Psychological Inflexibility, Cognitive Fusion, and Thought–Action Fusion as a Transdiagnostic Construct: Direct Comparisons Among Major Depressive Disorder, Obsessive–Compulsive Disorder, and Healthy Controls
Article information
Abstract
Objective
Evidence suggests that acceptance and commitment therapy (ACT) processes function as transdiagnostic factors in both major depressive disorder (MDD) and obsessive-compulsive disorder (OCD) individually. However, few studies have directly compared these two clinical disorders. Therefore, this study aimed to identify potential transdiagnostic factors associated with ACT across MDD, OCD, and healthy control (HC) groups.
Methods
A total of 34 MDD patients, 43 OCD patients, and 46 HCs were recruited through subway advertisements and outpatient services at a university hospital. Participants completed the Acceptance and Action Questionnaire-II, Cognitive Fusion Questionnaire, thought-action fusion (TAF) scale, and symptom severity scales.
Results
Direct group comparisons revealed higher psychological inflexibility, cognitive fusion, and likelihood of TAF in the clinical groups compared to the HCs, with no differences between the MDD and OCD groups. These three transdiagnostic factors were variously correlated with both depressive and obsessive–compulsive (OC) symptoms in all groups. Regression analyses demonstrated that the three transdiagnostic factors accounted for 26% of the depressive symptoms in the MDD group (R2=0.26, p=0.028) and 27% of the OC symptoms in the OCD group (R2=0.27, p=0.014).
Conclusion
These findings from the direct group comparisons of the three groups confirmed that psychological inflexibility, cognitive fusion, and likelihood of TAF are potential transdiagnostic factors that moderately contribute to the primary symptoms of both MDD and OCD. From another perspective, these results also highlight the need to consider how ACT addresses disorder-specific variations beyond what is explained by these transdiagnostic factors in the future.
INTRODUCTION
Seminal work by Aaron Beck introduced theoretical models in the field of cognitive-behavioral therapy, which have proliferated and focused on particular disorders that specify the cognitive and behavioral processes [1]. However, this “disorder-focus” trend has shifted towards an “across-disorder” or transdiagnostic perspective [1]. The transdiagnostic approach is an empirical work demonstrating potential psychological factors identified as important across different psychiatric disorders [2]. The transdiagnostic perspective offers several advantages [1,3]. First, it may explain the high rates of comorbidity observed in clinical practice. Second, it encourages the transfer of theoretical and treatment advances between disorders, aiding treatment development. Third, it can elucidate intriguing findings, such as exhibiting that disorders not directly targeted by interventions can respond similarly to the treatment.
One such attempt was performed by Hayes et al. [4], who have proposed that research and clinical practice should be grounded in functional diagnostic dimensions or common processes of etiology or maintenance. They demonstrated the usefulness of the functional approach by showing that experiential avoidance (EA), or attempts to avoid inner experiences, is a common process across several diseases and is related to the maintenance of the disease. This tendency may result from excessive cognitive fusion, which is reacting to ideas or concepts as if they were literal facts. These two factors are fundamental in the psychopathology of acceptance and commitment therapy (ACT), which is a part of a larger group of acceptance– and mindfulness-based therapies [5]. ACT applies six interrelated therapeutic processes to foster psychological flexibility: acceptance, cognitive defusion, contact with the present moment, self-as-context, values, and committed action [5,6]. Meanwhile, whereas cognitive fusion refers to the extent to which a person is entangled with inner experiences, thought-action fusion (TAF) refers to the extent to which an individual conflates inner experiences with observable behaviors [7]. TAF, originally derived from the clinical observation of obsessive-compulsive disorder (OCD), is also observed in various anxiety and depressive disorders and is considered a transdiagnostic process [8]. Although TAF overlaps with cognitive fusion, each has its own distinct nature in the context of obsessive-compulsive (OC) symptoms [7].
In this context, major depressive disorder (MDD) and OCD are representative examples of where this transdiagnostic approach is necessary due to their high comorbidity rate [9,10]. OCD is characterized by recurrent intrusive ideas, impulses, or urges (obsessions) along with overt or covert behaviors (compulsions) while MDD is defined by a combination of primary (depressed mood or loss of interest) and additional symptoms (impaired cognition, vegetative symptoms, low energy) [11]. Among the common symptoms of both disorders, rumination may be relevant to this study, as the potential transdiagnostic factors explored here may underlie this ruminative thinking style. Rumination has been identified as a form of repetitive thinking about negative content, along with worry, perseverative cognition, and obsessions [12].
Despite this clinical relevance, surprisingly, few studies directly compared the MDD and OCD ideally showing that ACT processes may be transdiagnostic factors across these two disorders. Previous research has often shown positive associations between psychological inflexibility or cognitive fusion and various forms of psychopathology, including depression and OC symptoms. But most of these studies focus on non-clinical (e.g., university students [7,13-15], community population [15-18]) or single-disorder samples (either MDD15 or OCD [19-21]), limiting their “across-disorder” scope. Some studies have divided a specific disease group into a group with and without a comorbid condition to test the between-group difference in transdiagnostic factors [22]. Additionally, a few investigators have explored ACT-related psychological processes as transdiagnostic factors in clinical samples, encompassing both MDD and OCD groups. For instance, Levin et al. [23] interviewed 972 college students, identified 202 people with current psychiatric disorders, and compared psychological inflexibility for each disorder. They showed significantly greater psychological inflexibility for current MDD and OCD groups (partial η2 values=0.04 in both conditions) relative to the no current disorder control group. Faustino [16] demonstrated that the degree of cognitive fusion was higher in the clinical sample than in the non-clinical. However, the clinical sample included only ten patients with MDD and three with OCD, and the researchers did not provide separate disorder data. Furthermore, some studies directly compared mindfulness facets between MDD and OCD [24,25]. However, since ACT views mindfulness as an emerging phenomenon through four processes (acceptance, defusion, self as context, and context with present moment) of psychological flexibility [26], it is difficult to categorize these studies solely as examining the pathology of ACT.
Therefore, this study initially aimed to verify potential transdiagnostic factors associated with ACT by directly comparing the MDD, OCD, and healthy control (HC) groups, using the Acceptance and Action Questionnaire-II (AAQ-II) [27,28], Cognitive Fusion Questionnaire (CFQ) [15,29], and Thought-Action Fusion Scale (TAFS) [30,31] for psychological inflexibility (or EA in the narrow sense), cognitive fusion, and TAF, respectively. Subsequently, we investigated whether the extent to which these factors contribute to each disease is consistent or varies. Furthermore, another significant issue in this field pertains to the necessity for disorder-specific measures for ACT processes. Current frequently used questionnaires for ACT processes are considered too general to adequately address the distinct contexts of particular disorders. Thus, by utilizing the Acceptance and Action Questionnaire for Obsessions and Compulsions (AAQ-OC) [32,33], which is specific for OCD, instead of the AAQ-II, we examined how the disorder-specific questionnaire differently explains MDD and OCD compared to a general questionnaire. Unfortunately, however, there is no corresponding AAQ scale specific for depression.
We hypothesized as follows: first, in comparing between the three groups, the potential transdiagnostic factors would be higher in the two clinical groups compared to the HC group. Second, these factors would predict depressive and OC symptoms in the MDD and OCD groups, respectively, to a similar extent. Third, the use of a disorder-specific questionnaire may change the contribution of the transdiagnostic factors to each disease.
METHODS
Participants
Participants aged 18–45 years were recruited via advertisements in the subway and online bulletin boards at Kyungpook National University and outpatient psychiatric services at Kyungpook National University Hospital.
Patients with OCD and those with MDD as the primary diagnosis were included. Psychiatric diagnosis was conducted by one experienced psychiatrist (L.S.J.) using the Structured Clinical Interview for Diagnostic and Statistical Manual of Mental Disorders-Fifth Edition-Clinical Version. Patients with acute medical or neurological conditions, a history of brain trauma, or other current comorbid psychiatric illnesses, such as schizophrenia/other psychotic disorders, bipolar disorders, anxiety disorders, and the presence of intellectual disabilities, were excluded. In addition, patients with both MDD and OCD were excluded from the OCD group if their score on the Center for Epidemiologic Studies Depression Scale (CES-D) [34-36] was 25 or higher (clinical depression cutoff for the Korean version) [37] and from the MDD group if their scores on the Yale-Brown Obsessive Compulsive Scale (Y-BOCS) score was 16 or higher (moderate level) [38,39].
For the HCs, psychiatric interviews were conducted to exclude existing psychiatric pathologies, psychotic symptoms, intellectual disabilities, neurological diseases, and history of head injury or medical illness. Additionally, only participants with CES-D scores below 21 (community population cutoff for the Korean version) [37] and Y-BOCS scores below 15 [38], indicating they were within the normal and mild ranges, were included.
The final analysis included 123 participants: 34 MDD patients (23 females), 43 OCD patients (21 females), and 46 HCs (24 females). The study procedures were conducted following written informed consent from all participants and adhered to the approved protocols by the Institutional Review Board of Kyungpook National University Hospital (2021-04-032).
Psychological measures
AAQ-II
The AAQ-II is a 10-item unidimensional scale that assesses EA and psychological inflexibility, central constructs within the ACT model [27]. Individual items are rated on a scale from 1 to 7, with higher scores indicating greater psychological inflexibility. The Korean version of the 8-item AAQ-II has been previously validated, demonstrating adequate psychometric properties (Cronbach’s α=0.85) [28].
AAQ-OC
The AAQ-OC was developed to measure EA specific to OC symptoms [32]. It also measures the ability to emotionally detach oneself from the explicit content of these intrusive thoughts. It consists of 13-items, with two factors: valued action and willingness. Individual items were questioned on a 7-point Likert scale ranging from 1 to 7, with higher scores indicating greater EA. This study used the Korean version of the AAQOC, which had good internal consistency (α=0.87 and 0.94 for university students and OCD, respectively) [33].
CFQ
The CFQ is a self-report instrument to measure excessive attachment to the literal content of thoughts [15]. Seven items are rated on a 7-point Likert scale, with higher scores indicating more fused thoughts. In the present study, the Korean version of the CFQ has good reliability and validity (Cronbach’s α=0.91) [29].
TAFS
The TAFS is a 19-item self-reported measure that assesses the tendency to fuse thoughts and actions [30]. The original scale had three subscales: TAFS-Morality (TAFS-M, 12 items), TAFS-Likelihood-for-Others (TAFS-LO, 4 items), and TAFSLikelihood-for-Self (TAFS-LS, 3 items). The TAFS-LO and TAFS-LS subscales can be combined into TAFS-Likelihood (TAFS-L). Each item is rated on a 5-point Likert scale from 0 (disagree strongly) to 4 (agree strongly). The recently validated Korean version of TAFS, which has excellent internal consistency (Cronbach’s α=0.92–0.93), was used in this study [31].
CES-D
The CES-D is a 20-item self-reported scale to measure the current degree of symptomatology and the initial screening of depression [34,35]. The score ranges between 0 and 60, with an established internal consistency and validity [36]. This study used a validated scale from the Korean version of the CES-D (Cronbach’s α=0.85–0.90), with cut-off points of 25 for clinical cases and 21 for the community population [37].
Y-BOCS–self report
The self-report version of the Y-BOCS is a 10-item questionnaire designed to measure the severity of OC symptoms [38]. It mirrors the traditional Y-BOCS but is completed individually by respondents. The 10-item symptom severity scale, rated from 0 to 4, evaluates the domains of obsession (items 1–5) and compulsion (items 6–10). Each domain is assessed through 5 items, measuring time consumption, interference, distress, resistance, and efforts to control. The Korean self-report version based on Baer’s version was used in this study [39]. It demonstrated adequate internal consistency in non-clinical (α=0.89) and clinical (α=0.87) samples.
Statistical analysis
The demographic and clinical characteristics of the groups were analyzed using one-way analysis of variance (ANOVA), t-test, and chi-squared test. One-way ANOVA and Tukey’s honesty significant difference (HSD) post-hoc tests were also conducted to compare psychological measures among the three groups. Within each group, Spearman’s correlations among psychological measures were analyzed. To examine the contributions of the three scales (AAQ-II, CFQ, and TAFSL) in predicting depressive symptoms and OC symptoms, respectively, a series of regression analyses were performed for each group. The same regression analyses were repeated using AAQ-OC as a variable instead of AAQ-II. The IBM SPSS Statistics software for Windows, version 23 (IBM Corp., Armonk, NY, USA), was used to conduct all statistical analyses, and statistical significance was set at p<0.05.
RESULTS
Demographic and clinical characteristics
Table 1 displays the demographic and clinical data for the MDD, OCD, and HC groups. The mean±standard deviation age of the MDD, OCD, and HC groups was 25.4±5.1, 26.4± 6.4, and 27.9±6.1 years, respectively. In terms of age, sex, and level of education, there were no differences between groups. Age at onset of illness was 20.7±6.0 years for the MDD group, and 18.9±5.4 years for the OCD group (F=1.8, p=0.185). The duration of illness was 5.1±4.6 years for the MDD group and 7.5±5.9 years for the OCD group (F=4.0, p=0.050).
Comparison of potential transdiagnostic measures among groups
ANOVA with post-hoc HSD analysis revealed that the MDD and OCD groups reported higher AAQ-II, CFQ, and TAFS-L scores than the HC group (Table 1). However, no differences were found between the MDD and OCD groups in these variables. Meanwhile, the moral TAF subscale did not differ between any groups. In terms of the AAQ-OC scores, patients with OCD scored the highest, followed by those with MDD, and lastly, by the HCs who scored the lowest (Table 1).
Correlations between potential transdiagnostic measures and symptom measures
In the MDD group, the CES-D scores were significantly correlated with CFQ and TAFS-L. The Y-BOCS scores, despite not reaching statistical significance, still showed an observed trend in the correlation with the AAQ-II, CFQ, and TAF-M. Moreover, the AAQ-OC scores did not correlate with any symptom measures (Table 2).
In the OCD group, the CES-D scores were positively correlated with AAQ-II, CFQ and AAQ-OC and the YBOCS scores were correlated with AAQ-II, CFQ, and AAQ-OC (Table 2). Meanwhile, the CES-D scores in the HCs were significantly correlated with the AAQ-II, CFQ, and TAFS-L scores, while the Y-BOCS scores were associated with the TAFS-L (Table 2).
Regression analysis predicting depression
Table 3 presents the results of the regression analyses predicting the CES-D score. In all three groups, three transdiagnostic measures explained about 30% of the variance (R2=0.26, p=0.028 for the MDD group; R2=0.31, p=0.002 for the OCD group; R2=0.34, p=0.001 for the HC group). The CFQ emerged as a significant individual predictor only in the HC group.

Summary statistics for transdiagnostic variables (AAQ-II included) predicting depressive and obsessive-compulsive symptoms
When the AAQ-OC variable was added instead of the AAQII in the regression model, almost identical results were obtained for the OCD and HC groups, but the model was not significant for the MDD group (Table 4).
Regression analysis predicting OC symptoms
Results of the regression analyses predicting the Y-BOCS score are presented in Table 3. The OCD and HC groups showed that three transdiagnostic measures explained 27% and 23% of the variance (R2=0.27, p=0.014 for the OCD group; R2=0.23, p=0.013 for the HC group). However, three variables in the MDD group explained a very small and nonsignificant amount of variance (R2=0.12, p=0.338). Finally, only the CFQ emerged as a significant individual predictor in the OCD and HC groups.
When the AAQ-OC variable replaced the AAQ-II in the regression model, similar results were obtained for the OCD and MDD groups (Table 4).
DISCUSSION
The current study conducted transdiagnostic measures on three clinically well-defined groups: MDD, OCD, and HC. The main findings are as follows. First, AAQ-II, CFQ, and TAF-L were identified as transdiagnostic factors. The two clinical groups had higher scores than the HC group, with no differences between the two clinical groups in these three factors. Second, correlation analyses showed that these transdiagnostic factors were variously correlated with both depressive and OC symptom scores in all groups. Third, regression analyses revealed that the three transdiagnostic factors explained 26% of the depression scores in the MDD group and 27% of the OC symptoms in the OCD group. Fourth, when the AAQ-OC variable replaced the AAQ-II in the regression model, the predictive power for depression became nonsignificant in the MDD group, while the model’s explanatory power for OC symptoms was maintained in the OCD group.
Before delving into the results, it is essential to outline the characteristics of the two clinical groups in this study. The MDD group exhibited severe depression levels, as indicated by the CES-D score. However, concerning OC symptoms, al-though statistically higher than those in the control group, their average YBOCS score was 6.5, which is still below the score of 7, indicating subclinical symptoms [40]. Meanwhile, the OCD group exhibited moderate to severe OC symptoms, while based on the commonly used cutoff of 16 points [34] in the CES-D, patients with OCD corresponded to subclinical or mild depression. Thus, the MDD group exhibited negligible OC symptoms, while the OCD group showed mild levels of depres-sion, where depressive symptoms were not entirely excluded.
In this study, we examined several potential transdiagnostic factors. As a result, psychological inflexibility (measured by the AAQ-II), cognitive fusion (by the CFQ), and likelihood TAF (by the TAFS) emerged as significant transdiagnostic factors. Two distinctly different clinical groups in this study exhibited higher levels of these three factors than the HC group, with no significant difference observed between the two patient groups. Additionally, these three factors demonstrated varying correlations with depression and OC symptoms within each group. These findings are consistent with previous ones, in which the OCD and MDD groups exhibited higher psychological inflexibility than nonclinical individuals, and the clinical symptoms in each group contributed a similar amount of psychological inflexibility [20]. Generally, many crosssectional studies have shown positive associations between psychological inflexibility and cognitive fusion, and various forms of psychopathology including depression and OC symptoms [13-17,19,20,41], although only a few head-to-head studies have compared MDD and OCD. Studies also have found a relationship between TAF and other disorders or symptoms [8]. In terms of moral versus likelihood TAF, moral TAF did not show any differences between any groups in this study. This finding is consistent with a previous study [42]. The moral TAF facet does not generally show a specific relationship to OC symptoms [43].
Spearman’s correlation analyses revealed that, across all groups, three transdiagnostic factors were associated with both depressive and OC symptoms. Notably, some correlations between transdiagnostic factors and OC symptom scores in the MDD and HC groups only reached a trend level, possibly reflecting the relatively low levels of OC symptoms within both groups. Another interesting finding is that, in contrast to AAQII, the AAQ-OC, another AAQ measure specific to intrusive thoughts from OCD, only showed a correlation with both depressive and OC symptoms in the OCD group. Relatedly, in the mean comparison, the AAQ-OC score was highest in the OCD group, as expected, although it was still higher in the MDD group than in the normal group. These findings showed that the AAQ-OC can still be useful in capturing psychological inflexibility, even in individuals with MDD. Thus, even if an AAQ questionnaire focuses on a specific disorder, it may still measure psychological inflexibility of different disorders to some degree, thereby indicating that psychological inflexibility is a transdiagnostic factor.
Alternatively, these findings underscore the need for disorder-specific measures for various psychiatric disorders, as the AAQ-II is too general to capture disorder-specific thoughts and responses [32,44]. The development of disorder-specific measures has been crucially important in the transdiagnostic research field since context-specific measures performs better than a generic measure of psychological flexibility in terms of incremental validity and treatment sensitivity [45]. For instance, if the AAQ-II is incorporated into a regression model to predict OC symptoms, the subsequent addition of the AAQ-OC shows additional incremental validity [33]. In fact, we put the AAQ-OC variable, instead of the AAQ-II, in the regression model and found that the predictive power for depression became nonsignificant in the MDD group, while the model’s explanatory power for OC symptoms was maintained in the OCD group.
In this study, three transdiagnostic measures explained about 30% of the variance of depressive symptoms across all three groups. An initial validation study of the AAQ-II reported the strongest correlation between the AAQ-II and the Beck Depression Inventory-II (r=0.071) [27]. Gillanders et al. [15] reported that the AAQ-II accounts for about 50% of depression, and the CFQ shows additional incremental validity, predicting further 12% of the variance in depression. Moreover, the interactive effect of two factors was found in depression with different measures [17,18]. In a general population sample, two factors, psychological inflexibility and cognitive fusion, explained 59% of the variance in the depression scores for the Depression, Anxiety, and Stress Scale-21 [17]. Thus, from the initial reports, AAQ and CFQ showed a high correlation with depression. However, existing studies have various differences in target groups, depression measurement tools, regression models, etc., making direct comparisons with the results of this study difficult.
In terms of predicting OC symptoms, three transdiagnostic measures significantly explained 27% and 23% of the variance in the OC symptoms in the OCD and HC groups, respectively. A previous study using the DOCS showed that AAQ-II and CFQ explained 29% of the variance in the unacceptable thoughts domain and 25% in the responsibility for harm domain in patients with OCD [20]. However, in patients with MDD, three measures did not predict OC symptoms. This negative result may be reflected by a lack of significant relationships between the Y-BOCS and transdiagnostic measures in the MDD group, although some measures showed trend-level significance. In addition, instead of TAFS-M, which showed trend-level significance in correlating with CES-D, TAFS-L was entered in the regression model. Overall, regression analysis in this study showed that the three transdiagnostic factors explained 26% of the depression scores in the MDD group and 27% of the OC symptoms in the OCD. That is, the same three variables explained approximately 25% of the primary symptoms of each disorder, irrespective of whether it was MDD or OCD, thereby indicating that these factors are transdiagnostic and of considerable importance. However, interpreting these results should be done with caution, as although these factors are associated with symptoms in both disorders, the evidence does not fully support broad generalizations about their transdiagnostic nature. In our regression models, among the three transdiagnostic measures, only the CFQ emerged as a significant individual predictor. Therefore, there may be different contributions among our measures. Additionally, further research is needed to support our preliminary transdiagnostic findings, as our measures explain only a small portion of disorder-specific symptoms.
Several limitations in this study should be noted. First, although comorbid OCD or MDD symptoms were excluded based on Y-BOCS and CES-D scores, it was practically difficult to completely eliminate even mild levels of comorbid symptoms from both disorders due to their high overlap. Additionally, the relatively small sample size may limit the generalizability of our results. Therefore, further studies with more stringent exclusion criteria and larger, more representative samples across various age groups are needed. Second, because this study is a cross-sectional study, it cannot guarantee a relationship between transdiagnostic factors and symptoms over time, especially during a period of treatment. In particular, it would be valuable to determine how the transdiagnostic factors and symptoms change in the two groups after receiving the same transtherapeutic treatment. Third, measurement validity should be considered. Although all the scales in this study have been standardized and published in peer-reviewed journals, this does not necessarily mean that they fully account for all cultural and linguistic nuances.
In conclusion, these findings from the direct group comparison of the three groups confirmed psychological inflexibility, cognitive fusion, and likelihood TAF as potential transdiagnostic factors that moderately contribute to primary symptoms in both MDD and OCD. These results support the effectiveness of ACT across different psychiatric disorders. From another perspective, they highlight the need to consider how ACT can address disorder-specific variations that go beyond these transdiagnostic factors in future studies.
Notes
Availability of Data and Material
The datasets generated or analyzed during the study are available from the corresponding author on reasonable request.
Conflicts of Interest
Seung Jae Lee, a contributing editor of the Psychiatry Investigation, was not involved in the editorial evaluation or decision to publish this article. All remaining authors have declared no conflicts of interest.
Author Contributions
Conceptualization: Seung Jae Lee, Sang Won Lee. Data curation: all authors. Formal analysis: Seung Jae Lee, Sang Won Lee. Funding acquisition: Seung Jae Lee. Investigation: all authors. Methodology: Seung Jae Lee, Sang Won Lee. Writing—original draft: Seung Jae Lee, Sang Won Lee. Writing—review & editing: Seung Jae Lee.
Funding Statement
This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korean government (MSIP) [grant numbers 2021R1A2C2004720].
Acknowledgements
None