The Relationship Between Expressive Suppression, Therapeutic Bond, and Treatment Outcomes of a Positive Affect Intervention for Adults with Anxiety and/or Depression

Abstract

Expressive suppression (ES; reducing emotional expression) is linked with reduced social connectedness in individuals with anxiety or depression. One implication is that people who use ES may have difficulty establishing a bond with their therapist which may impede clinical improvement. We examined this hypothesis in 33 adults with clinically elevated anxiety or depression receiving treatment focused on enhancing positive thoughts, emotions, and behaviors. At baseline, participants rated ES for positive and negative emotions during a standardized conversation task designed to generate connectedness. They also rated measures of early (session 3) perceived therapeutic bond and treatment outcomes (i.e., positive affect and social connectedness). ES of positive (r=−.39, p=.018), but not negative (r=.06, p=.747), emotions was negatively associated with therapeutic bond. Therapeutic bond mediated the relationship between greater ES of positive emotions during affiliation and lower post-treatment positive affect, 95% bias-corrected bootstrap confidence interval [−0.021, −0.000], adjusted for pre-treatment positive affect, as well as lower post-treatment social connectedness [−0.397, −0.015]; however, the indirect effect was not significant when accounting for pre-treatment social connectedness (p>.05). ES of positive emotions may be an important factor in the development of therapeutic bond and therefore treatment outcomes for individuals with anxiety or depression.

The capacity to connect with others is important for psychological well-being (Baumeister & Leary, 1995). Unfortunately, social disconnection and impairments in social functioning are commonly reported by individuals with anxiety or depression (Saris et al., 2017). Expressive suppression (ES) – an emotion regulation strategy that involves attempts to block observable signs of one’s felt emotional experience – is particularly important for relationship formation (Gross & John, 2003). Individuals with anxiety or depression report more frequent use of ES compared to individuals without either condition (Campbell-Sills et al., 2006; Dryman & Heimberg, 2018). Greater use of ES is consistently associated with worse social outcomes, such as fewer positive social events, less perceived social support, and poorer quality friendships in samples of individuals with anxiety or depression (Dryman & Heimberg, 2018; Farmer & Kashdan, 2012; Kashdan & Breen, 2008; Spokas et al., 2009).

Previous studies have used dyadic paradigms involving unacquainted conversational partners to investigate the link between ES and social affiliation. In these paradigms, one participant in each pair was instructed to use a given emotion regulation strategy (i.e., ES, cognitive reappraisal, or natural response) in the context of discussing distressing topics (e.g., nuclear war) whereas the other participant remained uninstructed (Butler et al., 2003; Butler et al., 2007; Peters et al., 2014). Results indicated that individuals in the ES conditions perceived their partners as less responsive to them (Peters & Jamieson, 2016) and their partners reported fewer feelings of affiliation (Butler et al., 2003; Butler et al., 2007) and rated them as worse communicators (Peters et al., 2014). Findings from these studies support the notion that ES of emotions may precipitate negative social outcomes.

Both positive and negative emotions play unique roles in social connection (Alexander et al., 2021; Fredrickson, 2001; Kuppens et al., 2008). Whereas positive emotions promote approach behaviors (e.g., exploration/curiosity, play, social initiation) and guides individuals towards potentially pleasurable or rewarding outcomes (e.g., social interactions, accomplishments), negative emotions promote avoidance behaviors (e.g., withdrawal, isolation) and guides individuals away from potentially aversive or threatening outcomes (e.g., social rejection, loss; Alexander et al., 2021; Fredrickson, 2001; Kuppens et al., 2008). It is thus plausible that ES of positive and negative emotions may be differentially related to social functioning. Consistent with this, we found that ES of positive emotions was more strongly related to fewer feelings of connectedness to one’s partner from both sides of the dyad compared to ES of negative emotions during a standardized relationship formation task in adults with anxiety and/or depression (Hoffman et al., 2022). Results suggest that ES of positive emotions, in particular, may bode poorly for relationship formation.

One fundamental relationship to develop in treatment-seeking individuals is that of the patient and therapist. While the therapeutic relationship shares some characteristics with general social relationships, it possesses unique qualities that are considered intrinsically healing (Wampold & Imel, 2015). The therapeutic bond is one aspect of the therapeutic relationship that can be characterized as the affective (e.g., mutual positive feelings, trust, respect) and collaborative alliance (e.g., agreement on therapeutic tasks, shared decision-making) that develops between a patient and therapist. Positive therapeutic alliance is one of the most reliable predictors of psychotherapy outcome regardless of therapy type (Ardito & Rabellino, 2011; Flückiger et al., 2012; Horvath et al., 2011; Krupnick et al., 1996; Norcross & Wampold, 2019), and especially when it is assessed at early stages of treatment (Horvath & Symonds, 1991; Klein et al., 2003; Martin et al., 2000; Zuroff & Blatt, 2006). This empirical evidence aligns with patient perspectives describing the therapeutic relationship as a key determinant of treatment success (Johansson & Eklund, 2003). Evidence from meta-analyses further supports the notion that forming a warm emotional bond and collaborative attachment is essential for a strong therapeutic relationship (Flückiger et al., 2019) and that the expression of affect during psychotherapy sessions by both therapists and clients is significantly and moderately-to-largely predictive of an improved therapeutic process and therapy outcomes (Peluso et al., 2018). Taken together, these data suggest it would be valuable to identify the factors that contribute to the formation of a strong therapeutic relationship.

Given established associations between ES and social affiliation, ES of positive and negative emotions could play a role in the development of the therapeutic bond for those undergoing treatment for anxiety and depression. This in turn could influence treatment outcomes. There have been no studies, to the best of our knowledge, to examine these associations. To address this, we conducted a secondary analysis of a clinical trial on a positive valence system intervention for anxiety and depression (Taylor et al., 2023). The analysis was pre-registered and is available at Open Science Framework (https://osf.io/wt3up)¹. We analyzed pre-treatment data from a standardized, confederate-controlled dyadic paradigm intended to induce social affiliation between unacquainted partners during which participants rated their ES of positive and negative emotions. The current study investigated the following research questions:

1. Is participant ES of positive and/or negative emotions during social affiliation associated with participant/therapist perceived therapeutic bond early in treatment? Given evidence suggesting ES is linked to negative social outcomes and because treatment sessions included discussion of positive and negative valence experiences, we hypothesized ES of either emotion during social affiliation would be significantly, negatively related to early perceived therapeutic bond.

2. Does participant/therapist perception of therapeutic bond (M) early in treatment account for the link between ES of positive/negative emotions during affiliation (X) and treatment outcomes (Y)? Given the established link between therapeutic alliance and treatment response, we hypothesized indirect effects of ES of positive and/or negative emotions on treatment outcomes via therapeutic bond.

Methods

Participants

The sample included 36 individuals between the ages of 18 and 55 enrolled in a clinical trial (ClinicalTrials.gov Identifier: NCT04945239) that selected for participants with clinically elevated symptoms of depression and/or anxiety, defined as scoring 10 or higher on the Patient Health Questionnaire-9 (PHQ-9; Kroenke et al., 2001) and/or scoring eight or higher on the Overall Anxiety Severity and Impairment Scale (OASIS; Norman et al., 2006), respectively. Participants were also required to score 90 or lower on the Social Connectedness Scale Revised (SCSR; Lee et al., 2008) and five or higher on the Sheehan Disability Scale–Social Domain (SDS-SD; Leon et al., 1997) – reflective of moderate or greater social disconnection and impairment. Participants were recruited through clinical referrals as well as posted announcements in community and online settings (e.g., ResearchMatch.org).

Assessments to determine principal and comorbid diagnoses were conducted using the Mini International Neuropsychiatric Interview (MINI; Sheehan et al., 1998 Version 5.0.0 or 7.0.0). Exclusionary criteria used to determine parent study eligibility were: (a) active suicidal ideation with intent or plan; (b) moderate to severe alcohol or marijuana use disorder (past year); (c) all other mild substance use disorders (past year); (d) bipolar I or psychotic disorders; (e) moderate to severe traumatic brain injury with evidence of neurological deficits, neurological disorders, or severe or unstable medical conditions that might be compromised by participation in the study; (f) inability to speak or understand English; (g) concurrent psychotherapy (unless 12-week stability criteria had been met for non-empirically supported therapies only); (h) concurrent psychotropic medication (e.g., selective serotonin reuptake inhibitors, benzodiazepines); and (i) characteristics that would compromise safety to complete an magnetic resonance imaging (MRI) scan (e.g., metal fragments in body). The MRI safety exclusion criterion (i) was specific to the primary objectives of the parent study and was not a consideration in the selection criteria for the current secondary analysis. The sample had a mean age of 30.25 years (SD = 9.46), was predominantly female (69.4%), and demographically diverse. See Table 1 for additional participant demographics.

Table 1.

Participant Demographics

	M(SD) or %
Age	30.25 (9.46)
Gender
Female	69.4
Male	27.8
Did not identify as female or male	2.8
Education (years)	15.44 (2.18)
Ethnicity
Hispanic/Latino	16.7
Not Hispanic/Latino	83.3
Race
Asian	30.6
Black	2.8
White	58.3
More than one race/Other	5.6
Native Hawaiian or Pacific Islander	2.8
Principal Clinical Diagnosis
Major Depressive Disorder	54.3
Social Anxiety Disorder	28.6
Generalized Anxiety Disorder	14.3
Obsessive Compulsive Disorder	2.9

Note. N = 36

Procedure

All procedures performed involving human participants were in accordance with the ethical standards of the University of California San Diego Human Research Protection Program and with the Code of Ethics of the World Medical Association (Declaration of Helsinki). All participants provided informed written consent before participating in the study. Within two weeks of completing the eligibility assessment, participants completed self-report measures and the social affiliation paradigm on a separate, baseline visit. Approximately one week later, participants started the positive valence system treatment (described below). Ten weeks after the baseline visit, participants completed a post-treatment assessment during which they completed measures of PA and social connectedness.

Social Affiliation Paradigm

The current study utilized a modified version of the task developed by Aron and colleagues (1997). The task reliably induces social connectedness and demonstrates strong psychometric properties which are reported in Hoffman et al. (2021). Participants and trained confederates alternated responses to a series of three 6-minute question sets with each question set gradually increasing in intimacy level. An experimenter informed the participant they would be getting to know an assistant who worked in the lab (i.e., the confederate) before the start of the task. With the confederate present, the experimenter stated that the purpose of the task was to get to know one another by answering a series of questions about themselves. Confederates and participants alternated responses and completed ratings following each question set (described below) on a separate form and out of their partner’s sight. The interaction lasted 18 minutes. A description of personnel and confederate training is presented in the Supplemental Materials.

Treatment

Participants completed a positive valence system intervention, Amplification of Positivity (AMP), focused on enhancing social reward sensitivity through targeting positive thoughts, emotions, and behaviors (Taylor et al., 2017; 2023). Consistent with the central outcomes in this line of work (Taylor et al., 2017; Taylor et al., 2020; Taylor et al., 2023), PA and social connectedness were selected as the primary and secondary outcomes in the current study, respectively. AMP is an individual, manualized, and clinician-delivered psychotherapy comprised of three core elements: (1) increasing exposure and responsiveness to positive events; (2) practicing gratitude; and (3) engaging in kind or generous acts towards others. Participants were randomized to a 5-session or 10-session treatment dose, or a waitlist control group. Only participants who were randomized to the 5-session (n = 23) or 10-session (n =22) treatment arm were analyzed in the current study. Both 5- and 10-session protocols were identical in the initial four treatment strategies (noticing and amplifying positive events; gratitude reflection; acts of kindness; scheduling pleasurable, engaging, and meaningful activities) – the point at which early perceptions of the therapeutic bond were assessed. The AMP protocol consisted of 1-hr, in-person sessions delivered by doctoral-level and master’s level clinicians. See Taylor et al. (2017; 2023) for a full description of the treatment design and procedure.

Measures

Social Affiliation Paradigm Measures

Expressive Suppression.

Immediately after each 6-minute question set during the social affiliation task with the confederate, participants rated the following questions: “When I was feeling positive emotions, I was careful not to express them” and “When I was feeling negative emotions, I was careful not to express them.” Participants rated their response on a scale ranging from 0 (not at all) to 100 (extremely; Gross & John, 2003; Kashdan & Breen, 2008).² Average levels of positive and negative ES during the social affiliation task were computed separately for each participant. The Cronbach’s alpha for ES of positive emotions was α = .92 and α = .95 for ES of negative emotions.

Treatment Measures

Therapeutic Bond.

The Working Alliance Inventory – Short Revised (WAI-SR; Hatcher & Gillaspy, 2006) is a measure of the therapeutic alliance. The WAI-SR contains 12 items that measure three aspects of the therapeutic alliance: agreement on the tasks of therapy (Task), agreement on the goals of therapy (Goal), and development of an affective bond between patient and therapist (Bond). Items are rated on a 5-point Likert scale ranging from 1 (seldom) to 5 (always). Participants and therapists completed the WAI-SR at the end of treatment sessions 3 and 8. Because we were interested in early relationship formation and participants across both treatment doses completed the WAI-SR at session 3 (n = 21 in 5-session arm; n = 15 in 10-session arm), only the WAI-SR completed at session 3 was analyzed in the current project. The total score on the Bond subscale was used as a measure of therapeutic bond in the current project because it reflects the perceived connection with one’s therapist/patient (cf. agreement on tasks or goals). The Cronbach’s alpha was α = .90.

Positive Affect.

Positive affect (PA) was the primary treatment outcome of interest for the current study. To create a more valid and reliable index of self-reported PA, scores from three established self-report measures were combined. Specifically, PA was measured using the Positive and Negative Affective Schedule (PANAS; Watson et al., 1988), Modified Differential Emotion Scale (mDES; Fredrickson et al., 2003), and Patient-Reported Outcome Measurement Information System Positive Affect Short Form 15a (PROMIS PA; Cella et al., 2010). The PANAS is a widely used measure comprising 20-items assessing activated forms of PA and negative affect (NA) using 5-point scales ranging from 1 (very slightly/not at all) to 5 (extremely). The mDES asks participants to assess how intensely they are experiencing 20 discrete emotions. Ten positive emotions (e.g., joy, love, awe) and 10 negative emotions (e.g., guilt, anger, fear) are rated on a five-point scale ranging from 0 (never) to 4 (most of the time). The PROMIS PA asks participants to rate how much they are feeling positive emotions using 15 items on a 5-point scale ranging from 1 (not at all) to 5 (very much). Each scale was administered and pre- and post-treatment. Participants were instructed to rate each of the measures according to how they felt during the past week with higher scores indicating greater PA levels. The raw PA scores for each scale were converted into z-scores and averaged to create pre-treatment (α = .81) and post-treatment (α = .94) composite scores.

Social Connectedness.

The Social Connectedness Scale – Revised (SCS-R; Lee et al., 2008) was used as a secondary treatment outcome measure of interest. The SCS-R is a 20-item self-report measure of the degree of connectedness to others and belongingness in a social environment, and thus social relationship functioning. Items are scored on a 1 (strongly disagree) to 6 (strongly agree) scale. Higher scores on the SCS-R reflect a stronger sense of social connectedness. The SCS-R was administered and pre- and post-treatment and the total score on the SCS-R assessed global social connectedness. The Cronbach’s alpha was α = .93.

Anxiety and Depression Symptoms

Pre-treatment anxiety symptoms were measured using the OASIS (Bragdon et al., 2016; Moore et al. 2015; Norman et al., 2006; Norman et al., 2011). The OASIS is a transdiagnostic measure of the frequency and severity of anxiety symptoms, as well as the level of anxiety-based avoidance and interference during the previous two weeks. Five items range from 0 (never/none/not at all) to 4 (constantly/extreme/all the time). The sample’s average OASIS score (M = 10.11, SD = 3.55) and the Cronbach’s alpha was α = .88.

Pre-treatment depression symptom severity during the past two weeks based on DSM-5 symptom criteria were measured using the PHQ-9 (Kroenke et al., 2001). The PHQ-9 consists of nine items ranging from 0 (not at all) to 3 (nearly every day). The sample’s average PHQ-9 score (M = 12.97, SD = 4.26) and the Cronbach’s alpha was α = .72.

Data Analysis

All analyses were conducted in IBM SPSS, Version 26. Descriptive statistics of key variables were first examined. Correlation coefficient effect sizes were interpreted as small (r = .10), medium (r = .30), and large (r = .50; Cohen, 1988). There were no significant differences in baseline PA (t = .88, p = .383) or social connectedness (t = .53, p = .603) scores between the 5- and 10-week treatment arms. Three participants did not complete post-treatment data and thus were excluded from post-treatment analyses³. The pre-registration of analyses intended to examine both participant and therapist perception of early therapeutic bond. However, therapist perception of early therapeutic bond was removed from analyses upon inspection of the data due to violations of the assumptions of linearity and homoscedasticity (assessed via scatterplots and residual plots), normality (assessed via Q-Q plots, P-P plots, and the Kolmogorov-Smirnov, D = .333, p <.001 and Shapiro-Wilk, W = .658, p < .001, tests), and low variance (possible range of scores: 4 to 20; actual range of scores: 15 to 20). Bivariate correlations between positive and negative ES, participant perception of therapeutic bond, post-treatment PA and social connectedness, demographic variables (age, gender, years of education, ethnicity [Hispanic/Latino vs. non-Hispanic/Latino]), and anxiety and depression symptoms were first examined. Then, ES of positive and/or negative emotions was contingently entered as an independent variable into separate multiple linear regression models predicting participant perception of therapeutic bond, while adjusting for any significant demographic covariates and/or symptoms of anxiety and depression related to the outcome. The assumptions of linearity and homoscedasticity (assessed via scatterplots and residual plots), normality (assessed via Q-Q plots, Q-Q plots, and the Kolmogorov-Smirnov, D = .132, p = .155, and Shapiro-Wilk, W = .912, p = .011, tests), and multicollinearity (assessed via variance inflation factor < 5.00) were tested and confirmed. Multivariate outliers of study variables were screened by examining standardized residuals (± 3), Mahalanobis distance (χ²(3) > 16.27) and Cook’s distance (>1; Tabachnick & Fidell, 2007). No multivariate outliers were identified. Separate simple unmoderated mediation models were run using the PROCESS macro (Hayes et al., 2012) to assess the effect of positive and/or negative ES (X) on treatment outcomes (Y) through participant perception of therapeutic bond (M). The indirect effects were tested using a bootstrap estimation approach with 5,000 samples.

Results

Bivariate correlations and statistical properties between study variables are presented in Table 2. Demographic (age, years of education, ethnicity) and symptom (anxiety, depression) variables were not significantly associated with therapeutic bond or ES of positive emotions (ps >.05). ES of negative emotions was positively correlated with anxiety symptoms (r = .38, p = .021). ES of positive and negative emotions were moderately correlated (r = .37, p = .028). There were no other significant associations between demographic or symptom variables and treatment outcomes (ps > .05).

Table 2.

Bivariate Correlation Between Study Variables

Variable	M	SD	1	2	3	4	5	6	7
1. ES of positive emotions	17.92	17.82	—
2. ES of negative emotions	49.61	31.29	.37*	—
3. Therapeutic bond	16.39	3.55	−.39*	.06	—
4. Post-treatment positive affect	−0.01	0.82	−.37*	−.02	.65***	—
5. Post-treatment social connectedness	72.93	18.94	−.36*	.06	.48**	.85***	—
6. Anxiety symptoms	10.11	3.55	.33	.38*	.19	.03	−.15	—
7. Depression symptoms	12.97	4.26	.16	.08	.05	−.26	−.48**	.51**	—

Note. ES = expressive suppression. ES of positive/negatives emotions was measured prior to treatment (baseline). Therapeutic bond was measured during treatment (session 3). Post-treatment positive affect and social connectedness were measured after all participants completed treatment (10 weeks following baseline).

^*p < .05.

^**p < .01.

^***p < .001.

Therapeutic bond was inversely associated with ES of positive emotions (r = −.39, p = .018), but was not significantly associated with ES of negative emotions (r = .06, p = .747). A mediation analysis indicated a statistically significant indirect effect of ES of positive emotions on post-treatment PA through participant perceived therapeutic bond when adjusting for pre-treatment PA, 95% bias-corrected bootstrap confidence interval [−0.0212, −0.0002]. Greater ES of positive emotions during social affiliation was associated with lower participant perceived therapeutic bond early in treatment, which was associated with less PA following treatment when accounting for pre-treatment levels of PA (see Figure 1). A second mediation model showed a significant indirect effect of ES of positive emotions on post-treatment social connectedness through participant perceived therapeutic bond, [−0.3971, −0.0153]. That is, greater ES of positive emotions during social affiliation was associated with lower participant perceived therapeutic bond early in treatment, which was associated with less social connectedness following treatment (see Supplemental Figure 1). The indirect effect was not significant when accounting for pre-treatment levels of social connectedness, [−0.2975, 0.0071].

Figure 1. Adjusted Indirect Effects Test of Expressive Suppression of Positive Emotions on Post-treatment Positive Affect via Therapeutic Bond.

Note. ES = expressive suppression; PA = positive affect. Model is adjusted for pre-treatment positive affect scores. ES of positive emotions was measured prior to treatment (baseline). Therapeutic bond was measured during treatment (session 3). Post-treatment PA was measured after all participants completed treatment (10 weeks following baseline).

Discussion

Expressive suppression (ES) disrupts social affiliation in individuals with anxiety and depression (Dryman & Heimberg, 2018; Farmer & Kashdan, 2012; Kashdan & Breen, 2008; Spokas et al., 2009), suggesting it may undermine formation of the bond between a patient and their therapist. The current study is the first, to our knowledge, to examine this possibility – using a standardized dyadic task to examine whether use of ES during relationship formation is relevant for early perceived therapeutic bond and thus treatment outcomes for adults with anxiety and/or depression. Consistent with hypotheses, ES of positive emotions during social affiliation was related to participants’ early perception of therapeutic bond and treatment outcomes. Simple mediation analyses demonstrated that lower perceived therapeutic bond early in treatment accounted for the association between greater use of ES for positive emotions during social affiliation and poorer treatment outcomes for PA and social connectedness. Findings of the separate mediation analyses remained consistent when adjusting for pre-treatment levels of PA, but not for pre-treatment levels of social connectedness.

Engaging in ES of positive emotions might be particularly important for therapeutic bond and thus treatment outcomes for several reasons. Relationship formation occurs through a dynamic, transactional process between two individuals in contexts that promote mutual engagement and self-disclosure. This process involves sharing one’s own experiences, feelings, and perspectives while also providing the opportunity for the other person to respond (Aron et al., 1997, Reis and Shaver, 1988). Positive emotions serve as signals of rapport and connection during social affiliation (Fredrickson, 2013). The expression of positive emotions (e.g., smiling, laughing) facilitates relationship development by eliciting reciprocal affiliation behaviors (Reis & Shaver, 1988). This may be especially relevant in the therapeutic context of AMP, which emphasizes sharing of positive experiences by patients as part of their home practice review. The therapist may respond to a patient’s expression of positive emotions via validation, affirmation, and their own expression of positive emotions. Reciprocal exchange of positive emotions between patient and therapist could in turn promote the mutual support, safety, and trust inherent to a strong therapeutic bond.

ES in general may impair one’s ability to interpret other’s emotions based on their facial expressions (Schneider et al., 2013). ES of positive emotions may thus make it challenging for therapists to accurately perceive patients’ signals which could create missed opportunities for the dyad to mutually engage in the affiliation process. Furthermore, ES of positive emotions may be inversely related to the experience of PA (Nezlek & Kuppens, 2008) which has been linked to increases in perceived connectedness (Taylor et al., 2017). In this way, individuals who engage in ES of positive emotions may feel less connected to their therapist, hampering the development of therapeutic bond.

ES of negative emotions during social affiliation was not significantly related to early perceived therapeutic bond. This finding is consistent with prior findings demonstrating an association between feelings of connectedness during initial affiliation and ES of positive, but not negative, emotions (Hoffman et al., 2022). It should be noted, however, that the current and previous study examined ES of positive and negative emotions in a social interaction context wherein social relationships are encouraged to develop (i.e., unacquainted pairs instructed to get to know one another). It is possible that ES of negative emotions may be differentially related to feelings of connectedness in other social contexts (e.g., existing relationships or in discussions of distressing topics) and research that examines theses associations is needed. Importantly, the current study investigated the role of ES for positive and negative emotions within a psychotherapy specifically focused on enhancing positive thoughts, feelings, and behaviors (Taylor et al., 2020; Taylor et al., 2023). This therapeutic approach may have created a context in which the expression or suppression of negative emotions was less central to the treatment goals.

Several study limitations should be considered. The current findings may not generalize to non-Western samples, as cultural norms and values influence the use of emotion regulation strategies, including ES. A meta-analysis suggested that use of ES may be more congruent with Eastern cultural social norms (Hu et al., 2014). ES was measured during a controlled conversation with a single interaction partner and may therefore not generalize to other types of interaction partners or contexts. Future research could benefit from assessing the use of ES within therapeutic interactions to gain a more nuanced understanding of its direct effect on the therapeutic relationship. Nonetheless, in conjunction with our previous research (Hoffman et al., 2022), findings of the current study suggest ES of positive emotions during the social affiliation paradigm with one interaction partner (the confederate) predicted affiliation in another interpersonal context with a different interaction partner (the therapist), indicating the paradigm could be useful for understanding the social consequences of ES more broadly. An additional limitation to consider is the potential for bias towards more positive evaluations in therapist’s self-assessments, which could have contributed to the low variability observed in their ratings of therapeutic bond. Consequently, we were unable to assess the association between ES and therapist perception of therapeutic bond in the current study, limiting the comprehensiveness of our understanding of factors relating to the therapeutic bond. The social connectedness task was designed as an experimental paradigm used to model initial affiliation processes rather than as a clinical tool to directly improve connectedness. Exploring the effects of the task and therapeutic bond in less interactive treatment formats (e.g., online self-guided interventions) may be a valuable avenue for future research. Our study focused on a specific emotion regulation strategy and therapeutic approach, highlighting the need for future research to replicate the link between ES and therapeutic bond across different therapies. Future research that also explores the role of other emotion regulation strategies (e.g., cognitive reappraisal) in the therapeutic relationship would be valuable as they may differentially influence treatment outcomes. The mediation analyses conducted within the current study are preliminary and cannot establish the causal mechanisms of the relationships. Given the modest sample size of the current study, the power to detect subtle yet potentially meaningful associations is constrained and results should be interpreted with caution. Research within larger samples is needed to replicate and extend these preliminary findings.

Findings from the current study suggest that the use of ES may be a potentially valuable clinical variable of interest. Future research should explicitly examine ES within the therapeutic context as a mediator in the relationship between early therapeutic bond and treatment outcomes. Actively promoting the experience and expression of positive emotions through activities that elicit PA may allow for greater processing and expression of positive emotions, ultimately facilitating enhanced perceived social connectedness (Carl et al., 2013; Taylor et al., 2020; Taylor et al., 2023), including strengthened therapeutic bond. Although preliminary, results of the current study suggest that ES of positive emotions, in particular, may play an important role in the development of therapeutic bond and psychotherapy outcomes for individuals with anxiety and/or and depression.