Two Decades of Data Show Deep Brain Stimulation Can Deliver Lasting Relief for Severe OCD

For patients whose obsessive-compulsive disorder has resisted every medication and every course of therapy, deep brain stimulation has long offered a tantalizing but imperfect promise. Now, one of the longest-running analyses of the technique ever assembled suggests that promise is real and durable: nearly three-quarters of patients followed for up to 21 years after surgery met the accepted threshold for clinical response, and once improvement took hold, it rarely slipped away. The findings, published in Frontiers in Psychiatry by a team at University Hospital Cologne, offer the most extended window yet into how the brain responds when electrodes are implanted in one of the most stubborn circuits in neuropsychiatry.

The study followed 30 consecutive patients, 17 of them women, who underwent deep brain stimulation targeting the nucleus accumbens and the anterior limb of the internal capsule, a combined target abbreviated NAcc/ALIC, between 2004 and 2022. All of them had severe, functionally disabling OCD that had failed to respond adequately to established pharmacological and psychotherapeutic treatments, and all had been approved for surgery by a multidisciplinary team. The researchers assessed symptom severity with the Yale-Brown Obsessive Compulsive Scale, or Y-BOCS, at three time points: before surgery, one year after implantation, and at a final follow-up at least three years after the operation, which in some cases stretched to 21 years.

The technical logic of the procedure is rooted in the cortico-striato-thalamo-cortical circuits, the looping networks that connect the prefrontal cortex to deep gray-matter structures and are thought to be pathologically overactive in OCD. By placing electrodes so that one contact sits within the nucleus accumbens, one at the transition to the internal capsule, and the two most dorsal contacts within the capsule itself, surgeons can modulate activity across this circuit with millimeter-scale precision. In this cohort, leads were implanted bilaterally in 25 patients and unilaterally on the right side in five, using Medtronic model 3387 or SenSight leads with 1.5-millimeter contact spacing. Standard stereotactic coordinates were individualized to each patient’s anatomy, with trajectories planned to avoid blood vessels, cortical sulci, and the ventricular system.

Stimulation began within days of implantation and was titrated over the first two weeks to therapeutic amplitudes, typically between 1.5 and 3.0 volts or milliamperes. Parameters were adjusted every four to six weeks during the first three months, then roughly every three months until the first postoperative year, after which follow-up intervals were individualized. The response criterion was a reduction of at least 35 percent in Y-BOCS score from baseline, in line with international consensus definitions, although the retrospective design meant that the companion Clinical Global Impression ratings required for the full consensus categories were not systematically available.

The statistical picture was unambiguous. Repeated-measures analysis of variance showed that Y-BOCS scores fell significantly across all three time points, with a large effect size, and every Bonferroni-adjusted pairwise comparison, from baseline to one year, from one year to final follow-up, and from baseline to final follow-up, reached significance. Mean scores dropped from roughly 30 at baseline to about 20 at one year and about 14 at the latest assessment. Expressed as percentages, the average reduction was 32.4 percent at one year and 51.6 percent at final follow-up, meaning that on average patients continued to improve well beyond the first anniversary of their surgery.

At the final follow-up, 22 of 30 patients, or 73.3 percent, met the response criterion. That figure exceeds the pooled response rates of 66 percent and 57.9 percent reported in earlier meta-analyses, although the authors caution that excluding four operated patients who lacked complete follow-up data may have inflated the proportion. Response was seen across all baseline severity categories: 80 percent of patients whose symptoms had been rated extreme at baseline, 66.7 percent of those rated severe, and two of the three patients with moderate total scores, whose eligibility rested on severe, disabling obsessions despite minimal compulsions.

Perhaps the most clinically consequential finding concerns timing. Eight patients, 26.7 percent of the cohort, were delayed responders: they had not crossed the 35 percent threshold at one year but did so later, sometimes years later. One patient with extreme baseline symptoms showed only a 12.5 percent reduction at one year but a 95 percent reduction at 21 years. Conversely, among the 15 patients who responded by one year, 14 retained their response and only one lost it. The authors suggest that delayed improvement could reflect gradual neural adaptation, long-term optimization of stimulation settings, changes in concomitant treatment, psychosocial shifts, or natural fluctuation in OCD severity, and the retrospective data cannot distinguish among these explanations. But the practical implication is clear: judging a patient a non-responder at the one-year mark may be premature.

The study is candid about its limits. It is retrospective, uncontrolled, and single-center, which precludes causal conclusions about DBS itself or about the temporal course of improvement. Follow-up duration varied enormously, from 3 to 21 years, and was treated as a categorical assessment rather than a continuous variable, so the analysis cannot estimate a yearly rate of change. Medication and psychotherapy histories were not systematically recorded, adverse events were not captured with a standardized framework, and the inclusion period spanned 18 years, during which surgical techniques, hardware, and programming practices inevitably evolved. The researchers also note that some patients experienced transient symptom worsening during transitions between voltage-controlled and current-controlled devices, particularly around generator replacements, observations documented mainly in narrative clinical notes rather than standardized assessments.

Those caveats notwithstanding, the trajectory data add an important dimension to a field that has often been judged on one-year outcomes. Long-term cohorts from Amsterdam and other centers have previously shown sustained benefit, but a dataset extending over two decades, with individual stimulation parameters and Y-BOCS scores documented for every patient, is rare. The finding that improvement continued to accrue between the first year and the final assessment, and that the cumulative reduction exceeded either interval-specific reduction, suggests that the therapeutic effect of NAcc/ALIC stimulation may mature slowly, in a way that shorter trials cannot capture.

For the roughly one in four people with OCD who do not respond to conventional treatment, the study offers a measured but meaningful message. Deep brain stimulation remains a last-resort intervention reserved for highly selected patients treated in specialized multidisciplinary centers, and no one should read these results as a guarantee. But they do provide descriptive evidence that clinically meaningful improvement can persist, and in some cases deepen, over very long follow-up, and that patience with the therapy, and with the patients, may itself be part of the treatment. The authors call for prospective studies with standardized follow-up intervals, systematic recording of concomitant treatments and adverse events, and analyses that incorporate exact follow-up duration to confirm what this two-decade cohort so strongly hints at.

Subject of Research: Long-term outcomes of deep brain stimulation for treatment-resistant obsessive-compulsive disorder

Article Title: Long-term outcomes of nucleus accumbens/anterior limb of the internal capsule deep brain stimulation in treatment-resistant obsessive-compulsive disorder

Article References: Cid-Rodríguez, F. X., Schmahl, R., Márquez-Franco, R., Heiden, P., Soriano-Ursúa, M. A., Wagemann, W., Huys, D., Baldermann, J. C., Kuhn, J., Visser-Vandewalle, V., & Andrade, P. (2026). Long-term outcomes of nucleus accumbens/anterior limb of the internal capsule deep brain stimulation in treatment-resistant obsessive-compulsive disorder. Frontiers in Psychiatry, 17, Article 1905823. https://doi.org/10.3389/fpsyt.2026.1905823

Image Credits: AI Generated

DOI: 10.3389/fpsyt.2026.1905823

Keywords: deep brain stimulation, obsessive-compulsive disorder, nucleus accumbens, anterior limb of the internal capsule, Y-BOCS, neuromodulation, treatment-resistant OCD, neurosurgery, psychiatry, long-term follow-up, cortico-striato-thalamo-cortical circuits, Frontiers in Psychiatry