Emotional Changes during Imagery Rescripting of Aversive Social Memories in Social Anxiety Disorder: A Randomized Controlled Trial

The experience of distorted self-images, frequently associated with past aversive social experiences, is a common feature of social anxiety disorder (SAD) [1‒3]. Changing the memory representation of these experiences to alter mental images has become a goal of psychotherapeutic interventions, such as imagery rescripting (ImRs) [4‒6]. During ImRs, patients reexperience their aversive memory from their child perspective (phase 1), then observe and intervene from their adult perspective to support and meet the needs of their child self (phase 2), and finally, reexperience the scene again from their child perspective observing the actions of their adult self (phase 3) [5, 6].

From a learning theory perspective, the intervention is assumed to modify the original or create a competing representation of the aversive event in memory, while cognitive perspectives consider changes in the meaning of the aversive memory and related schemata to be a central mechanism [4, 7, 8]. Both theories are underpinned by findings of ImRs resulting in stronger positive affective responses during aversive memory recall after ImRs, as well as positive changes in memory appraisals, associated core beliefs, and/or perceived mastery [9‒16].

The use of imagery in psychotherapeutic interventions has gained increasing attention, as imagery compared to verbal processing results in a stronger emotional activation [17], which in turn is theorized to be a necessary condition for emotional change (“activation of fear structure” [18]). In patients with SAD, ImRs/imagery-enhanced cognitive behavioral therapy resulted in a stronger modulation of physiological responses associated with emotional activation (e.g., attenuated heart rate [HR] during baseline [19] or increased heart rate variability [HRV] during a subsequent speech task [20]). HRV, reflecting the variation of intervals between heartbeats, represents the interplay of sympathetic (HR increases) and parasympathetic activity (HR decreases) [21, 22]. One prominent indicator of HRV, the root mean square of successive differences (RMSSD), predominantly indicates parasympathetic activation (while frequency-domain HRV parameters also reflect sympathetic activity) [23]. The flexible interaction of both branches of the autonomic nervous system is of particular importance for an adaptive physiological response to environmental demands and is, as such, also closely linked to the ability to regulate one’s emotions [22]. Accordingly, previous studies have demonstrated that increases in HRV are associated with emotion regulation [24], as well as with changes in activity of brain regions associated with emotion regulation processes [25]. Therefore, the abovementioned increased HRV during a speech task following ImRs probably indicates increased regulatory control over autonomic responses [20]. Promoting such regulatory processes might be crucially important, as previous research demonstrated higher HR during fearful imagery in SAD [26‒28], indicating increased sympathetic/decreased parasympathetic activity [21, 29, 30].

While previous findings provide valuable insights into affective, cognitive, and physiological changes as a result of ImRs, processes that occur during the intervention have rarely been investigated. One recent study [31], however, showed a large increase in positive affective valence from phase 1 to phase 2 of ImRs in patients with SAD, indicating that ImRs supports patients in adaptively regulating their emotions during the intervention. According to Gross, emotion regulation comprises “all processes by which individuals influence which emotions they have, when they have them, and how they experience and express these emotions” ([32], p. 275). In line with this, emotion regulation might be a key process underlying emotional changes during ImRs [33], probably leading to the abovementioned affective, cognitive, and physiological effects of ImRs. Autonomic responses (e.g., HR, HRV) during ImRs as indicators of underlying emotional and regulatory processes (e.g., initial activation of the aversive memory during reexperiencing, emotion regulation during rescripting) have, however, not been examined in patients with SAD so far.

The current study aimed to fill this gap by investigating physiological correlates (HR, HRV) as well as changes in self-reported positive and negative feelings in the course of one session of ImRs compared to a verbally-based control intervention in patients with SAD. We hypothesized a larger increase in negative feelings/decrease in positive feelings during reexperiencing (baseline – phase 1), as well as a larger decrease in negative/increase in positive feelings in response to the actual rescripting phases (phase 1 – phase 2, phase 2 – phase 3) for ImRs compared to the control intervention. On the physiological level, ImRs is expected to result in higher mean HR/reduced HRV (RMSSD) during reexperiencing of the aversive memory in phase 1, as well as higher HRV (RMSSD) during phase 2 and phase 3 (each compared to baseline and the control intervention), indicating more adaptive emotion regulation and parasympathetic modulation of autonomic arousal.

Seventy-seven patients with SAD recruited through the local university (outpatient clinic, postings) were randomly assigned to ImRs (n = 38) and the control intervention (n = 39; matched for sex and SAD symptom severity by a predetermined table, see online Supplement A; for all online suppl. material, see https://doi.org/10.1159/000539402). Participants were blind to treatment assignment. Baseline demographics and clinical characteristics did not differ significantly between groups (see Table 1). A priori power analysis revealed a sample size of 54 participants to be sufficient to detect significant within-between interactions at power of 0.95 and medium effect sizes (G*Power, Version 3.1.9.2) [34]. All participants exceeded cutoff values in the clinician-rated version of the Liebowitz Social Anxiety Scale (LSAS) [35, 36] and the Social Phobia Inventory (SPIN) [37, 38] and fulfilled the DSM-5 criteria for a primary diagnosis of SAD according to a structured diagnostic interview (see online Supplement B). The trial was preregistered at the German Clinical Trial Register (DRKS00016729).

On the first day, diagnostic interviews and questionnaires were conducted to assess study eligibility. In addition, the aversive social memory targeted by the interventions was selected (see online Supplement C). On the second day (9 to 23 days later), a 5-min baseline measurement was conducted. Immediately afterward, participants received either one session of ImRs or a control intervention. Participants rated the intensity of positive and negative feelings after baseline and during each of the three intervention phases directly after each phase, and electrocardiogram (ECG) was measured throughout the interventions. This study was part of a larger project with additional measures and three additional days (not including further psychotherapeutic interventions, see online Supplement D).

The protocol was subdivided into three phases [5, 6]. During phase 1, participants reexperienced their aversive social memory from their former perspective. In phase 2, participants imagined the same event from an observer’s perspective as their adult self. They were encouraged to carry out changes to the situation to support their former selves. The procedure was repeated if participants did not experience a decrease in negative feelings or if they were not satisfied with the new ending of the situation. During phase 3, participants were again asked to reexperience the event from their former perspective, but with all the changes introduced by the adult self in phase 2.

The control intervention was adapted from the ImRs protocol and consisted of three questions concerning the aversive social memory (see [39]). First, participants were asked to recall every detail regarding the aversive social event in general, second, regarding the location where the event took place, and third, regarding the persons involved in the event. During each phase, they were asked to write down what they remembered regarding the respective question while the experimenter left the room for 3 min. After that, the answers were discussed with the experimenter.

After each phase, participants rated the intensity of positive and negative feelings they experienced at the moment after baseline and during each of the three intervention phases on 9-point scales from 0 (“not at all”) to 8 (“very strong”).

The ECG was recorded throughout the baseline measurement, as well as the experimental interventions. For details regarding ECG recording and analyses, see online Supplement E. In addition to mean HR as an indicator for sympathetic/parasympathetic activity, root mean square of successive differences (RMSSD) was selected as the HRV measure and most likely reflects parasympathetic activity (for further parameters, see online suppl. Table 1).

Statistical analyses were conducted by means of IBM SPSS Statistics, version 24. The level of significance was determined at α = 0.05 (two-tailed). Shapiro-Wilk test revealed a violation of normal distribution for self-reported feelings, HR and HRV data, which is why non-parametric testing was applied. Therefore, difference scores between each successive pair of phases for self-reported feelings (phase 1 – baseline, phase 2 – phase 1, phase 3 – phase 2), as well as for each phase compared to baseline (phase 1 – baseline, phase 2 – baseline, phase 3 – baseline) for HR and HRV were computed. Differences between the intervention groups for these difference scores, as well as separately for each phase (follow-up tests), were analyzed by means of Mann-Whitney U tests. Follow-up Wilcoxon signed-rank tests were applied to compare differences between phases within groups.

Results (see Table 2, Fig. 1, online suppl. Table 2) showed a larger increase in negative feelings from baseline to phase 1 in the ImRs compared to the control intervention group, as well as for both groups separately. The groups did not differ significantly in negative feelings during baseline and phase 1.

From phase 1 to phase 2 of ImRs, negative feelings more strongly decreased and positive feelings more strongly increased compared to the control intervention. These changes were also evident in both groups separately. Moreover, the groups did not differ significantly in negative feelings during phase 2, but ImRs compared to the control intervention resulted in stronger positive feelings during phase 2, but not during phase 1.

The interventions also differed regarding changes in self-reported feelings from phase 2 to phase 3: while no significant changes in self-reported feelings were found from phase 2 to phase 3 of ImRs, the control intervention resulted in an increase in negative feelings and a decrease in positive feelings. Moreover, ImRs compared to the control intervention resulted in lower negative feelings and stronger positive feelings during phase 3.

Analyses revealed an increased HR during phase 1 compared to baseline for ImRs compared to the control intervention, as well as within the ImRs and the control intervention group, respectively (see Table 2, Fig. 1). The groups, however, did not differ significantly in HR during baseline and phase 1 (see online suppl. Table 2). Furthermore, there were no significant group differences in HR for phases 2 and 3, each compared to baseline.

ImRs compared to the control intervention was associated with higher HRV during phase 3 (but not during phases 1 and 2) compared to baseline (see Table 2, Fig. 1). Moreover, HRV was higher for phase 3 versus baseline within the ImRs but not within the control intervention group. However, HRV did not differ significantly between groups during baseline and phase 3, respectively (see online suppl. Table 2).

ImRs is a promising intervention in the treatment of SAD [40], leading to prominent changes in affective, cognitive, and physiological responses toward the aversive memory and/or stressful social situations [10, 13, 41]. Extending these previous findings, the results of the current study demonstrate a specific pattern of affective and physiological responses unfolding during ImRs compared to a verbally-based control intervention. Increased HR and negative feelings during reexperiencing (phase 1 compared to baseline) might represent an activation of the fear structure associated with the aversive memory, which is postulated to be a necessary condition for emotional change [18]. Subsequently, ImRs resulted in a significant decrease in negative and increase in positive feelings from phase 1 to phase 2, replicating previous findings [31]. Extending these findings, affective changes were significantly larger (phase 1 vs. phase 2) and more stable (phase 2 vs. phase 3) compared to a verbally-based control intervention. Moreover, phase 3, compared to baseline, was associated with higher HRV (RMSSD) in the ImRs compared to the control intervention group. This finding aligns with a previous study that reported increased HRV during a speech task after a single group session of ImRs, indicating enhanced parasympathetically mediated control over physiological responses as a result of ImRs [20]. As HRV has previously been associated with emotion regulation processes [24, 42, 43], a larger increase in HRV during phase 3 might point to regulatory processes taking place during the intervention (e.g., reappraisal [33]), probably leading to the abovementioned alterations in positive and negative feelings. This might be related to the theorized modification of the original or the creation of a new representation of the aversive event in memory and associated changes in the meaning of the memory [4, 7, 8, 18], which has to be investigated in future studies.

Despite these new and promising findings, the current study has some limitations. Competing processes taking place within each phase of the intervention (e.g., emotional activation and regulation) might contribute to nonsignificant differences in physiological measures between ImRs and the control intervention, e.g., regarding HRV during phase 2 (vs. baseline). Furthermore, we used a novel control intervention that has only been tested in HC so far [39]. In addition, our study investigated an unrepresentative sample (higher level of education), as well as the implementation of a single session of ImRs under laboratory conditions (not in routine care).

It nevertheless provides new insights into the potential relevance of psychophysiological biomarkers to investigate change processes during psychotherapeutic interventions such as ImRs. Understanding the underlying processes might further improve the effect of the intervention by identifying novel potential treatment targets that target physiological processes and positive feelings.

This study protocol was reviewed and approved by the Local Ethics Committee of the Faculty of Psychology and Sport Science at the Justus Liebig University Giessen, 2018-0036. Written informed consent was obtained from all participants.

The authors have no conflicts of interest to declare.

This work was supported by a PhD scholarship of the Verhaltenstherapeutische Ambulanz to R.J. Seinsche, Justus Liebig University Giessen. The study was further supported by grants from the German Research Foundation (DFG) to A. Hermann (HE 7013/2-1), and R. Stark (STA 475/20-1) and by “The Adaptive Mind,” funded by the Excellence Program of the Hessian Ministry of Higher Education, Science, Research and the Arts. The Verhaltenstherapeutische Ambulanz, DFG, and Hessian Ministry of Higher Education, Science, Research and the Arts had no further role in study design, collection, analysis, and interpretation of data, in the writing of the manuscript, or in the decision to submit the paper for publication.

Rosa J. Seinsche: conceptualization, methodology, formal analysis, investigation, and writing – original draft. Susanne Fricke: conceptualization, investigation, and writing – review and editing. Marie K. Neudert and Raphaela I. Zimmer: conceptualization and writing – review and editing. Rudolf Stark: supervision, funding acquisition, and writing – review and editing. Andrea Hermann: conceptualization, methodology, formal analysis, supervision, funding acquisition, resources, and writing – review and editing.

The data that support the findings of this study are not publicly available as they contain information that could compromise the privacy of research participants but are available from the corresponding author (A.H.) upon reasonable request.