Executive Dysfunction and Working Memory in Borderline Personality Disorder and Their Association with Impulsivity and Self-injurious Behavior: A Cross-sectional Study

Participants and Recruitment

Thirty participants meeting DSM-5 criteria for BPD were recruited from outpatient and inpatient settings. ¹³ While efforts were made to reduce selection bias by approaching consecutive patients, the specific gender distribution (60% male, 40% female) and the mix of outpatient and inpatient participants should be considered when interpreting generalizability. The sample size was determined via a priori power analysis using G*Power software (version 3.1). ¹⁴ Assuming a medium effect size (ρ = 0.5) for Spearman’s correlations between cognitive deficits and impulsivity, a power of 0.80, and an alpha of 0.05 (two-tailed), a minimum of 29 participants was required. Of 38 individuals screened, 33 were eligible; 5 were excluded (three for severe substance use disorders, two for neurological conditions), and three declined due to time constraints, yielding 30 participants (78.9% participation rate). ¹⁵ The recruitment flow is summarized as follows: Screened: n = 38; Eligible: n = 33 (excluded: n = 5; 3 substance use, 2 neurological); Consented: n = 30 (declined: n = 3; time constraints); Analyzed: n = 30. Inclusion criteria included ages 18–45 years, confirmed BPD diagnosis, at least one SIB episode, basic literacy, and consent willingness. Exclusion criteria comprised major neurocognitive or neurological disorders, severe substance use disorders (except nicotine/caffeine), and sedative use impairing cognition. ¹⁶ To minimize selection bias, consecutive patients meeting DSM-5 criteria were approached, reducing cherry-picking. ¹³

Measures

Participants completed validated instruments to assess BPD symptoms, impulsivity, SIB, and cognitive function, administered in a controlled clinical setting. ¹⁷ While robust exclusion criteria were applied to minimize confounding factors, potential influences of untreated or subclinical comorbidities (e.g., depression, anxiety) or the impact of ongoing psychotropic medication status on cognitive performance were not explicitly controlled for as separate variables in the analysis. This should be considered in the interpretation of results and is a point for future research. The Borderline Symptom List (BSL-23) is a 23-item self-report scale measuring BPD symptom severity, including emotional instability and behavioral dysregulation, with high reliability (Cronbach’s α = 0.94) and validity in psychiatric populations. ¹⁸ The Barratt Impulsiveness Scale (BIS-11) is a 30-item questionnaire assessing motor, attentional, and non-planning impulsivity, widely used in BPD research, with strong internal consistency (Cronbach’s α = 0.83). ¹⁹ The Deliberate Self-Harm Inventory (DSHI) is a 17-item measure capturing SIB frequency, type, and severity, demonstrating good reliability (Cronbach’s α = 0.89) and sensitivity to self-harm behaviors. ²⁰ Neurocognitive assessments included the Wisconsin Card Sorting Test (WCST), a standardized task evaluating ED through perseverative errors and categories completed, validated for assessing cognitive flexibility in BPD (test-retest reliability ≈ 0.80). ²¹ The Letter-Number Sequencing (LNS) task, from the Wechsler Adult Intelligence Scale (WAIS-IV), measures WM by requiring participants to reorder mixed sequences, with high reliability (Cronbach’s α ≈ 0.85) in clinical settings. ²² All tools were administered per standardized protocols, with no missing data for primary variables (WCST perseverative errors, LNS scores, BIS-11 totals, DSHI totals).

Procedure

Assessments lasted approximately 60 minutes per participant in a distraction-free environment to mitigate performance bias, though assessors were not blinded to BPD diagnosis due to the clinical context. ¹⁷ This study received no specific funding from public, commercial, or not-for-profit sectors, and was conducted as part of routine clinical research activities at the hospital.

Data adhered to STROBE guidelines, uploaded as supplementary online material. ²³

Statistical Analysis

Descriptive statistics summarized demographics and clinical variables. Independent t-tests compared neurocognitive performance by recent SIB status, and ANOVA assessed impulsivity across ED levels. Spearman’s correlations explored relationships between WCST errors, LNS scores, BIS-11, and DSHI scores. Impulsivity levels were categorized using a median split of BIS-11 total scores (median = 65) into high (>70), moderate (60–70), and low (≤60) groups to explore distinct cognitive profiles, complementing continuous regression analyses (see Table 4, Supplementary Material).

Multiple linear regression identified SIB predictors. Analyses used SPSS Statistics 30.0.0 (IBM Corp, 2023) ^{2 4} with significance at p < .05.

ED, Impulsivity, and Neurobiological Mechanisms

The significant correlation between BIS-11 impulsivity and WCST perseverative errors (ρ = 0.45, p = .02) highlights cognitive inflexibility’s role in impulsive decision-making, a mechanism also implicated in addiction, where prefrontal deficits drive compulsive behaviors. ⁶ This finding is consistent with previous research demonstrating impaired response inhibition and cognitive flexibility in BPD populations. Our study builds upon this by showing a direct predictive relationship where greater ED is associated with higher impulsivity. This aligns with neuroimaging evidence of dorsolateral prefrontal cortex (DLPFC) dysfunction in BPD, where reduced DLPFC activation during inhibitory tasks correlates with perseverative errors. ⁷ Similarly, DLPFC hypoactivity has been linked to impaired error monitoring, potentially explaining repetitive maladaptive behaviors in our high-impulsivity subgroup. ²

WM and Impulsivity

The negative correlation between LNS scores and BIS-11 totals (ρ = −0.38, p = .04) supports the frontoparietal network hypothesis, where WM relies on DLPFC and posterior parietal coordination. ⁸ Disruptions here, observed in emotionally charged tasks, ³ may impair task-relevant information retention under stress, precipitating impulsivity akin to patterns in addiction. ⁹ Our finding that poorer WM significantly predicts higher impulsivity further corroborates studies showing WM deficits in BPD and extends this by demonstrating its predictive utility for a core behavioral symptom. ⁷

SIB: Cognitive Versus ED

Although SIB correlated with emotional instability via BSL-23 scores, ¹⁵ its lack of a strong association with WM (ρ = −0.20, p = .29) or significant ED linkage (ρ = 0.41, p = .03) in the regression model challenges models positing cognitive deficits as primary SIB drivers. ¹⁰ Instead, our data support a dual-process theory, where limbic hyperactivity and prefrontal hypoactivity interplay, with emotional triggers dominating, ²³ a dynamic also relevant to addiction relapse. ²⁴ This suggests that while cognitive deficits contribute to the broader impulsivity seen in BPD, the more direct and immediate triggers for SIB may be rooted in emotional dysregulation, with cognitive factors potentially playing a more indirect or moderating role. Future research should explore the complex interplay between these cognitive and emotional pathways, perhaps through mediation or moderation analyses, to better understand their combined influence on SIB.

Clinical Implications

Identifying ED and WM as significant predictors of impulsivity suggests that cognitive remediation therapy (CRT)—using adaptive WCST or n-back tasks—could serve as a valuable adjunct to existing treatments like dialectical behavior therapy (DBT). In addiction contexts, CRT has shown promise in reducing relapse by strengthening cognitive control, indicating potential cross-applicability for BPD’s impulsive and SIB. Practically, CRT could be integrated into DBT by incorporating specific cognitive exercises into individual therapy sessions or group skills training. For instance, after teaching emotion regulation skills, a therapist might use WCST-based exercises to help patients improve cognitive flexibility in applying these skills under emotional distress. Combining DBT’s emotion regulation focus with CRT may thus address both cognitive and affective deficits more comprehensively. ²⁵

Strengths and Limitations

This study’s strength lies in its multimodal approach, merging neurocognitive and behavioral measures to explore cognition-impulsivity links in BPD. The a priori power analysis ensures an adequate sample size and complete data collection, enhancing the reliability of our findings. However, several limitations should be considered. First, the cross-sectional design precludes inferences about causality, meaning we cannot determine if ED and WM deficits precede or result from impulsivity and SIB. Second, the relatively small sample size (N = 30) limits the generalizability of these findings to broader BPD populations. Third, the absence of an age and gender-matched control group is a significant limitation, preventing direct comparisons of cognitive functions between individuals with BPD and healthy controls. Therefore, our findings should be interpreted with caution regarding the absolute severity of cognitive impairments in BPD. Fourth, while the study focused on cognitive deficits, the neurocognitive assessments did not directly incorporate emotional factors. Future research could benefit from more integrated emotional-cognitive assessments to offer a fuller picture of BPD’s neurocognitive profile, given the centrality of emotional dysregulation in BPD. Additionally, while standardized administration protocols were followed, assessors were not blinded to the BPD diagnosis due to the clinical context, which may have introduced minor assessment bias. However, efforts were made to mitigate this.

Future Directions

To establish causality, future research should prioritize longitudinal studies to determine if ED and WM deficits precede or are a consequence of chronic impulsivity and SIB in BPD. Specifically, these studies could track cognitive development and behavioral outcomes over time, potentially using growth curve modeling to identify predictive relationships. Functional neuroimaging studies (e.g., fMRI) could further map the neural correlates of these cognitive deficits and their dynamic interplay with emotional circuits during impulsive or self-injurious urges. ⁷ Furthermore, randomized controlled trials are critically needed to robustly test the efficacy of cognitive remediation therapy in reducing impulsivity and SIB in BPD, potentially informing addiction treatment strategies as well. Such trials could compare CRT as a standalone intervention, as an adjunct to DBT, or against other active control conditions to determine optimal treatment approaches. ²⁴