Blinding and the placebo problem in psychiatric DBS

Psychiatric neuromodulation trials are unusually vulnerable to expectation effects. Patients are severely ill and highly hopeful, primary outcomes are self-reported and subjective, the surgery itself is a powerful therapeutic ritual, and active stimulation can produce noticeable acute sensations or evoked affect that unblind the assignment.

Several pivotal randomized sham-controlled trials of DBS for depression — targeting the subcallosal cingulate and the ventral capsule/striatum — failed to separate active stimulation from sham on their primary short-term endpoints, even though open-label benefit later emerged in the same programs. This pattern may reflect large placebo and response variance, an underpowered or too-short blinded phase, heterogeneous patients, or a genuinely slow-building effect. Resolving what truly works requires rigorous designs with adequately long blinded phases, delayed-onset or discontinuation designs, and objective biomarkers that do not depend on self-report.