Local field potential beta activity in the subthalamic nucleus of patients with Parkinson's disease is associated with improvements in bradykinesia after dopamine and deep brain stimulation

Abstract

Parkinson's disease is treated pharmacologically with dopamine replacement medication and, more recently, by stimulating basal-ganglia nuclei such as the subthalamic nucleus (STN). Depth recordings after this procedure have revealed excessive activity at frequencies between 8 and 35 Hz (Brown et al., 2001, Kuhn et al., 2004, Priori et al., 2004) that are reduced by dopamine therapy in tandem with improvements in bradykinesia/rigidity, but not tremor (Kuhn et al., 2006). It has also been shown that improvements in motor symptoms after dopamine correlate with single unit activity in the beta range (Weinberger et al., 2006). We recorded local field potentials (LFPs) from the subthalamic nucleus of patients with Parkinson's disease (PD) after surgery to implant deep brain stimulating electrodes while they were on and off dopaminergic medication. As well as replicating Kuhn et al., using the same patients we were able to extend Weinberger et al. to show that LFP beta oscillatory activity correlated with the degree of improvement in bradykinesia/rigidity, but not tremor, after dopamine medication. We also found that the power of beta oscillatory activity uniquely predicted improvements in bradykinesia/rigidity, but again not tremor, after stimulation of the STN in a regression analysis. However improvements after STN stimulation related inversely to beta power, possibly reflecting the accuracy of the electrode placement and/or the limits of STN stimulation in patients with the greatest levels of beta oscillatory activity.

Introduction

Parkinson's disease (PD) is characterised by a poverty and slowness of voluntary movements (akinesia and bradykinesia), muscle rigidity and tremor at rest. Depth recordings in the basal ganglia of patients with PD undergoing functional brain surgery have revealed an excessive synchrony between neuronal populations in specific frequency bands. Pathological activity arises after degeneration of dopaminergic projections from the substantia nigra pars compacta to the striatum, but the exact mechanisms are unclear, as is how such activity contributes to Parkinsonian symptoms. Recently however, research has revealed some important relationships between such activity and both the symptoms of PD and its' treatment (Kuhn et al., 2006, Weinberger et al., 2006, Priori et al., 2004.

Currently PD is most effectively treated pharmacologically, with medications that increase dopamine levels in the basal ganglia, such as levodopa. However long-term use of these drugs is associated with motor fluctuations and dyskinetic movement abnormalities. In such cases deep brain stimulation (DBS) of basal-ganglia nuclei helps to ameliorate PD symptoms, allowing patients to use lower doses of dopaminergic medication. In such patients, and 1-Methyl-4-phenyl-1, 3, 4, 6-tetrahydropyridine (MPTP) treated primates, basal-ganglia nuclei show aberrant synchrony between frequencies within the alpha and/or beta bands (8–35 Hz) (Bergman et al., 1994, Brown et al., 2001, Kuhn et al., 2004, Priori et al., 2004), referred to from here on as the beta band.

Confirmation for a causal role for beta oscillatory activity in Parkinsonian motor symptoms would require evidence of increased motor symptoms in response to increased beta oscillatory activity. Two studies report increases in bradykinesia after stimulating the STN at beta frequencies (Chen et al., 2007, Eusebio et al., 2008), however the effects are small and can only be seen in specific patients. Nonetheless, research tends to show that treatments for PD reduce beta synchrony in the basal ganglia. One such report demonstrated a positive correlation between improvements in motor symptoms and the degree of beta power suppression after dopamine intake; measured via local field potential (LFP) recordings in the subthalamic nucleus of patients after surgery to implant DBS electrodes (Kuhn et al., 2006). The relationship existed for reductions in scores that indexed the degree of bradykinesia and rigidity, but not tremor. Similar relationships have been reported for reductions in cortico-cortical beta coherence in the scalp electroencephalogram (EEG) (Silberstein et al., 2005).

Using microelectrodes to record the activity of single units within the subthalamic nucleus during DBS surgery Weinberger et al. (2006) found that the percentage of cells manifesting oscillatory activity in the beta range correlated with the degree by which PD motor symptoms improved after dopamine replacement therapy. Surprisingly however, the incidence of oscillatory units did not correlate with the severity of symptoms before treatment. This suggests that beta oscillatory activity does not fully characterise motor impairment in PD, perhaps especially in the advanced stages reached by patients undergoing deep brain stimulation (DBS). However it might instead relate to the magnitude of the response of the basal ganglia to dopaminergic agents.

It is not yet clear if DBS of the STN also reduces beta oscillatory activity. The artefact recorded during stimulation precludes accurate measurement of neuronal oscillatory activity. Also, studying the time period immediately after stimulation while the effects of treatment persist has yielded conflicting results (Wingeier et al., 2006, Foffani et al., 2006). Therefore, deciding whether the relationship reported by Kuhn et al. (2006) between beta oscillatory activity and benefits from dopamine also exists for STN stimulation is not possible with current data. It is possible however to test the relationship reported by Weinberger et al. (2006) showing that beta oscillatory activity correlates with the degree to which symptoms respond to dopaminergic medication, as this does not require measurements of beta oscillatory activity during or after DBS.

DBS of the subthalamic nucleus has been suggested to improve Parkinsonian symptoms by disrupting pathological beta oscillatory activity. Therefore given the relationships between beta oscillatory activity and the dopaminergic response, it is reasonable to expect that beta power will also relate to the response to STN stimulation. Determining what this relationship is could help us to predict which patients will benefit most from surgery. However the effectiveness of STN stimulation depends on many factors, including the degree of the pre-surgery dopaminergic response and the accuracy of the electrode placement. Any investigation into the relationship between beta oscillatory activity and the efficacy of surgery should therefore control for the influence of such variables on improvements in symptoms.

The related but different associations between activity in the beta band and dopaminergic medication found by Kuhn et al. (2006) with local field potentials, and Weinberger et al. (2006) with single unit activity is yet to be demonstrated in the same set of patients. Since the unit activity was also found to be coherent with the simultaneously recorded LFP in the Weinberger et al., study, we should also be able to detect the relationship they found between the incidence of oscillating units and dopaminergic response using LFP data. In the present study we were able to corroborate both these findings in a single sample of patients using LFP recordings.

We also report evidence to support a negative association between power in the beta range and the efficacy of STN stimulation.