Personalized striatal targets for deep brain stimulation in obsessive-compulsive disorder

Background Psychiatric conditions currently treated with deep brain stimulation (DBS), such as obsessive-compulsive disorder (OCD), are heterogeneous diseases with different symptomatic dimensions, indicating that fixed neuroanatomical DBS targets for all OCD cases may not be efficacious. Objective/hypothesis We tested whether the optimal DBS target for OCD is fixed for all patients or whether it is individualized and related to each patient's symptomatic content. Further, we explored if the optimal target can be predicted by combining functional neuroimaging and structural connectivity. Methods In a prospective, randomized, double-blinded study in 7 OCD patients, symptomatic content was characterized pre-operatively by clinical interview and OCD symptom-provocation during functional MRI. DBS electrode implantation followed a trajectory placing 4 contacts along a striatal axis (nucleus accumbens to caudate). Patients underwent three-month stimulation periods for each contact (and sham), followed by clinical evaluation. Probabilistic tractography, applied to diffusion-weighted images acquired pre-operatively, was used to study the overlap between projections from the prefrontal areas activated during symptom provocation and the volume of activated tissue of each electrode contact. Results Six patients were classified responders, with median symptomatic reduction of 50% achieved from each patient's best contact. This was located at the caudate in 4 cases and at the accumbens in 2. Critically, the anatomical locus of the best contact (accumbens or caudate) was related to an index derived by combining functional MRI responses to prevailing symptom provocation and prefronto-cortico-striatal projections defined by probabilistic tractography. Conclusion Our results therefore represent a step towards personalized, content-specific DBS targets for OCD.