Common therapeutic mechanisms of pallidal deep brain stimulation for hypo- and hyperkinetic movement disorders

Abstract

Abnormalities in cortico-basal ganglia (CBG) networks can cause a variety of movement disorders ranging from hypokinetic disorders, such as Parkinson's disease (PD), to hyperkinetic conditions, such as Tourette syndrome (TS). Each condition is characterized by distinct patterns of abnormal neural discharge (dysrhythmia) at both the local single-neuron level and the global network level. Despite divergent etiologies, behavioral phenotypes, and neurophysiological profiles, high-frequency deep brain stimulation (HF-DBS) in the basal ganglia has been shown to be effective for both hypo- and hyperkinetic disorders. The aim of this review is to compare and contrast the electrophysiological hallmarks of PD and TS phenotypes in nonhuman primates and discuss why the same treatment (HF-DBS targeted to the globus pallidus internus, GPi-DBS) is capable of ameliorating both symptom profiles. Recent studies have shown that therapeutic GPi-DBS entrains the spiking of neurons located in the vicinity of the stimulating electrode, resulting in strong stimulus-locked modulations in firing probability with minimal changes in the population-scale firing rate. This stimulus effect normalizes/suppresses the pathological firing patterns and dysrhythmia that underlie specific phenotypes in both the PD and TS models. We propose that the elimination of pathological states via stimulus-driven entrainment and suppression, while maintaining thalamocortical network excitability within a normal physiological range, provides a common therapeutic mechanism through which HF-DBS permits information transfer for purposive motor behavior through the CBG while ameliorating conditions with widely different symptom profiles.

Keywords: Parkinson's disease; Tourette syndrome; deep brain stimulation; globus pallidus; nonhuman primate.

Fig. 1.

Representation of the intrinsic pathways of the basal ganglia. The schematic shows the “indirect” and “direct” pathways in the normal condition (black) and compares the relative strength (line thickness) and sign (inhibitory vs. excitatory) of the information flow as predicted by “box-and-arrow” models for hypokinetic/parkinsonism (red) and hyperkinetic/tourettism (green). Note the relative changes coming from the output nucleus of the basal ganglia [globus pallidus internus (GPi)] and the effects on motor cortical excitability. GPe, globus pallidus externus; STN, subthalamic nucleus. B: schematic illustrating how firing patterns, e.g., synchrony and oscillatory changes, are proposed to contribute to pathological behavior in hypo- and hyperkinetic disorders. In the normal GPi (top), neurons encoding physiological movement (white) transiently decrease activity, while surrounding neurons (black) maintain asynchronous firing patterns and normal levels of burstiness. In the parkinsonian condition (middle), there is an increase in low-frequency oscillatory activity (indicated by sinusoids), synchronous spike-to-spike firing, and burstiness, with a reduction in neurons encoding for normal movement. With tourettism (bottom), there is a large increase in the number of neurons simultaneously encoding an extended action-release signal.

Fig. 2.

Examples of pathological activity in the pallidum for hypo- and hyperkinetic disorders. A: raw neuronal traces from the GPe and GPi taken from a parkinsonian nonhuman primate. Note the increased power in the alpha (7–12 Hz) and beta (12–30 Hz) frequency bands (red shading). These spike trains often show pronounced cross-correlated activity in the temporal (bottom left) and spectral (bottom right) domains. Data from McCairn et al. 2011; all experimental procedures were completed with IACUC approval. B: example of myoclonic tic activity in the arm of a nonhuman primate, with associated EMG traces and raw neuronal recording of pallidal activity. Note the propensity for increases in the GPe (bottom left raster) and decreases in the GPi (bottom right raster) (data from McCairn et al. 2013c).

Fig. 3.

Therapeutic effects in nonhuman primates of pallidal deep brain stimulation. A: example of a measure of muscular rigidity (i) acquired through a reactive torque motor on and off GPi-DBS. The trace shows the angular displacement of a manipulandum that moves the arm ±20°, plus a measure of instantaneous torque required to move the arm through the angular displacement. During GPi-DBS there is a significant decrease in the “work” required by the torque motor to move the arm through the angular displacement (ii). This reduction in torque is an indirect measure of muscular rigidity (adapted from McCairn and Turner 2009 with permission). B: GPi-DBS induced reduction in the amplitude of myoclonic tics induced by bicuculline injection to the sensorimotor putamen. Top: raw rectified EMG trace during 8 episodes of 30 s of stimulation (gray bars) from 1 monkey. Bottom: mean amplitude of the tic-associated EMG off (black trace) vs. on (gray trace) stimulation. Note the large reduction in tic amplitude during the on-stimulation condition (adapted from McCairn et al. 2012 with permission).

Fig. 4.

Effects on pallidal neuronal activity following GPi-DBS. A: raw neuronal traces acquired from the pallidum in monkey models of parkinsonism and tourettism. Traces illustrate the different types of response that can be observed in response to GPi-DBS. Perievent histograms aligned to the onset of GPi-DBS for each of the examples shown in A are presented in B–E. B: Parkinson's disease (PD)-GPe unit shown in A that shows a partial reduction in firing rate activity. C: example of almost complete cessation of firing activity observed in the PD-GPi. D: example of Tourette syndrome (TS)-GPe that shows an initial increase in activity that subsides while stimulation is active. E: a TS-GPi neuron that shows an initial inhibition and then increased firing rate that persists beyond the cessation of GPi-DBS. Note that each of the different firing pattern responses to GPi-DBS can be found in each segment of the pallidum in both disease models (data from McCairn et al. 2012; McCairn and Turner 2009).

Fig. 5.

Short-latency entrainment of the pallidum following GPi-DBS. A: examples of interpulse changes in firing rate for individual neurons (color matrix) and population-scale changes in firing rate for each segment of the pallidum in a parkinsonian model. Note the prominent peaks and troughs in the mean averages that indicate stimulus-driven excitation and inhibition. B: multiphasic entrainment of pallidal neurons in a model of tourettism (same format as A); note the distinct peaks and troughs in both GPe and GPi, similar to parkinsonian models, that indicate stimulation-driven modulation of firing activity (data from McCairn et al. 2012; McCairn and Turner 2009).

Fig. 6.

GPi-DBS-mediated suppression of low-frequency oscillations in a parkinsonian model. A: perievent (DBSon) measure of torque (top) aligned to perievent spectrograms of GPe (middle) and GPi (bottom) firing activity. Note the suppression of oscillatory (∼10 Hz) activity in both pallidal segments as GPi-DBS starts and the subsequent reduction in torque required to move the arm (reduced yellow and green intensity) in the torque trace (data from McCairn et al. 2011; McCairn and Turner 2009). PSD, power spectral density. B: PSD plot showing the induction of high-frequency resonance induced by short-term entrainment with the stimulation pulse and a subsequent reduction of low-frequency oscillations. C: GPi-DBS suppresses low-frequency coherence, a key marker of synchrony, between pairs of simultaneously recorded neurons (data from McCairn et al. 2011; McCairn and Turner 2009).

Fig. 7.

GPi-DBS mediated suppression of tic-related phasic changes in the pallidum. Top: plots show response of each tic-responsive cell in the GPe (A) and GPi (B) in the off-stimulation (top matrix) and on-stimulation (bottom matrix) conditions. Note the prevalence of excitatory (GPe) and inhibitory (GPi) responses in the off-stimulation condition and their reduction during GPi-DBS. Middle: mean activity changes for the population of GPe and GPi neurons during off-stimulation and on-stimulation. Bottom: P values to indicate significant differences between the 2 stimulation conditions (data from McCairn et al. 2012).

Fig. 8.

GPi-DBS mechanism of action. Schematic illustration of a proposed mechanism of action for GPi-DBS. Top: pallidal outflow, when corrupted by synchronous, low-frequency pathological dysrhythmia, leads to hypo- or hyperkinetic condition that is underlined by abnormal thalamocortical excitability. Line thickness represents relative weight of inhibitory/excitatory output from each node. Bottom: GPi-DBS, through stimulation-driven entrainment, induces a high-pass, low-cut filter in basal ganglia output (GPi), followed by an induction of least pathological encoding in GPi-thalamic pathways, leading to a normalization of thalamocortical excitability and a maintenance of functional kinematic encoding for normal behavior. MCx, motor cortex.