SEARCH

SEARCH BY CITATION

Deep brain stimulation (DBS) has been an established treatment for Parkinson disease (PD) for the past decade, with strong evidence for efficacy.[1-3] Does it need to be improved? Yes. Although the technology for deep brain stimulators arose from cardiac pacemakers, brain stimulators for movement disorders are much less sophisticated. The current technique is open loop stimulation, in which the device delivers therapy at constant stimulation settings. The clinician or patient can adjust the stimulation parameters, but this is often a labor-intensive, trial-and-error process, in which each new setting is determined based on how the symptoms responded to the last setting. Furthermore, the open loop design results in high current drain, requiring battery units that are physically relatively large and require surgical replacement more frequently than desirable. Finally, not all patients who appear initially to be good candidates for DBS derive the expected benefit despite appropriately located leads. Incorporation of a feedback control system could address these problems. Until recently, there was little understanding of what specific brain signal might best represent the circuit dysfunction in PD, precluding the development of a “smart” feedback-controlled system in this disorder.

The article by Little et al in this issue describes an approach to feedback control in DBS for PD.[4] Starting in the 1990s, it was found that PD is a disorder of excessive brain synchronization throughout the basal ganglia nuclei, and more recently this has been shown for motor cortex as well. In prior publications, Professor Brown's group has hypothesized that the local field potential (LFP) recorded from the therapeutic subthalamic nucleus (STN) DBS electrode can be used as a measure of pathological synchronization in the parkinsonian brain. LFPs represent the summed pre- and postsynaptic activity of the neuronal elements close to the recording electrode, with synchronized neuronal activity as the largest contributor to the signal. Previously, the authors have shown that synchronized neuronal activity in the beta band is easily recorded in the STN LFP, and that beta band oscillations are reduced when parkinsonian symptoms are improved by oral L-dopa or by acute therapeutic stimulation.[5, 6] Now they have taken these findings to the next level by showing that STN beta band oscillations can be used to control the amplitude of therapeutic stimulation, and that acutely, this feedback-controlled stimulation is both more effective and less power-consuming than standard open loop stimulation. The increase in effectiveness provided by “closing the loop” is clinically meaningful; in this study, it produced a further 27% improvement in contralateral limb Unified Parkinson Disease Rating Scale scores. The trigger mechanism that initiated stimulation pulses in their study was tuned to detect changes in beta power that occur on a short time scale (<1 second), raising the possibility that fluctuations in beta power, rather than its average amplitude, are of greatest importance in PD pathophysiology.

As other approaches to feedback-controlled stimulation are tested, will basal ganglia LFP beta power prove to be the most effective control signal? In the work described here, adaptive control was accomplished not through a fully internalized DBS system, but rather via externalized electrodes whose signal was processed by hardware and software located outside the patient. The approach relies on delivery of therapeutic stimulation and signal detection in the same pea-sized brain structure. The STN LFP control signal is small (10–30 μV) and the stimulation artifact is large (1–3 V), such that sophisticated electronics are needed for artifact suppression. Miniaturization of such a system for full implantability might not be trivial. Furthermore, beta oscillations are a normal feature of the motor system, and the amplitude of beta oscillations may be a less-sensitive biomarker of the parkinsonian state than the tendency of unit activity to occur at a consistent phase of the beta rhythm.[7, 8] Motor cortex may be an alternative site for signal detection that mitigates these issues. In a nonhuman primate model of PD, spiking activity in primary motor cortex, recorded from penetrating microelectrodes, is an efficient driver of therapeutic stimulation.[9] In humans, it is now possible to detect pathological synchronization of population spiking using a cortical electrode that does not penetrate brain tissue. This measure of synchronization is also minimally disrupted by artifact from concurrent basal ganglia stimulation.[8] New generations of brain stimulation devices that are bidirectional (they sense and record data as well as therapeutically stimulate) will soon be employed to test these different algorithms in PD.

The paradigm presented by Little et al could make an already successful therapy even better. It is especially intriguing that even over the short stimulation times in this study (minutes), the duration of “stimulus on” epochs in the adaptive stimulation paradigm decreased. In the future, devices that “program themselves” will help simplify the therapy and increase its penetrance in the PD patient population, currently limited by the high complexity of the therapy for clinicians. Newer applications of DBS, such as in psychiatric disorders, may benefit even more from closed loop approaches, as symptoms in these disorders are more difficult to assess by clinicians and may fluctuate greatly in time.

Potential Conflicts of Interest

  1. Top of page
  2. Potential Conflicts of Interest
  3. References

P.A.S.: consultancy, Boston Scientific; grants/grants pending, MRI Interventions; honoraria, Medtronic; patents, detection of a cortical biomarker in movement disorders using a nonpenetrating electrode (UCSF). J.L.O.: grants/grants pending, Medtronic; patents, detection of a cortical biomarker in movement disorders using a nonpenetrating electrode (UCSF).

  • Philip A. Starr, MD, PhD

  • Department of Neurosurgery

  • University of California, San Francisco

  • San Francisco, CA

  • Jill L. Ostrem, MD

  • Department of Neurology

  • University of California, San Francisco

  • San Francisco, CA

References

  1. Top of page
  2. Potential Conflicts of Interest
  3. References