behavioural output, and subsequently manipulated by reconsolidation-specific memory enhancement. 19. Barreiro KA, Sua´rez LD, Lynch VM, Molina VA, Delorenzi A: Memory expression is independent of memory labilization/ reconsolidation. Neurobiol Learn Mem 2013, 106:283-291. 20. Balderas I, Rodriguez-Ortiz CJ, Bermudez-Rattoni F: Retrieval and reconsolidation of object recognition memory are www.sciencedirect.com
independent processes in the perirhinal cortex. Neuroscience 2013, 253:398-405. 21. Mamou CB, Gamache K, Nader K: NMDA receptors are critical for unleashing consolidated auditory fear memories. Nat Neurosci 2006, 9:1237-1239. A rodent fear-conditioning study showing the necessity of NMDA receptors within the amygdala for memory labilization, independent from the expression of fear memory. 22. Suzuki A, Josselyn SA, Frankland PW, Masushige S, Silva AJ, Kida S: Memory reconsolidation and extinction have distinct temporal and biochemical signatures. J Neurosci 2004, 24:4787-4795. 23. Osan R, Tort ABL, Amaral OB: A mismatch-based model for memory reconsolidation and extinction in attractor networks. PLOS ONE 2011, 6:e23113. 24. Sevenster D, Beckers T, Kindt M: Prediction error demarcates the transition from retrieval, to reconsolidation, to new learning. Learn Mem 2014, 21:580-584. A human fear-conditioning study showing that prediction error (PE) — measured by threat expectancies — is not sufficient to trigger reconsolidation, and that multiple PE’s triggered an intermediate phase (limbo state) between reconsolidation and extinction, which is insensitive to amnestic agents. 25. Merlo E, Milton AL, Gooze´e ZY, Theobald DE, Everitt BJ: Reconsolidation and extinction are dissociable and mutually exclusive processes: behavioral and molecular evidence. J Neurosci 2014, 34:2422-2431. 26. Cassini LF, Flavell CR, Amaral OB, Lee JLC: On the transition from reconsolidation to extinction of contextual fear memories. Learn Mem 2017, 24:392-399. 27. Eisenberg M, Kobilo T, Berman DE, Dudai Y: Stability of retrieved memory: inverse correlation with trace dominance. Science 2003, 301:1102-1104. 28. Bos MGN, Beckers T, Kindt M: The effects of noradrenergic blockade on extinction in humans. Biol Psychol 2012, 89:598605. 29. Agustina Lo´pez M, Jimena Santos M, Cortasa S, Ferna´ndez RS, Carbo´ Tano M, Pedreira ME: Different dimensions of the prediction error as a decisive factor for the triggering of the reconsolidation process. Neurobiol Learn Mem 2016, 136:210219. 30. Almeida-Correˆa S, Amaral OB: Memory labilization in reconsolidation and extinction — evidence for a common plasticity system? Neurobiol Learn Mem 2014, 108:292-306. 31. Balaban PM, Roshchin M, Timoshenko AK, Gainutdinov KL, Bogodvid TK, Muranova LN, Zuzina AB, Korshunova TA: Nitric oxide is necessary for labilization of a consolidated context memory during reconsolidation in terrestrial snails. Eur J Neurosci 2014, 40:2963-2970. 32. Jarome TJ, Ferrara NC, Kwapis JL, Helmstetter FJ: CaMKII regulates proteasome phosphorylation and activity and promotes memory destabilization following retrieval. Neurobiol Learn Mem 2016, 128:103-109. 33. Lee S-H, Choi J-H, Lee N, Lee H-R, Kim J-I, Yu N-K, Choi S-L, Lee S-H, Kim H, Kaang B-K: Synaptic protein degradation underlies destabilization of retrieved fear memory. Science 2008, 319:1253-1256. 34. Lee S-H, Kwak C, Shim J, Kim J-E, Choi S-L, Kim HF, Jang D-J, Lee J-A, Lee K, Lee C-H et al.: A cellular model of memory reconsolidation involves reactivation-induced destabilization and restabilization at the sensorimotor synapse in Aplysia. Proc Natl Acad Sci U S A 2012, 109:14200-14205. 35. Lee JLC, Flavell CR: Inhibition and enhancement of contextual fear memory destabilization. Front Behav Neurosci 2014, 8:144. 36. Sevenster D, Beckers T, Kindt M: Prediction error governs pharmacollogically induced amnesia for learned fear. Science 2013, 339:830-833. A human fear-conditioning study uncovering an independent index for prediction error (PE) by utilzing threat expectancy rating, and Current Opinion in Behavioral Sciences 2018, 19:116–122
122 Emotion–cognition interactions
demonstrating that PE is necessary for memory reconsolidation, determined by the interaction between original learning and retrieval. 37. Pedreira ME, Pe´rez-Cuesta LM, Maldonado H: Mismatch between what is expected and what actually occurs triggers memory reconsolidation or extinction. Learn Mem 2004, 11:579-585. A context-signal memory study in crabs demonstrating the necessary role of prediction error (PE) for reconsolidation and extinction of aversive memory. 38. Frenkel L, Maldonado H, Delorenzi A: Memory strengthening by a real-life episode during reconsolidation: an outcome of water deprivation via brain angiotensin II. Eur J Neurosci 2005, 22:1757-1766. 39. Santoyo-Zedillo M, Rodriguez-Ortiz CJ, Chavez-Marchetta G, Bermudez-Rattoni F, Balderas I: Retrieval is not necessary to trigger reconsolidation of object recognition memory in the perirhinal cortex. Learn Mem 2014, 21:452-456. 40. Milton AL, Merlo E, Ratano P, Gregory BL, Dumbreck JK, Everitt BJ: Double dissociation of the requirement for GluN2Band GluN2A-containing NMDA receptors in the destabilization and restabilization of a reconsolidating memory. J Neurosci 2013, 33:1109-1115. 41. Rossato JI, Ko¨hler CA, Radiske A, Bevilaqua LRM, Cammarota M: Inactivation of the dorsal hippocampus or the medial prefrontal cortex impairs retrieval but has differential effect on spatial memory reconsolidation. Neurobiol Learn Mem 2015, 125:146-151. 42. Holland PC, Schiffino FL: Mini-review: prediction errors, attention and associative learning. Neurobiol Learn Mem 2016, 131:1-9. 43. Exton-McGuinness MTJ, Lee JLC, Reichelt AC: Updating memories — the role of prediction errors in memory reconsolidation. Behav Brain Res 2015, 278:375-384. 44. Pe´rez-Cuesta LM, Maldonado H: Memory reconsolidation and extinction in the crab: mutual exclusion or coexistence? Learn Mem 2009, 16:714-721.
Current Opinion in Behavioral Sciences 2018, 19:116–122
45. Dı´az-Mataix L, Martinez RCR, Schafe GE, LeDoux JE, Doye`re V: Detection of a temporal error triggers reconsolidation of amygdala-dependent memories. Curr Biol CB 2013, 23:467472. 46. Yau JOY, McNally GP: Pharmacogenetic excitation of dorsomedial prefrontal cortex restores fear prediction error. J Neurosci 2015, 35:74-83. 47. Alfei JM, Ferrer Monti RI, Molina VA, Bueno AM, Urcelay GP: Prediction error and trace dominance determine the fate of fear memories after post-training manipulations. Learn Mem 2015, 22:385-400. A rodent fear-conditioning study showing the impact of temporal parameters of initial learning in determining the duration of CS exposure required to trigger distinctive memory processes: retrieval, reconsolidation, limbo state or extinction. 48. Sengupta A, Winters B, Bagley EE, McNally GP: Disrupted prediction error links excessive amygdala activation to excessive fear. J Neurosci 2016, 36:385-395. 49. Soeter M, Kindt M: Erasing fear for an imagined threat event. Psychoneuroendocrinology 2012, 37:1769-1779. 50. Lee JLC, Milton AL, Everitt BJ: Reconsolidation and extinction of conditioned fear: inhibition and potentiation. J Neurosci 2006, 26:10051. 51. Pedreira MaE, Maldonado H: Protein synthesis subserves reconsolidation or extinction depending on reminder duration. Neuron 2003, 38:863-869. 52. Bustos SG, Maldonado H, Molina VA: Disruptive effect of midazolam on fear memory reconsolidation: decisive influence of reactivation time span and memory age. Neuropsychopharmacology 2008, 34:446-457. 53. Pin˜eyro ME, Ferrer Monti RI, Alfei JM, Bueno AM, Urcelay GP: Memory destabilization is critical for the success of the reactivation–extinction procedure. Learn Mem 2014, 21: 46-54.
www.sciencedirect.com
ScienceDirect Emotional memory expression is misleading: delineating transitions between memory processes Leonidas Faliagkas1, Priyanka Rao-Ruiz1 and Merel Kindt2 The hypothesis that fear memory is not necessarily permanent but can change when retrieved opens avenues to develop revolutionary treatments for emotional memory disorders. Memory reconsolidation is however only one of several mnemonic processes that may be triggered by memory reactivation and subtle environmental differences may cause a transition from a malleable to a stable state. This poses a major challenge to translating the reconsolidation intervention to clinical practice. Here we review recent advances in understanding the transitions between memory processes in animals and humans, and discuss how the cognitive expression (i.e. threat expectancies) of fear memory in humans may serve as read-out to delineate the underlying processes necessary for memory reconsolidation, independent from the emotional expression of fear memory. Addresses 1 Department of Molecular and Cellular Neurobiology, Center for Neurogenomics and Cognitive Research, Amsterdam Neuroscience, VU University Amsterdam, Amsterdam, The Netherlands 2 Department of Clinical Psychology, University of Amsterdam, Amsterdam, The Netherlands Corresponding author: Faliagkas, Leonidas ([email protected]), Kindt, Merel ([email protected])
Current Opinion in Behavioral Sciences 2018, 19:116–122 This review comes from a themed issue on Emotion–cognition interactions Edited by Mara Mather and Michael Fanselow
https://doi.org/10.1016/j.cobeha.2017.12.018 2352-1546/ã 2017 Published by Elsevier Ltd.
Introduction For more than a century a widely accepted view in cognitive science was that memories are only initially labile and sensitive to disruption, after which they become imprinted or consolidated into the physical architecture of the brain. At the turn of this century, a major breakthrough in neuroscience was achieved with the (re) discovery that fear memory is not inevitably permanent, but can change when retrieved [1,2]. Nader and colleagues [1] demonstrated in rats that the administration of a protein synthesis blocker (i.e. anisomycin) upon memory Current Opinion in Behavioral Sciences 2018, 19:116–122
reactivation eliminated the expression (i.e. freezing) of a previously formed fear memory. Further observations of post-reactivation amnesia have drastically changed the neuroscience literature on memory by generating a novel and influential conceptual framework — usually referred to as the memory reconsolidation hypothesis. The reconsolidation hypothesis states that memory is an intrinsically dynamic process allowing modification of an established memory trace, should conditions require such adaptation [3]. Upon recall, the memory trace may transfer to a transient destabilized state, requiring time-dependent restabilization to persist further (Box 1). Gene transcription and protein synthesis are necessary for reconsolidation and offer a window of opportunity to fundamentally change the memory trace [1,4]. The fear-conditioning paradigm is widely used across different species to study the process of memory reconsolidation [5,6]. In the lab, fear memories can be established through Pavlovian fear conditioning, which involves the repeated pairing of an initially neutral cue (e.g. a tone; conditioned stimulus, CS) with an inherently aversive stimulus (e.g. an electrocutaneous shock; unconditioned stimulus, US). As a result, the representation of the CS and US will become connected in the brain, such that a later presentation of the CS will retrieve the US representation and elicit a conditioned fear response. This so-called associative fear memory represents both cognitive and emotional aspects of fear learning: it involves contingency learning between the originally neutral or ambiguous stimulus (CS) and the reinforcer (US), while the CS becomes also imbued with the affective properties of the reinforcers (US) they predict. Upon a reminder trial of the CS, a conditioned fear response (CR) (i.e. freezing) is usually taken as evidence that a CS-US association has been formed, but this behavioral read-out does not distinguish between the cognitive and emotional aspects of fear memory in animals. Yet, this distinction can easily be made in human fear-conditioning studies where the cognitive expression of fear memory is typically assessed by threat expectancies and the emotional expression by the fear-potentiated startle reflex (e.g. [7,8]). In order for reconsolidation to occur, memory has to be reactivated (Box 1) and destabilized during a generally brief reminder session. Once the memory is rendered labile, reconsolidation can take place and may be experimentally modified by pharmacological [1,9,10] or behavioral manipulations [11–13]. In particular, it has been shown that fear memories can be enhanced or weakened, www.sciencedirect.com
Delineating transitions between memory processes Faliagkas, Rao-Ruiz and Kindt 117
Box 1 Definitions of core processes and concepts of fear memory.
Memory reactivation: The return of memory from a neurobiologically inactive to an active state in the presence of salient reminder cues. Reactivation of memory may take place without any behavioral output. Memory retrieval: The complex process by which previously stored information is recalled, which leads to behavioral output. Match/mismatch: An estimation of how consistent the environment is with previous experience. Prediction error: For fear conditioning, the magnitude of the outcome or the outcome itself is not being fully predicted. Memory destabilization: The transient labile state that a memory may return to upon its reactivation. Memory reconsolidation: The proces of restabilization of memory, once it has been destabilized after memory reactivation. Current Opinion in Behavioral Sciences
depending on the manipulation used [6,14,15]. Since maladaptive memory processing lies at the core of emotional memory disorders (i.e. anxiety disorders, post-traumatic stress disorder, addiction) [8], targeting the process of memory reconsolidation opens avenues to develop a revolutionary treatment. Even though the findings obtained from basic neuroscience are promising, the critical conditions to target complex and pervasive emotional memories typically encountered in clinical practice are still unknown. Very often basic findings from animal literature are translated to clinical trials without a full understanding of the mechanisms of change, leading to disappointing and confusing results. A thorough understanding of the necessary and boundary conditions to trigger and observe memory reconsolidation is essential before we can witness a paradigm shift in clinical practice [6,8,10,16,17]. The reactivation of fear memory is often considered to be synonymous with its behavioral expression, but fear memory may also be reactivated without an observable mnemonic read-out. Hence, a reminder stimulus may return memory to a malleable state necessitating reconsolidation, even when it is not behaviourally expressed [18,19,20,21]. Furthermore, there is substantial evidence demonstrating that a memory is not labilized every time it is retrieved [7,20,22]. If retrieval (Box 1) is not sufficient to trigger memory reconsolidation, then the critical question arises: How do we know whether a particular fear memory actually requires an adaptation? Memory reactivation is thought to trigger memory reconsolidation when the reminder stimuli are similar but not identical (match/mismatch, Box 1) to the original learning environment [5,23]. Yet, a reminder session that is too different www.sciencedirect.com
from the original learning procedure might not trigger memory reconsolidation, but instead initiate other memory processes such as an intermediate pharmacologically insensitive state of limbo [24,25,26] or the formation of a new memory, such as in extinction learning [27,28]. Without an independent index of the necessary and sufficient conditions to trigger memory reconsolidation, other than the memory enhancing or amnesic effects of the manipulations themselves, determining the degree of similarity (or dissimilarity) between the original learning and reminder session presents a challenge to empirical falsifiability. In this review we will discuss different post-reactivation memory processes and illustrate how the transition between these processes depend on subtle changes in the reactivation procedures in interaction with the learning history. Given that there is no single method of memory reactivation that always triggers memory reconsolidation, we claim that an independent read-out is imperative to delineate the underlying processes necessary for memory reconsolidation. In particular, we will briefly review a selection of animal and human studies and describe how memory expression is neither necessary nor sufficient to trigger reconsolidation, but that assessment of a match/mismatch experience (Box 1) may indicate whether a certain reminder trial triggers reconsolidation [6,16,29]. In fear-conditioning studies the match/ mismatch between the expected and actual outcome during memory reactivation is operationalized as prediction error, meaning that the magnitude of the aversive outcome or the outcome itself (i.e. US) is not fully predicted (PE, Box 1) [6,16,29]. Furthermore, we will argue that PE is also not a sufficient condition for Current Opinion in Behavioral Sciences 2018, 19:116–122
118 Emotion–cognition interactions
reconsolidation, but under specific circumstances, it may in fact be a boundary condition. In animal studies the occurrence of PE can be defined by the experimental parameters of the reactivation trial. Even though molecular indicators of memory destabilization have also been identified, the assessment of these markers requires either the firstly, post hoc analysis of brain tissue immediately after the reactivation session, or secondly, pharmacological manipulation of specific proteins/pathways involved in the destabilization of memory [30–34]. While the former method of analysis cannot be used for longitudinal measurements of memory in the same animals, pharmacological interventions may be used to successfully block/enhance [35] the destabilization of memory. The molecular markers of memory destabilization may indicate the effect of PE on downstream memory process, but they do not evaluate the degree of PE itself. In humans, however, the cognitive expression (i.e. threat expectancies) of fear memory may serve as read-out to indicate the degree of PE itself. In a series of studies on fear memory reconsolidation in which the emotional (i.e. fear potentiated startle reflex) and cognitive (i.e. threat expectancies) expression of fear memory were measured in parallel, we have demonstrated a double dissociation between these different response systems [7,24,36]. By capitalizing upon this double dissociation, we revealed that the cognitive expression of fear memory might serve as read-out to demarcate the underlying processes necessary for memory reconsolidation, independent from the emotional expression of fear memory.
Memory expression is neither sufficient, nor necessary to trigger memory reconsolidation In animal models, the reactivation of fear memories is typically assessed by behavioral output to salient cues, while there is convincing evidence that memory can be reactivated without behavioral responding [18,19,20,21]. However, when a reminder stimulus effectively elicits behavioral memory expression (i.e. retrieval, Box 1), this is not sufficient to destabilize memory and trigger subsequent reconsolidation [7,36,37,38]. In humans, we [7] demonstrated that the mere reactivation of a previously formed fear memory did not lead to reconsolidation. We addressed this issue by eliminating the very possibility of reinforcement of the reminder cue. A previously established associative fear memory (pictureshock pairings) was reactivated with shock-electrodes either attached or unattached. Reconsolidation was not triggered when nothing could be learned during the reminder trial (i.e. shocks unattached). Only when the outcome of the reactivation was not fully predictable (i.e. uncertainty about receiving a shock), memory reconsolidation was engaged. We concluded that memory reconsolidation does not take place when there is nothing to be learned during memory reactivation that would be relevant to the previously formed memory. It is also striking that memory reactivation, either with or without Current Opinion in Behavioral Sciences 2018, 19:116–122
anticipation of the threat (US), did not affect the emotional expression of the fear memory (startle response). Memory can also be reactivated and destabilized at a neurobiological level without being expressed at all [18,19,20,21]. In an elegant set of experiments, utilizing a context-signal memory model in crabs, Caffaro and colleagues [18] found that scopolamine-induced amnesia resulted in severe memory expression deficits. Interestingly, even under these amnestic conditions this nonexpressible memory could be reactivated, and the deficits were subsequently reversed by reconsolidation-specific memory enhancement. Thus, under defined circumstances memory can be destabilized at a neurobiological level and trigger reconsolidation without being behaviourally expressed. This dissociation has also been observed in other memory paradigms [20,39], and further evidenced by studies showing differences in glutamate receptor activity for memory expression and reconsolidation in the amygdala, with AMPAR antagonists specifically affecting the former and NMDAR antagonists the latter process [21,40] (but see [19]). Taken together, these data indicate that the expression of fear memory is not synonymous with its reactivation and destabilization. While reactivation is needed for reconsolidation, behavioral output is not sufficient or necessary to drive this adaptive process (but see [41]). This inevitably leads to the subsequent question: if memory expression during reactivation is not a necessary or sufficient condition for destabilization and reconsolidation, what triggers these processes? The studies discussed above [18,20,39] induced a discrepancy between the training (reinforced CS) and the reminder session (unreinforced CS presentation) to trigger reconsolidation. Once reactivated, memory might follow several paths based on the detection of a match/mismatch experience.
Prediction error is necessary, but not sufficient to trigger memory reconsolidation Drawing upon the idea that memory reconsolidation serves as a means of incorporating new information into memories to adapt to ever changing environments [2,3], it may be postulated that the reminder session should not simply convey new information, but it should also be strongly related to past experience. For associative learning paradigms, a discrepancy between expectation and actual outcome (match/mismatch) during reactivation can signal a PE, which in turn may serve to destabilize memory necessitating restabilization in order to persist further [29,42,43]. Pedreira and colleagues [37] provide one of the first reports of PE as a necessary condition for fear memory destabilization and subsequent reconsolidation. Since then, PE has been used in several experimental settings, across different species, to trigger destabilization and reconsolidation [7,18,19,24,29,36,38,44– 46,47,48]. For fear-conditioning paradigms, PE signals www.sciencedirect.com
Delineating transitions between memory processes Faliagkas, Rao-Ruiz and Kindt 119
may be experimentally generated during a reminder session by a non-reinforced CS exposure (no US), a reinforced CS exposure (positive PE) [36], uncertainty [49] or a shift in CS-US temporal association [45]. This gives rise to unmatched expectation that triggers memory destabilization and subsequent reconsolidation. Destabilization can be measured most reliably at a molecular level [30–34], while reconsolidation can be quantified at a molecular, cellular and behavioral level (blocking/ enhancing reconsolidation) [29,36,45]. In animal fearconditioning studies PE itself cannot be reliably measured and can only be pre-defined by the specific experimental parameters introduced during the memory reactivation trial. The analysis of molecular markers of PEdriven memory destabilization [25,32,33] or modified memory expression [25,29,36,37] at a later retention test may be used to indirectly test if a match/mismatch occurred, but these indicators require post hoc analysis and cannot be used to quantify the extent of mismatch generated during the reminder trial. Translational approaches from animal to humans offer a unique advantage in observing and quantifying the match/mismatch experience by threat expectancy ratings. Whereas freezing and startle potentiation basically reflect an expression of fear memory, threat expectancy is not a mere reflection of what has previously been learned: It also incorporates information from the current situation. As such, the threat expectancy ratings can be used to estimate how consistent the environment is with previous experience. The potential of expectancy ratings in unveiling the necessary conditions for memory reconsolidation is demonstrated in a series of human fear-conditioning studies [7,36,49]. An amnesic agent (b-adrenergic receptor blocker, propranolol) presented before or after memory reactivation specifically affected the emotional expression of fear memory (i.e. fearpotentiated startle response), while the cognitive level (i.e. threat expectancies) remained unaffected. This striking dissociation between the emotional and cognitive expression of fear memory in humans created an opportunity to develop a novel measure of PE-driven memory reconsolidation that does not rely on manipulation induced changes in the conditioned fear response. In particular, we demonstrated in a fear-conditioning study that reconsolidation was induced by reactivation conditions that led to a subsequent change in ratings of threat expectancy upon CS presentation (i.e. match/mismatch). Reconsolidation occurred in both unreinforced and reinforced reminder trials where the aversive stimulus (US) was not fully predicted by the CS, even though no reinforcement resulted in a decrease in threat expectancy from learning to trial while reinforcement resulted in an increase [36]. In contrast, reminder trials that left reported US expectancies unaffected did not induce fear memory reconsolidation. Notwithstanding these new insights on PE as a necessary condition to trigger memory reconsolidation, a critical www.sciencedirect.com
question is whether PE is a sufficient condition for reconsolidation, since PE can also give rise to new learning (e.g. extinction learning) and we know that extinction training puts a constraint on reconsolidation [27,50]. Pedreira et al. [51] (also, see [27]) demonstrated that the duration of reactivation may act as a switch to trigger either reconsolidation of the original fear memory or the formation of an extinction memory. A short unreinforced reminder session has been repeatedly found to induce reconsolidation, whereas longer exposure leads to the formation of an extinction memory that will be subsequently expressed [22,51–53]. Although reconsolidation and extinction are both vulnerable to amnestic agents, distinctive features of the reminder session in interaction with the original learning determines [24,36] which memory trace is predominant and subsequently more sensitive to disruption [51]. The studies of Sevenster et al. [24] and Merlo et al. [25] investigated the transition from reconsolidation to extinction by utilizing multiple match/mismatches (CS-no US) during the reminder session. Repeated or prolonged unreinforced exposure generated multiple PEs, which eventually reduced both threat expectancy and fear responding resulting in extinction learning at the behavioral [24,25] and molecular level [25]. Intriguingly, a CS-exposure of intermediate duration has been observed to lead to an intermediate phase (limbo state), insensitive to amnestic agents [24,25], presumably because neither of the two opposite processes of memory reconsolidation or extinction is engaged. In addition to factors such as the age and strength of memory [22,52], Alfei and colleagues [47] found that the specific temporal conditions under which initial learning took place are also critical in determining the duration of CS exposure needed to trigger one of four mnemonic processes: retrieval, reconsolidation, an intermediate state of limbo, or extinction. The existence of a pharmacologically-insensitive limbo state, as well as the indefinite transitional boundaries between these distinctive memory processes is a challenge to unequivocally trigger reconsolidation in both pre-clinical and clinical settings. In sum, behavioral studies in animals and humans have shown that the degree of similarity between encoding and reactivation demarcates the transition from retrieval, to reconsolidation, to limbo, to extinction. If reminder conditions fully match with encoding conditions, the memory is just retrieved, without further updating the memory. If conditions are quite similar, yet hold an element of novelty or surprise, memory destabilization and subsequent reconsolidation is triggered and the memory will be updated. If, however, conditions are too dissimilar, the reminder episode will transform into either a limbo state, or new learning by creating a new memory that coexists with the old, unaltered memory. The assumption made here is that minor changes in threat expectancy ratings in humans may serve to delineate the transitions between these memory processes (Figure 1). Current Opinion in Behavioral Sciences 2018, 19:116–122
120 Emotion–cognition interactions
Figure 1
humans/animals
humans/animals
humans
humans/animals
post reactivation memory processes
match/mismatch
change in expectancy during reactivation
change in fear expression during reactivation
Retrieval
no
no
no
Reconsolidation
yes
no
no
Limbo state
yes
yes
no
Extinction
yes
yes
yes Current Opinion in Behavioral Sciences
Overall boundary conditions that affect the post reactivation fate of memory: Behavioral expression of fear during memory reactivation is not necessary or sufficient to trigger post-reactivation mnemonic processes. However, a match/mismatch based on prior learning, introduced during a reactivation trial, results in prediction error that is necessary for destabilization of memory. Depending on the number of errors generated during memory reactivation, as well as the nature of the original learning trial, memories may enter one of several states, namely; sole retrieval, reconsolidation, an intermediate limbo state or extinction. Delineating the transition between these different states may be possible in humans using a read-out of threat expectancy, but is more difficult in animals where the only quantifiable read-out is the expression of the conditioned fear response, which only changes when extinction learning takes place. The quantifiable changes mentioned in the figure may be bi-directional in the case of reconsolidation and limbo, but always result in a decrease for extinction.
Conclusion and future directions Translational research across species is critical in fostering our understanding of the necessary and sufficient conditions to trigger memory reconsolidation in clinical practice. A selection of research demonstrates that behavioral output during a reminder trial is not necessary or sufficient to induce memory reconsolidation. A match/mismatch experience, generally referred to as PE, is emerging as a necessary condition to return memory to a labile state. However, multiple PE’s during memory reactivation can lead to other mnemonic states like limbo or extinction, which are clear boundary conditions on the induction of memory reconsolidation. The subtle and undefined transitions between the distinctive memory processes pose a huge challenge to research on memory reconsolidation. Although fear-conditioning research in humans offer a unique advantage in quantifying the match/mismatch experience by threat expectancy ratings, animal research does not differentiate between the affective and cognitive properties of fear memory. So far, the occurrence of PE in animals is typically defined by specific experimental parameters of the memory reactivation trial and inferred from the effect that a match/mismatch experience has on downstream memory processes like destabilization and reconsolidation. However, this form of analysis does not provide an assessment of the degree of match/mismatch Current Opinion in Behavioral Sciences 2018, 19:116–122
generated during reactivation, which can profoundly affect the subsequent fate of memory. To solve this problem of empirical falsifiability in animal research, future studies on PE might dissociate the cognitive and emotional representations of fear memory by combining a declarative memory task with a fear-conditioning procedure. Alternatively, dissecting the intricacies of a match/mismatch experience at a neurobiological level could also advance our understanding of the dynamic balance between the stability and malleability of fear memory. Identifying brain regions that signal match/mismatch experiences in combination with in vivo techniques like optogenetics would enable us to directly manipulate PE signals with an unprecedented level of precision.
Conflict of interest statement Nothing declared.
References and recommended reading Papers of particular interest, published within the period of review, have been highlighted as: of special interest of outstanding interest 1.
Nader K, Schafe GE, Le Doux JE: Fear memories require protein synthesis in the amygdala for reconsolidation after retrieval. Nature 2000, 406:722-726. www.sciencedirect.com
Delineating transitions between memory processes Faliagkas, Rao-Ruiz and Kindt 121
A seminal fear-conditioning study in rodents launching the reconsolidation hypothesis by showing a post-retrieval protein-synthesis dependent restabilization of memory. 2.
Sara SJ: Retrieval and reconsolidation: toward a neurobiology of remembering. Learn Mem 2000, 7:73-84.
3.
Dudai Y: Predicting not to predict too much: how the cellular machinery of memory anticipates the uncertain future. Philos Trans R Soc B 2009, 364:1255-1262.
4.
Mamiya N, Fukushima H, Suzuki A, Matsuyama Z, Homma S, Frankland PW, Kida S: Brain region-specific gene expression activation required for reconsolidation and extinction of contextual fear memory. J Neurosci 2009, 29:402.
5.
Finnie PSB, Nader K: The role of metaplasticity mechanisms in regulating memory destabilization and reconsolidation. Neurosci Biobehav Rev 2012, 36:1667-1707.
6.
Elsey JWB, Kindt M: Tackling maladaptive memories through reconsolidation: from neural to clinical science. Neurobiol Learn Mem 2017, 142(Part A):108-117.
7.
Sevenster D, Beckers T, Kindt M: Instructed extinction differentially affects the emotional and cognitive expression of associative fear memory. Psychophysiology 2012, 49:14261435.
8.
Kindt M: A behavioural neuroscience perspective on the aetiology and treatment of anxiety disorders. Behav Res Ther 2014, 62:24-36.
9.
De?biec J, Ledoux JE: Disruption of reconsolidation but not consolidation of auditory fear conditioning by noradrenergic blockade in the amygdala. Neuroscience 2004, 129:267-272.
10. Kindt M, Soeter M, Vervliet B: Beyond extinction: erasing human fear responses and preventing the return of fear. Nat Neurosci 2009, 12:256-258. 11. Schiller D, Monfils M-H, Raio CM, Johnson DC, LeDoux JE, Phelps EA: Preventing the return of fear in humans using reconsolidation update mechanisms. Nature 2010, 463:49-53. 12. Rao-Ruiz P, Rotaru DC, van der Loo RJ, Mansvelder HD, Stiedl O, Smit AB, Spijker S: Retrieval-specific endocytosis of GluA2AMPARs underlies adaptive reconsolidation of contextual fear. Nat Neurosci 2011, 14:1302-1308. 13. Monfils M-H, Cowansage KK, Klann E, LeDoux JE: Extinctionreconsolidation boundaries: key to persistent attenuation of fear memories. Science 2009, 324:951-955. 14. Tronson NC, Wiseman SL, Olausson P, Taylor JR: Bidirectional behavioral plasticity of memory reconsolidation depends on amygdalar protein kinase A. Nat Neurosci 2006, 9:167-169. 15. Lee JLC: Memory reconsolidation mediates the strengthening of memories by additional learning. Nat Neurosci 2008, 11:1264-1266. 16. Elsey JWB, Kindt M: Breaking boundaries: optimizing reconsolidation-based interventions for strong and old memories. Learn Mem 2017, 24:472-479. 17. Kindt M: The surprising subtleties of changing fear memory: a challenge for translational science. Philos Trans R Soc B 2017 http://dx.doi.org/10.1098/rstb.2017.0033. (in press). 18. Caffaro PA, Sua´rez LD, Blake MG, Delorenzi A: Dissociation between memory reactivation and its behavioral expression: scopolamine interferes with memory expression without disrupting long-term storage. Neurobiol Learn Mem 2012, 98:235-245. A context-signal memory study in crabs showing that non-expressible memory could still be reactivated without a behavioural output, and subsequently manipulated by reconsolidation-specific memory enhancement. 19. Barreiro KA, Sua´rez LD, Lynch VM, Molina VA, Delorenzi A: Memory expression is independent of memory labilization/ reconsolidation. Neurobiol Learn Mem 2013, 106:283-291. 20. Balderas I, Rodriguez-Ortiz CJ, Bermudez-Rattoni F: Retrieval and reconsolidation of object recognition memory are www.sciencedirect.com
independent processes in the perirhinal cortex. Neuroscience 2013, 253:398-405. 21. Mamou CB, Gamache K, Nader K: NMDA receptors are critical for unleashing consolidated auditory fear memories. Nat Neurosci 2006, 9:1237-1239. A rodent fear-conditioning study showing the necessity of NMDA receptors within the amygdala for memory labilization, independent from the expression of fear memory. 22. Suzuki A, Josselyn SA, Frankland PW, Masushige S, Silva AJ, Kida S: Memory reconsolidation and extinction have distinct temporal and biochemical signatures. J Neurosci 2004, 24:4787-4795. 23. Osan R, Tort ABL, Amaral OB: A mismatch-based model for memory reconsolidation and extinction in attractor networks. PLOS ONE 2011, 6:e23113. 24. Sevenster D, Beckers T, Kindt M: Prediction error demarcates the transition from retrieval, to reconsolidation, to new learning. Learn Mem 2014, 21:580-584. A human fear-conditioning study showing that prediction error (PE) — measured by threat expectancies — is not sufficient to trigger reconsolidation, and that multiple PE’s triggered an intermediate phase (limbo state) between reconsolidation and extinction, which is insensitive to amnestic agents. 25. Merlo E, Milton AL, Gooze´e ZY, Theobald DE, Everitt BJ: Reconsolidation and extinction are dissociable and mutually exclusive processes: behavioral and molecular evidence. J Neurosci 2014, 34:2422-2431. 26. Cassini LF, Flavell CR, Amaral OB, Lee JLC: On the transition from reconsolidation to extinction of contextual fear memories. Learn Mem 2017, 24:392-399. 27. Eisenberg M, Kobilo T, Berman DE, Dudai Y: Stability of retrieved memory: inverse correlation with trace dominance. Science 2003, 301:1102-1104. 28. Bos MGN, Beckers T, Kindt M: The effects of noradrenergic blockade on extinction in humans. Biol Psychol 2012, 89:598605. 29. Agustina Lo´pez M, Jimena Santos M, Cortasa S, Ferna´ndez RS, Carbo´ Tano M, Pedreira ME: Different dimensions of the prediction error as a decisive factor for the triggering of the reconsolidation process. Neurobiol Learn Mem 2016, 136:210219. 30. Almeida-Correˆa S, Amaral OB: Memory labilization in reconsolidation and extinction — evidence for a common plasticity system? Neurobiol Learn Mem 2014, 108:292-306. 31. Balaban PM, Roshchin M, Timoshenko AK, Gainutdinov KL, Bogodvid TK, Muranova LN, Zuzina AB, Korshunova TA: Nitric oxide is necessary for labilization of a consolidated context memory during reconsolidation in terrestrial snails. Eur J Neurosci 2014, 40:2963-2970. 32. Jarome TJ, Ferrara NC, Kwapis JL, Helmstetter FJ: CaMKII regulates proteasome phosphorylation and activity and promotes memory destabilization following retrieval. Neurobiol Learn Mem 2016, 128:103-109. 33. Lee S-H, Choi J-H, Lee N, Lee H-R, Kim J-I, Yu N-K, Choi S-L, Lee S-H, Kim H, Kaang B-K: Synaptic protein degradation underlies destabilization of retrieved fear memory. Science 2008, 319:1253-1256. 34. Lee S-H, Kwak C, Shim J, Kim J-E, Choi S-L, Kim HF, Jang D-J, Lee J-A, Lee K, Lee C-H et al.: A cellular model of memory reconsolidation involves reactivation-induced destabilization and restabilization at the sensorimotor synapse in Aplysia. Proc Natl Acad Sci U S A 2012, 109:14200-14205. 35. Lee JLC, Flavell CR: Inhibition and enhancement of contextual fear memory destabilization. Front Behav Neurosci 2014, 8:144. 36. Sevenster D, Beckers T, Kindt M: Prediction error governs pharmacollogically induced amnesia for learned fear. Science 2013, 339:830-833. A human fear-conditioning study uncovering an independent index for prediction error (PE) by utilzing threat expectancy rating, and Current Opinion in Behavioral Sciences 2018, 19:116–122
122 Emotion–cognition interactions
demonstrating that PE is necessary for memory reconsolidation, determined by the interaction between original learning and retrieval. 37. Pedreira ME, Pe´rez-Cuesta LM, Maldonado H: Mismatch between what is expected and what actually occurs triggers memory reconsolidation or extinction. Learn Mem 2004, 11:579-585. A context-signal memory study in crabs demonstrating the necessary role of prediction error (PE) for reconsolidation and extinction of aversive memory. 38. Frenkel L, Maldonado H, Delorenzi A: Memory strengthening by a real-life episode during reconsolidation: an outcome of water deprivation via brain angiotensin II. Eur J Neurosci 2005, 22:1757-1766. 39. Santoyo-Zedillo M, Rodriguez-Ortiz CJ, Chavez-Marchetta G, Bermudez-Rattoni F, Balderas I: Retrieval is not necessary to trigger reconsolidation of object recognition memory in the perirhinal cortex. Learn Mem 2014, 21:452-456. 40. Milton AL, Merlo E, Ratano P, Gregory BL, Dumbreck JK, Everitt BJ: Double dissociation of the requirement for GluN2Band GluN2A-containing NMDA receptors in the destabilization and restabilization of a reconsolidating memory. J Neurosci 2013, 33:1109-1115. 41. Rossato JI, Ko¨hler CA, Radiske A, Bevilaqua LRM, Cammarota M: Inactivation of the dorsal hippocampus or the medial prefrontal cortex impairs retrieval but has differential effect on spatial memory reconsolidation. Neurobiol Learn Mem 2015, 125:146-151. 42. Holland PC, Schiffino FL: Mini-review: prediction errors, attention and associative learning. Neurobiol Learn Mem 2016, 131:1-9. 43. Exton-McGuinness MTJ, Lee JLC, Reichelt AC: Updating memories — the role of prediction errors in memory reconsolidation. Behav Brain Res 2015, 278:375-384. 44. Pe´rez-Cuesta LM, Maldonado H: Memory reconsolidation and extinction in the crab: mutual exclusion or coexistence? Learn Mem 2009, 16:714-721.
Current Opinion in Behavioral Sciences 2018, 19:116–122
45. Dı´az-Mataix L, Martinez RCR, Schafe GE, LeDoux JE, Doye`re V: Detection of a temporal error triggers reconsolidation of amygdala-dependent memories. Curr Biol CB 2013, 23:467472. 46. Yau JOY, McNally GP: Pharmacogenetic excitation of dorsomedial prefrontal cortex restores fear prediction error. J Neurosci 2015, 35:74-83. 47. Alfei JM, Ferrer Monti RI, Molina VA, Bueno AM, Urcelay GP: Prediction error and trace dominance determine the fate of fear memories after post-training manipulations. Learn Mem 2015, 22:385-400. A rodent fear-conditioning study showing the impact of temporal parameters of initial learning in determining the duration of CS exposure required to trigger distinctive memory processes: retrieval, reconsolidation, limbo state or extinction. 48. Sengupta A, Winters B, Bagley EE, McNally GP: Disrupted prediction error links excessive amygdala activation to excessive fear. J Neurosci 2016, 36:385-395. 49. Soeter M, Kindt M: Erasing fear for an imagined threat event. Psychoneuroendocrinology 2012, 37:1769-1779. 50. Lee JLC, Milton AL, Everitt BJ: Reconsolidation and extinction of conditioned fear: inhibition and potentiation. J Neurosci 2006, 26:10051. 51. Pedreira MaE, Maldonado H: Protein synthesis subserves reconsolidation or extinction depending on reminder duration. Neuron 2003, 38:863-869. 52. Bustos SG, Maldonado H, Molina VA: Disruptive effect of midazolam on fear memory reconsolidation: decisive influence of reactivation time span and memory age. Neuropsychopharmacology 2008, 34:446-457. 53. Pin˜eyro ME, Ferrer Monti RI, Alfei JM, Bueno AM, Urcelay GP: Memory destabilization is critical for the success of the reactivation–extinction procedure. Learn Mem 2014, 21: 46-54.
www.sciencedirect.com