More than Meets the “I”: A Panoramic View of Epistemic Trust in Psychotherapy

Human interactions are like an intricate Web of threads. One such subtle yet powerful thread that shapes the most complex relational patterns is recognized as epistemic trust (ET) [1]. ET pertains to the trust we place in the credibility of others’ knowledge and our confidence in the integrity of their words and actions. It is regarded as a critical element of interpersonal relationships and assumes great importance in psychotherapy. The term “trust” frequently appears in the literature but is often used without a detailed explanation, assuming a common understanding. However, trusting others spans various contexts, from impersonal trust in knowledge, like asking directions, to deeply personal trust in judgment, like when considering someone’s perspective on our own experiences [2]. Similarly, trust in therapy encompasses a broad spectrum, from its fundamental role in facilitating the therapeutic process, where social learning is rekindled, to its ideal outcome, where this learning extends and applies to relationships beyond the therapeutic setting [3]. Understanding the multidimensionality of trust in therapy requires consideration of its broader context, which can greatly enhance our perspective when applying it to the patient-therapist relationship. Erikson’s developmental theory, for example, emphasizes basic trust or mistrust as shaped by early experiences with caregivers and is critical in establishing trust later in life [4]. In this sense, psychotherapy can help individuals explore trust as a sense of continuity and sameness [5], fundamental to personality and identity development. Similarly, Buber’s “I-Thou relationship” highlights trust as an authentic, reciprocal encounter between two beings who acknowledge each other’s wholeness and humanity contrasts with the “I-It relationship,” characterized by objectification and instrumental use. Such a perspective forms the basis for understanding deeper relational trust in psychotherapy, where both therapist and patient engage in a meaningful, genuine relationship [6].

Within this broader context, ET is understood as the specific kind of trust essential to any effective therapy [7] that enables patients to feel that the information they receive is personally meaningful [8], making subsequent communications feel relevant and potentially applicable to future situations [1, 7]. However, maintaining some vigilance is a necessary adaptive tool, protecting us against misinformation until we are assured that it is safe to be more open. Thus, while ET facilitates social learning, its effective application involves a nuanced process of discerning who can be trusted as a reliable source of knowledge. But some individuals, who were maltreated in their early environment, remain persistently vigilant as an adaptive response to untrustworthy communications. Later in life, these individuals might develop severe mental pathologies, such as personality disorders, and struggle to overcome their epistemic hypervigilance. Consequently, they may mistrust social communication, including therapy, and perceive it as less personally relevant or applicable to their future [3, 9]. Hence, exploring ET in psychotherapy is essential.

The role of ET in psychotherapy has been extensively explored since its relevance to the therapeutic relationship was first proposed by Fonagy and Allison [1]. The existing literature encompasses theoretical analyses exploring the origins of ET within attachment relationships [10], its association with the ability to mentalize [3], its potential role as a catalyst for therapeutic change [7], and its broad importance across various therapeutic approaches [8]. This transtheoretical perspective on ET suggests it is crucial for benefiting from psychotherapy and that changes in ET may act as a mechanism of change, an active ingredient therapeutic in itself, especially for those who are initially mistrusting.

Despite the extensive theoretical discussions, empirical evidence remains scarce. Most empirical studies are based on clinical case studies. For instance, in a post hoc qualitative analysis of interview data initially collected for a broader study, Li et al. [11] focused on 15 depressed adolescents. Their aim was to create a typology of various trust and mistrust issues that emerge during therapy. The findings showed that while some adolescents moved from epistemic mistrust to trust, others consistently reported mistrust over a 2-year period. Additionally, several case studies have shown the significant role of ET in therapeutic work with individuals diagnosed with borderline personality disorder, particularly in the context of mentalization-based therapy (e.g., [12]). Furthermore, a case study that combined qualitative and quantitative analyses highlighted the importance of understanding and monitoring ET from the beginning of treatment and throughout the therapy process, underscoring its potential link to positive therapeutic outcomes [13].

Together, these case studies provide promising support for the theoretical models of ET in psychotherapy. Yet there remains a significant gap in comprehensive empirical evidence regarding the constituents of ET in therapy and its related effects. Whether we view ET as an inherent characteristic of patients or as a facilitator of therapeutic change, we must consider: what do we understand about a patient exhibiting high or low levels of ET? Drawing on attachment theory can help us clarify this inquiry. Extensive empirical research and clinical experience have shown that by the end of their first year, infants typically develop a distinct attachment style, whether secure, anxious avoidant, anxious ambivalent, or disorganized [14]. The formation of a specific attachment style is influenced by various factors, including the infant’s inherent traits (such as temperament or other innate characteristics), the parent’s qualities (such as sensitivity, availability, and responsiveness), and the dynamics of their relationship [15]. Applying these factors to ET raises additional questions. Does ET represent a persistent characteristic of a patient that appears across different relationships, including the therapeutic setting? Could it instead be influenced by the assets and competencies of the therapists involved? Or does it develop uniquely within the specific therapeutic relationship? Unlike attachment and other related concepts, we are not yet able to clearly differentiate these three aspects: the patient’s attributes, the therapist’s contribution, and their unique therapeutic relationship.

Given the limited empirical evidence on the specific role of each of these aspects, our understanding of ET remains incomplete. Therefore, the primary goal of this study is to explore the constituents of ET in therapy and its related effects. To accomplish this, this study examines ET constituents and related effects from various lenses, thereby adopting a panoramic view. The first lens focuses on the constituents of a snapshot of ET. The second lens explores the implications of this ET snapshot for the therapeutic process. The third lens further explores the dynamics within the therapeutic process as changes in ET occur. Together, the three lenses capture relatively stable individual differences in ET that shape a patient’s capacity to benefit from therapy (the trait-like component) and the potential for ET to change during therapy, acting as a mechanism of change that facilitates therapeutic change (the state-like component). A conceptual framework that systematically examines each of these lenses is outlined here. Figure 1 illustrates comprehensive or “panoramic” views of the three distinct perspectives.

A snapshot refers to a cross-sectional portrayal of a patient’s personal and interpersonal attributes by focusing on a single variable at a specific time, typically at the beginning of therapy [16]. Most traditional psychotherapy research uses this between-individual lens as it is believed to provide a detailed, complex view of all the elements that comprise the individual [17]. In the context of ET, this comprehensive method may highlight the intricate complexity of ET, demonstrating its vulnerability to various influences. This approach aims to capture a complete depiction of the three aspects of ET: the patient’s attributes, the therapist’s contributions, and the unique patient-therapist relationship at a specific time. For instance, the snapshot lens might capture the patient’s enhanced receptiveness to new information, increased willingness to accept support during challenging times, and greater ability to engage in the therapeutic relationship to safely explore their emotions and experiences (see Fig. 1a). It might also include the therapist’s inputs, such as their choice of therapeutic interventions, empathy, warmth, or responsiveness. Additionally, it may capture the unique dynamics of the patient-therapist therapeutic alliance, such as synchronicity and collaborative partnership. This perspective also considers within-patient factors, such as daily fluctuations and the specific timing of measurements, which are often regarded as “noise” [16]. As a result, this lens may reveal a blend of trait-like characteristics (individual differences between patients) and state-like changes (therapeutic processes within the patient), which could be challenging to separate when using a single snapshot [18].

If ET is significant in therapy, it implies prognostic value. This value manifests in two models of change: compensatory and complementary [18]. The compensatory model suggests that patients with certain deficits may benefit more from treatments aimed at mitigating these deficits [19]. For instance, patients exhibiting low ET, as shown by their limited openness to new information, difficulty accepting support, or reluctance to explore their experiences in therapy, are likely to show greater positive changes when these specific vulnerabilities are addressed [20]. This model also indicates that patients with low ET might improve their ability to form relationships by developing a positive therapeutic alliance, achieving synchrony with the therapist, and enhancing collaborative interactions.

Conversely, the complementary model suggests that patients with relative strengths will benefit more from treatments that emphasize these strengths [19]. Therefore, patients with high ET, characterized by their openness to new information, robust ability to accept support, and readiness to explore their experiences in therapy, are well positioned to further enhance these qualities [21]. According to this model, high ET patients may also strengthen their capacity to establish effective relationships by cultivating a strong therapeutic alliance, synchronizing with the therapist, and fostering collaborative partnerships.

Among these models, the complementary model particularly fits the context of ET. The theory posits that individuals with high levels of ET are likely to display strengths in their personal attributes and in establishing collaborative relationships over time [7]. This suggests that one’s openness to new knowledge is a resilience factor, enhancing the patient’s ability to benefit from treatment as it introduces new, personally relevant information [22]. This lens seeks to understand the trait-like characteristics (between-patient effects) to better tailor expected state-like changes (within-patient effects) in the therapeutic process. Therefore, the panoramic view from this lens may highlight changes in the patient’s attributes and the dynamics within the unique patient-therapist relationship while not focusing on the therapist’s contributions (see Fig. 1b).

While the previous two lenses focused on the constituents within an ET snapshot and the changes it predicts, the third lens concentrates on changes within ET itself. Decades of psychotherapy research have shown that patients evolve along different trajectories at varying rates and in distinct ways [23]. Each patient’s journey is inherently unique and evolves based on the specific relationship they develop with their therapist [13]. According to this within-patient perspective, changes in ET are expected when a patient’s understanding of their experiences aligns with the mental model provided by the therapist. At this “epistemic match” point [7], a patient is thought to perceive the therapist as recognizing them, thereby opening their mind to new ways of acting and reacting within their social environment. Changes in ET could indicate a growing recognition by both the patient and the therapist that their partnership and alliance have strengthened, surpassing the contributions of each individual [24]. Thus, establishing ET throughout treatment is crucial for fostering successful collaboration between the patient and therapist as they pursue shared goals and tasks, creating a strong emotional bond between them. As this lens describes the state-like changes in therapeutic processes that occur during treatment, the comprehensive panoramic view may capture shifts in the unique patient-therapist relational dynamics, focusing less on the patient’s attributes and the therapist’s contributions (see Fig. 1c).

The present study aims to elucidate the constituents of ET and its related effects within the context of psychotherapy. If the snapshot and change perspectives of ET have distinct theoretical, clinical, and empirical implications [18], clarifying them may shed light on ET’s complex nature. To date, no study has empirically differentiated between trait-like (between individuals) and state-like (within individual) ET in psychotherapy. This study seeks to distinguish these aspects by examining ET through the three specific lenses previously proposed. First, from the snapshot lens (blended trait-like and state-like effects), we hypothesize that ET at an early session (session 4) would correlate with patient attributes, therapist contributions, and their unique relationship. Second, from the prognosis lens, we hypothesize that the patient’s ET at an early session (session 4) would correlate with subsequent changes from session 4 to session 8 in patient attributes and the patient-therapist relationship. Third, from the changes lens (state-like within-patient effects), we hypothesize that changes in ET from session 4 to session 8 would correlate with within-individual changes in the unique patient-therapist relationship during the same timeframe.

To test these hypotheses, we utilized data from a randomized controlled trial (RCT) [25] that provided short-term psychodynamic psychotherapy for depression [26]. The choice of session 4 as a focal point is based on the specific treatment protocol implemented in the current study [27], in which the first three sessions serve to collect information and form insights regarding repetitive maladaptive relational patterns that characterize the individual patient. Session 4 is a significant time point at which the therapists share with their patients their clinical formulations aimed at explaining their suffering. The patient’s reactions to these clinical formulations introduced by the therapist offer valuable insights into their inclinations toward receiving and integrating new information. The rationale for focusing on session 8 stems from its significance as a midpoint in the therapy process. By this midpoint, substantial progress has already been made in enhancing the patient’s ability to identify and work through the clinical formulation dynamics with the therapist and implement tangible changes.

Participants were recruited through advertisements for treatment at the psychotherapy research laboratory clinic [25, 28]. Individuals who met predetermined inclusion and exclusion criteria underwent 16 weekly individual sessions of manualized psychodynamic psychotherapy, offered in either a supportive-focused or supportive-expressive-focused format [26]. Assignment to the respective treatment arms was conducted by an external institution with no direct involvement in the research. During the active phase of the trial, each patient completed self-report questionnaires, some of which were administered weekly, before or after the therapy session, and others at predetermined intervals (the start, middle, and conclusion of therapy).

Data were collected from a total of 118 patients, which included all intent-to-treat participants from both the RCT and the pilot preliminary phase. Due to technical issues during the recording of two sessions, 2 patients were excluded, resulting in data for 116 patients in session 4. An additional 3 patients dropped out before the eighth session, leaving a final count of 113 patients in session 8. Participants were eligible for the study if they met the following criteria: (a) a diagnosis of major depressive disorder (MDD) based on structured clinical interviews in accordance with the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition, with scores above 14 on the 17-item Hamilton Rating Scale for Depression (HRSD) [29] at two evaluation points, 1 week apart, and a diagnosis of MDD based on the Mini International Neuropsychiatric Interview (MINI) [30]; (b) if the patients were on medication, their dosage must have been stable for at least 3 months before entering the study, and they had to be willing to maintain a stable dosage for the duration of the treatment; (c) patient’s age must be between 18 and 65 years; (d) Hebrew language fluency; and (e) provision of written informed consent. Patients were also screened for the following exclusion criteria: (a) current risk of suicide or self-harm (HRSD suicide item >2); (b) current substance abuse disorder; (c) current or past schizophrenia or psychosis, bipolar disorder, or severe eating disorder requiring medical monitoring; (d) history of organic mental disease; and (e) currently in psychotherapy. Table 1 provides a detailed overview of the demographic and clinical attributes of the participants.

Nine therapists participated in the study, consisting of 6 women and 3 men, with an average of 14.42 years of clinical experience. These therapists completed a 20-h training workshop on supportive and expressive therapeutic techniques. Before participating in the trial phase, each therapist treated two pilot patients – one for each treatment type – and had to demonstrate sufficient adherence to the treatment protocol. Two therapists did not continue to the RCT active phase after the pilot phase: one secured a full-time position elsewhere, and the other showed low adherence levels. Each therapist participated in weekly group supervision led by two supervisors and also received weekly individual supervision from one of the supervisors, with feedback based on videotaped sessions.

The epistemic trust rating system (ETRS) is an observer-based measure designed to assess a patient’s ET within psychotherapy sessions. This measure was developed while rigorously integrating top-down and bottom-up processes, drawing on the triadic theoretical model for ET in therapy [33]. Utilizing verbatim transcriptions of therapeutic sessions, trained raters use a 5-point Likert scale ranging from 0 (very low/absent) to 4 (very high/pervasive) to evaluate three elements: sharing – the patient’s tendency to seek or avoid discussing significant experiences and the accompanying emotions and ideas with others; “we-mode” – the ability of the patient and therapist to collaboratively focus their attention on exploring the patient’s subjective experience; and learning – the patient’s receptivity to new ways of acting and reacting in social contexts [33].

The team of raters consisted of 10 individuals with diverse backgrounds, including undergraduate students, master’s and Ph.D. students in clinical psychology, and licensed clinical psychologists. Raters underwent comprehensive training on the ET concept and its associated rating system, applying their knowledge to rating examples in a team and then independently. Once they exhibited acceptable adherence to the rating procedure, each rater was assigned to rate real sessions for independent evaluation. Weekly meetings were scheduled to address any discrepancies in their ratings, with any deviations of 2 points or more considered discrepancies requiring thorough discussion to achieve consensus. Examples of low to high ends for each ETRS subscale can be found in the online supplementary materials (for all online suppl. material, see https://doi.org/10.1159/000541667) (for more detailed tables with ratings for each level, see [33]). The ETRS shows high validity and an intraclass correlation coefficient (ICC) of 0.86–0.90 within this dataset. Following Fisher et al. [33], we focus on the ETRS total scores, which are computed by averaging the scores across the three elements. Higher scores indicate higher levels of ET.

This aspect refers to an individual’s relatively stable characteristics and patterns that influence their perception and engagement in close relationships. As measured by the experience in close relationship (ECR), attachment orientation reflects the long-term effects of early experiences with caregivers on someone’s expectations, emotions, defenses, and behavior in all close relationships [34]. The IIP assesses a person’s typical interpersonal difficulties and problematic behaviors in relationships over time [35]. Both may shape someone’s propensity to foster or dampen ET.

This is a 36-item self-report measure of adult attachment. Participants rate their thoughts, feelings, and experiences in interpersonal relationships using a 7-point Likert scale from 1 (disagree) to 7 (highly agree) [36]. The scale generates scores for two dimensions: anxious and avoidant attachment. Higher scores on the anxious dimension indicate a greater fear of rejection and abandonment, while higher scores on the avoidant dimension suggest a greater discomfort with dependence on others and closeness. The reliability of the scores has been previously affirmed [34, 37], with reliability coefficients in the present study of 0.92 for the anxiety dimension and 0.89 for the avoidant dimension.

This is a widely used 32-item self-report measure designed to assess interpersonal distress and social adjustment level [35, 38]. It utilizes a 5-point Likert scale ranging from 0 (not at all) to 4 (extremely). An average of the 32 items is calculated to produce a total score reflecting the degree of difficulties experienced in relationships. Higher scores indicate more distressed interpersonal functioning, such as difficulties in expressing affection, being assertive, or being overly disclosing, among other challenges. Adequate validity and reliability for the IIP-32 have been established [39], with a reliability coefficient of 0.87 in the current study.

This aspect pertains to the quality and intensity of various interventions used by the therapist in a given session, assuming that the therapist’s actions drive change in therapy [40, 41]. The MULTI-30 common-factor (CF) subscale measures the therapist’s ability to foster a supportive, collaborative, and empathetic therapeutic relationship, which is crucial for the positive fostering of ET. The MULTI-30 person-centered (PC) subscale evaluates the therapist’s ability to authentically convey genuine care for the patient’s well-being, making the patient feel they can safely believe in the sincerity of their intentions [41]. Conceptually, ET in the therapeutic relationship is expected to develop when the therapist and patient successfully establish the “we-mode,” a state where the patient recognizes that the therapist truly sees them [13].

This is a 30-item self-reported measure designed to assess the usage of interventions from eight therapeutic orientations (cognitive, behavioral, process-experiential, PC, psychodynamic, interpersonal, dialectical-behavioral, and common factors) from the patient’s perspective [40, 41]. Participants rate items on a 5-point Likert scale based on how representative each item is of the therapy session they have just completed, with 1 indicating “not at all typical of the session” and 5 indicating “very typical of the session.” In this study, we focused specifically on the CF and PC subscales. The CF subscale assesses the therapist’s warmth, acceptance, and attunement, with items like “my therapist listened carefully to what I was saying.” The PC subscale assesses the therapist’s ability to unconditionally regard the patient’s subjectivity with items like “my therapist seemed interested in trying to understand what I was experiencing” [40]. These two subscales closely reflect the concept of we-mode, which emerges when a patient’s open disclosure aligns with a therapist’s sincere, empathetic engagement and understanding.

Other Multitheoretical List of Therapeutic Interventions (MULTI) subscales, such as psychodynamic and interpersonal techniques, emphasize supporting changes in interpersonal functioning by making unconscious patterns conscious. Items include “my therapist made connections between my current situation and my past” (psychodynamic) or “my therapist pointed out recurring themes or problems in my relationships” (interpersonal) [40]. For some patients, implicit personal narratives may be inaccessible, making it challenging to gain insight from interpretations of their early and current interpersonal tendencies. Thus, these two subscales do not capture the concept of we-mode, which may be impeded even when personal narratives are appreciated. The reliability coefficient in the current study was 0.81 for the CF subscale and 0.79 for the PC subscale.

This aspect refers to the patient-therapist’s collaborative effort to address the patient’s distress and maladaptive behaviors [42]. A strong alliance enables both parties to recognize agreed tasks while acknowledging each one’s unique roles and responsibilities when collaborating toward mutual goals [43]. The WAI captures these collaborative efforts and conceptualizes it as what “makes it possible for the patient to accept and follow treatment faithfully” ([44], p. 2), thereby fostering ET.

This 12-item self-report instrument measures the therapeutic alliance [45] based on Bordin’s [46] working alliance theory. Respondents rate items on a 7-point Likert scale from 1 (never) to 7 (always). The instrument provides a cumulative general alliance score, with higher scores indicating a stronger working alliance. The reliability coefficient in this study was 0.93.

Data were analyzed using SPSS software version 28. For each patient, two scores were computed: a “snapshot” score of trait-like individual differences between patients, using the raw scores of each measure as recorded in session 4, and a “change” score of state-like changes within an individual, calculated as the delta between the scores from session 4 and session 8. This methodology aligns with the literature on the operationalization of trait-like between-individual differences versus state-like within-individual changes [47] and reflects our assumptions about the role of ET, suggesting it is crucial for benefiting from psychotherapy and that changes in ET may function as a mechanism of change [15, 48].

Intensive longitudinal designs typically involve sequential measurements over five or more time points, during which a change process is expected to unfold within each subject [49]. However, in the current study, we focused on modeling within-individual changes between session 4 and session 8, rather than modeling the trajectory of change. The “snapshot” and “change” scores were utilized to test our three hypotheses regarding (1) the extent to which state-trait-like blended snapshot ET in session 4 correlates with variables related to the therapist, patient, and therapeutic relationship at the same session (hypothesis 1); (2) the extent to which state-trait-like blended snapshot ET in session 4 correlates with state-like changes in the patient and therapeutic relationship between sessions 4 and 8 (hypothesis 2); and (3) the extent to which state-like changes in ET between sessions 4 and 8 correlate with state-like changes in the therapeutic relationship during the same period (hypothesis 3).

Session 4 was chosen as the snapshot point for theoretical and empirical reasons. Theoretically, the fourth session provides an early opportunity for ET to manifest in therapy, marking a significant moment in the treatment process [11]. According to the supportive-expressive treatment protocol used in this study [26], therapists are instructed to explicitly articulate the patient’s clinical formulations aimed at explaining their suffering during the fourth session, potentially revealing insights into the patient’s long-term dynamics and openness to new information. Empirically, the fourth session has been identified as crucial in shaping the trajectory of the therapy [50].

Supportive expressive-focused therapy includes expressive techniques, such as interpretation, confrontation, and clarification [26], while supportive-focused therapy utilizes supportive techniques, such as affirmation and empathic validation. The supportive condition includes all supportive techniques detailed in Luborsky’s [26] manual but forbids the use of any expressive techniques. The decision to include both treatment arms in the present study stems from conceptual and methodological considerations. Conceptually, ET is considered a trans-theoretical construct. As such, ET is not expected to be confined to a single therapeutic approach but rather focuses on an individual’s readiness to perceive new information as personally meaningful beyond a specific context. Since we did not expect any variations in ET manifestations among the different treatment arms, we included patients from both arms in the findings. Methodologically, the small sample size relative to the number of variables (N = 116) did not allow for separate analyses for each arm.

Assessments of the ETRS in relation to the ECR, IIP, MULTI-30, and Working Alliance Inventory (WAI) provided partial support for the hypothesis, as detailed in Table 2. Regarding the patient’s attributes, the total score of the ETRS at session 4 showed a significant positive correlation with the ECR anxiety scale and the total score of the IIP. This indicates that individuals with higher ETRS scores are more likely to report greater anxiety in their attachment orientation and more interpersonal problems. Although these findings are significant, they suggest weak associations, contrary to our initial hypothesis concerning the patient’s attributes. Concerning the therapist’s contributions, the ETRS total score at session 4 was significantly positively correlated with both the CF and PC subscales of the MULTI-30. These results suggest that higher ETRS scores at session 4 are associated with more frequent use of both CF and PC techniques by the therapist at session 4. Regarding the therapeutic relationship, the ETRS total score at session 4 also showed a significant positive correlation with the WAI score at session 4, indicating that higher ETRS scores are positively associated with stronger therapeutic alliances, as measured by the WAI.

Assessments of the ETRS in relation to changes in the ECR, IIP, and WAI scores from session 4 to session 8 support the hypothesis, as detailed in Table 2. Regarding changes in the patient’s attributes, the ETRS total score at session 4 showed a significant negative correlation with changes in the ECR avoidance scale and the IIP score from session 4 to session 8 but not with the ECR anxiety scale. This indicates that patients with higher ETRS scores at session 4 are more likely to experience a decrease in avoidance attachment orientation as well as a reduction in interpersonal problem scores from session 4 to session 8. Concerning changes in the therapeutic relationship, the ETRS total score at session 4 was significantly positively correlated with the change in WAI scores from session 4 to session 8. This suggests that patients with higher ETRS scores at the beginning of the treatment are more likely to show improvements in their therapeutic alliances. These findings confirm our hypothesis that higher ET scores are associated with positive changes in patient attributes and the patient-therapist relationship, underscoring ET as a factor in resilience.

Assessments of changes in ETRS scores from session 4 to session 8, in relation to changes in WAI scores during the same period, provided full to partial support for the hypothesis, as detailed in Table 2. Specifically, an improvement in ETRS total score from session 4 to session 8 demonstrated a significant positive correlation with improvements in WAI total scores during the same interval. This indicates that patients who show a positive change in ETRS scores are also likely to experience a positive change in their overall therapeutic alliance as measured by the WAI. Contrary to expectations, improvement in ETRS total score from session 4 to session 8 also showed a substantial positive correlation with changes in ECR avoidance scores. This finding suggests that patients who exhibit improvements in ET may simultaneously show an increase in avoidance attachment experiences over time.

Three sensitivity analyses were conducted to verify the robustness of our findings. The analyses were made within the same dataset, each excluding a different subset to rule out the possibility of specific groups or biases unduly influencing the results. The first analysis was performed by repeating the tests for all three hypotheses with a subset of all 113 patients who had complete data for both session 4 and session 8 instead of 116 for hypothesis 1 and 2 and 113 for hypothesis 3 as in the original analysis. The results of this analysis are shown in online supplementary Table S1 in the online supplementary materials. The second analysis was performed by repeating the tests for all three hypotheses with only 103 patients out of the original N = 116 dataset. The N = 103 subset was due to excluding 13 individuals who received online psychotherapy during the COVID-19 pandemic in the same RCT and were included in our original dataset. The results of this analysis are shown in online supplementary Table S2 in the online supplementary material. The third analysis was performed by repeating the tests for all three original hypotheses while controlling for the treatment arm to rule out confound. This analysis aimed to verify that the treatment arm did not influence the results, that is, that ET is not confined to a single therapeutic approach. The results of this analysis are shown in online supplementary Table S3 in the online supplementary materials. The results of the three sensitivity analyses showed similar patterns to the original analysis, affirming the stability of our findings across different sample compositions.

The present study aimed to explore the constituents of ET in therapy and its related effects. By adopting alternating lenses, we sought to attain a panoramic view of the patient’s attributes, the therapist’s contributions, and the patient-therapist therapeutic relationship that constitutes ET. This approach considered both trait-like differences between patients, emphasizing how individual differences shape patients’ capacity to benefit from therapy, and state-like changes within patients, highlighting ET’s potential to act as a mechanism of change in therapy. The first lens offered an inclusive snapshot, blending trait-like and state-like effects. The second lens examined how this ET snapshot influences state-like changes throughout the therapeutic process. The third lens provided a focused view of state-like changes in ET. The results generally support the theoretical assumptions, with some nuances observed in the context of experiences of attachment relationships.

The results confirmed the first hypothesized lens that the fourth session constitutes a blended snapshot of trait- and state-like components. We hypothesized that ET at the early stages of therapy would correlate with three aspects – the patient, the therapist, and the patient-therapist therapeutic relationship. Our findings align with the model suggesting that therapy initiation entails a broad perspective that integrates both trait-like and state-like aspects [18]. Thus, significant correlations were found with all three aspects: the patient, the therapist, and the therapeutic relationship. A closer examination of the findings revealed interesting trends. Regarding the therapist’s contributions, higher levels of receptivity to new information were associated with greater readiness to see the therapist’s display of sensitivity, responsiveness, and empathic abilities. Regarding the therapeutic relationship, higher levels of the patient’s receptivity to new information correlated with a stronger therapeutic alliance. However, concerning the patient’s attributes, higher levels of receptivity to new information contradicted our hypothesis of a positive correlation with indications of a secure attachment style. Specifically, increased receptivity to new information was associated with pronounced anxious tendencies during distressing times and elevated interpersonal difficulties. On self-report measures, ET typically correlates negatively with both ECR avoidance and anxiety [51]. This blend of results, with some conforming to expectations and others diverging, reflects a nuanced perspective that incorporates both trait-like and state-like dynamics simultaneously. Perhaps on reflection, it is not surprising, given the context in which ET was assessed. Individuals who manifest a high level of trust early on in a therapeutic relationship may do so prematurely in other contexts, may be more likely to encounter interpersonal problems, and may experience greater anxiety in such contexts. This finding is not unexpected and emphasizes the importance of not assuming that between-person effects and within-person effects will align [52]. It aligns with previous studies suggesting that findings derived from between-patient and within-patient effects are substantially different and may sometimes even contradict each other [17]. The challenge of extrapolating from one effect to the other should motivate psychotherapy researchers to prioritize disentangling these distinct phenomena (e.g., [47]).

The results also confirmed the second hypothesized lens, which regarded the trait-state-like snapshot as a prognostic indicator of state-like changes in the patient’s attributes and the patient-therapist therapeutic relationship. We hypothesized that high ET at the beginning of therapy would predict positive changes in the patient’s attributes and the therapeutic relationship. The findings supported this assumption, showing that higher ET at the start of treatment was associated with positive improvements in the patient’s reported interpersonal difficulties, reduced avoidance in interactions with significant others, and an enhanced ability to establish a positive therapeutic alliance with the therapist. These findings align with the complementary model, which suggests “the rich get richer” [19], rather than the compensation model. In other words, our results support the notion that ET acts as a resilience factor [53]. The more open an individual is to social Learning, as indicated by high ET scores, the greater their capacity to engage in meaningful communication, thus enhancing their ability to benefit from social relationships even if, in a nontherapeutic context, premature openness to learning may create interpersonal difficulties and generate attachment anxiety. Consequently, as individuals successfully participate in the ongoing cycle of learning and teaching within human society, they are generally more resilient to mental disorders in a broader sense [24].

The results also confirmed the third hypothesized lens, which regarded state-like changes in ET as related to changes in the patient-therapist therapeutic relationship. We hypothesized that an increase in ET from session 4 to session 8 would correlate with the strengthening of the therapeutic alliance. The findings supported this assumption, indicating that, as expected, improvement in ET was associated with improvements in the therapeutic working alliance. This finding is consistent with ET theory that suggests increases in ET are expected to occur as an “epistemic match” is achieved [7, 13]. As humans’ ability to collaborate embodies a unique social role in sharing mutual recognition of intents [24], it enables mind-meeting moments between therapist and patient in psychotherapy. These so-called “we-mode” moments, where “two heads are better than one” ([3], p. 635), are enabled as the partners perceive their minds as distinct yet connected [54]. Together, we-mode moments hold the potential to prompt patients’ improved mental health [22], thus positioning ET as a promising mechanism of change in therapy [13].

However, one finding was contrary to our expectations. This unexpected finding was the positive correlation between change in ET and increased levels of avoidance in attachment. This suggests that improvements in ET coincide with increases in levels of avoidance. This finding becomes particularly intriguing when combined with our hypothesis 2 results, which indicated that patients with higher initial ET scores showed a decrease in avoidance attachment orientation from session 4 to session 8. That change in ET should signal increased avoidance and could reflect greater discernment in relation to interpersonal relationships associated with improved functioning of the system that underpins ET. In several studies [51, 55], we have observed a counterintuitive association between epistemic mistrust and epistemic credulity (unwarranted openness to belief in informants). Lower mistrust may signal lower credulity in those who are too ready to trust sources that are inherently untrustworthy, such as social media [56]. Thus, change in avoidance in those who are insufficiently cautious in their interpersonal interchanges may signal a positive rather than negative change. Taken together, these contrasting trends suggest that unique change processes occur in different patients. Specifically, while a positive improvement in ET may lead some patients toward decreased avoidance, for others, it may result in heightened defensive mechanisms as part of an adaptive organizational process. Nevertheless, caution is warranted when interpreting these results. Future comprehensive research exploring these trends could provide deeper insights into the complexities of these processes.

The panoramic exploration of ET and the use of varied lenses provide clinicians with the tools to view each patient as a distinct individual, akin to a “one-time phenomenon.” This approach suggests that the initial phase of treatment might include a combination of trait-like characteristics, state-like changes, and “noise” [16]. This complexity requires clinicians to develop flexible case formulations that can adapt to conflicting or evolving trends in a patient’s ET during the early stages of therapy. Analogous to developmental theories, where an infant’s attachment style becomes clear by the end of the first year, the early sessions in psychotherapy may represent a critical period to foster ET within the therapeutic relationship [53]. Filled with explicit cues [57], these initial sessions have the potential to reassure the patient that the therapist and the therapeutic environment offer a secure base for navigating challenging personal, interpersonal, and social experiences, as well as a safe haven conducive to learning.

The other two lenses focus more specifically on the implications of understanding the ETRS in terms of therapists’ abilities to enhance treatment effectiveness. One perspective offers a prognostic view, emphasizing the significant benefits that individuals with high ET are likely to derive from therapy [8]. The other perspective suggests that tailored treatment should prioritize the restoration of ET as a core component of the therapeutic process, potentially leading to broader changes within the patient.

While the present study offers fresh empirical insights into ET in therapy, it has several limitations. First, although the sample size is adequate, it is small compared to other studies, which have utilized larger samples for developing coding systems. This limitation makes it difficult to detect subtle nuances and generalize findings to a broader population. Additionally, the current sample does not include measurements of ET prior to the start of treatment, which hampers the complete differentiation of state-like components from trait-like ones. Future research could remedy this by measuring ET at multiple time points before, during, and after treatment and with multiple partners besides the therapist to foster a more comprehensive understanding of its dynamics. This approach might enable the identification of ET “signature” characterizing the patient beyond time in different relationships [16]. The methodological paradigm enabling extraction of individual-specific signatures is currently in the process of development [16].

While all participants in the present study had a diagnosis of MDD, there were also very high rates of personality disorder comorbidity (76.4%). As such, our sample may closely mirror the broader MDD population, which frequently exhibits high rates of comorbid personality disorders [58]. Moreover, ET is conceptualized as a transtheoretical and transdiagnostic construct, suggesting that the social learning process it facilitates may be temporarily or permanently disrupted in many forms of psychopathology [9, 22]. This model underscores adverse constitutional factors alongside aversive early experiences as risk factors for an individual’s compromised capacity to seek and receive help, belong to a social network, and maintain openness to change [3]. Such impairments may increase one’s susceptibility to developing psychopathology, especially personality disorders, through the effects that epistemic impairments have on social functioning [3].

Due to power considerations, we did not control for personality disorders in the current analysis. Future studies, however, could explore possible differences in ET among individuals with and without comorbid personality disorders. Since theory suggests ET is influenced by early and ongoing interpersonal relationships [9], we can speculate that therapy for individuals with personality disorders may work on a compensatory basis, enhancing mental flexibility through interpersonal implicit learning [59]. In contrast, therapy for individuals without personality disorders may work on a complementary basis, building on their intact social learning capacities. Understanding each group’s specific needs and challenges can lead to more personalized treatments.

While this study focuses on the role of ET in psychotherapy, other theoretical perspectives also offer valuable insights into the processes described. For instance, a robust body of research highlights the importance of relational and affective experiences in therapy, such as forming a strong therapeutic alliance [60], mentalizing [61], and therapist responsiveness [62]. These factors may form part of an interconnected network of change mechanisms that work alongside ET, mutually reinforcing one another. Future research could empirically explore how these processes function as an interconnected network, providing a deeper understanding of how and why psychotherapy is effective, specifically identifying the mechanisms through which treatments bring about change [63].

ET theory typically emphasizes the patient’s openness to learning from both the therapist and the therapeutic relationship. This study expands on this perspective by exploring what therapists can learn about their patients, thereby providing a panoramic view of the therapeutic process. Utilizing three distinct perspectives, as depicted in Figure 1, helps therapists understand the clinical landscape reflected by the patient across different treatment phases. This expansive outlook is prognostically significant and encourages clinicians to tailor treatment strategies to meet the dynamic needs of their patients.

The study was conducted in accordance with APA ethical standards. The study has undergone Institutional Review Board evaluation and has received approval at the relevant institution (the University of Haifa Ethical Committee approved the execution of the study) (Approval No.: 118/15, date: October 10, 2015). All participants, both patients and therapists, provided written and oral informed consent by signing forms authorizing participating in the study, the videotaping of their treatment sessions, and the use of all quantitative and qualitative data for research purposes, in which personal information was to be anonymized.

The authors have no conflicts of interest to declare.

The research was supported by two grants from the Israel Science Foundation (ISF) (Grant No. 186/15 and 395/19; principal investigator: Sigal Zilcha-Mano). The funders had no role in data collection, analysis, or interpretation.

Shimrit Fisher led the research project and contributed to the conceptualization, methodological design, data collection, and analysis and wrote the first draft of this paper. Peter Fonagy contributed to the conceptualization, methodological design, interpretation, and validation of the work and writing, revising, and editing the manuscript. Sigal Zilcha-Mano contributed to the conceptualization, methodological design, funding acquisition, and interpretation of the work and writing, revising, and editing of the manuscript.

Research data are inaccessible to the public due to ethical considerations. Further inquiries can be directed to the corresponding author.