Parkinsonian beta oscillations in the external globus pallidus and their relationship with subthalamic nucleus activity

Abstract

Inappropriately synchronized beta (beta) oscillations (15-30 Hz) in the subthalamic nucleus (STN) accompany movement difficulties in idiopathic Parkinson's disease (PD). The cellular and network substrates underlying these exaggerated beta oscillations are unknown but activity in the external globus pallidus (GP), which forms a candidate pacemaker network with STN, might be of particular importance. Using a clinically relevant rat model of PD, we demonstrate that oscillatory activity in GP neuronal networks becomes excessively and selectively synchronized at beta frequencies in a spatially widespread and brain state-dependent manner after lesion of dopamine neurons. Although synchronization of GP unit activity increased by almost 100-fold during beta oscillations, the mean firing rate of GP neurons decreased compared with controls. Importantly, in parkinsonian animals, two main types of GP neuron were identified according to their distinct and inversely related firing rates and patterns. Moreover, neurons of the same type tended to fire together, with small phase differences, whereas different types of neuron tended not to do so. This functional dichotomy in temporal coupling persisted across extreme brain states, suggesting that maladaptive interactions are dominated by hardwiring. Finally, the precisely timed discharges of GP and STN neurons indicated that rhythmic sequences of recurrent excitation and inhibition in the STN-GP network, and lateral inhibition between GP neurons, could actively support abnormal beta oscillations. We propose that GP neurons, by virtue of their spatiotemporal synchronization, widespread axon collaterals and feed-back/feed-forward mechanisms, are well placed to orchestrate and propagate exaggerated beta oscillations throughout the entire basal ganglia in PD.

Figure 1.

Single-cell and network activity in the globus pallidus of control and 6-OHDA-lesioned rats during cortical slow-wave activity. A, B, Simultaneous recordings of GP neurons during cortical SWA in a typical control rat (A, GP neurons 1–4) and a typical lesioned rat (B, GP neurons 5–8). Cortical activity (ECoG) is dominated by a slow oscillation at ∼1 Hz. Calibration: 200 μV (ECoG), 100 μV (units). All subsequent panels refer to these same GP neurons. C, D, Auto-correlograms (2 ms bins) of GP neurons in the control (C) and lesioned rat (D). Peaks and troughs indicate the fast oscillatory nature of single-cell firing (decreased activity at time ≈ 0 is due to the refractory period). E, F, Activity histograms of GP neurons in relation to the active (Act) or inactive (In) components of the cortical slow oscillation. In control rats, GP neurons typically fire independently of the slow oscillation components and are thus not modulated (NM) (E). Two distinct populations of GP neuron were identified in lesioned rats (F). Type A neurons (TA) discharge preferentially during the active component (see GP 5) whereas type I neurons (TI) discharge preferentially during the inactive component (GP 6–8). G, H, In control rats (G), pairs of GP units are typically not correlated, as exemplified by flat cross-correlograms (CC, bottom left). Coherence (Coh, top right) values for pairs are typically below significance at slow oscillation frequencies (p = 0.05, dashed line). Long time-scale auto-correlograms (10 ms bins, AC) are also flat (except at t ≈ 0). In contrast, in lesioned rats (H), pairs of GP units are typically synchronized in a slow oscillatory manner. Coherence values are significant and peak at the predominant SWA frequencies (0.4–1.6 Hz). Auto-correlograms (10 ms bins) indicate the slow oscillations in the spike trains of individual GP neurons.

Figure 2.

Population analyses of globus pallidus neuron spike rate and timing in control and 6-OHDA-lesioned rats during cortical slow-wave activity. A, Proportions of the three populations of GP neuron (type I [TI], Type A [TA] and not modulated [NM]) in control and lesioned rats. B, Mean activity histograms for all GP neurons in control rats (171 neurons) and lesioned rats (483 neurons) during cortical SWA. Data are means ± 1 SEM. C, Mean coherence of all GP unit pairs tested in control rats (612 pairs) and lesioned rats (1871 pairs) during SWA. Data are means ± 1 SEM. Dashed line is p = 0.05. D, Plot of instantaneous phase values for all significantly coherent pairs of GP units in control rats (108 pairs) and lesioned rats (1255 pairs). E, Instantaneous phase relationships of coherent pairs of GP units classified as type I or Type A in lesioned rats.

Figure 3.

Brain state-dependency of the power and frequency of local field potentials in cortex and globus pallidus in control and 6-OHDA-lesioned rats. A, B, Average power spectra of simultaneously recorded local field potentials in cortex (ECoG) and GP (GP-LFPs) in control and lesioned rats during SWA (A) and cortical activation (B). During SWA, ECoG and GP-LFPs power spectra are similar in control and lesioned rats. Prominent β oscillations emerge in cortex and GP during activated brain state in lesioned animals only. Weak power at 50 Hz in GP-LFP power spectra is line noise. C, Quantitative comparison of ECoG power in the SWA frequency band (0.4–1.6 Hz). Asterisks denote p < 0.01. D, Quantitative comparison of ECoG power and GP-LFP power in the β band (15–30 Hz) during activation. Cortical activation is accompanied by a significant decrease in SWA band power in control and lesioned rats, and a significant increase in β-band power in lesioned rats only. Data in A–D are means ± 1 SEM. Asterisks denote p < 0.01 compared with the three other groups.

Figure 4.

Single-cell and network activity in the globus pallidus of control and 6-OHDA-lesioned rats during cortical activation. A, B, Same GP neurons as in Figure 1 during subsequent epochs of activated brain states. Power spectra of field potentials in cortex (ECoG) and globus pallidus (GP-LFPs) show the excessive β oscillations (∼20 Hz) in the lesioned rat compared with the control. Calibration: 200 μV (ECoG), 100 μV (units). C, D, During activated state in control rats (C) pairs of GP units are typically not correlated (flat cross-correlograms; CC). Peaks and troughs in auto-correlograms (AC, 2 ms bins) indicate the fast oscillatory nature of single-cell firing. Coherence (Coh) values for pairs were typically below significance (p = 0.05, dashed line). In contrast, in lesioned rats (D), pairs of GP neurons tend to synchronize and coherence values peak at β frequencies (∼20 Hz).

Figure 5.

Excessive synchronization of globus pallidus ensemble activity after 6-OHDA lesion is associated with decreases in the firing rate and regularity of single neurons. A, B, Average firing rates and coefficients of variation of all GP neurons during SWA and cortical activation in control rats (171 and 149 neurons, respectively) and lesioned rats (487 and 478 neurons, respectively). Asterisks denote p < 0.01. C, Mean coherence of all GP unit pairs tested in control rats (491 pairs) and lesioned rats (2028 pairs) during cortical activation. Note appearance of significant coherence peak at β frequencies during activation in lesioned rats. D, Mean coherence of all GP unit pairs with significant coherence in the β-frequency band (15–30 Hz) in control rats (3 pairs) and lesioned rats (897 pairs) during cortical activation. Dashed line in C and D is p = 0.05. E, F, Correlation between the separation of significantly coherent GP unit pairs and peak coherence of pairs in the β band (E), and between separation and the percentage of significantly coherent GP pairs (F), in lesioned animals. Data in A–E are means ± 1 SEM.

Figure 6.

Dichotomy of globus pallidus unit activity in 6-OHDA-lesioned rats is maintained across slow (∼1 Hz) and β (∼20 Hz) network oscillations. A, B, Simultaneous recordings of three GP neurons [GP 1 and GP 2 are type I (TI), and neuron GP 3 is Type A (TA)] in a lesioned rat during SWA (A) and a subsequent activated state dominated by β oscillations (B). Gray lines centered on peaks of the cortical β oscillation exemplify spike timing relationships. Calibration: 500 μV (ECoG, SWA), 200 μV (ECoG, β), and 100 μV (units). C, D, Auto-correlograms (AC), cross-correlograms (CC) and coherence plots (Coh) for the same GP neurons demonstrating the dependency of synchronization and the frequency of oscillatory activity on brain state. 10 ms bins for AC and CC in C; 2 ms bins in D. Note that temporal relationships established between types of GP unit during SWA are maintained during β oscillations. E, F, Average coherence plots for all significantly coherent GP unit pairs during SWA (E) or during β oscillations prevalent in the activated brain state (F) in all lesioned rats. Note significant peaks of coherence for all GP pairs, all TI versus TI pairs, all TI versus TA pairs, and all TA versus TA pairs were at ∼1 Hz during SWA (E) and were at ∼20 Hz during activated state (F). G, H, Plots of instantaneous phase values for all pairs of GP neurons (same pairs as in E and F) during SWA (G) and β oscillations (H). Given mean peak frequencies of ∼1 Hz and ∼20 Hz, the phase bins of 10 degrees in (G) and (H) represent ∼28 ms and 1.4 ms, respectively. The clustering of phase relationships between types of GP neuron is maintained across the two extreme brain states. I, J, Average firing rates and coefficients of variation for each type of GP neuron (TI, 280 neurons; TA, 62 neurons) during SWA (I) and β oscillations (J). Asterisks denote p < 0.01. Data in E, F, I, and J are means ± 1 SEM.

Figure 7.

Spike timing of subthalamic nucleus and globus pallidus neurons in relation to cortical β oscillations in 6-OHDA-lesioned rats. A, Simultaneous recording of a single STN neuron and two GP neurons of different types (GP-TA and GP-TI) during cortical β oscillations (see ECoG) in a lesioned rat. Gray lines centered on peaks of the cortical β oscillation exemplify spike timing relationships. Calibration: 100 μV (filtered ECoG), 500 μV (STN unit), and 200 μV (GP units). B, Cross-correlograms (CC) show oscillatory synchronization of the STN versus GP-TA neuron pair and STN versus GP-TI pair at β frequencies (∼20 Hz). C, Linear phase histograms for neurons in A (left) and for all tested STN neurons, GP-TA neurons, and GP-TI neurons (right). For clarity, two cortical β oscillation cycles are shown. Data in population histograms (right) are means ± 1 SEM. The vertical black lines indicate the mean phase angles (error bars are 2 SEM). D, Mean coherence for all STN versus STN unit pairs, STN versus all GP unit pairs, STN versus GP-TI unit pairs, and STN versus GP-TA unit pairs (p = 0.05, dashed lines). E, Plots of instantaneous phase values for all pairs of neurons shown in D.