Mental Representations in OCD

Obsessive-compulsive disorder (OCD) entails a sequence of distressing, unwanted thoughts or doubts (obsessions), and time-consuming efforts to neutralise or avert further distress (compulsions; American Psychiatric Association, 2022). The most evidence-based psychological treatment for OCD is cognitive-behavioural therapy (CBT), with a focus on exposure and response-prevention (McKay et al., 2015; Reid et al., 2021). Although research supports the efficacy of CBT, most people experience some residual symptoms following treatment, and a minority of people do not experience clinically significant change (Fisher et al., 2020). To address these limitations, researchers continue to explore how we can enhance treatment for OCD.

One promising approach is to directly target underlying self-schemas in OCD. Self-schemas are transdiagnostic, multimodal models of self, also known as “mental representations”. These may involve semantic knowledge (e.g. “I might be a bad person”), as well as memories, emotions, imagery, and action tendencies (Arntz, 2020). Cognitive-behavioural models of OCD include self-schemas in describing how symptoms emerge. For example, core beliefs and life experiences influence appraisals of intrusive thoughts (Salkovskis et al., 1998; Wilhelm & Steketee, 2006), and a fear of possible self may contribute to obsessional themes (Aardema & Wong, 2020). One technique used to address self-schemas is imagery rescripting.

Imagery rescripting involves evoking a mental image, typically with implicit self-related meaning, and imagining a different outcome that promotes an adaptive response, such as self-compassion or a sense of agency (Holmes et al., 2007). Emerging clinical research suggests that this approach is an effective adjunct to other cognitive-behavioural therapies for OCD (Maloney et al., 2019; Tenore et al., 2020; Veale et al., 2015). There are variations to how imagery rescripting is implemented (Arntz & Weertman, 1999; Hackmann, 2011; Holmes et al., 2007), but their relative efficacy has not been empirically compared. Two variations of particular interest, when applied to OCD, are whether the target imagery is set in the past or future and whether the rescripted image involves an attachment figure or simply a more adaptive outcome.

Should Rescripting Target Memories or Future Imagery?

Most research into imagery rescripting employs a protocol that targets aversive memories (Morina et al., 2017). The rationale for targeting memories is that early or significant life experiences play a causal role in developing a maladaptive self-schema (Arntz & Weertman, 1999). Addressing memories of these (often traumatic) experiences may allow a person to re-evaluate their implicit meaning or at least provide a competing alternative (Strachan et al., 2020).

Cross-sectional research among people with OCD indicates that aversive experiences often feature in their life histories (Cromer et al., 2007), predict lower treatment effectiveness (Gershuny et al., 2008), precipitate onset of OCD symptoms (Pinciotti & Fisher, 2022), and feature in intrusive obsessional imagery (Speckens et al., 2007). Emerging research also indicates that when these aversive experiences are addressed, OCD symptoms reduce. Specifically, in three empirical case series studies, totalling 44 participants, three quarters of participants reported clinically improvement following rescripting (Maloney et al., 2019; Tenore et al., 2020; Veale et al., 2015).

On the other hand, episodic future mental imagery may also be an effective target for rescripting in the context of OCD. The content of obsessions commonly represents feared possibilities (Aardema & Wong, 2020; Keen et al., 2008; Storch et al., 2010), and the same systems that are used to recall episodic autobiographical imagery (e.g. memories) are likely also used to construct episodic future imagery (Buckner & Carroll, 2007). Specifically, the reconstructive process of remembering appears to involve similar activation of the medial temporal lobe system as simulation or imagining the self in prospective imagery. Thus, provided that a person’s maladaptive view of self is implicit in the episodic future imagery, rescripting may still provide a rich cognitive and affective experience to update their self-schemas that influence OCD.

Indeed, recent research has tested rescripting for future-based imagery for generalised anxiety (Ovanessian et al., 2019), social anxiety (Landkroon et al., 2022), and for some participants with persecutory delusions (Taylor et al., 2020). In addition, we experimentally demonstrated that evoking and rescripting episodic future imagery facilitated immediate changes in self-related beliefs, state emotions, and imagery characteristics among individuals who reported moderate-severe symptoms on an OCD screener (Cooper et al., 2023). Although there is compelling rationale for rescripting both past and future imagery in OCD, studies have not yet compared their relative efficacy.

Must an Attachment Figure Intervene?

Within cognitive behavioural and schema therapies, the most common protocol for imagery rescripting typically involves bringing an older version of the client, or another safe, responsive person, into the image to provide caregiving and support (Arntz & Weertman, 1999). Early pioneers in imagery rescripting explicitly noted how its theoretical basis includes attachment theory and serves to “establish a therapeutic secure base… and develop more effective self-calming and self-nurturing strategies” (Smucker & Dancu, 1999, p. 22). Recent qualitative evidence suggests that many patients and therapists consider the process of receiving care or protection in the rescripted imagery as the most important factor for change (Bosch & Arntz, 2023).

In OCD, meta-analytic evidence indicates that attachment insecurity is linked with more severe symptoms (van Leeuwen et al., 2020). One explanation for this link is that attachment relationships are a key context in which self-concept develops (Wallin, 2007). A recent systematic review implicated a “fractured, feared, or morally questionable” view of self in OCD (Jaeger et al., 2021), and compulsive rituals may be an effort to prevent the expression or fulfilment of a feared possible self (Aardema et al., 2021). Similarly, cross-sectional evidence suggests that the interaction between attachment insecurity and fear of self predicts symptom severity in OCD (Doron, 2020). Given that attachment-based interventions can help to correct a maladaptive a view of self (Johnson & Campbell, 2021; Wallin, 2007), an attachment figure may be an important component of imagery rescripting for OCD.

Alternative approaches to rescripting focus on other imagery modifications rather than introducing an attachment figure. Possibilities include introducing factual information to a distorted image of adult trauma, simple perspective-taking of a traumatic event, or transforming imagery of a perpetrator (Hackmann, 2011). Examples from experimental research include rescripting distressing life events in any realistic or unrealistic way to simply reduce the associated distress (Strohm et al., 2019) or writing about how one would move forward after their worst-case scenario became true (Ovanessian et al., 2019). To our knowledge, no research has directly compared the effects of imagery rescripting with and without an attachment figure.

Aims and Hypotheses

This study aimed to explore whether varying the target imagery (memory/future) or rescripting approach (attachment/mastery) influenced the rate of change in self-related beliefs, emotions, imagery characteristics (vividness and intensity), and action tendencies (urge to suppress). We recruited participants who were pre-screened as high in OCD traits and facilitated variants of imagery rescripting using an online audio-guided self-administered format. We have previously piloted an audio-guided format (Cooper et al., 2023) and researchers have also tested a range of rescripting tasks without a therapist (Lee et al., 2020; Moritz et al., 2018; Ovanessian et al., 2019; Pennesi & Wade, 2018; Zhou et al., 2020). Participants were assigned to one of four conditions reflecting a combination of target imagery and rescripting approaches.

In memories, participants taking the perspective of a younger self may have felt less able to meet their own needs; we predicted that when the target imagery was a memory, participants assigned to the attachment-based rescripting condition would tend to report greater reductions in strength of negative beliefs and negative emotions following rescripting (after controlling for pre-rescripting scores) than those in the mastery-based rescripting condition. Conversely, we predicted that when the imagery was of a future event, participants in the mastery-based rescripting condition would report greater reductions in negative emotions and self-belief following rescripting (after controlling for pre-rescripting scores) than those in the attachment-based rescripting condition. This is because participants imagining themselves in the future may have more difficulty identifying an attachment figure relevant to their future self. In addition, we sought to explore qualitative feedback to determine issues in the acceptability of the online format.

Methods and Materials

Design and Participants

The study was a 2 (time) × 2 (target imagery) × 2 (rescripting type) design. We used Prolific.co to recruit participants who reported moderate levels of OCD tendencies or higher (scores ≥ 11) on a well-validated screener (Abramovitch et al., 2021), were 18 years or older, and lived in the UK.

Measures

Demographics and Psychological Symptoms

Participants were asked to report their gender, age, ethnicity, education, and employment status. The OCI-12 (Abramovitch et al., 2021) was used to measure OCD tendencies. The Depression, Anxiety and Stress Scales (DASS-21; Lovibond & Lovibond, 1995) was used to measure the general psychological distress of the sample.

Trait Attachment

The Experiences in Close Relationships–Revised–General Short Form (ECR-R-GSF; Hao et al., 2019; Wilkinson, 2011) was used to measure general attachment. The scale comprises two 10-item subscales: attachment anxiety (e.g. “I often worry that other people close to me don’t really love me.”) and avoidance (e.g. “I prefer not to show others how I feel deep down.”). Ratings (1 = Strongly agree, 5 = Strongly disagree) are averaged to generate subscale scores. Secure attachment is indicated by low scores on both items. The scales had good internal consistency (anxiety α = .92, avoidance α = .84).

State Attachment

Changes in momentary attachment security were measured using the Security subscale of the State Adult Attachment Measure (Gillath et al., 2009) and administered pre- and post-manipulation. Participants rated their agreement on items such as “I feel like I have someone to rely on.” using a Likert-type scale (1 = Disagree strongly, 7 = Agree strongly), and scores were averaged to obtain a subscale total. In this study, the scale had good internal consistency at baseline (α = .96).

Emotional State and Imagery Experience

Participants rated their state emotions (anger, anxiety, disgust, and sadness), imagery vividness and intensity, and urge to neutralise imagery pre- and post-manipulation. Anchors specific to each item and all items were rated out of 100. For example, “How much anxiety do you feel right now?” (0 = No anxiety, 100 = Extreme anxiety). The full list of questions is provided in the Supplementary Materials.

Certainty of Feared Self

To elicit and rate idiopathic meaning of the imagery, participants were instructed to complete the following sentence in a way that is most true for them, “This image has some personal significance because it means that I might be a(n) ______ person”, and to type an adjective in a free-text field (open-ended). On the following page, participants were asked “How likely is it that you are, deep down, [adjective]?” and to type a number to rate the strength of this belief (0 = Impossible, 100 = Certain). On the following pages, participants were then asked the same questions, but the open-ended question was replaced by a closed-ended list of adjectives from the feared self section of the Selves Questionnaire (Wong et al., 2020). The rating questions were administered pre- and post-manipulation.

Tasks

Generating Target Imagery

To identify and generate target imagery, we adapted an audio-guided exercise from standard protocols often used in schema therapy (Arntz & Weertman, 1999; van der Wijngaart, 2021). The guidance provided a description of an OCD-related experience, without explicitly mentioning OCD, then asked participants to think of a recent similar experience. It then facilitated an “affect bridge” (van der Wijngaart, 2021) to identify target imagery. Participants completed one of two versions to identify either a memory or episodic future imagery, depending on the condition to which they were randomly assigned. Consistent with standard protocols, and to prevent inadvertent imaginal exposure, the audio guidance explicitly guided participants to not “let [their image] play out completely, just the beginning”.

Imagery Rescripting

We adapted an audio-guided rescripting exercise from the same standard protocols (Arntz & Weertman, 1999; van der Wijngaart, 2021). Participants were randomised to one of two rescripting approaches. In the attachment condition, the audio guided participants to observe the imagery from the perspective of someone who may be able to meet their needs in the image and intervene in the narrative (or their older, kinder, self in the absence of another figure). In the mastery condition, participants were guided to construct a new ending that would be more positive and empowering for them. When targeting memories, this attachment figure was an “older, wiser, kinder” self, and when targeting future imagery, this was an “stronger, wiser, kinder” part of themselves. In all conditions, participants were then guided to re-imagine this alternative imagery from the original first-person perspective. Transcripts for the audio guidance are provided in the supplementary materials.

Procedure

Ethical approval was obtained from the University of New South Wales Human Research Ethics Committee (HC220254). After providing informed consent, participants were randomly assigned to one of two target imagery conditions (memory/future) and rescripting approaches (attachment/mastery) using two respective random algorithms in Qualtrics.com. All participants then completed the same questionnaires regarding demographics, trait attachment, and psychological symptoms. Next, they engaged in audio-guided exercises to identify their target imagery (memory/future, respective to condition).

All participants then answered the same questions regarding emotional state and imagery experience, state attachment, and certainty of feared self. Participants then engaged in rescripting according to their assigned condition (attachment/mastery), followed by the same measures. At the end of the study, we invited participants to type any open-ended general feedback if they wished. The median completion time was 31.5 minutes, and participants were recompensed £4.50 for their time.

Data Screening and Analytic Plan

Results were first screened to remove duplicate attempts. Hypothesis tests were performed using mixed linear modelling with the lme4 package (Bates et al., 2015) in R (R Core Team, 2022). For each dependent variable, a model was defined with a three-way interaction of time (pre/post) × target imagery (memory/future) × rescripting approach (attachment/mastery) as fixed effects and participant as a random effect. Plots of residuals and predicted scores were inspected for deviations from model assumptions. For skewed response variables, results were compared with and without transformations. In this study, transformation did not change interpretation of the results and so all variables are reported untransformed. Estimated marginal means produced by the reference grid from each model were calculated and contrasted to produce F-ratios for the joint contribution of model terms (including the interaction) for each dependent variable using emmeans (Lenth, 2022).

To determine adherence to the assigned imagery type (memory/future), we subtracted the age of the participant in the imagery from their current age. To determine if the target imagery involved a maladaptive meaning for the participant, the first author coded the self-related adjectives from the open-ended imagery meaning question (asked prior to rescripting) as either “maladaptive” and “adaptive”. These codes were then reviewed by the senior author.

Qualitative feedback was also analysed using a grounded thematic approach (Braun & Clarke, 2006). The first and second authors separately familiarised themselves with the data and generated initial codes and searched for themes, then met to reach consensus on themes to be coded. Both authors then separately coded all feedback using a structured tabular method (Robinson, 2022) and met to iteratively refine the themes and resolve disagreements in coding. Remaining differences were resolved in discussion with the senior author. Before reporting qualitative themes, 13 responses that mentioned nothing other than nominal positive feedback (e.g. “Thank you.”) were first removed.

Results

Sample Characteristics

Figure 1 illustrates that, of 258 participants who began the experiment, 250 completed it across the four conditions. The average age was 38.19 years (SD = 12.12), and there was an almost equal identified gender distribution (46.8% female). Although participants had opportunity to enter their own gender description, all participants identified as either male or female. Participants typically identified as white (87.2%), worked fulltime (61.2%) or part time (16%), and reported that their highest level of education was an undergraduate degree (41.2%) or high school (26.8%). On average, the sample reported mild stress, moderate depressive symptoms, and mild-to-moderate anxiety, according to severity benchmarks (Lovibond & Lovibond, 1995). Similarly, the sample reported OCD symptoms of moderate severity (Abramovitch et al., 2021). Full demographic results, scale descriptive statistics, and correlations are provided in the supplementary materials.

Fig. 1
figure 1

Participant flow. Although participants were randomly allocated, unequal group sizes emerged. Two iterations of randomisation were applied consecutively (for target imagery and imagery rescripting type), neither produced perfectly equal results and these imbalances were compounded to result in unequal group sizes. Further technical explanation is provided in the supplementary materials

Full size image

Protocol Adherence

Regarding target image, our analyses showed that 32 (29%) of the 111 participants assigned to rescript episodic future imagery reported that their age in the mental imagery was younger than their current age, suggesting that they targeted a memory rather than an episodic future image. In contrast, all participants assigned to rescript a memory reported an age in their imagery that was less than or equal to their current age. Regarding implicit meaning in the imagery, 43 (17%) participants reported adaptive self-related adjectives in the open-ended prompt, which was not consistent with the instructions. However, to maintain fidelity to the analysis plan, all records were retained in the hypothesis tests.

Hypothesis Tests

Table 1 provides a summary of the joint test results from contrasting estimated marginal means produced from linear mixed effects modelling for each variable. These are also illustrated in Fig. 2. On average, participants experienced an adaptive change in all study variables over time, regardless of condition. There was a large effect of time for attachment security, certainty of feared self (using both open- and closed-ended prompts), aversive emotions (anxiety, anger, disgust, sadness), and urge to suppress the imagery. Similarly, there was a medium effect of time on intensity of imagery and a small effect of time on vividness.

Table 1 Between-condition differences in changes in attachment security, beliefs, emotions, and imagery experience
Full size table
Fig. 2
figure 2

Estimated marginal means before and after imagery rescripting by condition

Full size image

In contrast, there was no difference in rates of change between target images (memory/future). Regarding rescripting approach (attachment/mastery), there was no effect on any dependent variable except attachment security. Participants assigned to attachment-based rescript reported an average increase in state-based attachment security of 2.21 scale points (95% CI 0.44, 3.98, on a scale ranging from 7 to 49) following rescripting, more than those who completed the mastery-based imagery rescripting (t(254) = 2.46, p = .015). Given the consistent effect of time across variables (and lack of between-group effects), the pre- and post-rescripting scores are reported in Table 2, along with contrasts tests and within-group effect sizes.

Table 2 Estimated marginal means and contrasts for pre- and post-manipulation
Full size table

Contrary to hypotheses, there was no interaction effect of target image × rescripting approach on changes in strength of beliefs or negative emotions, with one exception. Among the participants who were assigned to rescript an episodic future image, those who completed the mastery-based rescript reported an average decrease in state anxiety of 7.91 points (95% CI 0.78, 15.04, on a scale ranging from 0 to 100) following rescripting, more than those who completed the attachment-based imagery rescripting (t(222) = 2.19, p = .030).

Qualitative Feedback

The themes extracted are provided in Table 3 with descriptions and example responses under each. Overall, 58 participants provided substantive feedback, from which we extracted 89 codes across eight themes. Exploratory analyses revealed no evidence that the target image or rescripting approach influenced the likelihood of providing feedback (p’s of all chi-squared tests > .1). However, among participants who gave any substantive feedback (n = 58), we found evidence of a difference between groups for one theme. Participants who re-scripted a memory were more likely to give feedback regarding “Adaptive Change or Insight” than those who re-scripted an episodic future image (χ2(1) = 8.44, p = .003).

Table 3 Themes in qualitative feedback
Full size table

Discussion

This study was designed to examine whether variations in the target imagery (memory/future) or rescripting approach (attachment/mastery) affected immediate changes in belief, emotion, or imagery characteristics when rescripting OCD-salient imagery. The first hypothesis, that attachment-based rescripting for memories would lead to greater reductions in strength of negative beliefs and emotions than mastery-based rescripting for memories, was not supported. The second hypothesis, that mastery-based rescripting for episodic future imagery would lead to greater reductions in strength of negative beliefs and emotions than attachment-based rescripting, was supported only by the results with respect to state anxiety.

Consistent Effects Across Conditions

Across conditions, participants tended to report immediate pre-post benefit on all study variables following rescripting, including large effects for implicit self-related meaning and emotions. There were generally no differences in change between target imagery (memory/future) or rescripting approaches (attachment/mastery). One exception was that the attachment approach to rescripting increased state attachment security more than the mastery approach. This adds to substantial existing evidence that guided imagery can be used to prime attachment security (Gillath & Karantzas, 2019), but the differential benefits of the attachment approach did not extend to implicit meaning or emotional variables.

One explanation for the consistent improvements across conditions is that neither a memory nor an attachment figure appears to be essential elements in rescripting for OCD. Put differently, the underlying mechanisms of imagery rescripting outperform the minor variations in our experiment. These conditions share a core process in that mental representations were experientially evoked and then modified in some way. Given that the same brain networks are likely involved in recalling memories and episodic future imagery (Buckner & Carroll, 2007), it is possible the temporal orientation of the imagery does not matter—so long as it reflects a salient mental representation. Once evoked, the imagery is modified in a way that addresses self-related implicit meaning. The modified imagery might then compete with the original representation in a memory trace, or the rescripting process might allow the original representation to be updated (Strachan et al., 2020).

It is notable that the changes in certainty of feared self occurred following only a brief task. The audio-guided rescripting was 11 minutes long in all conditions, which is typically much shorter than imaginal exposure interventions, and so changes are unlikely to be a by-product of habituation. Rather, the imagery rescripting process appears to involve a relatively rapid and direct route to experientially modifying multi-modal mental representations of the self. This is supported by the finding that changes in certainty of feared self were accompanied by reductions in aversive emotions, vividness, intensity, and urge to neutralise. In contrast, briefer imaginal exposures typically facilitate less within-session habituation (van Minnen & Foa, 2006), and in-session habituation is not a strong predictor of outcome (Craske et al., 2008).

Mastery Reduces Anxiety from Episodic Future Imagery

When rescripting episodic future imagery, anxiety reduced more using the mastery approach than the attachment approach. The change in anxiety is consistent with previous experiments that targeted future imagery using a mastery approach for OCD (Cooper et al., 2023, 2024). However, this study also extends these findings to tentatively suggest that, in episodic future imagery, priming a sense of mastery has a more direct impact on anxiety than priming secure attachment. One possible explanation is that a mastery-based approach leads to more realistic episodic future imagery because the outcome does not depend on the serendipitous or fanciful presence of an attachment figure (or older self). Given that the modified imagery may be more realistic, it may then reduce the perceived severity or likelihood of a catastrophic imagined event, which in turn could reduce anxiety. Similarly, imagining mastery in a future event may increase a sense self-efficacy, which may be a more effective way to reduce anxiety than imagining receiving help from an attachment future.

Acceptability in Qualitative Feedback

The analysis of qualitative feedback indicated that the interventions were generally acceptable. The most common themes were adaptive benefits following the exercise, pleasant emotions such as positive mood during the exercise, or enjoying the audio delivery of instructions. The predominance of these themes suggests that respondents found the exercise generally helpful or pleasant. Adaptive change was reported more frequently among participants assigned to rescript a memory than a future image, which may suggest that memories are more available or developed forms of mental imagery and therefore more accessible for re-appraisal.

Conversely, a quarter of statements coded reflected criticisms or difficulty with the exercise. The most common negative theme was criticism or difficulty that could have been overcome with therapist guidance, and a small number of participants expressed difficulty forming mental imagery. These comments suggest that the standardised approach failed to facilitate engagement for several participants. Importantly, only six statements indicated that the exercise prompted a negative emotional response, and this mirrors research finding that imagery rescripting produces less distress and upsetting emotions than other techniques such as imaginal exposure (Siegesleitner et al., 2019). Three statements suggested low belief in the value of the exercise. Overall, these findings indicate that the exercises were acceptable to most participants, aligning with findings that imagery rescripting tends to attract acceptable to high acceptability and credibility judgements (Clarke et al., 2021; Ghaderi et al., 2022; Moritz et al., 2018).

Limitations

Aspects of the study design limit its generalisability. First, we did not seek to assess symptom change or include follow-up measurement. To sustain its effects, rescripting may also need to be repeated, extended, combined with other techniques, or used in conjunction with therapist-facilitated imagery rescripting in session. Second, this was an analogue sample, and greater differential effects may emerge in a clinical sample.

Third, potential variance in protocol adherence may have reduced between-group effects. We observed that 29% of participants assigned to the future imagery condition imagined themselves as younger. If some of these participants imagined a memory, differences in dependent variables between target images (future/memory) may have not been detected. We performed follow-up analyses, removing these cases in question, and results remained similar with respect to the meaning of the imagery. However, to confirm these findings, follow-up studies targeting future imagery should adapt the instructions to ensure more consistent adherence.

Fourth, notwithstanding the largely positive qualitative feedback, the negative themes highlighted that any standardised protocol will not be suitable for everyone, and the role of the therapist remains important. Some participants would benefit from personalised guidance in imagining an attachment figure, constructing an adaptive rescripted image, or simply elaborating on the rationale for the exercise to support credibility. For research purposes, personalised guidance may also assist in addressing issues with protocol adherence observed in the results. In addition, while an estimated 4% of people cannot generate visual mental imagery (Dance et al., 2022), expert opinion suggests that rescripting the narrative itself may still facilitate a change in meaning (Stopa, 2011). In such cases, audio-guidance which emphasises visual aspects of the imagery would not be suitable. A therapist who is competent in imagery rescripting could aptly overcome these challenges and provide a more personalised and targeted intervention.

Implications and Conclusion

To our knowledge, this is the first experimental study of imagery rescripting to directly compare the effects of temporal orientation (memory/future) and approach (attachment/mastery) on participants’ response (including meaning, emotion, imagery experience). The lack of difference among conditions highlights the essential role of the core processes in imagery rescripting, which affords a certain level of flexibility in application. Imagery rescripting approach is increasingly being applied across diagnostic categories (Morina et al., 2017); researchers might consider this freedom to vary the protocol as appropriate when testing rescripting interventions.

Regardless of protocol, it is notable that people high in OCD traits tended to report emotional changes combined with shifts in implicit self-related beliefs. We consider the combination of variables included in this study to approximate aspects of mental representations. Future research into mechanisms of change might consider how changes in these variables affect changes in symptoms. Finally, the pre-post results for this online audio-guided task suggest that may be a helpful experiential exercise to include in internet-based interventions for CBT. However, additional studies involving an inert control and clinical sample are needed prior to including a similar exercise in online interventions. We invite other researchers to use the materials and analysis code as a basis to replicate and extend this design to include clinical populations and other disorders.