Commentary on Tuch

In “Problems Applying Neuroscientific Research to the Clinical Setting,” Richard Tuch challenges the idea that neuroscientific findings on memory reconsolidation, the process by which memories can become fragile upon retrieval (Alberini 2011; Nader and Einarsson 2010), might be useful or even correctly interpreted in psychoanalysis and, in general, in the context of psychotherapeutic applications. He expresses “concern about a tendency to accept neuroscientific findings without rigorously studying the methods used to arrive at the conclusions reached and to rush to find ways of applying such findings to the clinical setting. There are clear tendencies on the part of some to uncritically rely on neuroscientific studies to explain more than they reasonably can” (p. 314). Regarding memory reconsolidation as a target method for disrupting recalled memories, Tuch then raises two specific and important critiques: (1) he questions whether research conducted for the most part using a single learning exposure to a threatening condition in an animal model (or even in humans) is relevant to complex individual human experience resulting from cumulative trauma; (2) he notes that in some cases reconsolidation shows temporal limits, with recent but not remote memories being changeable.

He notes also that “analysts typically treat patients suffering from older, well-consolidated memories that have been laid down and reinforced over time—not at all the sort of experiences these researchers were dealing with in the laboratory. It is possible that reconsolidation may end up playing an important role in the treatment of patients presenting posttraumatic conditions, but it seems unlikely that reconsolidation has much to do with the sorts of long-term emotional memories that plague our patients’ lives” (p. 313).

I will try to address these issues from the point of view of my dual training in neuroscience and psychoanalysis. As a neuroscientist, I have been working specifically on memory consolidation and reconsolidation in animal models; hence I hope to provide some clarification on this matter.

I could not agree more with Tuch that it is important to exert caution before drawing strong and general conclusions from the findings of new and often quite exciting fields, and that the translation of these findings from animal models to humans requires careful and solid evidence. However, the field of reconsolidation has expanded only in recent years, and should not be dismissed out of hand simply because it has not yet addressed all the important questions regarding complex individual behavior, especially in humans. Science generally proceeds by investigating simple systems first and then moving to complex ones. The reconsolidation field has made remarkable progress, starting with understanding simplified memory paradigms in animals and controlled memory tasks in humans. These results are critical for understanding how memory processes work and for formulating future questions regarding memories in humans. I believe that knowing the temporal dynamics and mechanisms underlying memory retrieval, stabilization, destabilization, processing, and updating with new associations will significantly aid in designing more precise and individuated treatments in psychoanalytic settings. For example, this knowledge can help determine when to intervene, what type of information should be used, and what modalities are most effective.

The first point I would like to clarify is that most neuroscientists do not credit “the idea that pathogenic memories can entirely be erased” (p. 314); they in fact eschew any use of the term “memory erasure” when referring to reconsolidation interference. Although the term has inadvisably been used by some, the studies on memory reconsolidation have in fact yielded no data showing that memories can entirely be “erased”; this idea comes mostly from the sensationalism generated by the exciting findings that memories can reconsolidate. What these studies have actually shown is that interfering with post-retrieval memory mechanisms can decrease (not erase) the behavioral expression of the memory. In some studies, this behavioral decrease has been found to correlate with the loss of brain activation or activity signatures—hence the idea of disrupting (hence “erasing”) memory substrates. Whether the decrease in memory expression is due to memory modulation, changes in the emotional association, or removal of the memory trace representation at subsequent retrievals remains to be determined.

I must also take issue with Tuch’s conclusion that disrupting memories (he understands this as complete erasure) “is a romantic notion one only wishes were true. Until a more efficacious form of treatment can be found, we analysts are going to have to keep trudging through the muck on our way to a ‘cure’ that is far from perfect” (p. 314). In fact, as mentioned above, I believe that understanding the mechanisms of memory consolidation and reconsolidation has provided, and will continue to provide, valuable information for when and how to intervene in psychoanalytic settings. Thus, such knowledge may offer valuable tools for altering pathological states and alleviating suffering. Neglecting or dismissing it entirely would be as big a mistake as believing that reconsolidation can disrupt or “erase” all unwanted memories. Please note that I have referred to both memory consolidation and reconsolidation as tools for modulating memory strength because, although the rediscovery of reconsolidation was instrumental for changing the way we think about memory formation and storage, I do believe—as the work of my laboratory and others has demonstrated—that reconsolidation does not always occur upon retrieval, and that older memories can become insensitive to reconsolidation. This issue has been correctly noted by Tuch, and I will be discussing it below. I also believe that memory reconsolidation itself is not the mechanism by which memories modify their content, but that new memory traces are formed following retrieval (Alberini, Milekic, and Tronel 2006; Alberini 2011; Alberini, Johnson, and Ye 2013). These new traces that associate old and new experiences, together with traces that may undergo reconsolidation, lead to a modification of the memories and knowledge of the past experiences, thus providing new choices for processing mental responses. I will expand on these points below. I will summarize the current knowledge on memory reconsolidation research specifically relevant to Tuch’s comments, as it will not be possible to review in a commentary all of the vast literature related to memory reconsolidation.

The rediscovery that recalled memories can become fragile and changeable, occurring in neuroscience around the year 2000, transformed the accepted view of memory processes and storage. Studies over the past fifteen years have established very clearly that in many species, from worms to humans, memories are stored not in a fixed form, as previously believed, but in a highly dynamic manner (Sara 2000; Nader and Einarsson 2010; Alberini 2011; Alberini, Johnson, and Ye 2013; Dudai 2012). It had been known for more than a century that a newly formed memory, initially labile, undergoes a process of stabilization, known as consolidation, allowing the memory to be stored long-term. Until recently it was thought that this consolidation occurs only once, and that memories, once stabilized, cannot return to a state of fragility (McGaugh 2000; Squire and Alvarez 1995); the implication was that once they are stored, memories are fixed. This, however, has recently been shown to be incorrect: recalling, retrieving, or in general reactivating established memories returns them temporarily to a labile state. During this post-retrieval fragile phase the memory undergoes a process of restabilization called memory reconsolidation (Sara 2000; Nader, Schafe, and LeDoux 2000). This period of reconsolidation offers a remarkable opportunity for intervention: a temporal window during which memories can not only be disrupted by interfering with reconsolidation mechanisms, but can also be enhanced or updated. In other words, reconsolidation offers a window of post-retrieval malleability.

These findings have changed the way we think about memory storage and processes and have remarkable clinical implications: an obvious one is the opportunity to interrupt or prevent the reconsolidation of pathogenic memories contributing crucially to maladaptive behaviors. This opportunity has been enthusiastically embraced, and in the last decade and a half neuroscientists, psychologists, cognitive scientists, and psychiatrists have investigated memory reconsolidation in both healthy and maladaptive conditions using both animal models and humans. Conclusions reached by reconsolidation studies to date include the following.

In sum, reconsolidation research has investigated many different types of memories, as well as the contribution of distinct brain regions, mechanisms, and temporal progressions and provided important answers. From here, many other questions need to be addressed. An important one raised by Tuch’s paper is the contribution of reconsolidation regarding memories formed after multiple, cumulative traumatic events, as opposed to single, discrete ones. It is clear that these are not comparable conditions, and the underlying differences need to be systematically investigated. The field has started exploring the effect of a single threatening event or a few repeated ones, and in both cases found that reconsolidation can occur (Alberini 2011; Milekic et al. 2006; Lee 2008); however, whether reconsolidation takes place after chronic experience, especially traumatic experience, has yet to be determined. As discussed below, research on memory reconsolidation has been focusing mostly on adaptive memories, and has not yet sufficiently investigated models of severe or chronic trauma, except for the few studies done with PTSD patients.

Another important question still debated is the function of memory reconsolidation. Why does memory become temporarily labile after recall? Two main hypotheses have been tested: first, that reconsolidation mediates the updating of memories with novel information; second, that it mediates memory strengthening. Several authors, primarily myself, have proposed that reconsolidation mediates updating when the new information is similar to the original memory (e.g., memory strengthening [Inda, Muravieva, and Alberini 2011; Forcato, Fernandez, and Pedreira 2014]), but does not when the new information is incongruent with and distinct from the original. This matter is quite relevant for psychotherapy and psychoanalysis. What happens in the analytic work? Are recalled memories changing because information brought on by the new context and objects modifies the old memory content? Or because new memories are stored in parallel with the old ones, thus providing additional knowledge and the ability to change mental processing and, in train, behavioral responses? Our data in animal models based on mechanistic double dissociations and temporal analyses of memory reconsolidation have shown that updating old memories with new, distinct information occurs through the formation of new memory traces, and thus through the consolidation of new memories rather than the reconsolidation of old ones. The new trace coexists with, is linked to, the old memory, which upon retrieval may then undergo reconsolidation (Tronel, Milekic, and Alberini 2005). This conclusion accords with findings that the reconsolidation of complex episodic memories is a temporally limited process. How, given this temporal limit, can reconsolidation explain the updating of old memories? I must disagree, then, with the idea that reconsolidation fully explain the process of change in psychoanalysis and psychotherapy (Lane et al. 2014; Ecker, Ticic, and Hulley 2012). I see reconsolidation as a selective component that, together with the formation of new memories (which includes the extinction and reprocessing of emotional representations), leads to changes in the processing of implicit and explicit mental responses (Alberini, Johnson, and Ye 2013; Alberini, Ansermet, and Magistretti 2013). This expands our understanding of, without being in conflict with, the classical view of how the transferential process of working through is therapeutic. Based on the currently available data, I generally agree with Tuch’s conclusion that experiences cannot be entirely erased and that the core of “emotional memories are durable and continue to exert influence over an individual’s emotional life even after the transference has been ‘worked through’” (p. 311). I also agree that given the important contributions of repeated and/or old experiences a successful therapy may be the result of “the modification of unconscious associational networks brought about by ‘weakening links between mental processes that have become associatively linked. . . . [This involves] the creation of new associative linkages, or the strengthening of links that were previously weak’” (p. 311). It should be recognized, however, that the studies on memory consolidation and reconsolidation provide us scientific knowledge for use in designing more precise and effective therapeutic interventions targeting these long-lasting emotional memories.

Based on the data available thus far, I propose that in a psychoanalytic setting both reconsolidation and new memory formation occur upon recall. Recalling complex declarative or episodic memories, as in psychotherapy, entails the recall of components and memories from different ages, as well as their continual updating. Thus, recall likely reactivates, simultaneously, both recent and older memory traces with implicit and explicit components. Further, memory retrieval is always unique and complex; thus, its modalities will dictate whether or not reconsolidation and/or another retrieval-mediated process (e.g., extinction or consolidation) will take place. For example, retrieval without the conditioning contingency (the trauma), when repeatedly experienced, is known to induce extinction, a new learning that informs the subject that the trauma itself does not always happen, thus leading to a decrease in the conditioned response to the trauma. This new memory coexists with the old memory of the experienced trauma, but changes the behavioral response to it. In brief, memories are not fixed, but highly dynamic (Inda et al. 2011; Alberini, Johnson, and Ye 2013; Alberini, Ansermet, and Magistretti 2013), and their storage constantly changes though plasticity mechanisms evoked by reactivation of the memory traces (Finnie and Nader 2012). Reactivation can induce reconsolidation, extinction, or the consolidation of new memories. Targeted interference, potentiation, and co-regulation of each of these processes, such as the remarkable effect of extinction-mediated interference of reconsolidation (Schiller et al. 2010; Schiller, Raio, and Phelps 2012; Monfils et al. 2009; Jones and Monfils 2013), are specific tools available for modulating and modifying memories.

Given the dynamic storage of memory, why are memories of traumatic experiences or drug abuse so strong and persistent? This is an important question regarding clinical intervention, one raised in Tuch’s paper. Studies on consolidation of these strong and persistent memories suggest that they are overrepresented and thus are more strongly and widely stored. Further, they seem to hijack circuitry that normally is not involved (Pittenger and Duman 2008; Kauer and Malenka 2007). In brief, they differ from weaker, adaptive memories. Can these persistent memories be disrupted after retrieval? Both clinical and animal model studies show conflicting results. While animal models suggest that in some cases, but not all, these memories can indeed be disrupted by targeting retrieval (Taubenfeld et al. 2009, 2010; Gazarini et al. 2014; Hymel et al. 2014; Costanzi et al. 2014), clinical investigations have yet to find convincing effects. For example, PTSD patients treated with pharmacological interventions targeting reconsolidation did not show the desired effects (Wood et al. 2015; Spring et al. 2015; Schiller and Phelps 2011), and reconsolidation seemed to target implicit responses but not symptomatology (Lonergan et al. 2013).

However, these clinical data should not be taken as final proof that targeting reconsolidation is ineffective in humans. The lack of significant effects might be due to several factors still to be clarified and investigated. For example, the number of compounds tested to date is still limited, and the different modalities of memory reactivation need to be further explored. Also, as mentioned before, pathological memories are likely different from memories developed by healthy subjects. Although animal models have been used to reproduce the memorization of threatening experiences, it is becoming increasingly clear that memories of weak threats are not adequate models of traumatic memories. Thus, the information provided about memory reconsolidation must be expanded, and we need to know more about the mechanisms underlying pathogenic memories. Neuroscientists are currently developing better models (animal, computational, preclinical, etc.) of trauma and anxiety disorders, in order to investigate their underlying mechanisms and test potential therapies. Finally, whether targeting reconsolidation helps weaken symptoms in severely traumatized patients remains to be determined and studies are in progress. I would speculate, however, that early intervention supporting the reconsolidation of traumatic experiences may help in preventing the development of pathologies like PTSD and anxiety disorders.

In sum, findings important for psychoanalysis that have emerged from memory reconsolidation studies include the following:

This knowledge may be useful in developing more effective and shorter therapies, and psychoanalysis would benefit from tested and proven protocols.

I conclude with a comment on animal studies: Are they informative for human brain and mind physiology and pathology? Like any other function, behavior and brain responses evolve in complexity in parallel with the evolution of species: thus, understanding memories in animals will inform human studies and the design of better therapies. As I mentioned earlier, science typically begins by investigating the simple in order to access and understand the complex. Researchers are aware of what the conclusions, pitfalls, and future questions in memory reconsolidation are. Let’s not throw the baby out with the bathwater.

I conclude with a quotation from one of Freud’s letters to Fliess: “Our psychic mechanism has come into being by a process of stratification: the material present in the form of memory traces being subjected from time to time to a rearrangement in accordance to fresh circumstances—to a retranscription. Memory is present not once but several times over. . . . Successive registrations represent the psychic achievements of successive epochs of life . . .” (Freud 1896, p. 287).