Daily Self-Reflection Fluctuations Predict Borderline Personality Traits in Real-Time Study

A newly published investigation tracking young adults in real time reveals that day-to-day instability in how individuals understand their own inner thoughts and emotions is closely linked to borderline personality features, whereas confidence in interpreting others appears largely unrelated. Conducted across seven days in a non-clinical community sample, the study utilized ecological momentary assessment to capture the temporal dynamics of mentalizing—the social-cognitive capacity to interpret behavior through intentional mental states such as feelings, desires, and intentions. By issuing repeated smartphone surveys to participants in their daily routines, the research team discovered that individuals with stronger borderline traits experienced greater day-to-day volatility in self-certainty, alongside lower average confidence in their internal motives. In sharp contrast, neither the average level of certainty regarding others’ mental states nor its daily variability demonstrated an association with borderline accentuation once self-reflection was statistically controlled, providing new evidence on how self-perception vulnerabilities characterize borderline traits.

Borderline personality disorder is a severe mental condition characterized by pervasive impairments in interpersonal relationships, affective instability, poor self-control, and distress across social and occupational functioning. Epidemiological surveys indicate that while the disorder affects roughly one to two percent of the general population in Western societies, its prevalence exceeds three percent among young adults. Clinical theorists identify disruptions in mentalizing as a cornerstone of borderline pathology, positing that impaired capacity to reflect upon mental states destabilizes the coherent organization of the self and distorts perceptions of interpersonal interactions. When mentalizing falters, individuals are frequently overwhelmed by intense affect, heightened rejection sensitivity, and an unstable sense of self, contributing to the impulsive behaviors and volatile relationships typical of borderline personality disorder. In response to these core vulnerabilities, specialized clinical interventions such as mentalization-based therapy, developed by Anthony Bateman and Peter Fonagy, focus on strengthening patients’ capacity to understand mental states to foster emotional stability and personal functioning.

Despite the clinical relevance of mentalization, most research has treated the construct as a static trait measured through single-timepoint laboratory tasks or retrospective questionnaires. This static approach overlooks theoretical models showing that reflective capacity operates as a fluctuating state subject to substantial intra-individual variation across time and changing situations. Initial psychotherapy research examining expressions within treatment sessions demonstrated that mentalizing varies markedly from moment to moment. In a study by Möller and colleagues evaluating patients with borderline personality disorder and comorbid substance abuse, at least seventy-eight percent of the variation in reflective functioning was attributable to differences between statements within individual sessions, a pattern corroborated in single-case psychotherapy analyses. In addition, an initial ecological momentary assessment investigation led by Steinberg and colleagues recorded participants’ emotionally intense moments via daily smartphone audio entries, showing that sixty-seven percent of total variance in reflective functioning stemmed from within-person fluctuations, with higher borderline symptoms predicting greater day-to-day instability.

Capturing momentary fluctuations in mentalizing within daily life presents significant practical challenges that have long constrained quantitative investigations. The established gold-standard instrument, the Reflective Functioning Scale, depends on extensive narrative coding performed by highly trained evaluators, making frequent daily sampling prohibitively costly and labor-intensive. Furthermore, standard reflective functioning ratings assess generalized certainty and uncertainty at a metacognitive level without differentiating between reflection on oneself and reflection on others. To address these operational limitations, researchers developed the Certainty About Mental States Questionnaire, which systematically distinguishes self-certainty from other-certainty across diverse states including feelings, thoughts, intentions, and goals. The instrument also conceptualizes both deficient certainty (hypomentalizing) and excessive, overconfident assumptions (hypermentalizing) as maladaptive psychological patterns. In the present study, the authors adapted this validated framework into a four-item questionnaire optimized for repeated smartphone delivery, providing an economical self-report tool capable of capturing momentary intrapersonal and interpersonal mentalizing in everyday environments.

To test their hypotheses regarding daily mentalizing dynamics, the research team recruited one hundred and one participants from the University of Duisburg-Essen. Following strict data screening criteria, fifteen participants were excluded due to mobile protocol compliance rates falling below seventy-five percent, resulting in a final analytical sample of eighty-six individuals. The cohort was predominantly young and female, with seventy-four women accounting for eighty-six percent of the sample, an average age of 22.93 years, and a standard deviation of 5.36 years. Most participants were single (94.2 percent) and spoke German as their native language (87.2 percent). Additionally, 22.1 percent reported a history of past psychological illness, while 14 percent reported current mental health conditions. Prior to mobile testing, participants completed a thirty-minute computerized baseline assessment measuring demographic details and non-pathological borderline accentuation via the Spontaneous-Borderline Scale of the Personality Styles and Disorder Inventory, which demonstrated high internal consistency with a Cronbach’s alpha of 0.86.

Following baseline testing, participants entered a seven-day ecological momentary assessment phase administered through the mobile software application m-Path. The application delivered surveys six times daily at randomized intervals between 9 am and 9 pm, creating an average interval of two hours and a minimum interval of thirty minutes between notifications. Participants could postpone surveys by up to thirty minutes, with automated reminders dispatched after fifteen minutes of inactivity. Across the seven days, individuals completed up to forty-two assessments, achieving an impressive average compliance rate of 90.9 percent. Each mobile survey presented four items derived from the Certainty About Mental States Questionnaire, scored on a five-point Likert scale. Two items evaluated self-certainty regarding personal virtues and reasons for behavior, while two items assessed other-certainty regarding others’ emotional states and honesty. Reliability analyses using the Spearman-Brown formula confirmed excellent internal consistency across momentary responses, registering 0.93 for self-certainty and 0.97 for other-certainty.

Descriptive findings revealed that participants were, on average, more certain about their own internal states than about the mental states of others. A paired samples t-test confirmed that this difference between self-certainty and other-certainty was statistically significant and moderate in magnitude, yielding t(85) = 4.93 with p < 0.001 and Cohen’s d = 0.53, corresponding to a negative mean difference score of -0.28. To analyze temporal fluctuations, the researchers calculated the Mean Squared Successive Difference for each person across consecutive measurement points. This metric averaged 0.61 for self-certainty (standard deviation 0.56) and 0.94 for other-certainty (standard deviation 0.83). In line with the study’s first hypothesis, borderline accentuation correlated significantly and positively with the Mean Squared Successive Difference of self-certainty (r = 0.26, p = 0.008). In contrast, fluctuations in other-certainty showed no meaningful association with borderline traits (r = 0.06, p = 0.300), indicating that temporal instability in borderline personality is uniquely tied to fluctuations in self-reflection.

When examining average mentalizing levels across the observation period, self-certainty emerged as a substantial negative predictor of borderline personality style, supporting the authors’ second hypothesis. The zero-order correlation between self-certainty and borderline scores was moderately negative (r = -0.46, p < 0.001). To account for intercorrelation between scales, the authors computed semi-partial correlations; controlling for other-certainty, self-certainty remained significantly and negatively related to borderline accentuation (rsp = -0.36, p = 0.001), indicating that self-certainty alone accounted for thirteen percent of the variance in borderline features. In multiple regression analysis, self-certainty significantly predicted borderline accentuation with a standardized coefficient of beta = -0.49 (p < 0.001). Conversely, the third hypothesis was not supported: although other-certainty initially showed a weak negative zero-order correlation with borderline traits (r = -0.29, p < 0.001), this association disappeared completely once self-certainty was controlled (rsp = 0.03, p = 0.789; regression beta = 0.04, p = 0.786).

The researchers also evaluated whether borderline traits are associated with an other-self discrepancy, which theoretical accounts frame as a marker of hypermentalizing. To evaluate this discrepancy hypothesis without confounding individual predictor effects, the team utilized condition-based regression analysis. Although the difference between the standardized regression coefficients of other-certainty and self-certainty was positive (beta difference = 0.53), statistical testing of the auxiliary discrepancy parameter revealed no significant effect (abs = 0.08, p = 0.393). As a result, the fourth hypothesis was not confirmed, demonstrating that higher borderline accentuation is not associated with an elevated discrepancy between other-certainty and self-certainty in daily life. This outcome suggests that an asymmetry where other-certainty matches or exceeds self-certainty may not represent an inherently maladaptive hypermentalizing tendency in naturalistic settings. The authors observed that past findings linking hypermentalizing to borderline pathology often derived from artificial laboratory assessments rather than momentary evaluations embedded in participants’ lived experiences.

In considering the clinical implications, the researchers emphasized that psychotherapeutic interventions such as mentalization-based therapy could benefit from prioritizing self-oriented mentalizing to stabilize identity and reduce symptoms. Digital ecological momentary interventions delivered via smartphones could provide real-time prompts guiding attention toward personal internal states to diminish affective and cognitive volatility throughout the day. However, several critical limitations qualify the generalizability of these findings. The study evaluated a non-clinical, sociodemographically homogeneous sample consisting predominantly of female university students with relatively low baseline borderline scores. In addition, borderline tendencies were measured via the Personality Styles and Disorder Inventory, an instrument assessing non-pathological personality styles rather than clinical disorder criteria. Finally, because the brief four-item questionnaire relied strictly on time-based prompts without capturing specific interpersonal encounters or social interactions, the authors underscored that future research must employ event-contingent sampling and multidimensional measures within diverse clinical populations to comprehensively evaluate mentalizing dynamics in borderline personality disorder.

Subject of Research: Ecological momentary assessment of momentary mentalizing and borderline personality traits

Article Title: Mentalizing on the edge: borderline traits and their association with real-time self-reflection – an ecological momentary assessment study

Article References: Höller, I., Schwitzky, S., Forkmann, T., Forster, S., & Santangelo, P. S. (2026). Mentalizing on the edge: borderline traits and their association with real-time self-reflection – an ecological momentary assessment study. Borderline Personality Disorder and Emotion Dysregulation, 13(1), Article 13. https://doi.org/10.1186/s40479-026-00347-1

Image Credits: AI Generated

DOI: 10.1186/s40479-026-00347-1

Keywords: Borderline Personality Disorder, Mentalizing, Ecological Momentary Assessment, Self-Certainty, Other-Certainty, Personality Styles and Disorder Inventory, Clinical Psychology, Psychopathology, edge, borderline, traits, association