Cognitive Flexibility as a Mediator Between Mentalization and Symptom Severity in Generalized Anxiety Disorder
Article information
Abstract
Objective
Mentalization and cognitive flexibility are key constructs in contemporary models of anxiety psychopathology; however, their interrelationship in clinically diagnosed generalized anxiety disorder (GAD) remains insufficiently delineated. This study aimed to examine whether cognitive flexibility mediates the association between mentalization and anxiety severity.
Methods
In this cross-sectional study, 110 patients with GAD, diagnosed according to DSM-5 criteria, completed the Mentalization Scale, Cognitive Flexibility Inventory, and Generalized Anxiety Disorder-7. Associations were assessed using Pearson correlation analyses, and mediation was examined using bias-corrected bootstrapping (5,000 resamples).
Results
Cognitive flexibility was moderately negatively associated with GAD severity (r=-0.421, p<0.001), whereas mentalization showed a weaker but significant negative association (r=-0.293, p=0.002). In addition, mentalization and cognitive flexibility were strongly positively correlated (r=0.636, p<0.001). Mediation analysis revealed a significant indirect effect of mentalization on anxiety severity through cognitive flexibility (B=-0.083, 95% confidence interval [-0.132, -0.042], p=0.001), while the direct effect was not statistically significant.
Conclusion
Cognitive flexibility significantly mediated the association between mentalization and anxiety severity; however, given the cross-sectional design, this finding should be interpreted as statistical mediation rather than evidence of a causal mechanism. Although causal inferences cannot be established due to the cross-sectional design, cognitive flexibility may represent a clinically relevant target for reducing anxiety symptoms. Future longitudinal and interventional studies are warranted to clarify the directionality and clinical implications of these findings.
INTRODUCTION
Generalized anxiety disorder (GAD) is a psychiatric condition characterized by excessive and persistent anxiety and worry across a range of events or activities for at least six months, which individuals find difficult to control and which result in clinically significant functional impairment [1]. Anxiety disorders represent a major global public health burden, with an estimated prevalence of 4.05% in 2019. Moreover, the global number of affected individuals increased substantially from 194.9 million in 1990 to 301.4 million in 2019 [2]. Beyond their high prevalence, anxiety disorders are associated with marked impairments in interpersonal and social functioning. Individuals with anxiety and related disorders have been shown to experience greater relationship distress and poorer relationship quality than non-clinical populations [3]. Given the widespread prevalence and substantial functional burden associated with anxiety disorders, there is increasing interest in identifying transdiagnostic mechanisms that may contribute to their persistence and clinical severity.
In recent years, growing attention has been directed toward higher-order social-cognitive processes, particularly mentalization, in efforts to clarify the mechanisms underlying anxiety psychopathology. Mentalization refers to the capacity to interpret and understand one’s own behavior and the behavior of others in terms of underlying mental states, including thoughts, emotions, desires, and intentions [4]. This construct is conceptualized as multidimensional, encompassing four principal domains: automatic versus controlled processing, internal versus external focus, self-related versus other-related orientation, and cognitive versus affective mentalizing, each representing distinct yet interrelated dimensions of mentalization [5]. Self-related mentalization refers to the capacity to reflect upon and understand one’s own internal mental states, whereas other-related mentalization involves inferring the thoughts and emotions of others to interpret their behavior. Although these domains are conceptually distinguishable, they are closely interdependent, and the ability to differentiate one’s own mental states from those of others is regarded as fundamental to both psychological and interpersonal functioning [6]. This multidimensional conceptualization has prompted increasing interest in determining which aspects of mentalization may be most relevant to specific forms of psychopathology.
Although the term “mentalization” is frequently used interchangeably with theory of mind in the literature, mentalization has been proposed to represent a broader construct. Whereas theory of mind primarily concerns the cognitive inference of mental states, mentalization additionally incorporates affective and interpersonal regulatory processes that extend beyond traditional theory of mind frameworks [3,7]. A substantial body of literature has examined the role of mentalization across diverse psychiatric conditions, including schizophrenia, eating disorders, depression, and borderline personality disorder [8-11]. In addition, related social-cognitive impairments have been documented in child and neurodevelopmental psychiatric populations, including attention-deficit/hyperactivity disorder, specific learning disorder, and autism spectrum disorder [12]. By contrast, investigation of mentalization in anxiety and related disorders remains comparatively limited. Nevertheless, available evidence suggests that mentalizing capacities may also be impaired in these conditions [13-15].
Cognitive flexibility refers to the capacity to adapt cognitive and behavioral responses in accordance with changing situational demands and is considered a central component of executive functioning [16,17]. In anxiety disorders, heightened threat sensitivity and persistent hyperarousal may constrain available cognitive resources, thereby promoting cognitive rigidity and reducing cognitive flexibility [18,19]. Empirical findings indicate that individuals with GAD demonstrate impaired adaptation to changing rules and reduced cognitive flexibility relative to healthy controls. Moreover, diminished cognitive flexibility has been associated with greater anxiety severity, suggesting that this construct may contribute to both the development and maintenance of anxiety symptoms [20].
From a theoretical standpoint, whereas mentalization reflects the capacity to interpret one’s own and others’ mental states, cognitive flexibility may facilitate the adaptive revision of these interpretations in response to changing contextual information. Effective mentalization may therefore require not only the accurate identification of mental states but also the capacity to update these interpretations flexibly when circumstances change. Accordingly, the ability to shift between alternative perspectives has been proposed as an important component of effective mentalization [21].
Despite accumulating evidence linking mentalization deficits to anxiety symptoms, the mechanisms underlying this association remain insufficiently understood. In particular, the contribution of cognitive processes to this relationship has yet to be clearly delineated. Although cognitive flexibility has been associated with anxiety symptomatology, its potential role in linking mentalization to symptom severity remains largely unexplored, particularly in clinically diagnosed GAD populations. To the best of our knowledge, this relationship has not previously been examined in a clinically diagnosed GAD sample using a mediation framework.
The present study aimed to investigate the relationships among mentalization, cognitive flexibility, and symptom severity in individuals with GAD. Specifically, this study examined whether cognitive flexibility statistically mediates the association between mentalization capacity and GAD severity.
METHODS
Participants
This cross-sectional study included 110 patients who were consecutively recruited to the outpatient psychiatry clinic of Kartal Dr. Lütfi Kırdar City Hospital between October 15, 2025 and April 15, 2026. All participants were diagnosed with GAD according to DSM-5 criteria1 based on clinical psychiatric evaluation. To minimize potential confounding effects of ongoing treatment, only patients who were not receiving psychotropic medication or psychotherapy at the time of assessment were included.
Participants were required to be aged 18 years or older, literate in Turkish, and capable of providing written informed consent. Individuals were excluded if they had significant sensory impairments that could interfere with participation, neurological or systemic conditions associated with cognitive impairment, comorbid psychiatric disorders other than specific phobias, or alcohol and/or substance use disorders.
Sample size was determined using an a priori power analysis conducted with G*Power 3.1 (https://www.psychologie.hhu.de/arbeitsgruppen/allgemeine-psychologie-und-arbeitspsychologie/gpower). Based on a medium effect size (f2=0.15), an alpha level of 0.05, and a statistical power of 0.80, a minimum of approximately 90 participants was required. Initially, 123 individuals were assessed for eligibility. After applying the predefined inclusion and exclusion criteria, 110 participants met the study criteria and were included in the final analyses.
The study was approved by the Ethics Committee of Kartal Dr. Lütfi Kırdar City Hospital (Approval No: 2025/010.99/20/15, dated September 29, 2025). All procedures were conducted in accordance with the ethical standards of the institutional research committee and the 1964 Helsinki Declaration and its subsequent amendments. Written informed consent was obtained from all participants, and all data were anonymized prior to analysis.
Data collection tools
Sociodemographic data form
A sociodemographic data form was developed by the researchers in line with the relevant literature. The form collected information on age, gender, education level, marital status, employment status, and psychiatric history.
Mentalization Scale
Mentalization capacity was assessed using the Mentalization Scale (MENTS), a 28-item self-report instrument rated on a 5-point Likert scale ranging from 1 (“completely incorrect”) to 5 (“completely correct”). The scale comprises three subdimensions: self-related mentalization (MentS-S), other-related mentalization (MentS-O), and motivation for mentalization (MentS-M). In the original study, the Cronbach’s alpha coefficient was reported as 0.84 for the total scale and between 0.76 and 0.77 for the subscales [22]. The Turkish version of the scale was validated by Törenli Kaya et al. [23], confirming a three-factor structure consistent with the original instrument. The internal consistency coefficients for the Turkish version were 0.84 for the total score, and 0.78, 0.80, and 0.79 for the MentS-S, MentS-O, and MentS-M subscales, respectively. In the present study, Cronbach’s alpha coefficients were 0.84 for the MentS-O subscale, 0.43 for the MentS-M subscale, 0.81 for the MentS-S subscale, and 0.82 for the total score.
Cognitive Flexibility Inventory
Cognitive flexibility was assessed using the Cognitive Flexibility Inventory (CFI), developed by Dennis and Vander Wal [24] and adapted into Turkish by Gülüm and Dağ [25]. The inventory consists of 20 items rated on a 5-point Likert scale ranging from 1 (“not at all appropriate”) to 5 (“completely appropriate”). The CFI includes two subdimensions: alternatives and control. Several items are reverse scored. Both total and subscale scores can be calculated, with higher scores indicating greater cognitive flexibility. The Turkish version demonstrated high internal consistency, with a Cronbach’s alpha coefficient of 0.90 for the total scale and 0.89 and 0.85 for the alternatives and control subscales, respectively. In the present study, the CFI also demonstrated good internal consistency, with Cronbach’s alpha coefficients of 0.86 for the alternatives subscale, 0.84 for the control subscale, and 0.87 for the total score.
Generalized Anxiety Disorder-7
Anxiety severity was assessed using the Generalized Anxiety Disorder-7 (GAD-7), developed by Spitzer et al. [26] and adapted into Turkish by Konkan et al. [27]. The scale consists of 7 items rated on a 4-point Likert scale ranging from 0 (“not at all”) to 3 (“nearly every day”). The Turkish version demonstrated good internal consistency (Cronbach’s alpha=0.85) and a unidimensional factor structure. In the present study, the GAD-7 also demonstrated good internal consistency, with a Cronbach’s alpha coefficient of 0.80.
Procedure
Participants completed the assessment battery in a single session in a clinical setting. Following the diagnostic evaluation, all self-report measures were administered under the supervision of trained researchers to ensure completeness and accuracy. The instruments were presented in a fixed order: sociodemographic data form, MENTS, CFI, and GAD-7.
Statistical analysis
All statistical analyses were pre-specified based on the study hypotheses and conducted using jamovi version 2.6 (https://www.jamovi.org/). The distribution of continuous variables was evaluated using both visual (histograms, Q–Q plots) and statistical methods. Values between -2 and +2 for skewness and between -7 and +7 for kurtosis were considered indicative of normal distribution [28]. Pearson correlation coefficients were used given that assumptions of approximate normality were met.
A simple mediation model was tested using the mediation analysis module in Jamovi to evaluate whether cognitive flexibility statistically mediates the association between mentalization and anxiety severity. Indirect effects were estimated using bias-corrected (BC) bootstrapping procedures (5,000 resamples) with 95% confidence intervals (CIs).
RESULTS
Sociodemographic and clinical characteristics
The sociodemographic and clinical characteristics of the participants are summarized in Table 1, including age, gender, and education level.
Descriptive statistics
The mean GAD-7 score was 14.96 (standard deviation [SD]=4.19). The mean total cognitive flexibility score was 68.44 (SD=12.40), with subscale means of 48.60 (SD=8.50) for alternatives and 19.84 (SD=6.45) for control. The mean total mentalization score was 87.96 (SD=12.64), with subscale means of 21.81 (SD=7.20) for self-related mentalization, 35.46 (SD=5.89) for other-related mentalization, and 30.69 (SD=4.03) for motivation-related mentalization.
Correlation analyses
Pearson correlation analyses revealed that GAD-7 scores showed a moderate negative correlation with cognitive flexibility (r=-0.421, p<0.001) and a weaker but significant negative correlation with mentalization (r=-0.293, p=0.002). In addition, total cognitive flexibility and total mentalization were strongly and positively correlated (r=0.636, p<0.001).
At the subscale level, GAD severity was significantly negatively correlated with both the alternatives (r=-0.259, p=0.006) and control (r=-0.468, p<0.001) dimensions of cognitive flexibility. Among the mentalization subdimensions, only self-related mentalization showed a significant negative correlation with GAD severity (r=-0.364, p<0.001), whereas other-related (r=-0.140, p=0.145) and motivation-related mentalization (r=-0.065, p=0.502) were not significantly associated with symptom severity (Table 2).
Mediation analysis
A simple mediation model was tested with mentalization as the independent variable, cognitive flexibility as the mediator, and GAD severity as the dependent variable. Indirect effects were tested using BC bootstrapping procedures (5,000 resamples) with 95% CIs. The total effect of mentalization on GAD severity was statistically significant (B=-0.097, standard error [SE]=0.030, β=-0.293, p=0.001). The direct effect of mentalization on GAD severity was not significant after inclusion of cognitive flexibility in the mediation model (B=-0.014, SE=0.037, β=-0.043, p=0.703). Importantly, the indirect effect of mentalization on GAD severity through cognitive flexibility was statistically significant (B=-0.083, SE=0.026, 95% CI [-0.132, -0.042], β=-0.250, z=-3.255, p=0.001). Examination of the component paths indicated that mentalization was positively associated with cognitive flexibility (B=0.624, SE=0.072, β=0.636, p<0.001), and cognitive flexibility was negatively associated with GAD severity (B=-0.133, SE=0.038, β=-0.394, p<0.001). These findings are consistent with a statistical mediation pattern, suggesting an indirect association between mentalization and GAD severity through cognitive flexibility (Figure 1).
DISCUSSION
The present study examined the relationships among mentalization, cognitive flexibility, and symptom severity in individuals with GAD. The findings indicated that cognitive flexibility was significantly associated with anxiety severity, whereas the direct association between mentalization and anxiety severity was no longer significant after inclusion of cognitive flexibility in the mediation model. The mediation analysis was consistent with an indirect association between mentalization and anxiety severity through cognitive flexibility. However, given the cross-sectional design, this finding should be interpreted as statistical mediation rather than evidence of a temporal or causal pathway. From a theoretical perspective, GAD has been conceptualized as involving maladaptive regulation of emotional processing across both cognitive and affective systems. In particular, Borkovec et al. [29] describe worry as a predominantly verbal-linguistic process that functions to modulate emotional responding by limiting engagement with affectively salient internal experiences. Although this process may attenuate immediate emotional activation, it may simultaneously interfere with full emotional processing, thereby contributing to the persistence of anxiety. Within this framework, such processing patterns may foster a more rigid mode of responding and may therefore be associated with diminished cognitive flexibility. Consistent with this perspective, cognitive models of GAD similarly characterize intolerance of uncertainty as involving rigid cognitive processing patterns marked by difficulty tolerating ambiguity and generating alternative interpretations [30].
Consistent with this framework, cognitive flexibility has been repeatedly implicated in anxiety and related symptomatology. Reduced cognitive flexibility has been associated with rigid cognitive styles, perseverative thought patterns such as rumination, and a diminished capacity to generate alternative interpretations, all of which are central processes in the maintenance of anxiety-related psychopathology. Cognitive inflexibility has also been linked to restricted coping repertoires, impaired problem-solving abilities, and heightened need for control, each of which may exacerbate emotional distress and perpetuate anxiety symptoms. In line with this literature, prior empirical evidence suggests that cognitive flexibility may represent an important factor associated with anxiety severity, with lower levels of flexibility associated with heightened anxiety symptoms [31]. The present findings further support this relationship by demonstrating an association between lower cognitive flexibility and greater anxiety severity in individuals with GAD. Although the findings were obtained in adults with GAD, they are consistent with a broader literature suggesting that cognitive flexibility is clinically relevant across different diagnostic contexts. For example, Şahin et al. [32] reported that autistic trait severity was independently associated with cognitive flexibility performance in adolescents with epilepsy, supporting the transdiagnostic relevance of cognitive flexibility beyond adult anxiety populations.
Mentalization has most commonly been operationalized in empirical research through the construct of reflective functioning. Although a substantial literature has examined reflective functioning across diverse forms of psychopathology, investigations specifically addressing anxiety disorders remain comparatively limited. Nevertheless, available evidence suggests that impairments in reflective functioning may be associated with anxiety-related symptomatology. For example, deficits in reflective functioning, particularly reduced certainty regarding mental states, have been associated with higher levels of anxiety and depressive symptoms in bereaved individuals [33]. More broadly, meta-analytic findings indicate that mentalizing capacities tend to be only modestly impaired across anxiety and internalizing disorders, suggesting that their contribution may be present but not necessarily strong or direct [34]. In the present study, mentalization demonstrated a significant bivariate association with anxiety severity, with the self-related dimension of mentalization showing a particularly notable relationship with symptom severity. These findings may suggest that difficulties in understanding and interpreting one’s own internal mental states are relevant to anxiety disorders such as GAD, in which excessive worry is closely linked to maladaptive self-focused cognitive processes. Accordingly, impairments in self-related mentalization may contribute to anxiety symptomatology, although this relationship may be influenced by related cognitive processes.
Empirical studies directly examining the relationship between mentalization and cognitive flexibility remain limited. One of the few studies to investigate this association was a cross-sectional study of graduate students in China, which found that multiple dimensions of mentalization were significantly associated with anxiety symptoms and that psychological flexibility partially mediated this relationship. These findings suggest that the association between mentalization and anxiety may be linked, at least in part, to cognitive and regulatory processes rather than reflecting an exclusively direct relationship [35]. The present findings may be interpreted as extending this line of research to a clinically diagnosed GAD sample. Specifically, although mentalization was associated with anxiety severity at the bivariate level, this association was no longer significant after cognitive flexibility was included in the mediation model. Moreover, the mediation analysis suggested that cognitive flexibility may partly explain the association between mentalization and anxiety severity at a statistical level. Taken together, these findings are consistent with an indirect association between mentalization and anxiety symptoms through cognitive flexibility, while no conclusions can be drawn regarding temporal or causal ordering. Neuroanatomically, mentalization has been associated with a distributed network including medial prefrontal and temporoparietal regions involved in mental state inference, and more broadly with default mode network activity during internally oriented social-cognitive processing [36,37]. Cognitive flexibility, in contrast, is supported by frontoparietal control networks involving prefrontal and parietal regions that enable adaptive updating and regulation of behavior [38,39]. Emerging evidence suggests partial overlap between these systems, particularly in prefrontal and temporoparietal regions, indicating shared neurofunctional substrates for perspective shifting and mental state attribution [40]. Within this framework, mentalization may involve the generation of mental representations, whereas cognitive flexibility may facilitate their adaptive updating in response to changing contextual demands. This overlap may be particularly relevant in GAD, where dysregulation of fronto-limbic circuitry affects both emotional processing and cognitive control [41,42].
Several limitations of the present study should be acknowledged. First, the cross-sectional design precludes causal inferences regarding the directionality of the observed associations. Although cognitive flexibility statistically mediated the association between mentalization and GAD severity, this finding reflects a statistical mediation model and should not be interpreted as evidence of a causal mechanism. Longitudinal and experimental studies are therefore needed to clarify the temporal ordering and causal nature of these relationships. Second, all psychological variables were assessed using self-report measures, which may introduce response-related biases, including social desirability effects and shared method variance. Additionally, the Cronbach’s alpha coefficient for the MentS-M subscale was 0.43 in the present sample; therefore, findings involving this subdimension should be interpreted with caution. Third, the sample consisted of patients recruited from a single clinical center, which may limit the generalizability of the findings to broader, more heterogeneous, or treatment-experienced populations. Another limitation is the predominance of female participants in the sample. Because the study included 93 women and 17 men, the findings may not be equally generalizable to male patients with GAD. Gender was therefore not included as a covariate in the main mediation model, as the markedly unbalanced distribution and small number of male participants could have resulted in unstable estimates. Future studies with larger and more gender-balanced samples are needed to examine whether these associations differ by gender. Fourth, mentalization was assessed using the MENTS, which is a self-report measure of perceived mentalizing capacity rather than a performance-based or interview-based assessment. Therefore, the findings may not be directly generalizable to other operationalizations of mentalization, such as reflective functioning interview measures. Finally, the absence of measures assessing related constructs such as intolerance of uncertainty and depressive symptom severity may have limited the ability to fully evaluate alternative explanatory pathways. Future investigations incorporating a broader range of relevant variables may help further clarify these relationships.
The present findings have several clinically relevant implications. Most notably, cognitive flexibility emerged as a key variable associated with GAD severity and showed a statistically significant indirect association in the relationship between mentalization and anxiety symptoms. These findings suggest that difficulties in adapting cognitive perspectives and generating alternative interpretations may be relevant to the association between mentalization and anxiety severity. From a therapeutic standpoint, cognitive flexibility may therefore represent a potentially relevant clinical target. Interventions designed to enhance cognitive flexibility directly or indirectly, including cognitive restructuring strategies within cognitive-behavioral therapy and perspective-taking components incorporated into mentalization-based approaches, may contribute to symptom reduction in individuals with GAD. These findings further underscore the potential value of integrating cognitive frameworks in clinical intervention. Addressing both the capacity to understand mental states and the ability to flexibly reinterpret situations may offer a more comprehensive therapeutic approach for GAD.
In conclusion, the present study demonstrates that both cognitive flexibility and mentalization are associated with the severity of generalized anxiety symptoms; however, cognitive flexibility appeared to be more closely associated with anxiety severity within the present statistical model. These findings suggest that cognitive flexibility may be an important cognitive correlate of the association between higher-order social-cognitive processes, such as mentalization, and clinical outcomes in anxiety disorders. Future longitudinal and experimental research is warranted to clarify the temporal direction of these associations and to determine whether interventions targeting cognitive flexibility can improve clinical outcomes in GAD.
Notes
Availability of Data and Material
The data supporting this study’s findings are available from the corresponding author upon reasonable request.
Conflicts of Interest
The authors have no potential conflicts of interest to disclose.
Author Contributions
Conceptualization: all authors. Data curation: Cansu Ünsal Mavi, Zeliha Büşra Durgungöz, Hatice Kübra Çiçek Hekimsoy. Formal analysis: Cansu Ünsal Mavi. Investigation: Cansu Ünsal Mavi, Zeliha Büşra Durgungöz, Hatice Kübra Çiçek Hekimsoy. Methodology: all authors. Writing—original draft: all authors. Writing—review & editing: Cansu Ünsal Mavi, Esma Çörekli Kaymakçı.
Funding Statement
None
Acknowledgments
None
