Adaptive (closed-loop) deep brain stimulation

Adaptive DBS senses a neural biomarker in real time and adjusts stimulation (amplitude, frequency, or on/off) accordingly, instead of delivering the fixed continuous output of conventional open-loop DBS. In Parkinson's disease the canonical control signal is subthalamic beta-band (roughly 13 to 30 Hz) power, which co-varies with bradykinesia and rigidity; the controller reduces stimulation when beta is low and increases it when beta rises, aiming to deliver only the stimulation the moment requires.

Motivations are fewer stimulation-induced side effects, less habituation, and longer battery life. The engineering is hard: sensing microvolt neural signals during ongoing millivolt-scale stimulation is contaminated by stimulation artifact; the biomarker must be stable across activities, medication states, and years; and the controller (simple threshold, proportional, or richer) must not oscillate. Whether well-designed aDBS clinically outperforms carefully titrated open-loop DBS is still being established in trials.