Uncertain reflective functioning, anxious attachment, and cumulative impact of early relational trauma as predictors of severity in borderline personality disorder

Abstract

Background:

The mentalization model of borderline personality disorder (BPD) utilizes a developmental psychopathology lens, emphasizing an unstable or reduced capacity to mentalize—stemming from early attachment disruptions and relational trauma—as the core feature of BPD. While the empirical evidence for the proposed intersections between mentalizing, attachment, and trauma and the development and manifestations of borderline personality is still limited, this knowledge is essential for developing effective assessments and interventions.

Aim:

To examine mentalizing, attachment, and early relational trauma as predictors of the severity of symptoms in BPD.

Materials and Methods:

The sample included 60 individuals diagnosed with BPD, aged 18–45 years, recruited from inpatient and outpatient services in a tertiary care mental health center. Participants completed measures of mentalizing (Reflective Functioning Questionnaire, Interpersonal Reactivity Index, and Reading the Mind in the Eyes test), attachment (Attachment Style Questionnaire), early relational trauma (Complex Trauma Questionnaire), and symptom severity (Borderline Symptom List).

Results:

The majority of the participants reported experiences of polyvictimization (93.3%) with the most common traumas being psychological abuse (93.3%), neglect (91.7%), and rejection (90.0%). Regression analyses revealed that uncertain reflective functioning, anxious attachment style, and the cumulative impact of trauma together predicted 53.1% of the variance in BPD symptom severity.

Conclusion:

Impaired mentalizing, characterized by inflexible understanding of mental states, is a critical target for therapeutic interventions in BPD. The significant role of anxious attachment and the necessity of assessing the perceived impact of traumatic experiences highlight the importance of trauma-informed and mentalization-based interventions for this vulnerable group.

The mentalization model of borderline personality disorder (BPD) uses a developmental psychopathology framework to understand symptom emergence and manifestation. According to this model, the core feature of BPD is an unstable or reduced capacity to mentalize – ‘make sense of self and of others in terms of subjective states and mental processes’.[1] Mentalization develops through secure attachment relationships and the parent’s ability to mentalize the child appropriately.[1] Disruptions in early attachment and trauma affect mentalizing abilities, the formation of a coherent self-structure, and the navigation of the interpersonal world. Evidence links insecure attachment (unresolved, preoccupied, and fearful) with BPD symptoms. However, most studies use categorical measures of attachment, with a few exploring different attachment patterns and their intersections with mentalizing abilities and BPD. The mentalization model proposes that in individuals with BPD when the attachment system is activated, there is heightened arousal accompanied by a rapid switch from controlled mentalizing to automatic mentalizing.[2,3] This results in a complete lack of understanding of the mental states of self and others or the development of complex models of people’s minds that are not anchored in reality. Subsequently, this manifests in difficulties in comprehending and navigating the interpersonal world.

There is an emerging focus on the links between early relational trauma and mentalizing difficulties, particularly among persons with borderline personality vulnerabilities. While trauma histories are neither necessary nor sufficient for the development of BPD, they have frequently been found in the narratives of individuals with BPD.[4] A review of studies on social cognition in BPD proposed that childhood trauma is associated with deficits in cognitive empathy, evaluation biases, reduced emotional empathy, and aberrant sending of social signals, which together could explain the behavioral problems seen among individuals with BPD.[5] However, most of the research on childhood adversities and BPD only considers the presence or absence of trauma. There is limited understanding of how multiple traumas cumulatively impact the severity and pattern of borderline vulnerabilities over time.

Studies on impairments in mentalizing and related constructs such as emotion recognition and cognitive and affective empathy have yielded mixed results. For instance, studies on emotion recognition and empathy in BPD show both enhanced[6,7,8] and poor abilities.[9,10,11] The ‘borderline empathy’ paradox[12] has been used to explain this phenomenon, according to which enhanced attention and perception of social and emotional stimuli interact with deficits in attentional control, emotion regulation, and attachment system activation, leading to dysfunctional interpersonal interactions.[13]

Mentalization is a complex and multidimensional construct which makes its assessment complex too – in terms of both the way it is assessed and what is being assessed.[14,15] This makes it important to examine its various aspects through different methodologies and understand the nature and extent of these vulnerabilities and their intersection with other variables to determine the severity of BPD manifestations.

Mentalization, attachment, and trauma are all implicated in the mentalization model of BPD. Some scattered studies demonstrate certain linkages, for example, between attachment and mentalizing abilities,[16] attachment and trauma experiences,[17] and mentalizing abilities and trauma experiences.[18] While therapeutic interventions using the mentalization model of BPD have shown efficacy in alleviating symptoms,[19] there is still limited empirical evidence for the proposed intersections between mentalizing, attachment, and trauma and the development and manifestations of borderline personality. This research was aimed at understanding the relationship among mentalizing, attachment, trauma, and BPD symptom severity to provide empirical support for the mentalization-based model of BPD, inform psychological assessment frameworks, and determine targets of therapeutic intervention.

MATERIALS AND METHODS

Sample

A sample of 60 individuals between 18 and 43 years with a diagnosis of BPD were recruited from the inpatient and outpatient services of a tertiary mental health care hospital in India. The sample size was determined in consultation with the biostatistician based on patient data available over the past 5 years and the number of variables in the study. Individuals with past or current psychotic symptoms, other comorbid personality disorders, current symptoms of mania/hypomania/severe depression, head injury, or neurological illnesses and those who had received electroconvulsive therapy or more than ten sessions of psychotherapy in the past 6 months were excluded.

Measures

The Reflective Functioning Questionnaire (RFQ-8)

The 8-item self-report measure, RFQ,[20] assesses certainty (RFQc) and uncertainty about mental states of self and others (RFQu). High RFQc scores indicate hypermentalizing (excessive certainty about mental states), while average scores indicate more genuine mentalizing. High RFQu scores suggest hypomentalizing (inadequate knowledge about mental states), with average scores indicating good mentalizing. The internal consistency of the RFQu and RFQc is 0.77 and 0.65, respectively.[20]

The Interpersonal Reactivity Index (IRI)

The Perspective-taking (PT) and the Empathic Concern (EC) subscales of the IRI[21] were used to assess cognitive empathy and affective empathy. The scale has good internal consistency with Cronbach’s α ranging from 0.68 to 0.79. The test–retest reliability (60–75-day interval) ranged from 0.61 to 0.81.[21]

The Reading the Mind in the Eyes Test (RMET)

The RMET[22] is a task-based measure assessing an adult’s ability to attribute beliefs, intentions, and desires to others. Participants view 36 black-and-white photographs of eyes and select the word that best describes the person’s thoughts or feelings from four displayed mental state words. The total score is based on the number of correct responses, with lower scores indicating difficulties in interpreting mental states. The measure has adequate internal consistency (Cronbach’s α =0.605), test–retest reliability (intraclass correlation coefficient = 0.833), and good concurrent validity with the Empathy Quotient Test.[22]

The Attachment Style Questionnaire (ASQ)

The 40-item ASQ[23] is a self-report measure that assesses five dimensions of adult attachment: Confidence, Discomfort with Closeness, Need for Approval, Preoccupation with Relationships, and Relationships as Secondary. Factor analysis revealed two latent factors – Anxiety and Avoidance. The 10-week test–retest reliability had coefficients ranging from 0.67 to 0.78. Both scales yield good coefficient alphas (anxiety: 0.90, avoidance: 0.84).

The Complex Trauma Questionnaire (ComplexTQ)

The 70-item self-report Complex TQ[24] assesses multiple types of relational trauma before age 15 and their associated frequency and impact, separately involving maternal, paternal, and other attachment figures. The nine domains of maltreatment and trauma include neglect, reject, role reversal, psychological abuse, physical abuse, sexual abuse, witnessing domestic violence, separation, and loss. For this study, the cumulative impact of trauma experiences was calculated by summing the average impact of different types of trauma experienced by each participant. Complex TQ-clinician version discriminated normative from high-risk and clinical samples, and convergent validity was reported with the Adult Attachment Interview (AAI) scales.[24]

The Borderline Symptom List – 23 (BSL-23)

The 23-item self-report BSL-23[25] assesses BPD symptomology over the past week on a 5-point Likert scale. Scores range from 0 to 92, with higher scores indicating higher severity. It has high internal consistency with Cronbach’s α ranging from 0.94 to 0.97 across various samples. The scale is strongly related to the presence of BPD as measured by semistructured interviews and discriminated adequately between those with BPD and those with Axis I disorders.[25]

Procedure

Ethics approval for the research proposal was obtained from the Institutional Ethics Committee, National Institute of Mental Health and Neurosciences. Approval number: No. NIMH/DO/IEC/(BEH. Sc. DIV)/2016-17. Written informed consent was obtained from all participants included in the study.

During the pilot phase, a professional translator from the Institute of Translation Studies translated and back-translated the measures into Kannada and Malayalam. Six patients with BPD reviewed the questionnaires for clarity and ease of understanding and reported no issues. Most participants (96.7%) recruited had completed at least secondary school education and were fluent in English, therefore completing the questionnaires in English.

Figure 1 illustrates the detailed procedure followed. Of the 75 participants who were approached for the study, only 60 were recruited as the others either did not meet research criteria or did not provide consent.

Figure 1.

Flow chart of procedure

Statistical analysis

The analysis was conducted using the IBM SPSS Statistics Version 22.[26] Descriptive statistics such as mean, standard deviation (SD), frequency, and percentages were calculated for all the variables, and normality was checked using the Shapiro–Wilk test. Correlational analysis was followed by univariate regression and stepwise linear regression to determine the independent variables that predicted the dependent variable. Multicollinearity was checked using the Variance Inflation Factor (VIF) while fitting the model, and the values for all the predictor variables were less than 4 (ranging from 1.04 to 1.31).

RESULTS

The sample, with a mean age of 26.18 years (SD 6.33), was predominantly women (86.7%), with a minimum of senior secondary education (96.7%). A majority had a comorbid Axis-I disorder (81.7%) and were on medication (71.6%). A large proportion had neither been admitted for inpatient care (76.7%) nor received psychotherapy (68.3%).

Tables 1 and 2 display the descriptive statistics [frequencies, means, and standard deviations of the dependent (BPD symptom severity) and independent variables (mentalization; attachment; early relational trauma].

Table 1:

Descriptive statistics of the study variables – Borderline Symptom List, Mentalization measures, and Attachment

Variable		M	SD
BSL	Total	49.57	24.05
RFQ	Uncertainty	1.62	0.79
	Certainty	0.63	0.67
IRI	Empathic Concern	21.83	5.12
	Perspective Taking	14.82	5.72
RMET	Total	23.18	5.46
ASQ	Anxious Attachment	54.48	12.84
	Avoidant Attachment	63.03	12.22

Note: (BSL – Borderline Symptom List -23; RFQ – Reflective Function Questionnaire; IRI – Interpersonal Reactivity Index; RMET – Reading the mind in the eyes test; ASQ – Attachment Style Questionnaire)

Table 2:

Descriptive statistics of study variables – Early relational Trauma

	n	%	Impact of trauma
			M	SD
Number of Trauma experiences
No Trauma	1	1.70	-	-
1-3 types of trauma	3	5.00	-	-
Low Polyvictimization (4-6 types)	23	38.30	-	-
High Polyvictimization (7 and above)	33	55.00	-	-
Types
Psychological Abuse	56	93.30	1.41	1.06
Neglect	55	91.70	1.76	1.29
Rejection	54	90.00	2.35	1.61
Role reversal	50	83.30	1.25	1.18
Physical abuse	47	78.40	1.64	1.45
Witnessing domestic violence	47	78.40	2.02	1.65
Sexual abuse	38	63.30	1.05	1.24
Loss	22	36.70	1.52	2.14
Separation	13	21.70	0.33	0.76
Cumulative Impact of Trauma	-	-	1.98	0.93

The majority (93.3%) of the sample reported polyvictimization, experiencing trauma across four or more domains, with over half (55%) indicating trauma across seven or more domains before the age of 14 years. The most frequently reported early relational trauma was psychological abuse (93.3%) involving verbal and demonstrative acts by caregivers intended to control or intimidate the child or cause psychological pain or fear (e.g. ridiculing, blaming, shaming, emotional cruelties such as threatening to hurt the child or the child’s loved one). Neglect (91.7%) was the second most common trauma, involving failure to meet the child’s developmental needs such as their physical, emotional, and school and educational needs. The third most reported trauma was Rejection (90.0%) characterized by cold and hostile caregiving that actively turned away the child’s expressions of need and attachment. Participants rated experiences of rejection, witnessing domestic violence, and neglect as the most impactful.

Table 3 displays the results of the correlational analysis of the variables in the study.

Table 3:

Pearson’s correlation between all the variables

Variables	1	2	3	4	5	6	7	8
1. BSL
2. RFQu	0.60**
3. RFQc	-0.46**	-0.64**
4. IRI_EC	0.06	0.15	0.04
5. IRI_PT	-0.20	-0.33**	0.32*	0.30*
6. RMET	-0.16	-0.09	0.03	0.20	0.01
7. ASQ_Anx	0.60**	0.47**	-0.52**	0.02	-0.28*	0.12
8. ASQ _Avo	0.39**	0.27*	-0.27*	-0.23	-0.27*	-0.11	0.42**
9. cTQ_impact	0.37**	0.18	-0.12	0.26*	-0.03	0.06	0.13	0.21

*p < 0.05, **p < 0.01. Note: (BSL – Borderline Symptom List -23; RFQu – Reflective Function Questionnaire- uncertainty; RFQc – Reflective Function Questionnaire - certainty; IRI_EC – Interpersonal Reactivity Index_Empathic Concern; IRI_PT – Interpersonal Reactivity Index_Perspective Taking; RMET – Reading the mind in the eyes test; ASQ_Anx – Attachment Style Questionnaire_Anxious attachment; ASQ_Avo – Attachment Style Questionnaire_Avoidant attachment; cTQ_impact – Complex Trauma Questionnaire_Impact cumulative)

The results shown in Table 3 demonstrate that while reflective function was significantly correlated with the severity of BPD symptoms, none of the other measures of mentalization had any associations with symptom severity. Higher levels of uncertainty (r = 0.60; P < 0.01) and lower levels of certainty (r = -0.46; P < 0.01) about the mental states of self and others were both found to be associated with greater BPD severity. High attachment anxiety (r = 0.60; P < 0.01), attachment avoidance (r = 0.39; P < 0.01), and cumulative impact of trauma (r = 0.37; P < 0.01) were all correlated with greater severity of BPD symptoms. The lack of awareness of mental states was observed to be associated with higher attachment anxiety (RFQu: r = 0.47; P < 0.01) and avoidance (RFQu: r = 0.27; P < 0.05). On the other hand, a certain degree of awareness of mental states of self and others was found to be correlated to lower levels of attachment insecurity (anxiety: RFQc: r = -0.52; P < 0.01; avoidance: RFQc: r = -0.27; P < 0.05). However, the cumulative impact of trauma was not significantly associated with mentalizing abilities or levels of anxious or avoidant attachment.

Based on the correlational and univariate regression analyses, the following variables were entered into the stepwise multiple regression analysis as potential predictors of severity of BPD symptoms: uncertain and certain reflective functioning, attachment anxiety, attachment avoidance, and the cumulative impact of traumatic experiences.

The findings [Table 4] indicated that uncertainty about mental states of self and others (B = 11.30, P = 0.001), an anxious style of attachment (B = 9.78, P = 0.001), and the cumulative impact of trauma (B = 6.41, P = 0.010) accounted for 53.1% of the variance in borderline symptom severity assessed over the past 1 week.

Table 4:

Regression model for Predictors of Borderline Symptom Severity (n=60)

Variable	B	SE B	B	t	P
Uncertainty	11.30	3.20	0.37	3.53	0.001*
Anxious Attachment	9.78	2.75	0.37	3.55	0.001*
Cumulative Impact of Trauma	6.41	2.41	0.25	2.66	0.010*

Note: R²=0.531

DISCUSSION

The findings support the mentalization model of BPD, which implicates mentalizing impairments arising from early attachment vulnerabilities and traumatic experiences in the development and manifestation of BPD.[27,28]

The findings indicate that attachment anxiety is predictive of BPD symptom severity and is also associated with mentalizing impairments. This is consistent with research indicating that attachment anxiety rather than attachment avoidance is strongly associated with BPD.[29,30] Attachment anxiety is associated with the use of hyperactivating strategies (e.g. vigilance, preoccupation, attention-seeking), which increase the frequency and intensity of negative emotions and impulsivity and are associated with greater symptom severity.[29] When the attachment system is activated for individuals with an avoidant style of attachment, deactivating strategies such as emotional withdrawal are prominent.[31] This may allow them to maintain mentalization for a longer duration in interpersonally stressful situations than individuals with an anxious style.[3]

In the present study, while a range of mentalizing and allied social cognition constructs were assessed through self-report and task-based measures, only uncertain reflective functioning significantly predicted the severity of BPD symptoms. Uncertain reflective functioning has been noted in several studies comparing individuals with BPD to healthy controls[16,32] and is associated with borderline psychopathologies such as emotional dysregulation, impulsivity, difficulties in attentional control, and interpersonal dysfunction.[1,33] The uncertain RF style observed in the sample suggests that mentalizing is concrete and inflexible, making it difficult to consider complex ways of understanding their own or someone else’s mind. This can deprive these individuals of appraisal processes that help reduce stress in social contexts, leaving them vulnerable and confused in terms of interpreting and expressing mental states in a socially acceptable manner.[34]

A key contribution of this study is the comprehensive assessment of the types and perceived impact of early relational trauma across several domains. In this study, 85% of the participants with BPD reported at least five different types of childhood trauma. Research suggests that survivors of multiple traumas experience greater and more complex psychological problems.[35] In BPD studies, it is often a combination of traumas across several domains, rather than the occurrence of a single trauma, which has explained more of the variance observed in symptoms and severity.[36,37]

The prominence and enduring impact of experiences of rejection and neglect dovetail with findings that early exposure to emotional abuse and neglect is strongly associated with BPD compared with community controls.[38] While neglect characterized by the caregiver’s withdrawal from caretaking and failure to respond to the developmental needs of the child (physical, emotional, supervisory, and educational needs) can make the child feel unloved, unnoticed or invisible, rejection, involving the caregiver’s active spurning of the child’s expressions of need and attachment, can make the child feel avoided, unwanted, or disliked.[24] Both experiences are strongly associated with the development of BPD.

While the association between early adversities and BPD is well documented,[18,38,39] the presence of multiple traumatic experiences and the perceived impact of these events on the individual have not been extensively studied. In the present study, it was the perceived cumulative impact of trauma that was predictive of the severity of BPD symptoms.

There are several theoretical, clinical, and practical implications of the findings of this study. The results of this study add to the emergent empirical evidence for the MBT framework of BPD.

Practice implications include detailed assessments of attachment patterns, trauma profiles, and mentalizing abilities to guide the conceptualization and treatment of BPD. Exploring trauma in depth—considering its timing, involved individuals, duration, perceived impact, and its relationship with mentalizing—can inform treatment approaches and enhance our understanding of BPD. Given the ubiquity of early relational trauma in BPD, there is debate over using complex PTSD (cPTSD) as a less stigmatizing and more validating diagnosis. Although no consensus exists, the significance of relational trauma in BPD underscores its importance in assessment and intervention.[40]

Therapeutic implications involve understanding attachment patterns and mentalizing profiles to develop personalized interventions. Identifying attachment vulnerabilities can inform the therapeutic relationship and help address potential alliance ruptures. Although the impact of trauma on BPD symptoms is recognized, standard interventions often do not address this directly. Our findings highlight the importance of early adversities and their intersections with mentalization, supporting emerging work on mentalizing traumatic experiences[41] and the development of trauma-focused mentalization-based therapy.[42] Given that trauma reported by individuals with BPD is often relational, involving families in treatment is crucial, especially in collectivistic societies like India, where family ties remain strong in adulthood and the treatment process. Early caregiving contexts, which influence attachment and relational trauma predictive of BPD symptoms, highlight the need for preventive approaches targeting family and community risk factors. Integrating trauma screening, parent–infant attachment interventions, parenting support programs, intimate partner violence interventions, community awareness campaigns, and public policy initiatives into health and social services as part of the multisystemic approach can help reduce risks and build resilience.

The limitations of the study include the use of a cross-sectional design and predominantly self-report measures and the absence of information on experiences of relational trauma in adulthood. In addition, the sample size precluded an examination of the impact of varied psychiatric comorbidities on the severity of BPD symptoms. The RFQ-8, one of the few reliable measures of RF at the time of the study, has faced criticism for its limited sensitivity to ‘hypermentalizing’, unequal distribution of items that assess self and other related mentalization, and potential overlaps with personality pathology.[43] Other measures with better validity may be more appropriate for future research. Though translated into regional languages, the measures lack validation for the Indian population, so results should be interpreted cautiously due to potential cultural and linguistic differences.

Future research should focus on in-depth measurement of personality vulnerabilities using a dimensional approach and the inclusion of emotional regulation, which is a significant predictor of BPD severity. Longitudinal studies with larger sample sizes may be better for mediational and other powerful statistical analyses. This would help establish stronger and more comprehensive empirical support for the mentalization-based model of BPD and inform the development of effective assessments and interventions for this vulnerable group.

Data availability statement

The participants of this study did not give written consent for their data to be shared publicly, so due to the sensitive nature of the research, supporting data is not available.

Authors’ contributions

Study conception and design: All authors. Material preparation, data collection, and initial analysis: KDD. Interpretation and detailed analysis of the data: KDD, PB. The first draft of the manuscript: KDD. Critical revision for important intellectual content: All authors.

Conflicts of interest

There are no conflicts of interest.

Funding Statement

The article is part of the doctoral work of the first author which was supported by funding from the Indian Council of Medical Research (ICMR). The funding source was not involved in the conception and design of the study, acquisition of data, analysis and interpretation of data, or drafting of the article.