To Expose or Not to Expose: A Comprehensive Perspective on Treatment for Posttraumatic Stress Disorder

Abstract

Trauma-focused psychotherapies, in particular Prolonged Exposure therapy, have been recognized as the “gold standard” for the treatment of PTSD. But effectiveness and implementation data show that a large proportion of patients who undergo exposure therapy retain their PTSD diagnosis, and implementation studies have shown low engagement and high dropout rates. Meanwhile, non-trauma-focused therapies have shown promise in treating PTSD. In this review, we aim to answer the question of whether exposure is necessary to treat PTSD by integrating clinical and research literature from multiple perspectives. We review the roots of exposure therapy in both psychodynamic and behavioral paradigms and their proposed mechanisms. We then review non-trauma-focused treatments and their proposed mechanisms. We conclude that the specific form of exposure required by PE is not necessary for symptom remission. Finally, common psychotherapy factors may facilitate patient self-directed exposure outside of the therapy context. These findings should alter the direction of clinical research to identify the therapy processes that most effectively promote processing of trauma memories. With respect to clinical practice, shared decision-making should allow for increased patient autonomy in choosing either trauma-focused or non-trauma-focused treatments.

Posttraumatic stress disorder (PTSD) is a prevalent, debilitating, and often chronic condition experienced by approximately 8% of the US adult population (Kilpatrick et al, 2013). While PTSD was long considered intractable, trauma-focused cognitive behavioral treatments developed in the 1980s and 90s reported significant therapeutic gains and thus were recommended as first-line treatments by numerous organizations (APA, 2017; ISTSS, 2018; NICE, 2018; VA/DoD; 2017). APA treatment guidelines state that “the shared element of controlled exposure may be the critical intervention,” and experts concur that exposure to trauma memories is a common factor underlying effective psychotherapy for PTSD (Schnyder et al, 2015).

However, there is a growing consensus that current treatment approaches are insufficient. While randomized controlled trials (RCTs) demonstrate evidence of clinically meaningful symptom improvement (Mclean et al., 2022), they also reveal high levels of dropout (Hembree et al., 2003), greater than 50% retention of PTSD diagnosis (Cusack et al., 2016), and high levels of residual symptoms (Larsen et al., 2019). Results show even lower engagement and remission in military veterans (Steenkamp et al., 2015). Exposure is viewed by clinicians as potentially aversive or damaging, especially for the complex, multiply traumatized populations they treat (Becker et al., 2004; Doran et al., 2019; Markowitz, 2010; Najavits, 2015). Further, the publication of the APA treatment guidelines in 2018 generated controversy among clinicians, including the leadership of the guideline development, who contended that the process relied on an overly narrow conception of empirical support that prioritizes RCTs over clinical experience (Curtis & Brown, 2019). RCTs tend to exclude clinical populations experiencing substance use, suicidality, and unstable living situations such as homelessness and domestic violence, biasing the accrual of evidence toward specific interventions and against the flexible and response application of clinical judgment to complex cases (Norcross & Lambert, 2018).

In addition, recent research findings have supported the application of more general, non-exposure-based treatments to PTSD, such as Present-Centered Therapy (PCT), Interpersonal Psychotherapy (IPT), Acceptance and Commitment Therapy (ACT), relapse prevention, and non-trauma-focused CBTs (Hamblen et al, 2019; Lenz et al, 2017; Shea et al, 2020). These findings accord with the general trend, observed for the last half-century, that bona fide treatments do not differ greatly in their efficacy and that specific therapy techniques account for a small proportion of treatment outcome variance (Luborsky et al., 1975; Wampold et al., 2010).

This paper therefore aims to examine the following question: given current knowledge about the development, maintenance, and treatment of PTSD, to what extent is exposure necessary to overcome trauma and achieve remission from PTSD? To address this question, we (1) define exposure, (2) review the history and conceptualization of exposure therapy for PTSD, (3) examine proposed biopsychosocial mechanisms of exposure, especially in light of neurobiological findings regarding traumatic memory, (4) address potential clinical rationales for non-exposure-based treatments. The paper is a qualitative review that does not include original empirical research, a quantitative meta-analysis, or previously unpublished data. We conclude that exposure to trauma memories remains a central component in the treatment of PTSD, albeit within a broader conceptualization of the PTSD syndrome and clinical change.

Defining Exposure

Exposure therapy has been defined as “deliberate, systematic confrontation with stimuli (eg. situations, objects) that are feared despite being safe or having low probability of producing harm” (Foa et al., 2007, p. 13). This fear-based model arises out of a paradigm of Pavlovian conditioning applied to the field of behavioral therapy for anxiety disorders. Exposure therapy seeks to correct pathological anxiety by repeatedly exposing the patient to the feared stimulus, using imaginal and in vivo exposure.

Prolonged Exposure therapy for PTSD employs the most prolonged, intensive, and thorough exposure to trauma memories and cues. While most trauma-focused treatments involve some elements of exposure, their proposed mechanisms of therapeutic action differ, leading to differences in the type, frequency, and dosage of exposure that is prescribed.

It is also important to note the distinction between exposure as a therapist technique and exposure as a patient change event. While exposure therapy is “a form of treatment in which people are introduced to situations they fear or experience as aversive,” the exposure process involves “experiencing and tolerating a feared or anxiety-provoking stimulus in the absence of the expected aversive outcome” (Muir & Hibberg, 2019, p. 86). Thus, therapist-directed exposure may not necessarily result in the exposure process, while the exposure process may occur outside of therapy. Exposure therapy may fail when therapist-directed exposure does not optimally bring about a patient change event, e.g., because of patient avoidance or overengagement.

Conversely, the process of exposure may occur outside of therapy. Foa and Kozak’s theory of natural recovery from PTSD posits that most people who experience a trauma overcome fear as they are exposed to trauma cues in the course of everyday life, as well as processing the trauma with trusted others (Foa & Kozak, 1986). This raises the possibility that psychotherapy for PTSD may work by facilitating patients’ exposure outside of therapy, without using exposure as a therapist technique.

History of Exposure Therapy

Exposure therapy for trauma originated in two different contexts that converged in the 1980s with the creation of PE. The first context is the treatment of psychological trauma, postulated by Freud. The second context is that of exposure therapy for the anxiety disorders, which originated in behavioral learning theory.

Psychodynamic Origins

The modern history of treatment for psychological trauma originated in nineteenth century Victorian Europe, where the neurologist Jean-Martin Charcot sought to understand and treat hysteria, the prevailing malady of the day (1877). Charcot suggested that the symptoms of hysteria were rooted in psychological trauma. His students, Pierre Janet and Sigmund Freud, devised parallel formulations and treatments for hysteria (Craparo et al., 2019).

Janet suggested that hysteria resulted from the inability to integrate traumatic memories into autobiographical narrative (1925). He postulated that the strong emotions experienced during a traumatic event disrupted the mind’s ability to synthesize and integrate new information. Traumatized patients are unable to narrate the traumatic event as a typical autobiographical memory, but memories intrude into consciousness in dreams or re-enactments. These features of traumatic memory that have been confirmed by current research (LeDoux, 1998; Brewin, 2001).

Janet proposed that in order for the patient to be cured of a fixed idea, the nature of the dissociated event must be brought back into consciousness, the vehement emotions processed, and the meaning of the event integrated into the personality. Processing of the trauma memory alone was not enough to achieve a cure. Rather, treatment proceeded in three stages: stabilization and symptom reduction, modification of traumatic memories, and relapse prevention (Janet. 1925). In the second, exposure-based, stage, traumatic memories were brought into consciousness and altered in what Janet termed liquidation, substitution, and reframing. In liquidation, the patient was hypnotized, which decreased resistance to revisiting the traumatic event, and a verbal memory of the event was constructed. In patients who could not tolerate this process, Janet used substitution, in which he evoked the traumatic memory and then transformed the image into a neutral or positive one. In reframing, which Janet used for patients who had difficulty accessing emotional experience, the focus was on reinterpreting the meaning of the traumatic event, alleviating self-blame and making the memory more acceptable to the patient.

Freud and Breuer built on Charcot’s and Janet’s work in their Studies in Hysteria, which contains the famous declaration that “hysterics suffer mainly from reminiscences” (1895/2004, p. 7). Like Janet, Freud noted that traumatic memories differ from typical autobiographical memories: they “persist for a long time with astonishing freshness and with the whole of their affective colouring,” yet the memory is “only present in a highly summary form” (1895/2004, p. 8–9). Freud posited that traumatic events cause an alteration in consciousness, which he called a “hypnoid state,” which leads to a “inhibition of reaction,” a paralysis of thought, emotion, and action. But Freud diverged from Janet in emphasizing the role of catharsis rather than synthesis in resolving trauma.

Freud proposed the inhibition of a response during a traumatic event led to the retention of its affective charge and the fixation on the past until catharsis or abreaction became possible. Catharsis could take place through words, as follows: “Each individual hysterical symptom immediately and permanently disappeared when we had succeeded in bringing clearly to light the memory of the event by which it was provoked and in arousing its accompanying affect, and when the patient had described that event in the greatest possible detail and had put the affect into words… it subjects it to associative correction by introducing it into normal consciousness.” (1895/2004, p. 6)

Though it comes from a different theoretical frame, this description bears remarkable similarities to modern exposure therapy. Freud’s emphasis on emotional activation and expressing affect through speech are similar to the principle of emotional processing in PE (For & Kozak, 1986). The emphasis on creating a coherent narrative from the raw, unformulated, trauma memory dovetails with the goals of PE and other narrative therapies (Foa & Kozak, 1986). Finally, the concept of associative correction bears similarity to cognitive approaches which seek to correct overaccommodated ideas, (Resick et al, 2016), as well as the correction of fear overgeneralization in PE.

Like modern trauma therapists, Freud found that the treatment of hysteria was not always resolved by catharsis, because patients resisted revisiting the affect associated with trauma memories. To combat this, Freud developed the technique of interpreting the resistance: “the doctor uncovers the resistances which are unknown to the patient; when these have been got the better of, the patient often relates the forgotten situations and connections without any difficulty” (1914, p. 150). Thus Freud came to focus not just on directing patients to recount the trauma memory, but also grappling with what is preventing the patient from confronting the trauma in the first place. Freud also noticed that in many cases, the patient was not able to consciously recall the traumatic event; rather, the relational dynamics associated with the trauma were recreated in the relationship between analyst and patient (1914).

After the political significance of hysteria receded in the nineteen-teens, the next major surge of interest in trauma occurred around soldiers in the World Wars. During World War II, psychiatrists began to use pharmacological agents, such as sodium amytal, that created an altered state of consciousness, in order to facilitate the recovery and abreaction of traumatic memories (Herman, 1992). Abram Kardiner, a psychoanalyst and student of Freud who treated soldiers from both wars, reported his conceptualization and treatment in The traumatic neuroses of war (1941). Kardiner posited that trauma overwhelmed and paralyzed the ego, inhibiting the usual functions of adaptation and response. Thus patients were stuck continually re-enacting the event in order to belatedly respond to it, as they withdrew from the world to protect themselves.

Kardiner posited that the time period immediately after the trauma was critical. Most traumatized people exhibit an acute trauma reaction, including hypervigilance and avoidance. But only if the patient indulged “his now budding tendencies to retreat,” would the reaction become chronic (1941, p. 217). Chronic withdrawal altered the patient’s “conception of the outer world” as well as “his conception of his own capacities to deal with it” (1941, p. 232). Kardiner considered Freud’s cure by abreaction to be effective only in acute cases of trauma, while in chronic cases, “the whole ego structure has been altered… and the abreaction, since alone it cannot alter the new ego structure of the traumatic neurosis, is quite irrelevant” (1941, p. 216).

Kardiner’s treatment consisted of a regimen of rest and isolation, followed by reconstruction of the traumatic memory. Through learning to tolerate reconstructing the details of the traumatic event, the patient was able to realize that his “defensive devices” (hypervigilance) were a response to the traumatic past, and not to the “actual world in which he was living” (1941, p. 224). Kardiner emphasized that therapy was effective not because of memory recovery or abreaction, but through more realistic program of rehabilitation. He was also aware of the potential for dropout in trauma treatment: “one must leave the initiative with the patient, for he knows his tolerance better than you do. His reactions must be respected, for they have a defensive intent, or the case will be lost to treatment” (1941, p. 220).

All of these clinicians posited the importance of exposure to the traumatic memory. However, they differed in their proposed mechanisms of cure. For Janet, dissociated memories needed to be brought into conscious memory so that they could be integrated into autobiographical narrative and the patient could achieve a more benign interpretation of the event. For Freud, exposure was necessary to achieve emotional catharsis. For Kardiner, recovery of the memory was necessary so that the patient could learn to distinguish between the traumatic past and the reality of the present. All three thinkers recognized the unique nature of traumatic memory, presaging modern neurobiological findings. They noted that traumatic memories were created in different mode of consciousness, leading to their being stored in a way that is narratively fragmented and incomplete, but affectively persistent and fresh. They recognized that the ability to narrate traumatic memory is a laborious and affectively charged process of reconstruction. These foundational theories also presaged the difficulties still encountered in exposure treatment: patient avoidance, dropout, and residual symptoms (Cusack et al., 2016; Hembree et al., 2003; Larsen et al., 2019).

The 1970s: Vietnam and Rape Trauma Syndrome

In the 1970s, American social movements renewed the political awareness of trauma, particularly combat trauma for men and sexual trauma for women. Within the VA system, Shatan and Lifton created “rap groups” for returning war veterans, in which veterans would share their experiences and raise awareness of the realities of war (Herman, 1992). This treatment prized social support and solidarity over psychiatric technique, but it did involve a form of exposure in the telling of war stories. In the 1970s, psychologists and sociologists collected the first data on the prevalence and devastation of sexual assault. In the wake of increased awareness of trauma, numerous clinical and research groups developed conceptualizations of and treatments for trauma-based disorders, culminating in the inclusion of PTSD in DSM-III in 1980. The 1990s saw the publication of several integrative summaries of clinical approaches (Herman, 1992; van der Kolk et al, 1996). Herman’s Trauma and recovery prescribed three stages resembling Kardiner’s: stabilization and the establishment of safety, processing trauma memories, and rehabilitation of the personality (1992).

Behavioral Treatments for Anxiety Disorders

The second context for the development of therapies for PTSD is the history of behavioral therapy for anxiety disorders. This field emerged from the advent of learning theory in the early twentieth century. In the 1950s and 1960s, therapists began to conceptualize anxiety disorders as based on classical and operant conditioning. At first applied to specific phobia and OCD, these principles were applied to the emerging field of PTSD in the 1970s.

In the 1960s, two behavioral therapies for anxiety disorders developed in tandem: systematic desensitization (Wolpe, 1964) and flooding/implosion (Stampfl & Levis, 1967). Both posited that systematic exposure to the feared object could promote new learning and diminish fear. Wolpe proposed the theory of reciprocal inhibition, which stated that exposure to the feared stimulus while in a state incompatible with fear (either relaxation or excitement) would preclude a fear response, allowing the patient to tolerate the stimulus and diminishing fear over time (1964). Wolpe initiated deep breathing to induce a state of relaxation, after which the patient would be exposed to a hierarchy of increasingly feared stimuli. He reported that over 90% of patients with specific phobia remitted using his technique (1958).

In the alternate paradigm of flooding/implosion, therapists exposed patients to the most-feared stimulus for a prolonged period of time, until anxiety diminished. The premise of the flooding techniques was that maximal similarity between the feared stimulus and the exposure cue would promote optimal extinction learning (Stampfl & Levis, 1967). The therapist exposed the patient to feared cues through therapist description, reenacting scenes with the patient, images and videos, or physical objects. A paradigm called implosive therapy integrated psychodynamic principles into the technique by exposing patients to anxiety-provoking areas of dynamic conflict, such as rejection/humiliation, aggression, and sexual material (Stampfl & Levis, 1967). Flooding/implosion advocates argued that flooding was just as effective as systematic desensitization, while being briefer and more efficient.

The paradigms of systematic desensitization and flooding competed until the mid-1970s, when a systematic review showed that the type, intensity, and order of exposures were not decisive: the critical element was conducting exposure until distress diminished (Marks, 1978). The finding that exposure alone was effective confirmed the therapeutic efficacy of both techniques, but it also undermined their theoretical underpinnings.

After the formulation of PTSD as a mental disorder in DSM-III in 1980, exposure therapy was applied to the treatment of trauma. Because PTSD was initially conceptualized an anxiety disorder, the use of anxiety treatment paradigms for PTSD was natural. Behavioral therapists found Mowrer’s two-factor theory of avoidance learning to be a natural fit for the PTSD diagnosis (1947). Classical conditioning led to the generalization of fear to neutral cues co-occurring with the traumatic event. As traumatized individuals found that avoidance of trauma triggers lowered anxiety, avoidance was reinforced through operant conditioning.

The first RCT of implosive therapy was published in 1989 (Keane et al., 1989). The therapy began with 2–3 sessions of relaxation training, which was posited to improve the patients’ engagement in treatment and the vividness of the imagery they would use during exposure. Subsequent sessions included 45 minutes of flooding bookended by 10 minutes of relaxation, followed by 10 minutes of “integrating any new information or emotions that were evoked as a function of the imagery just presented” (1989, p. 249). Patients in the treatment condition achieved clinically significant reduction in symptoms of hyperarousal, intrusion, and depression, and treatment gains were maintained at 6-month follow-up. The authors noted that symptoms of numbing and social avoidance did not improve, and they proposed social skills training as an adjunct to therapy. They conclude that “although the results of this study generally confirm the usefulness of implosive therapy in the treatment of PTSD, the precise mechanism of action responsible for these changes remains unclear” (1989, p. 256).

Thus current exposure-based therapies for PTSD developed along two parallel tracks. One started from the clinical phenomenon of trauma and its sequelae. Clinicians working with traumatized patients noticed the unique nature of traumatic memory. They noted that the emotional retelling of the story was essential for healing, but also that the therapeutic context and timing was of utmost importance. They typically found that additional behavioral and interpersonal work was necessary to achieve symptom remission. The second track started with behavioral techniques for anxiety disorders based on learning theory. This track conceptualized PTSD as a fear-based disorder, characterized by the overgeneralization of fear, the failure of extinction learning, and erroneous beliefs about the feared situation. Exposure was found to be effective in treatment PTSD, but its mechanisms remained unclear.

Mechanisms of Exposure Therapy

Clinical Literature on Mechanisms

Investigation into the mechanisms of exposure has continued to the present day, resulting in a wealth of theoretical and empirical contributions. The dominant conceptualization of the mechanism of exposure in PTSD is Foa and Kozak’s (1986) Emotional Processing Theory (EPT). Foa and Kozak proposed the existence of a “fear structure,” a neural network of associations that represent the feared situation. The fear structure can be conceptualized as a set of propositional statements concerning stimulus elements, response elements, and the meaning of the connection between the two. The authors proposed that exposure therapy worked by 1) activating the fear structure and 2) introducing new information that was incompatible with the fear structure, thereby disrupting and updating the fear structure. Their theory included not just the extinction of fear, but cognitive and meaning-based elements (Foa & Rothbaum, 1998). PE uses in vivo exposure, imaginal exposure, and processing to activate the fear structure and introduce incompatible information.

In their foundational work, Foa and Kozak made empirical predictions regarding mediators of symptom improvement according to the tenets of Emotional Processing Theory. The first is activation of the fear structure, measured by subjective report or by physiological measures such as heart rate and galvanic skin response. The second is within-session habituation, operationalized as a decrease in subjective distress over the course of imaginal exposure. The third is between-session habituation, operationalized as a decrease in peak or mean subjective distress levels between the initial and the final PE session.

EPT also predicts specific therapy techniques and client change events that are necessary and sufficient for therapeutic change. First, in order for the fear structure to be fully activated, the traumatic memory must be reconstructed in as much detail as possible, as “greater matching of input information with a preexisting fear memory enhances fear evocation” (Foa & Kozak, 1986, p. 24). Second, the patient must attend to the exposure cues sufficiently to activate the fear structure. The therapist therefore actively directs the patient to focus and emotionally engage in the trauma narrative; distraction, dissociation, and skipping over parts of the narrative are discouraged. Third, the exposure must be prolonged and uninterrupted to promote within-session habituation: “changes in S-R associations typically require habituation of feared responses and this process unfolds gradually. Accordingly, long duration exposure is most likely to permit habituation, and thereby to abolish these dysfunctional associations” (Foa & McNally, 1996, p. 225). Foa & Kozak specify exposure lengths that are necessary for various anxiety disorders.

Over the last 35 years, Foa and Kozak’s predictions have garnered mixed empirical support, with cognitive change and between-session habituation receiving more support than within-session habituation and emotional engagement. Within-session habituation and exposure duration have consistently failed to predict treatment outcomes (Craske, 2008; Bluet et al., 2014; Brown et al., 2019). Studies have also shown that distraction during exposure may either facilitate or detract from fear reduction, in contrast to the theoretical prediction that distraction is necessarily detrimental (Weisman & Rodebaugh, 2018). Emotional engagement has been found to be modestly related to treatment outcome (Cooper et al., 2017). Some studies have found a modest association between between-session habituation and treatment outcome, while others have shown treatment gains in the absence of between-session habituation, suggesting that patients may instead gain distress tolerance skills (Bluett et al, 2014; Cooper et al., 2017; Craske et al., 2008). In contrast, cognitive change has repeatedly been shown to precede and predict symptom reduction (Brown et al., 2019; Kleim et al, 2013; Zalta et al, 2014). In response to these findings, Foa and colleagues have updated their theory to de-emphasize habituation and emphasize incorporating incompatible information as the fundamental mechanism of change: “the critical factor in exposure therapy is the formation of new associations rather than within-session habituation or the duration of exposure per se”(van Minnen & Foa, 2006, p. 436).

Neurobiological Mechanisms of Exposure

The neurobiological literature on memory, learning, and exposure provides additional insight into the mechanisms of exposure. Translational research focuses on a model of traumatic memory using classical conditioning procedures in animal models and in humans (LeDoux 1998). Classical conditioning paradigms typically model traumatic memory through the creation of a conditioned fear response to a neutral stimulus (CS) by pairing it with an unconditioned feared stimulus (UCS). They then model psychotherapy through efforts to extinguish or decondition the fear response by decoupling the CS and the UCS. Neuroimaging studies have found that the amygdala, a part of the limbic system that plays a role in species-specific instinctive emotions, is essential to fear learning pathways that connect sensory experiences to emotional reactions and behavior (LeDoux, 1998). Fear learning may take place through a “fast” route that goes through subcortical regions (thalamus → amygdala → hypothalamic–pituitary– adrenal), or a ”slow” route that goes through cortical regions (thalamus → visual or auditory cortex → amygdala). The hippocampus plays a central role in registering the context of fear conditioning, such as the environment in which it took place (Phillips & LeDoux, 1992).

Studies of fear deconditioning have found that repeated presentations of the CS in the absence of the UCS decrease, or extinguish, the fear response to the conditioned stimulus. Neurologically, extinction does not actually change the original fear memory; rather, it creates a new memory involving cortical brain regions that inhibit the fear response (Bouton, 1993). Specifically, it has been shown that the ventromedial prefrontal cortex (vmPFC) is crucial to extinction learning (Phelps et al. 2004), and some have found that hippocampus involvement is also significant (Milad et al. 2007). These results imply that the vmPFC exerts an inhibitory effect on amygdala-based fear learning, while the hippocampus registers context in both acquisition and extinction.

The finding that extinction learning does not alter an existing fear association but rather lays down new, inhibitory associations, explains clinical and experimental findings regarding the reinstatement of fear after extinction. Such phenomena include the spontaneous recovery of fear after the passage of time, reinstatement of fear after being exposed to the unconditioned stimulus, and renewal of fear after being exposed to the context in which the initial learning was done (Hermans et al., 2005). These phenomena suggest that extinction learning may be transient and may not be powerful enough to modify original trauma memories associated with PTSD. Thus, interest has turned to post-retrieval extinction, which has the potential to alter the underlying fear memory. Post-retrieval extinction relies on the existence of a consolidation window following retrieval of a memory, which returns the memory to a labile state and allows it to be modified and reconsolidated in an altered form (Phelps & Hofmann 2019; Lee et al. 2017).

The first evidence for the possibility of post-retrieval extinction to “erase” the original fear memory was presented by Nader et al. (2000). They showed that blocking the synthesis of protein in the amygdala during the reconsolidation window (i.e. following retrieval of the fear-memory) caused a lasting decrease in fear response to the previous learned fear stimulus in rats (Nader et al. 2000). Schiller et al. demonstrated that in humans, a single retrieval of the feared memory prior to an extinction procedure caused a decrease in fear response to the stimulus 24 hours later (Schiller et al. 2010). This effect was only shown in a group that received the retrieval cue 10 minutes prior to extinction learning, but not in a group that received a cue 6 hours prior to extinction, suggesting that there is a narrow post-retrieval memory reconsolidation window between 10 minutes and 6 hours. Recent findings suggest that post-retrieval extinction does not involve the vmPFC, which is known for its inhibition role in extinction learning, suggesting an alternate mechanism (Schiller et al. 2013). While post-retrieval extinction is emphasized here, it is important to note that some competing or complementary theories exist as well. The idea of time-dependent memory consolidation suggests that initial memory consolidation is affected by multiple reactivations of the memory before permanent storage in the neocortex (Squire & Alvarez, 1995). Another theory suggests that while semantic memories are indeed consolidated in the neocortex, the hippocampus holds the episodic memory permanently (Nadel et al., 2000). Finally, the work of Tronson and colleges suggests that consolidation and reconsolidation processes differ in their molecular mechanisms, affecting the stage at which these processes occur (Tronson & Taylor, 2007). This body of work has implications for the time horizon and mechanism by which reconsolidation processes could be harnessed clinically.

The idea that retrieval of a trauma memory may return it to a labile state suggests novel intervention options (Kida, 2019). Studies have found support for post-retrieval interventions such as extinction, counterconditioning, and competing cognitive processes, in altering fear and appetitive processes (Lee et al. 2017). Potential clinical applications of reconsolidation to PTSD include incorporating positive emotion into the administration of exposure treatment in order to update trauma memories, as well as the use of pharmacotherapy agents that facilitate reconsolidation processes, thus reviving procedures used by Janet and Kardiner (Kida, 2019). For example, Duek et al. (2021) found that a single infusion of ketamine prior to massed PE therapy led to decreased amygdala and hippocampus activation, with no change in amygdala-vmPFC connectivity, implying that reconsolidation rather than extinction processes were operative.

The neurobiology of fear learning has also been used in cognitive neuroscience accounts of PTSD, which emphasize the unique features of traumatic memory and their clinical implications (Brewin, 2001, 2014). According to LeDoux, the neurological fear system is evolutionarily designed to create quick, automatic, and permanent responses to sensory cues that are reminiscent of survival-threatening, (1998). LeDoux argues that the release of cortisol, which amplifies amygdala activity and dampens hippocampal activity, enables this to take place. As noted above, the quicker, subcortical amygdala pathway encodes low-level perceptual features of the event, leading to automatic and rigid stimulus-response contingencies. The slower, hippocampal processing “results in the laying down of integrated, coherent representations of conscious experience, located in the appropriate temporal and spatial context” (Brewin, 2001). Thus the encoding of traumatic events privileges low-level perceptual memory over a more sophisticated representation of context. This reflects clinical observations that traumatic memory is often imagistic and fragmented, and also suggests an explanation for the phenomenon of flashback (Brewin, 2001). According to Yovell, these different pathways also explain the “dissociation between facts and feelings” in traumatic memory, and the isolation of trauma memories from other autobiographical memories (2000).

These clinical elements of traumatic memory, first noticed by Janet, lead to a different potential neural mechanism of exposure therapy: the integration of fear memory into autobiographical narrative. Brewin (2001) suggests that imaginal exposure works not through extinction, but through the process of reconstructing the memory, accruing more details and contextual memories over time. Brewin invokes the concept of “interleaved learning,” a process by which the neocortical system is introduced to new information combined with old (memory) data, to incorporate new and old information. Brewin contends that such a process may underlie emotional processing in PE as well as the process of assimilation in CPT. This perspective relies on the same neural basis, the assertion of cortical inhibitory control over fear memories, but emphasizes the process of integration of memory systems rather than the extinction of fear. Clinical implications of this view involve a de-emphasizing of repeated trials, presumed to be necessary for extinction learning, and an emphasis on narrative integration and meaning.

In sum, research on the mechanisms of exposure have led to a decreased emphasis on habituation and extinction processes. Mechanisms such as reconsolidation and the integration of memory systems have shown more promise. Clinical implications include decreased emphasis on the repetition and duration of the trauma memory, and increased emphasis on understanding the optimal conditions for memory reconsolidation and integration.

Trauma-Focused Therapies for PTSD

In this section, we review trauma-focused, evidence-based therapies that involve some element of exposure to traumatic memories, but not the type, duration, or frequency of exposure prescribed in PE. The strong evidence for efficacy of these therapies implies that an attenuated form of exposure to the traumatic memory is also effective in overcoming PTSD.

Cognitive Processing Therapy

Cognitive Processing Therapy (CPT) is considered a first-line treatment for PTSD by APA and international treatment guidelines (APA, 2017; ISTSS, 2018; NICE, 2018; VA/DoD; 2017). Overall, research has found treatment effect sizes around 1.5 and loss of PTSD diagnosis in about 50% of individuals in RCTs, comparable to PE (Cusack et al., 2015).

CPT is based on a cognitive model of PTSD, which posits that PTSD symptoms result from maladaptive cognitions related to a traumatic event. These distorted cognitions, or “stuck points,” lead to manufactured emotions of guilt, shame, and anger, and they prevent the patient from fully experiencing the natural emotions evoked by the trauma. CPT focuses less on reconstructing the traumatic event, and more on the manufactured emotions that prevent the patient from fully experiencing fear and loss. This formulation resembles Freud’s focus on interpretation of the resistance, as well as Janet’s emphasis on constructing a benign interpretation of the event. Although CPT involves discussion of the traumatic event, the patient is not required to narrate the traumatic event in detail, uninterrupted, or for a particular length of time. Its creators argue that CPT is not an exposure therapy: “One might argue that any discussion about traumas constitutes exposure to avoided memory. However, there is a difference between talking about why something happened and reexperiencing the memory of the trauma in graphic detail” (Schnyder et al., 2015).

EMDR

Eye Movement Desensitization and Reprocessing (EMDR) is another trauma-focused therapy that is considered a first-line treatment by most guidelines (ISTSS, 2018; NICE, 2018; VA/DoD, 2017) and a second-line treatment in APA guidelines (2017). A recent meta-analysis showed that EMDR has an effect size of 1.14, comparable to the rate observed in trauma-focused CBT (Bisson et al., 2013). EMDR was created based on the personal experience of its founder, Francine Shapiro, who experienced a release of difficult memories and emotions after engaging in saccadic eye movements (Shapiro, 2018). Shapiro later developed the Adaptive Information Processing model (AIP), which holds that pathology results from the state-specific processing of memories under conditions of stress. AIP proposes that when memories are processed and assimilated into the rest of the memory network, symptoms decrease. The EMDR protocol involves phases designed to educate the patient, process the traumatic memory and rework maladaptive cognitions, and develop coping techniques. In the desensitization phase, which most resembles exposure, patients are instructed to recall the most distressing aspect of a memory for 30–60 seconds, while engaging in bilateral stimulation such as eye movements. Like CPT, EMDR involves aspects of exposure to the trauma memory, but does not require the level of detail, duration, or focus as in PE. The patient is allowed to spontaneously associate to other memories, a feature that is reminiscent of the psychoanalytic free association method (Balbo et al., 2019). While this would be predicted by EPT to permit avoidance of aspects of the trauma memory, it may also allow patients to more freely access aspects of the fear network or relevant previous memories. Indeed, some clinicians have found “an increase in the effectiveness of exposure when they adopted a less directive approach and permitted their clients to follow associations” (Balboa et al., 2019, p. 52).

Thus, trauma-focused treatments other than PE propose slightly different mechanisms of therapeutic action. They revive aspects of early trauma treatments while eschewing the fear-learning conceptualization of EPT. CPT focuses on reinterpreting the traumatic event in a more benign way, resembling Janet’s approach to integrating the trauma memory. EMDR is a staged treatment that focuses on stabilization and rehabilitating the personality, as in Kardiner’s approach, and allows free association reminiscent of Freud’s talking cure.

Non-Trauma-Focused Treatments

In recent years, evidence has accrued for the efficacy of PTSD treatments that do not raise the trauma memory at all (Shea et. al., 2020). Treatment guidelines include several non-trauma-focused therapies as second-line treatments, including Present-Centered Therapy, non-trauma-focused CBT (NTF-CBT), Interpersonal Psychotherapy (IPT), and Acceptance and Commitment Therapy (ACT). Some analyses have found that there is no benefit of trauma-focused therapies over other bona fide treatments for PTSD, while non-trauma-focused therapies have lower dropout rates (Huge & Chard, 2018; Lenz et al., 2017; Wampold et al., 2010). However, others argue that trauma-focused therapies have consistently shown slightly higher effect sizes (Schnurr, 2019).

Non-trauma-focused therapies tend to include the skills-building and relational components of PTSD treatment described early therapists such as Kardiner, as well as common factors (1941). Present-Centered Therapy (PCT), initially designed as an active control condition for PTSD psychotherapy trials, includes establishing a supportive therapeutic relationship, providing psychoeducation about PTSD, and problem-solving strategies. A recent Cochrane review found that PCT led to a greater reduction in PTSD symptoms than treatment as usual or wait list conditions (Belsher et al., 2019). Dropout was lower in PCT than in trauma-focused CBT, and the difference in treatment gains was not clinically significant at 6 and 12-month follow-up.

Interpersonal Psychotherapy (IPT) for PTSD seeks to address the interpersonal sequelae of trauma, conceptualized as a threat to the sense of safety mediated by the attachment system (Markowitz et al., 2015; Bowlby, 1969). The first phase of the treatment addresses trauma-induced deficits in the perception of interpersonal risk. The second phase applies traditional IPT techniques to help patients access social support and gain skills in assertiveness and emotional expression. The evidence base for IPT for PTSD includes five open trials and four RCTs, including one in that showed non-inferiority to PE (Markowitz et al., 2015). Markowitz and colleagues note that “patients who improved in Interpersonal Psychotherapy seemed to gain confidence in daily social interactions, gathered social support, and then spontaneously – without therapist encouragement – exposed themselves to trauma reminders,” suggesting that IPT may facilitate recovery through exposure outside of the therapy context (Markowitz et al., 2015).

Acceptance and Commitment Therapy (ACT) is a treatment oriented toward reducing experiential avoidance and increasing psychological flexibility rather than reducing symptoms (Hayes et al., 2006). ACT for PTSD includes in vivo exposures; however, these are designed to reorient the patient toward personal values rather than to decrease fear (Harris, 2021). ACT also emphasizes acceptance, rather than reduction, of psychological distress, in line with research showing that patients in exposure therapy may learn to tolerate distress rather than experiencing habituation (Bluett et al, 2014). ACT’s focus on increasing patients’ ability to tolerate distressing thoughts and feelings may facilitate patient-directed exploration of trauma memories. ACT has shown emerging empirical support, including several uncontrolled trials and one RCT that showed comparable effectiveness to PCT (Shea et al., 2020).

Discussion

We accumulate memories throughout our lifetime; some of them are aversive. Because fear plays a crucial role in species survival, extinguishing its memory traces proves to be difficult. At the same time, one of the unique properties of memory is its ability to change over time. Usually this change involves forgetting (decay of the original memory traces) or changes in memory encoding via reconsolidation. Neuroscience research has shown that each time a memory is brought into consciousness, it is in a labile state and thus amenable to change. The everyday rethinking of aversive experience is incorporated into the reconsolidated memory, ameliorating traumatic memories by reworking their meaning. In PTSD, however, it appears that the original emotional encoding of the traumatic memory remains unchanged. Regenerating this process could be the core mechanism of change in PTSD treatment, but when and how to best reignite it remains unresolved.

In this review, we have raised the question of whether exposure is necessary to treat PTSD, by examining the history and evidence for exposure and non-exposure-based trauma treatments. As early clinicians found, exposure does not always result in PTSD remission. Within PE treatment, how to determine the optimal exposure “dose” for each individual, i.e., the duration and the number of trauma narrative repetitions necessary for meaningful therapeutic gain, remains a clinical art. In addition, the conceptualization of PTSD as a fear-based disorder has been challenged in recent years as overly simplistic and the field has pivoted to emphasize other emotions such as anger and guilt (Yehuda et al., 2016).

It has also become clear that non-exposure-based treatments can be effective in treating PTSD. We propose three reasons why this may be: 1) spontaneous exposure, 2) common factors, 3) the heterogeneity of PTSD. The psychotherapy literature tends to emphasize exposure as a therapist technique and to ignore the possibility of spontaneous exposure to trauma memories. Self-directed exposure is a normative response to trauma that mitigates PTSD development. When the severity of an event exceeds an individual’s capacity to engage in processing and reconsolidation of the memory, PTSD may result. The role of therapy may include changing representations of self and other in order to resolve defensive avoidance and facilitate confrontation with traumatic memories. For example, patients in IPT for PTSD spontaneously chose to expose themselves to traumatic reminders outside of the therapy context, suggesting that the therapy facilitated self-directed exposure (Markowitz et al., 2015). Thus, exposure may be the mechanism for change even in therapies that do not include therapist-directed exposure. Future trials should collect data on self-directed exposure in order to clarify the mechanisms for non-trauma-focused treatments.

Moreover, the efficacy of non-exposure-based treatments may be a special case of the more general finding that bona fide treatments do not differ in their efficacy, and that common factors account for a greater proportion of treatment gains than specific therapy techniques (Norcross & Wampold, 2019). In fact, PTSD patients whose sense of safety has been shaken may be especially sensitive to common factors such as warmth and trustworthiness in the therapist, and responsiveness to patient needs may be particularly crucial in the delicate art of approaching aversive memories (Norcross & Wampold, 2019).

In addition, PTSD is a heterogeneous condition, leading some to argue that the construct does not warrant a unified approach to treatment (Galatzer-Levy & Bryant, 2013). PTSD may be divisible into subtypes based on different symptom profiles, e.g., a depression-based subtype with elevated dysphoria and anhedonia, and a fear-based subtype with elevated re-experiencing and hypervigilance (Zoellner et al, 2013). The type and chronicity of trauma varies greatly, leading to the ICD formulation of the complex PTSD syndrome, which recognizes the differential effects of chronic developmental trauma (Cloitre, 2015; Naifeh et al., 2008). Finally, for many, if not for most, PTSD is only one part of a complex clinical picture complicated by other comorbidities (Steenkamp et al., 2015). Addressing comorbid conditions has been shown to lead to PTSD symptom improvement (Simpson et al., 2017). Clinical complexity has been shown to be a moderator of treatment effects, favoring specific treatments for noncomplex populations and nonspecific treatments for complex ones (Gerger et al., 2014). Thus, some suggest that “one size fits all models” are ill advised; treatments must address the holistic needs of the patient rather than targeting trauma memories alone (Cloitre, 2015).

Finally, it is worth reflecting on what we mean by “recovery” from PTSD. Non-trauma-focused treatments may facilitate enough symptom relief to lose the PTSD diagnosis. However, holistic clinical goals for PTSD patients usually include the ability to confront and integrate the traumatic event into their life narrative. This historical review has led us to broaden our conceptualization of what may successfully reduce symptoms of PTSD, to conclude that paradigmatic exposure may not be necessary, and to conceptualize a broader set of interventions that may engage the patient’s innate capacity for reconsolidating traumatic memories.

Public significance statement:

For over 30 years, exposure-based therapies have been considered to be the gold standard in treatment for PTSD, leading to widespread public investment in dissemination of these therapies. Recent findings have called into question both the effectiveness and the centrality of exposure in treating PTSD. This paper integrates clinical and research literature on the role of exposure in processing traumatic memories in order to elucidate this controversy and to further inform clinicians, researchers and trainees. It also provides a pathway for future directions in clinical and basic research on PTSD.