Amygdala neurofeedback as an adjunct to residential dialectical behavior therapy for borderline personality disorder: a pilot randomized controlled trial

Published October 16, 2026. Participants learned to reduce amygdala activity during training, showing that the procedure was feasible. The small trial did not find added improvement in daily-life instability or self-reported symptoms beyond intensive residential DBT.

Hofmanová M, Santangelo PS, Kleindienst N, Schmahl C, Paret C.
BMC Psychiatry · Publication date: 2026-10-16
Added to site: 2026-10-18
DOI: 10.1186/s12888-026-08767-7

Published October 16, 2026. Added to Neurofeedback.com October 18, 2026.

Modalities

fMRI neurofeedback, Real-time neurofeedback, DBT

Clinical focus

Borderline personality disorder, Affective instability, Adjunctive treatment

Population

Adults with borderline personality disorder and residual symptoms during residential dialectical behavior therapy

Sample

36 randomized: 23 to DBT plus amygdala neurofeedback and 13 to DBT alone; 24 completed post-treatment assessment, 23 completed three-month follow-up and 19 completed six-month follow-up

Intervention / method

The neurofeedback group received three real-time fMRI amygdala-neurofeedback sessions, each containing two training runs, alongside a 12-week residential DBT program. Ecological momentary assessment of affective instability was the primary outcome.

Comparator

Residential DBT plus amygdala neurofeedback versus residential DBT alone, without sham neurofeedback

Clinician summary

EMA negative affect, inner tension and affective instability improved over time without consistent neurofeedback superiority. Self-reports likewise showed no incremental benefit, and BSL-23 scores rose after discharge in both groups. Neural changes were transient and did not consistently align with clinical change.

How to interpret the findings

Small under-recruited pilot, imbalanced 23:13 allocation, attrition through six months, no sham-neurofeedback control, intensive concurrent DBT and baseline group imbalances. Neural target engagement does not establish added clinical benefit, and the study was not powered as a definitive efficacy trial.

Equipment and disclosures

A 3T Siemens MAGNETOM Prisma Fit scanner with 64-channel head coil was used; online processing used SPM12 and custom MATLAB scripts. This is source-reported equipment, with no BrainMaster attribution.

Funded by the AE Foundation. Christian Schmahl reported consultancy honoraria from Boehringer Ingelheim unrelated to the article; the other authors reported no conflicts. No indexed correction, retraction or concern notice appeared in the retained PubMed or Crossref records; this is a bounded check. AI-assisted review.

Related evidence

Compare findings from the same clinical area, including mixed results.