R1. Enduring change in psychotherapy

A major question raised by some of the commentaries is about the scope of what we intended to address with our model. Collectively, commentators wondered whether we were aiming to address all possible treatments for all mental disorders. The answer is, yes and no. The IMM serves as a starting point for understanding the common mechanisms of change among many psychotherapies, but it is not meant to apply to every kind of disorder or symptom. Our focus is on understanding how psychotherapies can result in enduring change, which requires learning. In the context of psychotherapy this learning involves the revision of previously acquired behaviors and emotional responses that became routine over time. This revision of previous learning can occur in several way, including extinction and reconsolidation. Extinction and reconsolidation differ in that only reconsolidation involves updating and revising what was previously learned, whereas extinction involves new learning that overrides, rather than changes, the old. In our view, updating the memory structure through reconsolidation provides enduring change that establishes new and adaptive ways to respond to novel situations.

Mancini & Gangemi question how we can account for the diversity of psychopathology with a “single-process explanation.” Although we discussed four psychotherapeutic perspectives – behavioral, cognitive-behavioral, experiential, and psychodynamic – we expect the basic tenets of the IMM to apply to many interventions, including those discussed by Kimbrel, Meyer, & Beckham (Kimbrel et al.) – dialectical behavior therapy, acceptance and commitment therapy, and interpersonal psychotherapy – and by Spanagel & Bohus, such as general interventions including psycho-education and relaxation training, and disorder-specific interventions such as anti-craving skills in substance dependence, shifting of attention in social phobia, and imagery rehearsal in the treatment of nightmares. To the extent that these therapies have as their primary goal bringing about enduring changes in situational construals and responses that are more adaptive, and to the extent that each of these modalities facilitates the regulation of emotion through a supportive relationship with a therapist, the ingredients are in place to reactivate old memories, try out new ways of experiencing the self in interaction with the external world, and consolidate more adaptive emotional and behavioral responses into an integrated memory structure.

Psychodynamic psychotherapy appears especially well-suited to address the issue of maladaptive behaviors and their enduring modifications. We were therefore somewhat surprised that Levin thought the processes we discussed were not relevant to the day-to-day activities of psychoanalysts. Rather than addressing the evidence we presented and the specific theoretical (sects. 2 and 3) and technical (sect. 9.4) implications for psychodynamic psychotherapy, Levin's commentary almost resembled a psychoanalytic interpretation: that we, the authors, were not consciously aware of the pervasive influence of culture on our ideas and that, unbeknownst to us, we were repeating the same old explanations others had for the past century, however much we might think otherwise. As we stated in section 9.4,

Levin states that we should move beyond the immediate need for methodological validation and bureaucratic credibility. This is not our view. It is important to appreciate that without empirical evidence supporting its efficacy and its mechanisms of action, psychoanalysis, like any other therapy, is in danger of stagnation and ultimate decline. We hope that in some small way this paper contributes not only to its survival but to its growth.

The notion of enduring change also addresses the suggestion of Kimbrel et al. that the common mechanism of change should be a biological one, because pharmacological interventions produce outcomes similar to psychotherapy. Our perspective is a bit different, in that we draw on considerable evidence that the combination of psychotherapy and pharmacotherapy is superior to either treatment alone for most major mental disorders (Cuijpers et al. Reference Cuijpers, Sijbrandij, Koole, Andersson, Beekman and Reynolds2014). This could be because the two classes of intervention target different brain mechanisms that together constitute a more comprehensive and synergistic package (Goldapple et al. Reference Goldapple, Segal, Garson, Lau, Bieling, Kennedy and Mayberg2004). Although pharmacotherapy is often an effective treatment for symptoms such as depression or anxiety, our emphasis has been on the role of psychotherapy in reducing vulnerability to recurrence by instilling enduring changes that are maintained after pharmacotherapy is discontinued. Supporting a biological mechanism, Marks-Tarlow & Panksepp point to the efficacy of deep brain stimulation treatments for depression that involve subcortical stimulation of positive affect as evidence that “bottom-up” (presumably, biologically based) treatments can be therapeutically effective. We would argue that the simple activation of positive affect alone addresses neither the importance of the context in which the positive experience emerges, nor the interpersonal meaning that the experience has for the client. In our view, updating memory structures through reconsolidation is both necessary and sufficient for enduring change to occur in psychotherapy.

Taking a cognitive perspective, Moyal, Cohen, Henik, & Anholt (Moyal et al.) suggest that acquisition of adaptive emotion regulation strategies is a more plausible mechanism of change in psychotherapy than reconsolidation of emotional memories. We agree that change in psychotherapy involves changes in emotion regulation. We would argue that changing the way an individual interprets situations and the way she responds to situations represents the development of antecedent-focused and response-focused emotion regulation strategies, respectively (Gross Reference Gross1998a). Therefore, our model for change in psychotherapy includes an important role for emotion-regulation strategies.

A related issue is whether our approach is relevant for disorders other than trauma-related disorders, a question raised by Kimbrel et al. Indeed, Moyal et al. assume that post-traumatic stress disorder (PTSD) is the primary focus of our paper, and that we assume a traumatic etiology for all psychiatric disorders. This is a misconception of our model. Fundamentally, we are addressing how situations are construed and how one responds to them. Early-life trauma is certainly an important topic here because it can induce strong biases in the ways people construe and respond to situations. However, IMM describes a general process whereby emotionally charged situations induce cognitive and emotional adaptations that have enduring effects on patterns of behavior, thought, and feeling. Moyal et al. suggest that what we mean by therapeutic change is a reduction in clinical symptoms. On the contrary, in our view the goal of psychotherapy is, to put it simply, to promote a more adaptive situational response. This may result in not only a reduction in symptoms, but also an improvement in the ability to achieve one's goals in major life contexts such as work, interpersonal relationships, and leisure pursuits.

Ainslie is in agreement with Moyal et al., as they state that the role of emotional trauma in psychopathology is limited and that putting traumatic memories at the core of pathogenesis overstates the case. Ainslie places less emphasis on how change occurs than on describing the cognitive mechanisms – specifically, the conflict between urges and self-control – that maintain and exacerbate mental symptoms such as phobias or addictions. But our paper is primarily about the mechanisms of change in psychotherapy, not the etiology of psychopathology. Our core assertion is that emotional arousal and memory reconsolidation are primary mechanisms by which enduring change occurs in psychotherapy. Adaptation to psychologically overwhelming experiences (trauma) is an important contributor to psychopathology, but not to all psychopathologies. Many other factors likely contribute to pathogenesis, including genetic and other biological factors, gene–environment interactions (Caspi & Moffitt Reference Caspi and Moffitt2006), and modeling and other contexts that are not traumatic in nature (Kendler Reference Kendler2012). The question of why one person becomes borderline whereas another becomes agoraphobic, for example, is unanswered at present, and is likely to remain so in the foreseeable future. As we said at the outset, it is not our intention to account for all psychopathology or to account for the pathogenesis of all mental disorders.

R2. On the nature of memory

Given the central role that memory and memory processes play in our model of therapeutic change, it is not surprising that many commentaries raised issues regarding this core element. Some commentators (Montemayor and Solms) focused on the question of what memory is for, each in his own way emphasizing the idea that memory provides the basis for predictions that guide future behavior, and that this gives prediction error, or novelty, a central role in learning (a point also emphasized by Ecker, Hulley, & Ticic [Ecker et al.]). We completely agree with this perspective and would even go beyond these comments to point out that all forms of memory, not just episodic memory, are used for predictive purposes; for example, semantic memory, too, plays a critical role in prediction.

As we discussed at length in our paper, and echoed by Klein & Markowitsch, the dividing line between episodic and semantic memory is getting harder and harder to see. Semantic knowledge can be conceptualized as probabilities, based on past experience, that the world works a certain way. We learn the statistics of the purely physical world of the sun, moon, and planets, and we also learn the statistics of the behavioral world, leading to predictions about the behavior of others, and to the knowledge, for example, that relationships at times do not end well. In our target article, we emphasized the interactive nature of semantic and episodic memories in interpreting the world around us and guiding our behavior. We note, as an historical aside, that O'Keefe & Nadel (Reference O'Keefe and Nadel1978), in laying out their thoughts about the hippocampus and its critical role in learning and memory, emphasized exploration and novelty as the key to understanding this system.

Several commentators thought we should pay less attention to memory and more attention to the self (Montemayor and Whelton), or to autonoetic consciousness (Klein & Markowitsch). We respectfully disagree. Both conscious and unconscious access to memories are important in understanding the impact of prior experiences on future behavior. In many forms of therapy, making the noetic become autonoetic may be helpful, but it does not appear to be necessary for psychotherapeutic change. Although the noetic–autonoetic distinction speaks to the level of subjective access to prior experiences, it does not undermine the basic distinction between semantic structures (knowledge) versus memory for unique past episodes, a dissociation well documented in patients with bilateral damage to the hippocampus (Nadel & Moscovitch Reference Nadel and Moscovitch1997). Our focus is on the processes and mechanisms of enduring change, not on the content of the change or how that content is conceptualized. It is not unreasonable to think that the self is constituted, at least in part, of memory for prior events and emotions, and that revising memories and emotions will revise aspects of the self as well.

R3. Updating memories: How, why, and when?

A number of commentators asked what is essentially the following core question: When a memory updating event occurs, is the original memory overwritten, or erased, or transformed (Brewin, Montemayor, and Roache)? In our view, each of those things can happen, depending on the circumstances and the kind of memory one is talking about. For the kind of autobiographical memories psychotherapy is most often concerned with, the most common outcome is transformation, which is exactly what the therapeutic process is intended to accomplish. Memories are not replaced; they are revised by incorporating new information into an integrated trace through the corrective experience. Transformation does not imply having “no trace” of the original memory, as Brewin suggests. Although the core of the original memory may remain, the details of the event and their connection to emotional responses may change considerably. In this way, the old memory becomes unavailable in its earlier form. But that is not the same as the trace being lost or overwritten. (An interesting incidental comment by Marks-Tarlow & Panksepp referred to therapists who “explicitly wish to promote reconsolidation.” We should make clear that memory reconsolidation occurs automatically without intentional promotion by the therapist.)

There is a wealth of evidence supporting the reconstructive nature of memory (reviewed in our target article), although the degree to which our memories can be transformed is sometimes difficult for people to accept. Montemayor raises the issue of the need for balance between accuracy and malleability of memories, which we agree is an area of inquiry that should be encouraged. Neisser's (Reference Neisser1981) notion of “repisodic memory” is relevant here. Neisser argued that though the details of an event may morph considerably over time, the core of a given memory remains intact.

The issue of “truth” in memory is often controversial. Is it ethical to “replace” a “true” memory with a transformed one that is clearly not the “truth?” Is it “adaptive” (Montemayor)? For example, Patihis suggests that memory reconsolidation is equivalent to memory distortion, essentially creating “false memories,” and questions whether it is ethical either to the client or to family members. That such processes can be mishandled and can be potentially damaging is certainly true, as demonstrated by numerous stories of falsely recovered childhood memories during psychotherapy, sometimes leading to terrible outcomes for the client and their families (Loftus & Ketcham Reference Loftus and Ketcham1996). This is a complex issue, but it is important to start from the appropriate point – there is no such thing as a “true” memory in the first place. Memory of a past life event is almost always a reconstruction, more or less correct and subject to significant distortion. Some memories, perhaps our most salient ones, are retained with apparently clear and truthful detail, but this is very much the exception. Some have speculated that these exceptional cases are retained in detail precisely because they are highly diagnostic (Klein et al. Reference Klein, Cosmides, Tooby and Chance2002), and these may or may not be the ones we care about in therapy. Transformation of memories happens all the time, like it or not, and the point of psychotherapy should be to harness this naturally occurring phenomenon for good purpose. Nevertheless, there is clearly a need for rigorous training in psychotherapy before one should be allowed to practice. The situation may be analogous to the potential benefits or harm that can result in general medicine or surgery depending upon the quality of the practitioner's training and level of expertise.

Ecker et al.'s commentary addresses the need to better understand the circumstances required for reconsolidation to take place – that is, why a memory becomes unstable, or to use their word, deconsolidated, setting the stage for transformation through reconsolidation. They point out that emotional arousal per se is not required for inducing reconsolidation, and we would agree with that statement. Our model allows for multiple routes to change, including semantic structures (as would be emphasized in cognitive behavioral approaches) and the experience of new, contradictory, episodic events. However, in the context of psychotherapy, where the focus is overwhelmingly on emotional experiences and emotional responses to the world, emotion becomes a key route for change. We marshal a great deal of evidence in support of the idea that emotional arousal is a key component of psychotherapy, and that the corrective emotional experience in therapy is a major part of what is reconsolidated.

Whether there are circumstances that optimally induce deconsolidation in other contexts is a very interesting research question that warrants investigation. Ecker et al. suggest it is only under circumstances of cognitive mismatch that memories are destabilized and available for transformation. Intuitively, it makes sense that memories are updated only when there is new information available to update them. Our view is that a corrective emotional experience represents a very salient kind of cognitive (and emotional) mismatch, such as when the therapist responds in a way that is clearly contrary to the expectations of the client. Levenson (Reference Levenson, Elkman and Davidson1994) long ago emphasized that emotion is activated whenever there is a change in the interaction between the person and his or her current situation that has implications for that person's needs, values, or goals. According to Diekelmann & Forcato, and Stein, Rohde, & Henke (Stein et al.) this is exactly the type of situation that facilitates memory reconsolidation. Physiological arousal may be an important attribute used by memory systems, not only to distinguish what is important to remember for survival or adaptation versus what is not, but also to determine what justifies revision of established memory. This certainly leaves open the possibility that purely cognitive mismatches contribute to memory reconsolidation, but it also suggests that emotional responses may provide a particularly potent type of mismatch that leads to memory revision.

In the end, the similarities between Ecker et al.'s formulation (Ecker et al. Reference Ecker, Ticic and Hulley2012) and ours are notable. They, like us, state that there are three essential steps to creating change. Their first step involves explicit recall of the problematic memory and its current expression and feeling it emotionally. Our first step involves activating the problematic memory but not necessarily explicitly recalling it. Their second step involves conscious recognition of mismatch or disconfirming information and feeling the experience of juxtaposing that information with the original memory in step one. Our second step involves activation of new emotional experiences (a critical step also emphasized by Pascual-Leone & Pascual-Leone) that allows the situation to be experienced and understood in a different way. Their third step involves repeated juxtapositions of the contradictory information in steps one and two. Our third step involves repeatedly experiencing and “working through” the emotional consequences of new learning in a variety of contexts. Further research that evaluates the similarities and differences in these models will likely be useful to the field.

Several other commentators discuss proposed mechanisms for memory change that have aspects in common with our IMM. We welcome the comparisons across models that may lead to a deeper understanding of this issue. For example, Llewellyn provides a nuanced discussion of the term reconsolidation, preferring instead reassociation,based on evidence derived from research on sleep and dreaming, and emphasizing the importance of both integration and segregation for understanding memory change. We do not disagree that reassociation, which appears to involve transformation of semantic knowledge structures, can be an important source of enduring change. But it is not the only kind of transformation possible. Our use of the term reconsolidation is meant to encompass a number of ways in which memories and knowledge structures can be changed. In this sense, our model is in line with Llewellyn's ideas, and it will be strengthened by considering the details of the processes involved in memory transformation and the impact of the corrective experience.

Relatedly, Reyna & Landa emphasize the considerable overlap and specific areas of separation between fuzzy trace theory and our proposed framework. One area of convergence is that our focus on the transition from episodic to semantic memory involves extracting common or gist elements from personal experiences. Although Reyna & Landa emphasized the distorting aspects of emotion on memory, their commentary raises the question of whether corrective emotional experiences consisting of new primary emotion might recast the gist of past experiences in a way that would promote development of a healthier sense of self.

R4. The emotional arousal component of the IMM

Several of the commentaries seem to suggest that we were proposing emotion as the only avenue to therapeutic change (e.g., Ecker et al. and Patihis). That is not the case. Our model is interactive and integrated, such that memories, semantic structures, and emotions are engaged equally when any one component of the model is accessed. We go to great lengths to point out that there are multiple routes to change, all leading to updating of the entire model. Nevertheless, activating or modulating emotions likely increases the efficiency with which change comes about, regardless of the emphasis of any particular psychotherapeutic method. So, emotions do matter.

Counter to our position, Patihis points out that cognitive-behavior therapy (CBT) does not aim at high emotional arousal, nor does it always induce it, and yet CBT works well. We are aware that CBT, and more specifically cognitive therapy (Beck Reference Beck1979), is the only one of the four modalities we discussed that often does not focus on emotional arousal as an ingredient of change. However, it is also true that the effectiveness of CBT treatment for anxiety and depressive disorders may not be enduring, at least in some cases. For example, Vittengl et al. (Reference Vittengl, Clark, Dunn and Jarrett2007) reported that the mean proportion of patients who experience relapse or recurrence after receiving acute phase cognitive therapy was 29% in the first year and 54% in the second year. In section 9.2 of the target article we explain how changes in the way problematic situations are construed as a result of CBT could lead to reduction in emotional arousal associated with those situations. We suggest that these changes in arousal (and valence) might be incorporated into the memory structure through reconsolidation. Thus, we propose that a change in arousal might be a mechanism of change in CBT rather than an outcome. We appreciate that it is often not customary to view emotional arousal as playing a causal role in how CBT works. We present this perspective through the lens of the IMM as a testable hypothesis for future research.

Marks-Tarlow & Panksepp suggest that we advanced a “top-down” approach to emotion and memory reconsolidation, asserting that we failed to give adequate credence to the subcortical origin of emotions. This was a surprising characterization to us and does not fit the mechanisms that we describe. Their position is actually consistent with classic psychoanalytic concepts of emotion (Solms & Panksepp Reference Solms and Panksepp2012), but it is one with which we take exception based both on the neuroscientific evidence that we reviewed in considerable detail in section 2 and for important clinical reasons. As we discussed in section 2, clients who have been abused may have emotional responses at the time of the trauma that are diffuse, undifferentiated, and high in arousal, and indeed, those clients are often unaware of how they felt at the time. It is often only in therapy that they can formulate for the first time, and experience for the first time, the full range of emotions related to the trauma (if one is cowering in fear, one does not feel capable of angrily fighting back and does not experience anger, although it may be activated to some degree). Therefore, we agree that emotions originate subcortically and in fact often have a bottom-up trajectory, but we would argue that in many cases emotional experience is formulated and created for the first time in therapy (Levine Reference Levine2012), not simply recovered by undoing defenses that conceal the presumably well-differentiated emotions.

We must point out that there is in fact very little difference between Marks-Tarlow & Panksepp's view of the subjective experience of emotion in animals and our view of implicit emotion in people. Our views on the nature of animal emotional experience are consistent with those of LeDoux (Reference LeDoux2012), who proposes that humans and other living beings share survival circuits (neural mechanisms mediating basic life functions such as feeding and reproducing), and that conscious emotional feelings occur when activation of these survival circuits interacts with the mechanisms for conscious processing. Panksepp's theory of what animals experience when subcortical affective circuits are activated (Panksepp & Biven Reference Panksepp and Biven2012) is nonspecific and consistent with the interoceptive awareness of bodily sensations that people can experience during the activation of implicit emotions. We would also point out that many of Panksepp's basic affective circuits include connections between subcortical nuclei and structures such as the anterior cingulate cortex (Panksepp & Biven Reference Panksepp and Biven2012), raising questions about why the subcortical structures alone are thought to mediate conscious experience when they are in direct communication with phylogenetically newer paralimbic structures that in human studies correlate with subjective experience (Medford & Critchley Reference Medford and Critchley2010).

Mancini & Gangemi question the fundamental importance of the transition from implicit to explicit emotional responses, because subjective reports of anxiety often do not correlate with objective measures of physiological arousal. Certainly, the topic of the variable relationship between subjective emotional experience and physiological indices of arousal in normative or clinical contexts is one that is widely debated and for which no clear consensus has been achieved (Critchley & Harrison Reference Critchley and Harrison2013). These authors also find it paradoxical that we espouse promoting emotional awareness when many clients, such as those with anxiety disorders, appear to suffer from an overawareness of their anxiety or related symptoms. We recognized this issue in the target article by pointing out the deleterious effects of repeatedly rehearsing the same distressing experiences from the past. Instead, we emphasized the importance of accessing and becoming aware of new emotional information, such as experiencing a therapist's encouraging responses when criticisms and negative judgments were anticipated. In short, our goal was to emphasize the usefulness of “corrective emotional experiences,” not emotional awareness of any and all kinds.

R5. The neural bases of psychotherapeutic change

Our position is that enduring changes occur in psychotherapy that are manifested emotionally, cognitively, behaviorally, and physiologically in the brain. Roache appreciated the coherent, multilevel nature of our model of change in psychotherapy. Several authors, however, commented on the pressing need to further our understanding of the level of brain mechanisms mediating these changes, and we agree. In his commentary, LaBar points out the limitations of our current understanding of the interaction of the neural circuitry of emotions, centered on the amygdala, and those mediating episodic memories. In psychotherapy, we understand that this intersection is key to real behavioral change. We agree with LaBar that the field would benefit greatly from a broader neural systems approach that integrates emotional memory, emotion regulation, and reconsolidation. Similarly, Ortu brings our attention to the specific neuromodulatory mechanisms that underlie behavioral learning following an aversive event. This leads to the intriguing idea that the therapeutic session may be ideally suited to positive change because the therapist can directly manipulate specific learning contingencies, contingencies that are unlikely to be manifest in the real world.

The influence of sleep on memory mechanisms is another area of cognitive neuroscience with great potential. We pointed out that one of the implications of our model is that napping or sleeping after a therapy session could potentially enhance reconsolidation of memories and that new information added to the original memory could include emotional information from the therapy session. Diekelmann & Forcato report evidence from exposure therapy treatment of spider phobia that sleep may specifically support the strengthening and updating of emotional memories. This point was further amplified by Stein et al., who report on a series of experimental studies demonstrating that emotional content not only enhances memory consolidation but also has a stronger effect than emotionally neutral content on boosting memory reconsolidation.

Taking the contrary position, Spanagel & Bohus question the utility of memory reconsolidation. They argue that reconsolidation effects have been difficult to demonstrate consistently in the laboratory, and that clearly defined determining factors and experimental conditions that might inform therapy have yet to be identified. Liberzon & Javanbakht point out that memory reconsolidation was initially described in animal models and in relationship to relatively simple memory traces of fear conditioning. They suggest that more complex memory systems in humans, particularly for traumatic events, might work differently. Patihis points out that there is insufficient evidence that changing memories causes improvement in therapy (see also Ainslie and Spanagel & Bohus for similar comments). We discuss the evidence for reconsolidation in humans in our article, including its relevance to PTSD. Although we recognize the challenges in studying reconsolidation, particularly as it applies to human memory, it is our opinion that further research on the neural mechanisms of reconsolidation and the relationship between reconsolidation and therapeutic improvement is important. One of the challenges in such research is that the usual short-term endpoints after treatment (e.g., three or six months) may not be sufficient to capture the enduring changes that we are seeking to explain.

R6. Misconceptions about the IMM

Moyal et al. point out that CBT, emotion-focused therapy (EFT), and other modalities focus on the here and now and do not specifically endorse recollection of past events for the purpose of understanding the developmental origins of current difficulties. Here we must reiterate that it is not necessary to explicitly recall the memory to engage in revision through reconsolidation, but rather it is necessary to reactivate the memory trace through reminders that reference the original memory, regardless of awareness. Thus, having experiences with the therapist that are counter to expectations, in which the expectations are a product of the influence of an old memory, is sufficient to engage reconsolidation.

An important question raised by Liberzon & Javanbakht is whether “contextualization” – that is, the process of updating old memories by adding novel contextual elements – is actually independent of reconsolidation, because it does not require that the old memories become unstable. We believe that recalling old memories with some details of the original context puts that memory into a labile state that allows new contextual elements to be added to the original memory through reconsolidation. The specific components or the extent of a retrieved context that is necessary to shift the memory into a labile state (referred to as deconsolidation by Ecker et al.) is an empirical question that has yet to be answered.

von Hecker, McIntosh, & Sedek (von Hecker et al.) note that depressed individuals show limitations in their ability to create and update mental models of their social world. They propose that mental model construction, not just memory, is a central feature of cognitive change in psychotherapy. Although they present no evidence to support their supposition, their approach is representative of cognitively based treatment modalities that focus exclusively on cognitive mechanisms as opposed to a reconsideration of the past. In our view, a mental model is a semantic memory structure. If it is to influence future behavior, a change in an integrated memory-emotion-semantic model must be retained so that it can affect construals of social situations in the future. Thus, we would take issue with the position of von Hecker et al. that a focus on memory excludes mental model construction. We would emphasize once again that memory structures can be activated and modified without explicit recall of the past.

R7. Enhancing therapeutic change

Finally, a number of commentaries provided varied and interesting discussions of factors and methods that are generally consistent with our formulation but may further enhance the therapeutic outcome. In this section, we briefly describe how these proposals mesh with the IMM model.

Mann, Cone, & Ferguson (Mann et al.) point out that implicit attitudes are highly relevant to the situational construals that are addressed in psychotherapy. Based on their experimental findings, they contend that implicit attitudes can be revised strongly, quickly, and durably if individuals believe that new information is true and predictive of what will likely happen in the future. In our model, we emphasize the need to expand these experiences to other contexts. One important way this expansion occurs is through the therapist's facilitation of the client's own experiences outside of therapy. The latter shifts the focus from the client's belief in the therapist to the client believing what her own newly articulated experience is telling her about herself. Of course, validation and support by the therapist is an important part of the process of the client coming to accept and integrate this new information, but the critical discovery and articulation of the new experiences often originate with the client, not the therapist.

The commentary by De Brigard & Hanna highlights the use of counterfactual thinking and the emotional changes that it induces, both in the context of CBT and in daily life; that is, the tendency to experience positive affect when repeatedly thinking about the implausibility of alternative outcomes of past personal experiences. This idea reinforces and extends our proposal that a critical ingredient of change in CBT is the emotional state that results from the alterations in thinking that are promoted by CBT. Thus, the phenomenon that we describe in relation to CBT may be a more general phenomenon linked to the post-processing of emotional events that helps healthy individuals stay healthy.

The concept that the IMM is relevant to normative functioning was expanded by Benga, Neagota, & Benga (Benga et al.), who point out an intriguing connection between our view and anthropology. These authors raise the possibility that the processes we describe as essential ingredients for enduring change in psychotherapy also apply to cultural rites of passage, which constitute culturally created methods of facilitating developmental transitions across the life cycle. Consideration of whether IMM applies to a range of cultures is potentially important, particularly in light of the comments of Levin, who expressed concern that our model might reflect a particularly Judeo-Christian orientation to human experience.

Several promising suggestions are discussed by Staniloiu & Zaretsky regarding alterations in retrieval operations, such as enhancing the capacity to retrieve positive emotional memories, and maintaining a first-person perspective during recollection of both positive and negative memories. Although our model focuses on the mechanisms of change (encoding), there is no doubt that considering both encoding and retrieval mechanisms will result in a more comprehensive understanding of how memory change occurs in psychotherapy, a point also emphasized by Brewin. Focused research that compares and contrasts the relative merits of competitive retrieval versus reconsolidation in a psychotherapy context would be very useful.

Whelton's commentary emphasizes the positive impact of interpersonal relationships, not only in real life but also in the client–therapist relationship (in contrast to Mancini & Gangemi's discussion of the role of negative interpersonal relationships in maintaining dysfunction). This is an intriguing idea; these social interactions may play a crucial role in memory transformation when stories are repeatedly recollected and shared with others, as we have argued elsewhere (Ryan et al. Reference Ryan, Hoscheidt, Nadel, Dere, Easton, Huston and Nadel2008c). A more thorough understanding of the qualities of social interaction that lead to positive transformation of painful and distressing memories could be extremely important for understanding not only the therapeutic process but also the social circumstances that allow individuals to cope with adverse life events.

Individual differences in the clinical response to psychotherapy is one of the most important, and poorly understood, issues in the field today. Kimbrel et al. raised the critically important question of how the proposed model accounts for individual differences in treatment response. Here we would focus on individual differences in the capacity to process emotional information, as addressed by Pascual-Leone & Pascual-Leone. The ability of clients to experience differentiated emotions and describe their experiences in words is a predictor of successful treatment for panic disorder either with CBT or manualized psychodynamic psychotherapy (Beutel et al. Reference Beutel, Scheurich, Knebel, Michal, Wiltink, Graf-Morgenstern, Tschan, Milrod, Wellek and Subic-Wrana2013). In our target article we discussed the inverted-U shaped functional relationship between arousal and the ability to articulate one's own thoughts and feelings as well as those of others. Shifting arousal level so that the client is functioning at the peak of the inverted-U function may contribute to greater awareness and emotion processing that can contribute to the propensity for corrective emotional experiences. The same may apply to understanding individual differences in vulnerability to disorders, such as PTSD. Only about 20% of traumatized individuals develop PTSD (Admon et al. Reference Admon, Milad and Hendler2013). Consistent with the inverted-U relationship just mentioned, it has been proposed that complex childhood trauma is associated with a “window of tolerance” between sympathetic-dominant hyperarousal and parasympathetic-dominant hypoarousal (Corrigan et al. Reference Corrigan, Fisher and Nutt2011). Functioning within this window may enable subjects to self-regulate more effectively than at the extremes on the arousal continuum and could conceivably play a preventative role. A more thorough understanding of the biological, psychological, and social factors that contribute to individual differences in resilience and vulnerability are needed for the field to make progress.