Ear Nerve Stimulation Eases Negative Feelings in Women with Borderline Personality Disorder

A single session of electrical stimulation applied to a small patch of cartilage in the outer ear was associated with lower self-reported negative affect in women with borderline personality disorder, according to a randomized, single-blind, sham-controlled trial conducted in Sweden. The technique, known as transcutaneous auricular vagus nerve stimulation, or taVNS, delivers gentle current through the skin of the ear to activate afferent fibers of the vagus nerve, a major conduit by which the body signals the brain. In the study, published in the journal Borderline Personality Disorder and Emotion Dysregulation, active stimulation of the left cymba conchae was compared against sham stimulation of the earlobe, a site not considered to provide effective vagal activation, while participants watched a series of emotionally charged video clips.

Borderline personality disorder is characterized by marked affective instability, chronic impulsivity, and recurrent self-destructive behaviors. Clinicians commonly describe an emotional vulnerability built on three interrelated features: heightened emotional sensitivity, exaggerated reactivity to emotional triggers, and a delayed return to baseline after arousal. Although psychotherapies such as Dialectical Behaviour Therapy are effective, access remains limited, and there is no clear first-line pharmacological option. Many patients continue to experience severe emotional distress despite treatment, which motivated the research team to probe whether acute modulation of vagal-afferent signaling could influence emotional responding in real time.

The investigators framed their hypothesis using an interoceptive-threshold framework. Interoception refers to the sensing, representation, and interpretation of signals originating within the body, and contemporary models of emotion hold that emotional experience depends not only on the magnitude of physiological change but on how bodily signals are detected, weighted, and integrated with context and expectations. Prior work in borderline personality disorder has reported reduced heart rate variability, elevated resting arousal in some studies, and differences in cortical representation of bodily signals, supported by heartbeat-evoked potential findings and neuroimaging implicating the insula, amygdala, anterior cingulate cortex, and prefrontal regions. Yet the literature remains heterogeneous, and physiological hyperreactivity has not been consistently demonstrated, raising the possibility that the core disturbance lies less in the magnitude of peripheral autonomic responses and more in the threshold at which those responses become affectively salient.

TaVNS offers a non-invasive way to test this idea. Stimulation of the auricular branch of the vagus nerve reaches brainstem nuclei, including the nucleus tractus solitarius, with downstream projections to central autonomic and interoceptive-affective networks involving the insula, cingulate, prefrontal, and limbic regions. Previous experimental studies, mostly in healthy participants or non-BPD samples, suggested possible effects on emotion recognition, mood, and stress-related responding, although meta-analytic evidence indicates that group-level effects of active versus sham taVNS on vagally mediated heart rate variability are inconsistent. According to the researchers, taVNS had not previously been investigated in patients with borderline personality disorder.

Thirty-four female psychiatric outpatients aged 18 to 50 with a current DSM-5 diagnosis of borderline personality disorder took part. Diagnosis was confirmed with the Structured Clinical Interview for DSM-5 Personality Disorders, and exclusionary psychiatric conditions were screened using the Mini International Neuropsychiatric Interview. Participants were randomized 1:1 to active taVNS or sham stimulation using sealed, opaque, sequentially numbered envelopes; participants were blinded to allocation, whereas investigators were not. Stimulation was delivered with a tVNS device at 25 Hz in alternating cycles of 28 seconds on and 32 seconds off, with intensity individually titrated until a clearly noticeable but not painful tingling sensation was reported. After the tolerated intensity was reached, stimulation continued for four minutes before the emotional task began, and then continuously throughout the roughly 45-minute experimental protocol.

The mood-induction paradigm consisted of one neutral and three negative four-minute film clips depicting domestic violence, a funeral scene, and sexual assault, presented in randomized order after the neutral video. Self-reported negative affect was measured repeatedly with the Negative Affect subscale of the Positive and Negative Affect Schedule at baseline, after a pre-task stimulation period, after the neutral video, after each negative video, and after each recovery period. The primary analysis modeled these ratings jointly across four collapsed phases, adjusting for baseline negative affect and stimulation intensity. Continuous electrocardiography and electrodermal activity were recorded throughout, allowing exploratory analysis of heart rate, the natural log of the root mean square of successive differences, phasic electrodermal response frequency, and tonic skin conductance level.

The primary result showed that active taVNS was associated with lower PANAS Negative Affect across the paradigm compared with sham stimulation, with a Treatment Group effect of F(1, 36.44) = 5.74, p = 0.022. Adjusted marginal means were 17.25 in the active group versus 21.22 in the sham group. Negative affect varied significantly across phases, rising from a low of 16.52 before induction to a peak of 23.70 after the negative video blocks and falling to 18.94 during recovery, confirming successful mood induction. Crucially, the Treatment Group by Phase interaction was not significant, meaning the benefit was a global downward shift across the whole paradigm rather than an effect confined to one phase. Sensitivity analyses using a compound symmetry covariance structure and adding recent nicotine use as a covariate yielded the same pattern. Among the registered outcome-specific analyses, the pre-induction comparison also favored active stimulation, and the recovery analysis showed a larger decrease from post-induction to recovery in the active group, whereas the registered post-induction analysis alone did not reach statistical significance.

The subjective improvement, however, was not mirrored in the body. Exploratory physiological analyses found no treatment-related effects on heart rate, lnRMSSD, or phasic EDA response frequency. Tonic skin conductance level showed no main effect of treatment group, but a significant Treatment Group by Epoch interaction, driven by higher tonic skin conductance during negative-video exposure in the active group, a pattern the authors interpret cautiously as possibly reflecting increased physiological engagement with the emotional challenge rather than peripheral calming. The researchers suggest this dissociation between subjective and physiological indices is compatible with predictive-coding accounts of interoception, in which emotional experience depends on the reliability weight assigned to ascending bodily signals rather than on autonomic output alone, though the study cannot test that mechanism directly.

The authors are careful about what the findings do not show. The single-session design precludes conclusions about durability, clinical efficacy, or longer-term symptom change, and the results should not be interpreted as evidence of lasting clinical benefit. The sample was modest, restricted to women, and lacked healthy or clinical control groups, limiting conclusions about BPD specificity. No direct measure of interoception was included, the video clips were not drawn from a standardized stimulus database, participants were a clinically complex and largely medicated outpatient sample, blinding success was not formally assessed, and menstrual cycle phase was not controlled. No serious adverse events occurred, and no participant discontinued because of stimulation-related discomfort. The trial was retrospectively registered at ClinicalTrials.gov on 7 June 2023.

Even with those caveats, the study is, to the authors’ knowledge, the first to examine taVNS in patients with borderline personality disorder, and it supports further investigation of the technique as a mechanistic probe of acute affective responding. Future work, the team writes, should incorporate direct measures of interoception such as behavioral accuracy tasks, the Multidimensional Assessment of Interoceptive Awareness questionnaire, and heartbeat-evoked potentials, along with mediation analyses testing whether interoceptive changes explain the emotional effects. Repeated-session studies are needed to determine whether the acute reduction in subjective negative affect can be extended over time and translated into clinically meaningful improvements, and studies including both healthy and clinical comparison groups will clarify whether the observed effects are specific to borderline personality disorder or reflect broader transdiagnostic mechanisms.

Subject of Research: A randomized sham-controlled trial of transcutaneous auricular vagus nerve stimulation and acute emotional responding in women with borderline personality disorder

Article Title: Transcutaneous auricular vagus nerve stimulation and emotional responding in borderline personality disorder: a randomized single-blind, sham-controlled study

Article References: Guerriero, G., Ruocco, A. C., Carlsen, H. K., Daros, A. R., Kosara, D., & Steingrimsson, S. (2026). Transcutaneous auricular vagus nerve stimulation and emotional responding in borderline personality disorder: a randomized single-blind, sham-controlled study. Borderline Personality Disorder and Emotion Dysregulation, 13(1), Article 24. https://doi.org/10.1186/s40479-026-00367-x

Image Credits: AI Generated

DOI: 10.1186/s40479-026-00367-x

Keywords: borderline personality disorder, taVNS, vagus nerve stimulation, emotion regulation, interoception, mood induction, PANAS, randomized controlled trial, autonomic nervous system, heart rate variability, electrodermal activity, neuromodulation