A Time-Domain Charge-Balancing Method for Neuromodulators

Closed-loop neuromodulation systems, featuring neurostimulation adapted to simultaneously recorded biomarkers, have shown tremendous potential in the treatment of neural disorders. When periodic stimulation is applied, considerations regarding patient and electrode safety are of utmost importance: Unbalanced stimulation can lead to residual charge, which can accumulate and lead to irreversible reactions, associated with electrode damage and cytotoxic byproducts. Current-controlled stimulation has improved control over the introduced charge, and biphasic stimulation allows for a first-order cancellation. However, unavoidable mismatches in the anodic and cathodic current branch will nevertheless result in imperfect symmetry and hence in residual charge on the electrode post-stimulation. Active or passive charge balancing methods have been explored; while passive discharge has large and uncontrollable initial current density, active cancellation requires significant hardware effort. In this paper, we propose to explore the availability of a neural recorder and employ an iterative method of altering the duration of the compensation pulse in biphasic stimulation in order to achieve excess charge minimization in the time domain with fine (μs) granularity. No excess hardware is required, as the neural recorder is utilized as a sensing path for charge imbalance via the triggered artifact. By relying only on the sign of the initial threshold crossing of the artifact voltage response, a robust implementation with minimal signal processing is achieved. The method requires no prior knowledge or model of the mismatch or the electrode-tissue interface and is verified via wet-electrode measurements.