Self-Tuning Deep Brain Stimulation Controller for Suppression of Beta Oscillations: Analytical Derivation and Numerical Validation
Open Access
- 30 June 2020
- journal article
- research article
- Published by Frontiers Media SA in Frontiers in Neuroscience
- Vol. 14, 639
Abstract
Closed-loop control strategies for deep brain stimulation (DBS) in Parkinson's disease offer the potential to provide more effective control of patient symptoms and fewer side effects than continuous stimulation, while reducing battery consumption. Most of the closed-loop methods proposed and tested to-date rely on controller parameters, such as controller gains, that remain constant over time. While the controller may operate effectively close to the operating point for which it is set, providing benefits when compared to conventional open-loop DBS, it may perform sub-optimally if the operating conditions evolve. Such changes may result from, for example, diurnal variation in symptoms, disease progression or changes in the properties of the electrode-tissue interface. In contrast, an adaptive or “self-tuning” control mechanism has the potential to accommodate slowly varying changes in system properties over a period of days, months, or years. Such an adaptive mechanism would automatically adjust the controller parameters to maintain the desired performance while limiting side effects, despite changes in the system operating point. In this paper, two neural modeling approaches are utilized to derive and test an adaptive control scheme for closed-loop DBS, whereby the gain of a feedback controller is continuously adjusted to sustain suppression of pathological beta-band oscillatory activity at a desired target level. First, the controller is derived based on a simplified firing-rate model of the reciprocally connected subthalamic nucleus (STN) and globus pallidus (GPe). Its efficacy is shown both when pathological oscillations are generated endogenously within the STN-GPe network and when they arise in response to exogenous cortical STN inputs. To account for more realistic biological features, the control scheme is then tested in a physiologically detailed model of the cortical basal ganglia network, comprised of individual conductance-based spiking neurons, and simulates the coupled DBS electric field and STN local field potential. Compared to proportional feedback methods without gain adaptation, the proposed adaptive controller was able to suppress beta-band oscillations with less power consumption, even as the properties of the controlled system evolve over time due to alterations in the target for beta suppression, beta fluctuations and variations in the electrode impedance.Keywords
Funding Information
- Irish Research Council
- Campus France
- H2020 European Research Council
This publication has 77 references indexed in Scilit:
- Effective Deep Brain Stimulation Suppresses Low-Frequency Network Oscillations in the Basal Ganglia by Regularizing Neural Firing PatternsThe Journal of Neuroscience, 2012
- Subthalamo‐pallidal interactions underlying parkinsonian neuronal oscillations in the primate basal gangliaEuropean Journal of Neuroscience, 2011
- Somatic Membrane Potential and Kv1 Channels Control Spike Repolarization in Cortical Axon Collaterals and Presynaptic BoutonsThe Journal of Neuroscience, 2011
- Deep brain stimulation can suppress pathological synchronisation in parkinsonian patientsJournal of Neurology, Neurosurgery & Psychiatry, 2010
- Conditions for the Generation of Beta Oscillations in the Subthalamic Nucleus–Globus Pallidus NetworkThe Journal of Neuroscience, 2010
- Modeling shifts in the rate and pattern of subthalamopallidal network activity during deep brain stimulationJournal of Computational Neuroscience, 2010
- The Wilson–Cowan model, 36 years laterBiological Cybernetics, 2009
- PyNN: a common interface for neuronal network simulatorsFrontiers in Neuroinformatics, 2008
- Feedback suppression of neural synchrony by vanishing stimulationPhysical Review E, 2007
- An Efficient Method for Computing Synaptic Conductances Based on a Kinetic Model of Receptor BindingNeural Computation, 1994