Summary
A medRxiv preprint reports that chronic brain recordings from 24 people receiving deep brain stimulation for treatment-resistant OCD distinguished likely responders before stimulation and tracked early treatment engagement. The findings describe a potential biomarker framework, but the study is still a preprint.
A preprint study reports that long-term recordings from implanted brain electrodes could help forecast who will respond to deep brain stimulation (DBS) for treatment-resistant obsessive-compulsive disorder (OCD), while also indicating whether treatment is beginning to engage the relevant neural circuits.
The researchers analysed nearly 120,000 patient-hours of chronic intracranial recordings from 24 people undergoing DBS targeting the ventral capsule/ventral striatum. Before stimulation began, circadian patterns in neural activity distinguished participants who later became responders from those who did not. During the first two weeks after therapy started, changes in the predictability of neural signals also separated the eventual response groups and were associated with the amount of symptom improvement measured later.
The work was posted on medRxiv on September 21, 2026. It is a preprint, meaning the manuscript has been publicly released before formal peer review.
A biomarker across three stages of treatment
DBS uses surgically implanted electrodes to deliver electrical stimulation to specific brain regions. In severe, treatment-resistant OCD, stimulation is used as a neuromodulation therapy, with clinical improvement often developing over weeks or months while clinicians adjust treatment.
That delay can make it difficult to tell early whether a treatment setting is beginning to have the intended effect. The study examined whether neural signals recorded continuously or repeatedly over an extended period could provide information at three points in the treatment course.
First, measurements taken before stimulation began contained circadian neural features that distinguished eventual responders from non-responders. Second, within the first two weeks of stimulation, reductions in neural predictability differentiated the groups and correlated with the magnitude of symptom improvement that appeared later. Neural predictability describes how regular or forecastable patterns in recorded brain activity are over time.
Finally, during chronic treatment, neural predictability tracked the participants' clinical response state. Taken together, the results suggest that the same broad signal property may provide an early forecast, an indication of therapeutic engagement after activation and a way to follow response over time.
Why the finding matters
OCD treatment response is usually evaluated through clinical symptoms, which may change gradually. A neural measure that shifts earlier could give clinicians an additional physiological signal while treatment is being optimised. It could also help researchers study why DBS benefits some patients more than others.
The authors describe the result as a framework for neural biomarkers spanning the course of DBS therapy. The study therefore links three different questions—who may respond, whether stimulation is beginning to engage the brain and whether response is being maintained—to recordings obtained from the implanted system.
The study involved 24 patients, making it an early clinical investigation rather than a validated test for routine treatment decisions. Its evidence comes from people already receiving implanted DBS for treatment-resistant OCD, so applying the findings to other patients, other stimulation targets or non-invasive treatments would require further study. The preprint status also means the findings have not yet undergone formal peer review.
The researchers state that deidentified participant-level neural and clinical datasets supporting the findings are available upon request through the Data Archive for the BRAIN Initiative. The work received ethical approval from Baylor College of Medicine and the University of Utah, and was funded by the NIH Common Fund and other listed organisations.