Transcranial direct current stimulation and online cognitive training to enhance cognitive function and emotional stability in borderline personality disorder: an open-labelled pilot study

Abstract

Background:

Despite the effectiveness of specialized therapies, people with borderline personality disorder (BPD) continue to face substantial psychosocial challenges, which may be partially attributed to neuropsychological deficits arising from imbalances in the corticolimbic system. Transcranial direct current stimulation (tDCS) targeting the dorsolateral prefrontal cortex (DLPFC) could enhance impulse control, emotional regulation, and cognitive functions; thus, we sought to explore the effectiveness of tDCS combined with online cognitive training on cognitive functions, BPD symptoms, and psychosocial functioning among patients with BPD.

Methods:

This open-label study recruited adults with BPD who were not undergoing psychotherapy. Participants completed informational psychoeducation sessions, followed by 10 daily sessions of 20-minute tDCS over 2 weeks. Stimulation involved a continuous 2-mA current with the anode over the left DLPFC and the cathode over the right DLPFC. During each session, participants simultaneously engaged in online cognitive training using the Lumosity app (aspredicted.org no. 206 001).

Results:

We included 29 participants. We noted significant improvements in cognitive functions, including the Towers of London task (Cohen d = −0.38 to −0.78), the Corsi Block-Tapping direct and total scores (d = −0.41 and −0.42, respectively), and the Stroop Interference and Alternance tests (d = 0.80 and 0.94, respectively). Emotional dysregulation showed a substantial reduction (d = 0.44), while impulsivity did not change significantly. Symptoms of BPD decreased (d = 0.69), while general functioning (d = 0.33) and the internal component of BPD functioning improved (d = −0.51).

Limitations:

Although these preliminary findings are encouraging, further controlled studies are necessary to validate the efficacy and long-term effect of the intervention.

Conclusion:

This combined approach appears to be well tolerated and produced promising short-term improvements in cognitive performance, BPD symptoms, and overall functioning. The results underscore the relevance of the left DLPFC in developing neuropsychologically integrative interventions for BPD.

Introduction

Borderline personality disorder (BPD) is a mental disorder that exhibits instability in several domains, such as relationships, emotions, and self-perception.1 This disorder is also characterized by substantial impulsivity.1 Its prevalence is estimated at 1%–2% within the general population and around 15%–28% among patients with psychiatric conditions.2–4 Clinical research has advanced considerably in the symptomatic treatment of BPD, primarily through psychotherapy.5 However, although these treatments show moderate effect sizes in reducing symptoms, their overall influence on psychosocial functioning is generally regarded as limited.5,6

One possible explanation for this limited improvement is that these treatments primarily focus on emotions, relational patterns, and dysfunctional cognitive schemas without particularly targeting neuropsychological impairments.7 Although these factors are connected to psychosocial functioning, they may not fully account for it.7 Compared with the general population, people with BPD exhibit substantial neuropsychological impairments across multiple domains, particularly in cognitive flexibility, inhibition, and executive control.8 A meta-analysis by D’Iorio and colleagues8 found moderate-to-strong effect sizes in areas like executive functioning, working memory, decision-making, and sustained attention. Processing speed and visuospatial abilities showed smaller deficits, while memory — primarily verbal and spatial — exhibited significant levels of impairment.9–11 A meta-analysis by Ruocco and colleagues11 and a systematic review by McClure and colleagues12 both found that patients with BPD also experience notable difficulties in planning.13 These neuropsychological impairments are associated with daily functional challenges.8,14

Numerous studies have highlighted specific brain regions or networks that support these neuropsychological functions. 15 In BPD, neurobiological anomalies are linked to symptoms and cognitive dysfunctions, particularly corticolimbic alterations.16,17 Therefore, a central aspect of the neurobiological explanation for BPD involves functional and structural alterations of the dorsolateral prefrontal cortex (DLPFC), which plays a critical role in impulse control, cognitive functions, and emotional regulation.18,19 Findings about the lateralization of the DLPFC are yet to be specific to 1 side.18,19 The left DLPFC has been directly implicated in several studies specific to BPD, although the results are somewhat contradictory.18,20,21 Hypoactivation of the left DLPFC in BPD has been observed during reward-based tasks, reflecting a potential dysfunction in processing motivational salience or future-oriented behaviour, which contributes to impulsivity and difficulty in decision-making.18 This reduced activation impairs the DLPFC’s ability to modulate limbic and subcortical activity, which is essential for controlling impulsive behaviour.17,20–22 Interestingly, the left DLPFC may show heightened activation during tasks involving negative emotional stimuli and behavioural inhibition (e.g., aggression regulation), suggesting an overengaged cognitive control system in emotionally charged contexts, further complicating emotional regulation.18,20,21 Another study reported that people with BPD exhibit reduced bilateral DLPFC activation during negative emotion processing, suggesting a reduced capacity for cognitive control and the use of emotion regulation strategies, which are key features of the disorder’s cognitive dysfunction.18,23 In contrast, healthy individuals show more functionally distinct lateralization in the DLPFC, with the left DLPFC consistently engaged in working memory, stimulus interference control, planning, and proactive control, often without right DLPFC involvement, while the right DLPFC is associated with behavioural inhibition and impulse control during reactive tasks such as go/no-go or stop-signal paradigms.18,23,24

These findings have paved the way for new therapeutic approaches like neuromodulation. In the past decade, these methods have emerged to treat various mental health conditions.25 Notably, many neuromodulation studies in BPD have targeted the left DLPFC, particularly in protocols aiming to enhance regulatory capacities and cognitive functions.25 This focus highlights the potential functional relevance of the left DLPFC in treatment-oriented approaches. Among all these techniques, transcranial direct current stimulation (tDCS) is the most cost-efficient and logistically simple to administer, with the added advantage of potential for supervised home use.26 It applies a low electrical current to the scalp to modulate neuronal excitability and enhance cognitive functions.26,27 It has shown effectiveness for several mental illnesses, including major depressive disorder, anxiety disorders, obsessive–compulsive disorder, and bipolar disorder.26

Thus far, 4 randomized controlled trials (RCTs) have examined the potential of DLPFC-targeted tDCS in BPD, providing preliminary evidence of improvements across different cognitive and symptomatologic domains, depending on electrode placements. These discrepancies may arise from variability in stimulation parameters and the number and frequency of sessions.28 The studies of Molavi and colleagues29 and Wolkenstein and colleagues30 both applied anodal stimulation to the left DLPFC and reported improvements in executive functioning, emotional regulation, and cognitive control. This aligns with broader findings in tDCS research on neurodegenerative and psychiatric disorders, where anodal stimulation of the left DLFPC has been associated with significant improvements in executive and cognitive performance.31 Anodal targeting the left DLPFC has shown reduction of depressive symptoms, particularly among patients with major depressive disorder, and may be a relevant intervention for BPD, given its high comorbidity with depression.3,32 In contrast, studies that have applied anodal stimulation to the right DLPFC have shown mixed outcomes. Lisoni and colleagues33 observed reductions in impulsivity and aggression, while Schulze and colleagues34 reported no significant effects. A thesis reported increased impulsivity after anodal stimulation of the right DLPFC (with the cathode on the left), supporting our decision to target the left DLPFC in the present study.35 Furthermore, the Molavi and Lisoni RCTs delivered tDCS over multiple sessions (10 and 15 sessions, respectively) and reported significant improvements in core cognitive and emotional symptoms of BPD.29,33 In contrast, single-session protocols used in the Schulze and Wolkenstein RCTs found limited or no significant effects on cognitive control.30,34 These results suggest that multiple-session protocols may produce more substantial and reliable effects than single-session approaches.29,33

Additionally, there is a need for standardized, noninvasive brain stimulation protocols designed explicitly for cognitive training that address executive dysfunctions in BPD, as well as more research on cognitive enhancement.28 Cognitive training typically involves nonstandardized, computerized, repetitive tasks aimed at enhancing various cognitive functions. 36 For other neurodegenerative and psychiatric disorders, studies that combined tDCS with online cognitive training showed significant improvements in cognition, attention, learning, and memory, all of which are crucial for daily tasks like planning, problem-solving, and multitasking. 31,37–39 This approach could enhance functional brain connectivity and optimize cognitive outcomes, particularly decision-making and cognitive processing.28,40 Tailoring tDCS to specific online cognitive tasks and applying it to the neural circuits engaged by those tasks during stimulation could ensure better consistency, effectiveness, and treatment outcomes.28,36,40 The neuropsychological effects and their connection to psychosocial functioning in BPD remain underexplored, despite the direct influence of neuromodulation on these outcomes, in contrast to the well-studied effects of psychotherapies.

This article presents a preliminary analysis from a larger study that also included a second phase involving cognitive remediation. The objectives described here are specific to the first phase, which examined the combination of tDCS with online cognitive training, alongside the use of more robust neuropsychological tests to establish conclusive outcomes. This exploratory open-label, single-arm pilot study primarily sought to investigate the potential beneficial effects of 10 daily sessions of tDCS that targeted the left DLPFC, paired simultaneously with 10 sessions of online cognitive training, on cognitive functions such as executive function, working memory, cognitive flexibility, and inhibitory control. Additionally, we sought to examine the effects of the stimulation on emotional instability and impulsivity, which are core symptoms in BPD. The main targets were determined based on previous studies that observed the effects of tDCS on BPD.28 The combination of tDCS with online cognitive training was not solely focused on exercises targeting a specific function but rather on fostering overall cognition. Our secondary objective was to observe the preliminary trends of this intervention’s potential effect on BPD symptoms and general function. We hypothesized that this intervention would significantly enhance cognitive functions, reduce emotional instability and impulsivity, and result in mild improvement of BPD symptoms and daily functioning.

Methods

Study design

This open-label study involved administering neuropsychological tests and questionnaires to assess cognitive functions, symptomatology, and overall functioning. Assessments were conducted before and after the tDCS sessions, which consisted of 10 tDCS sessions combined with online cognitive training (Figure 1). This study was part of a larger project that included 2 months of group cognitive remediation following the tDCS, with a follow-up after 3 months (aspredicted.org no. 206 001). However, this specific analysis focused solely on the 2-week follow-up after tDCS.

Figure 1.

Timeline of a 24-week longitudinal intervention study, with the current article focusing exclusively on phase 1, which evaluated the effects of the combined transcranial direct current stimulation (tDCS) and cognitive training intervention. The larger study includes 4 assessment points (V1–V4). The V1 data collection involved pre-intervention (baseline) questionnaires and neuropsychological assessments. The V2 data collection involved post-tDCS and cognitive training questionnaires and neuropsychological assessments. For the post-tDCS questionnaires, participants were instructed to report symptoms and functioning based solely on their experiences during and after the tDCS combined with cognitive training. The V3 data collection involved post-cognitive remediation questionnaires and neuropsychological assessments, whereby participants were instructed to report symptoms and functioning on their experiences during and after the cognitive remediation (2 previous months). The V4 data collection involved follow-up questionnaires, conducted 3 months after the end of the intervention, with participants instructed to report symptoms and functioning on their experiences after the study. Psychoeducation involved 1–2 sessions during the first 2 weeks of the study and tDCS plus cognitive training was delivered in 10 daily sessions over a 2-week period. Cognitive remediation was conducted as 8 weekly group sessions. Psychiatric follow-up and treatment as usual continued throughout the study period.

Participants

Between March 2023 and January 2024, we recruited patients with BPD from the Institut universitaire en santé mentale de Montréal. We included French-speaking patients aged 18 years or older with diagnoses of BPD based on the Structured Clinical Interview for DSM-IV Axis II Personality Disorders (SCID-II). To ensure treatment stability, participants were required to have no changes in their psychopharmacologic treatment (dosage, type, or frequency) during the 2 weeks of the study. We applied this criterion to minimize confounding effects and ensure consistency in clinical and cognitive outcomes. However, no specific pharmacologic treatments were excluded. The exclusion criteria were having a concurrent mental disorder that would hinder participation (e.g., mania, severe substance use disorder, intellectual impairment, psychosis), contraindications for tDCS (e.g., epilepsy, intracranial metal, scalp scars or skin issues, pacemaker), homelessness, and undergoing psychotherapy.

Intervention

Psychoeducation sessions

Before starting the tDCS, we conducted two 1.5-hour sessions to provide information on the psychological aspects of BPD. Participants who could not attend received a document to ensure they had the necessary information. Both the sessions and materials covered etiology, types of symptoms, treatments, and coping strategies for BPD to help participants better understand their disorder. This is a standard preparatory element commonly used in psychiatric treatment protocols to enhance participants’ understanding of their condition and support their engagement in the upcoming intervention.41 It did not involve therapeutic or skills-based training.

Transcranial direct current stimulation

The stimulation was performed using the DC-Stimulator device from neuroConn (neuroCare Group) with 25 cm² square electrodes. Given the preliminary results of past RCTs in BPD, we placed the anode on the F3 site (left DLPFC) and the cathode on the F4 site (right DLPFC).29,30 The stimulation parameters were set to 2 mA for 20 minutes, including a 20-second fade-in and fade-out period. Ten sessions took place over 2 weeks, with the administration days varying based on the participant’s availability.

Online cognitive training

During the 10 daily tDCS sessions, participants engaged in cognitive training using the Lumosity application. Each participant underwent 20-minute sessions, and although we recommended that the exercises be completed only during the 10 daily sessions, some participants continued beyond the suggested timeframe. Lumosity provided 5 daily cognitive games, along with a “strengthen training” option consisting of 5 additional games. The Lumosity exercises aimed to enhance working memory, problem-solving, cognitive flexibility, attention, and processing speed, each accounting for around 18% of the recommendation. The games also addressed other areas, such as mathematics and language, although Lumosity recommended these less frequently, around 5% each. Our team received forty 3-month trials as part of the Lumosity Clinical Access Research & Engagement Program.42

It is important to note that Lumosity is a general cognitive training application and was not specifically designed for people with BPD. Most studies using this tool have involved healthy adults, older adults, or people recovering from stroke, with inconsistent results regarding cognitive benefits.43,44 No studies have targeted psychiatric populations. Thus, its use in this study is exploratory and further research is needed to assess its clinical value in BPD and related conditions. Nevertheless, we chose Lumosity for its accessibility, gamified structure, and broad cognitive engagement, which aligned with the study’s aim to explore the feasibility and potential cognitive effects of concurrent cognitive stimulation in a population with BPD.

Measures

Neuropsychological tests

The Tower of London test assesses executive function specific to problem-solving and planning.45 This test consists of 12 situations, each lasting 60 seconds, where participants rearrange 3 balls on 3 towers of varying sizes to match a presented image. They have 3 attempts per problem and a limited number of moves. Various scores are assessed based on skills, such as initiation, execution, and total time, number of correct moves, and total number of moves.46 The test–retest reliability coefficients for planning accuracy over a 1-week interval range from 0.69 to 0.74, indicating a good score.46

The Corsi Block–Tapping test measures short-term visuospatial and working memory.47 In this task, participants are shown 9 blocks arranged on a flat surface and must tap them in a specific sequence. The sequence begins with 2 blocks and increases by 1, with each correct repetition lasting up to 9 blocks. The test concludes when the participant fails to replicate the sequence.44 It includes both a direct and reverse task for replicating the sequences. The average performance for adults ranges between 5 and 7 blocks.47 The test–retest reliability coefficients for this task over a 20-minute interval range from 0.84 to 0.95, indicating excellent reliability.48 Given that this range is based on a brief interval, the reliability may be slightly overestimated, but similar tasks show adequate stability over 2 weeks, supporting its use.

The Stroop test assesses cognitive flexibility, inhibitory control, processing speed, and attentional capacity.49 It measures participants’ ability to distinguish a word from its colour in a reading task that requires them to name the colour of the words quickly and accurately while ignoring the semantic aspect of the phrase.49 The test–retest reliability coefficients over a 1–2-week interval range from 0.67 to 0.83, indicating a good reliability.50

The Number Sequencing or Letter-Number Sequencing Task assesses auditory–verbal working memory.51 In these tasks, participants listen to sequences of numbers or alphanumeric series and then repeat them orally, with numbers in ascending order, followed by letters in alphabetical order. Difficulty increases with more characters added at each level; each correct response earns 1 point, for a maximum of 21 points.51 The test–retest reliability coefficients for over a 1-week interval is around 0.75, indicating a good score.52

Symptomatology questionnaires

The Short Urgency, Premeditation, Perseverance, Sensation Seeking, Positive Urgency, Impulsive Behavior Scale (S-UPPS-P) scale is a shortened version of the UPPS-P scale, comprising 20 items.53,54 This scale measures impulsivity across 5 dimensions: negative urgency, positive urgency, lack of premeditation, lack of perseverance, and sensation seeking, each with 4 items. Responses are rated on a 4-point scale (1 = strongly agree; 4 = strongly disagree). The total score for this scale varies from 20 to 80, with higher scores indicating greater impulsivity. The test–retest reliability of the subscales ranges from 0.84 to 0.92.54

The Difficulties in Emotion Regulation Scale (DERS) uses 36 items to measure 6 facets of emotion regulation: awareness, clarity, goals, impulse, nonacceptance, and strategies.55 Each item is rated from 1 (rarely) to 5 (almost always), with higher scores indicating greater difficulties in emotional regulation. Total scores range from 36 to 180, while subscale scores range from 6 to 30, calculated by summing the 6 items for each dimension of emotion regulation. 56 The test–retest reliability coefficient is 0.88, suggesting high reliability.57

The Borderline Symptom List (BSL)-23 is a 23-item version of the BSL-95, based on DSM-5 criteria for BPD. Each item reflects mental states common among patients with BPD and is rated for intensity over the past week on a 5-point Likert scale (0 = not at all; 4 = very strong). The questionnaire assesses symptom severity, with average scores categorized into 6 levels: none or low (0–0.3), mild (0.3–0.7), moderate (0.7–1.7), high (1.7–2.7), very high (2.7–3.5), and extremely high (3.5–4.0).58,59 The test–retest reliability is excellent at 0.84.59

General and daily function questionnaires

The Functional Assessment of BPD (FAB) is a 36-item questionnaire used to determine the general functioning level of people with BPD.60 It evaluates daily activities, community activities, social environment, and intrapersonal dimensions. Each question is rated on a 4-point scale (1 for unaccomplished or very disorganized, 2 for poorly organized with impulsivity, 3 for some instability, and 4 for well integrated into daily routine). The total score, obtained by summing all responses, indicates harmful daily functioning, with higher scores reflecting integrated functioning and a score close to 0 indicating more substantial dysfunction. The test–retest reliability coefficient is 0.92 for functioning and 0.87 for daily difficulty level, demonstrating excellent test–retest reliability.60

The World Health Organization Disability Assessment Schedule (WHODAS 2.0) assesses the functioning, health, and disability level of an adult living with a physical or mental condition.61 This questionnaire has 12 items across 6 domains: understanding, communication, daily activities, self-care, mobility, and social participation. Responses are rated on a 5-point scale (1 = none; 5 = extreme or cannot do). The total score, ranging from 0 to 100, is obtained by summing the scores, with higher scores indicating more substantial disability.61 The overall test–retest reliability coefficient is 0.98, with domain-specific reliability coefficients ranging from 0.93 to 0.96, indicating excellent reliability.61

Statistical analysis

For our primary objective, we used neuropsychological tests to evaluate cognitive abilities, the DERS to assess emotional regulation, and the S-UPPS-P to measure impulsivity. For our secondary objective, we measured symptomatology with the BSL-23 and general functionality using the FAB and WHODAS 2.0.

We began by calculating descriptive statistics, such as mean values and standard deviations (SDs). We then conducted a Shapiro–Wilk test to assess the normality of the data. For normally distributed data, we applied paired samples Student t tests. We used Wilcoxon signed-rank tests for non-normally distributed data for each dependent variable, including cognitive function, symptoms, and general functioning. We employed correction for false discovery rate (FDR) to adjust our results in hypothesis testing or multiple comparisons. Given the exploratory nature of this study and the relatively small sample size, we employed a more liberal FDR threshold (q < 0.1) to increase sensitivity and reduce the likelihood of overlooking potentially meaningful effects. Findings reported here should be interpreted as preliminary and will be subject to validation in subsequent, larger-scale studies. Additionally, we calculated effect sizes using Cohen d and correlation analysis between the cognitive, symptomatic, and general function variables. We conducted analyses using SPSS 29.

Ethics approval

The study received approval from the research ethics committee of the Centre intégré universitaire de santé et de services sociaux de l’est-de-Montréal (no. 2023-3263). Participation was voluntary; participants were fully informed about the study protocol and provided written consent after meeting the inclusion criteria.

Results

We recruited a total of 45 patients; 13 abandoned before starting the study and 3 after starting (Figure 2). Twenty-nine patients completed the study, including 22 women, 6 men, and 1 nonbinary person. The participants’ ages ranged from 18 to 63 years, with a mean age of 35.28 (SD 11.75) years. Years of education varied between 6 and 20 years, with an average of 12.69 (SD 3.40) years. All 29 participants completed the 10 tDCS sessions. Of these, 27 completed the sessions within the intended 2-week period, while the remaining 2 participants completed the intervention in slightly under 3 weeks because of scheduling constraints (Appendix 1, Table S3, available at www.jpn.ca/lookup/doi/10.1503/jpn.250041/tab-related-content). All 29 participants maintained stable psychopharmacologic treatment, with no changes in dosage, type, or frequency during the tDCS intervention. There was no tolerance issue reported for the current intensity during the intervention. Some mild adverse effects were observed, including tingling sensations (n = 22), which decreased as sessions progressed, headaches (n = 8), light nausea or dizziness (n = 2), and drowsiness (n = 1), primarily occurring after the first 2–3 sessions.

Figure 2.

Flowchart of the recruitment and completion of participants completion of transcranial direct current stimulation (tDCS) combined with cognitive training. See Related Content tab for accessible version. BPD = borderline personality disorder.

Cognitive functions

The analysis revealed a statistically significant increase from pre- to post-tDCS in percentiles for total correct moves and execution time for the Towers of London task (Table 1). The effect sizes varied from −0.38 to −0.78. However, the percentiles for initiation time significantly decreased.

Table 1.

Cognitive functions before and after 10 sessions of bilateral dorsolateral prefrontal cortex transcranial direct current stimulation (tDCS)

Measure	Domain	Mean ± SD		p value	Effect size, Cohen d (95% CI)
		Pre-tDCS	Post-tDCS
Tower of London, percentiles
Total correct moves	Problem-solving	57.17 ± 29.15	67.07 ± 26.72	0.047	−0.38 (−0.74 to −0.00)
Total moves	Problem-solving	52.79 ± 31.71	66.76 ± 20.24	0.04	−0.48 (−0.85 to −0.10)
Total initiation time	Planning	84.86 ± 14.85	79.45 ± 15.46	0.02	0.47 (0.09 to 0.84)
Total execution time	Executive function	39.76 ± 27.70	58.10 ± 25.48	0.004	−0.63 (−1.01 to −0.23)
Total time		24.38 ± 18.71	40.17 ± 24.31	< 0.001	−0.78 (−1.19 to −0.37)
Corsi’s Block-Tapping task
Direct	Working memory	7.86 ± 1.57	8.62 ± 1.97	0.07	−0.41 (−0.78 to −0.04)
Reversed	Concentration	7.38 ± 1.50	7.66 ± 1.47	0.4	−0.17 (−0.52 to 0.19)
Total score		15.24 ± 2.43	16.28 ± 2.96	0.08	−0.42 (−0.79 to −0.05)
Number Sequencing task
Direct	Working memory	11.28 ± 2.56	11.66 ± 2.44	0.4	−0.18 (−0.53 to 0.18)
Reversed	Concentration	8.93 ± 2.64	9.06 ± 2.25	0.5	−0.06 (−0.41 to 0.30)
Letter-Number Sequencing task	Cognitive flexibility	18.66 ± 2.04	19.24 ± 2.37	0.1	−0.37 (−0.74 to −0.00)
Total score		38.86 ± 5.79	39.97 ± 5.48	0.1	−0.34 (−0.70 to 0.03)
Stroop test, s
Inference	Interference	58.94 ± 20.98	50.72 ± 17.96	< 0.001	0.80 (0.38 to 1.20)
Alternance	Alternance	68.20 ± 15.49	57.88 ± 16.39	< 0.001	0.94 (0.50 to 1.36)

CI = confidence interval, SD = standard deviation.

For the Corsi Block-Tapping test, the direct task and total scores significantly increased from pre-tDCS to post-tDCS, with effect sizes of −0.41 and −0.42, respectively. Conversely, the reverse task did not differ substantially before and after tDCS. Working memory and sequencing tasks (direct, reverse, and letter–number) showed no significant differences between the 2 time points (Table 1).

The Stroop Interference test showed statistical significance in inhibiting cognitive interference after tDCS, with the average time decreasing and an effect size of 0.80 (Table 1). The Stroop Alternance test also demonstrated a significantly reduced average time between the 2 measures, with an effect size of 0.94.

Symptoms

The statistical analysis of the S-UPPS-P showed no statistically significant differences in impulsivity between the 2 time points (Table 2). However, we observed a statistically significant reduction in emotional dysregulation, measured by the DERS, after tDCS, with an effect size of 0.44. Statistically significant decreases were identified in specific domains such as goals, strategies, and clarity, with effect sizes of 0.41, 0.45, and 0.47, respectively. In contrast, we did not observe any statistically significant changes in other emotional regulation domains, such as nonacceptance, impulse, and awareness.

Table 2.

Symptoms before and after 10 sessions of bilateral dorsolateral prefrontal cortex transcranial direct current stimulation (tDCS)

Measure	Domain	Mean ± SD		p value	Effect size, Cohen d (95% CI)
		Pre-tDCS	Post-tDCS
BSL-23	BPD symptoms	2.25 ± 0.99	1.58 ± 1.12	< 0.001	0.69 (0.29 to 1.08)
DERS	Total score: difficulties in emotional regulation	128.97 ± 27.59	117.55 ± 32.51	0.047	0.44 (0.07 to 0.81)
	Nonacceptance	20.86 ± 7.40	18.83 ± 7.79	0.2	0.26 (−0.10 to 0.62)
	Goals	21.14 ± 4.32	19.28 ± 5.36	0.06	0.41 (0.03 to 0.77)
	Impulse	20.21 ± 5.56	19.24 ± 6.37	0.7	0.18 (−0.18 to 0.53)
	Awareness	20.45 ± 5.45	19.21 ± 5.22	0.1	0.31 (−0.06 to 0.67)
	Strategies	30.07 ± 7.14	26.76 ± 8.94	0.047	0.45 (0.07 to 0.82)
	Clarity	15.93 ± 5.08	14.24 ± 5.07	0.047	0.47 (0.09 to 0.84)
S-UPPS-P	Total score: impulsivity	53.93 ± 9.96	52.93 ± 9.47	0.7	0.18 (−0.18 to 0.53)
	Negative urgency	11.21 ± 3.30	11.37 ± 2.90	0.8	−0.05 (−0.41 to 0.30)
	Positive urgency	12.24 ± 2.92	12.13 ± 2.57	0.7	0.05 (−0.30 to 0.41)
	Lack of premeditation	10.00 ± 2.25	9.62 ± 2.93	0.7	0.12 (−0.24 to 0.47)
	Lack of perseverance	10.28 ± 2.99	9.69 ± 3.30	0.7	0.14 (−0.22 to 0.50)
	Sensation seeking	10.41 ± 3.16	10.10 ± 3.10	0.7	0.19 (−0.17 to 0.54)

BPD = borderline personality disorder; BSL-23 = Borderline Symptom List; CI = confidence interval; DERS = Difficulties in Emotion Regulation Scale; S-UPPS-P = Short Urgency, Premeditation, Perseverance, Sensation Seeking, Positive Urgency, Impulsive Behavior Scale; SD = standard deviation.

According to the BSL-23, the treatment significantly reduced BPD symptoms, as patients’ mean scores shifted from the high range to the moderate range, with an effect size of 0.69 (Table 2).

General and daily functions

The WHODAS 2.0, which evaluated general daily functions, showed a significant difference between the baseline and post-treatment scores, with an effect size of 0.33 (Table 3). However, general functioning specific to BPD, measured with the total FAB score, showed no statistically significant increase. Only the internal component domain showed a statistically significant increase, with an effect size of −0.51. No statistically significant differences were found in interpersonal relationships, daily activities, and community activities.

Table 3.

General functioning before and after 10 sessions of bilateral dorsolateral prefrontal cortex transcranial direct current stimulation (tDCS)

Measure	Domain	Mean ± SD		p value	Effect size, Cohen d (95% CI)
		Pre-tDCS	Post-tDCS
WHODAS 2.0	General daily function	54.20 ± 17.52	50.69 ± 18.43	0.03	0.33 (−0.04 to 0.69)
FAB	Total score: general daily function regarding BPD	60.00 ± 14.20	67.00 ± 16.30	0.2	−0.41 (−0.78 to −0.04)
	Daily activities	56.35 ± 15.83	64.35 ± 19.13	0.2	−0.37 (−0.74 to −0.00)
	Community activities	62.28 ± 14.07	64.44 ± 19.21	0.9	−0.13 (−0.49 to 0.22)
	Interpersonal relationships	71.42 ± 18.29	79.05 ± 15.58	0.1	−0.38 (−0.75 to −0.01)
	Internal component	50.00 ± 18.60	61.14 ± 21.82	0.009	−0.51 (−0.89 to −0.13)

BPD = borderline personality disorder; CI = confidence interval; FAB = Functional Assessment of Borderline Personality Disorder; SD = standard deviation; WHODAS = World Health Organization Disability Assessment Schedule.

Correlation analysis

Primary cognitive outcomes showed limited significant correlations with secondary outcomes, whereas symptom severity and general functioning were significantly associated with each other (Appendix 1, Table S1 and S2).

Discussion

This pilot study investigated the effects of combining 10 daily sessions of tDCS with online cognitive training for people with BPD. This safe, 2-week treatment led to small-to-significant improvements in cognitive functions, BPD symptoms, and certain aspects of psychosocial functioning. Notably, 91% of participants remained throughout the entire duration of the tDCS procedure, demonstrating high adherence to the study protocol. This indicates that the procedure was generally well received.

We found significant improvements in the cognitive domains of executive functions, visuospatial working memory and cognitive flexibility, and inhibitory control. We did not observe significant changes in impulsivity; however, we noted some improvements in emotional instability, particularly regarding goals, strategies, and clarity. We also observed a significant reduction of symptoms and improvement in general function specifically relating to the internal component of BPD.

Our findings are consistent with 2 pilot RCTs that used anodal stimulation of the left DLPFC in individuals with BPD.29,30 The RCT by Molavi and colleagues29 evaluated the effect of anodal stimulation of the left DLPFC and the cathode on the right DLPFC over 10 sessions at 2 mA and found significant improvements in cognitive functioning for the tDCS group compared with the sham group, with notable enhancements in response inhibition, working memory, emotional control, sustained attention, cognitive flexibility, and stress tolerance. However, cognition was assessed using a self-report instrument.29 Wolkenstein and colleagues’30 RCT involved single-session, 1-mA anodal stimulation targeting the left DLPFC, with the cathode placed on the contralateral mastoid bone and assessed cognition through 2 neurocognitive tests that revealed improvements in cognitive control. Similarly, but with a more comprehensive neuropsychological evaluation, we observed significant improvements in neuropsychological functioning particularly in executive function, problem-solving, cognitive inhibition and flexibility, and working memory.

In contrast, the RCT conducted by Lisoni and colleagues31 explored the effectiveness of anodal stimulation of the right DLPFC with the cathode placed on the left DLPFC (15 sessions at 2 mA) and demonstrated reduced impulsivity and aggression in the tDCS group compared with the sham group. However, our study showed no significant reduction in impulsivity. This signifies that distinct electrode placements can still result in efficient cognitive alterations that are beneficial for patients with BPD. Even though stimulation sites differed, our results were similar in the observed significant reduction in emotional dysregulation, measured using the DERS.33 However, their study observed substantial changes in the DERS for both the active and sham groups, which suggests a potential placebo effect related to the sham intervention, particularly regarding emotional instability.33 Schulze and colleagues’ 34 RCT (1 mA during 1 session, anodal right DLPFC, cathodal left deltoid muscle) evaluated cognitive control of negative stimuli using a delayed working memory task and reported no significant amelioration. Our study adds to the evidence that the anodal stimulation of the left DLPFC effectively benefits cognitive functions; further research will have to be done for emotional instability and impulsivity.

Our study and that of Wolkenstein and colleagues30 both showed improvements in cognitive control for patients with BPD after 1 session of tDCS with anodal stimulation of the left DLPFC.30 However, Wolkenstein and colleagues30 found no significant reduction in interference effects during a working memory task. This may be owing to differences in study settings, such as their placement of the reference electrode on the contralateral mastoid bone instead of the right DLPFC and lower stimulation intensity.30 The single session and lower intensity (1 mA) in the study by Schulze and colleagues34 also failed to show significant cognitive control improvements. These varying results emphasize the importance of stimulation frequency and intensity.

Although our study’s reliance on a baseline comparison without a control group may limit the robustness of its causal inferences, other research supports our findings, particularly regarding the role of the left DLPFC in cognitive control and executive functions.62 Neuroimaging BPD studies also confirm the hypoactivation of prefrontal regions related to emotional and executive disturbance, reinforcing the relevance of our results.63,64 Patients with BPD demonstrated improved cognitive flexibility and inhibition of interference after the tDCS, which could be linked to increased activation of the left DLPFC and bilateral DLPFC stimulation.24 In many psychiatric conditions, such as depression, there can be abnormal levels of activation in the frontal areas of the brain, which affects meso–cortical–limbic dopamine transmission.65 Bifrontal tDCS enhances dopamine release in the ventral striatum, which is linked to improved cognitive functions.65 Bifrontal tDCS could also help mitigate stress-induced cognitive impairment in psychiatric conditions by modulating the activity of the hypothalamo–pituitary–adrenal axis.66 Stimulation may help decrease stress-induced cortisol release, improving symptoms such as decision-making and stress resilience.66

Our pilot study used a dual-intervention approach to target executive functions by incorporating the Lumosity application. Cognitive games have been shown to enhance critical aspects of cognitive functioning, such as working memory, processing speed, visual learning, memory, executive control, and cognitive flexibility; these skills are essential for tasks like planning, decision-making, and adapting to new situations in daily life.36 This evidence supports the rationale behind combining tDCS with online cognitive training. On the one hand, this approach standardizes stimulation, increases task engagement through gamification, and targets improvements in executive functions and memory.28 However, whether the combination is superior to tDCS or cognitive training alone remains unclear. Data from RCTs have shown the specific effects of tDCS on these outcomes, but studies showing a significant effect of Lumosity on executive functions are lacking. Only 1 study has shown significant improvements in the Lumosity Performance Index and attention-switching tasks among healthy people who used the app, compared with a control group.67

We observed an improvement in BPD symptoms after the intervention, which may be attributed to the indirect effects of better cognitive functions. The functional impairment of the left DLPFC in BPD disrupts impulse control, decision-making, and emotion regulation, which may explain the cognitive and emotional challenges seen in individuals with this disorder. Therefore, the enhancement of cognitive abilities with tDCS and online cognitive training could play a key role in helping emotional dysregulation often associated with BPD.17,18,22 Our study also included psychoeducation sessions before the first evaluation, which may have partially influenced BPD symptoms by improving patients’ understanding of their disorder. Some pilot studies suggest a medium-term effect of such interventions.68

We evaluated functionality, our secondary clinical target, using the WHODAS and FAB. The FAB seems more sensitive to BPD-specific dysfunctions, such as BPD-related intra- and interpersonal issues.56 Conversely, the WHODAS provides a broader measure of general dysfunction, emphasizing overall physical and mental well-being. We found mild improvements with the WHODAS, suggesting some gains in general daily functioning. Changes in the FAB were limited, with only the internal component demonstrating notable improvements. This suggests that the treatment combination has some potential to target BPD-specific cognitive and internal emotional processes. Nevertheless, the clinical importance of these changes remains uncertain, necessitating further research to assess their effect on real-world functioning.

Limitations

The potential of these findings to open new therapeutic avenues is limited by several constraints. First, the absence of a sham tDCS or control group limits conclusions about intervention efficacy; future studies should include controls to distinguish training effects from genuine cognitive improvements. It also must be considered that the Lumosity app is primarily designed for the general population and is not standardized for people with severe mental disorders, including BPD. The follow-up assessment was also limited to the short term, as this study was part of a larger research project. We recognize that measuring functionality at more than just 2 time points over 2 weeks is necessary to thoroughly assess symptomatology and psychosocial functioning. Moreover, our reliance on self-report questionnaires to gauge emotional regulation, BPD symptoms, and general functioning may only partially reflect the outcomes, as these measures are subjective. Future studies should include qualitative and semi-structured interviews to provide a more comprehensive evaluation of the outcomes. There may be a practice effect when taking neuropsychological tests for cognitive functions 2 weeks apart, which could introduce bias and skew the results. Psychoeducation sessions may have introduced a symptomatologic bias for the participants who attended; to reduce this bias, we provided access to the same information to nonattendees and time to answer any further questions. On a technical level, the electrode dimensions were not precise enough to target specific regions, which may have led to the stimulation of areas outside the DLPFC. Finally, the placement was determined by external measurements; incorporating neuroimaging could improve the accuracy in targeting specific regions with high-definition tDCS, which could lead to focal neurostructural changes and longer-lasting neuromodulation effects.69

Conclusion

In this pilot study, we found that a combination of online cognitive training and tDCS targeting the left DLPFC significantly improved executive functions, emotional dysregulation, BPD symptoms, and overall functioning. These findings emphasize the potential of the left DLPFC as a neuromodulation site and tDCS as an effective intervention for BPD. Future research should utilize more rigorous methodologies, such as RCTs, and concentrate on optimizing stimulation parameters to maximize prospective therapeutic outcomes.