* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download Excitatory Postsynaptic Potentials Trigger a Plateau Potential in Rat
Patch clamp wikipedia , lookup
Node of Ranvier wikipedia , lookup
Caridoid escape reaction wikipedia , lookup
Activity-dependent plasticity wikipedia , lookup
Mirror neuron wikipedia , lookup
Development of the nervous system wikipedia , lookup
Clinical neurochemistry wikipedia , lookup
Multielectrode array wikipedia , lookup
Apical dendrite wikipedia , lookup
Neurotransmitter wikipedia , lookup
Synaptogenesis wikipedia , lookup
Central pattern generator wikipedia , lookup
Neural coding wikipedia , lookup
Neuroanatomy wikipedia , lookup
Neural oscillation wikipedia , lookup
Biological neuron model wikipedia , lookup
Neuropsychopharmacology wikipedia , lookup
Spike-and-wave wikipedia , lookup
Membrane potential wikipedia , lookup
Optogenetics wikipedia , lookup
Nonsynaptic plasticity wikipedia , lookup
Feature detection (nervous system) wikipedia , lookup
Premovement neuronal activity wikipedia , lookup
Action potential wikipedia , lookup
Resting potential wikipedia , lookup
Molecular neuroscience wikipedia , lookup
Chemical synapse wikipedia , lookup
Stimulus (physiology) wikipedia , lookup
End-plate potential wikipedia , lookup
Nervous system network models wikipedia , lookup
Electrophysiology wikipedia , lookup
Pre-Bötzinger complex wikipedia , lookup
Synaptic gating wikipedia , lookup
Channelrhodopsin wikipedia , lookup
Excitatory Postsynaptic Potentials Trigger a Plateau Potential in Rat Subthalamic Neurons at Hyperpolarized States TAKESHI OTSUKA,1 FUJIO MURAKAMI,2,3 AND WEN-JIE SONG1 Department of Electronic Engineering, Graduate School of Engineering, Osaka University, Suita 565-0871; 2 Division of Biophysical Engineering, Graduate School of Engineering Science, Osaka University; and 3 Core Research for Evolutional Science and Technology/Murakami Laboratory, Center for Advanced Research Projects, Osaka University, Toyonaka 560-8531, Japan 1 Received 5 March 2001; accepted in final form 26 June 2001 Otsuka, Takeshi, Fujio Murakami, and Wen-Jie Song. Excitatory postsynaptic potentials trigger a plateau potential in rat subthalamic neurons at hyperpolarized states. J Neurophysiol 86: 1816 –1825, 2001. The subthalamic nucleus (STN) directly innervates the output structures of the basal ganglia, playing a key role in basal ganglia function. It is therefore important to understand the regulatory mechanisms for the activity of STN neurons. In the present study, we aimed to investigate how the intrinsic membrane properties of STN neurons interact with their synaptic inputs, focusing on their generation and the properties of the long-lasting, plateau potential. Whole cell recordings were obtained from STN neurons in slices prepared from postnatal day 14 (P14) to P20 rats. We found that activation of glutamate receptor-mediated excitatory synaptic potentials (EPSPs) evoked a plateau potential in a subpopulation of STN neurons (n ⫽ 13/22), in a voltage-dependent manner. Plateau potentials could be induced only when the cell was hyperpolarized to more negative than about ⫺75 mV. Plateau potentials, evoked with a depolarizing current pulse, again only from a hyperpolarized state, were observed in about half of STN neurons tested (n ⫽ 162/327). Only in neurons in which a plateau potential could be evoked by current injection did EPSPs evoke plateau potentials. L-type Ca2⫹ channels, Ca2⫹-dependent K⫹ channels, and TEA-sensitive K⫹ channels were found to be involved in the generation of the potential. The stability of the plateau potential, tested by the injection of a negative pulse current during the plateau phase, was found to be robust at the early phase of the potential, but decreased toward the end. As a result the early part of the plateau potential was resistant to membrane potential perturbations and would be able to support a train of action potentials. We conclude that excitatory postsynaptic potentials, evoked in a subpopulation of STN neurons at a hyperpolarized state, activate L-type Ca2⫹ and other channels, leading to the generation of a plateau potential. Thus about half of STN neurons can transform short-lasting synaptic excitation into a long train of output spikes by voltage-dependent generation of a plateau potential. The subthalamic nucleus (STN) acts as a driving force of the basal ganglia by exerting glutamatergic excitatory effect on the output structures of the basal ganglia, the globus pallidus and the substantia nigra (see Kitai and Kita 1987 for a review). The significance of the STN in motor control has been implicated in a clinical observation that pathological changes in the nucleus causes hemiballism (Whittier 1947) and in animal experiments in which manipulation of the activity of the nucleus dramatically affects motor behavior (Hamada and Hasegawa 1996; Wichmann et al. 1994b). These observations indicate the importance of controlling outputs of the basal ganglia by STN neurons. It is therefore crucial to know how the activity of STN neurons is regulated for understanding basal ganglia function in motor control. The activity of a neuron is determined by extrinsic synaptic inputs and intrinsic membrane properties. The STN is known to receive excitatory input from the cortex (Fujimoto and Kita 1993; Hartmann-von Monakow et al. 1978; Kitai and Deniau 1981; Nambu et al. 1996) and the thalamus (Féger et al. 1994; Mouroux and Féger 1993) and inhibitory inputs from the globus pallidus (Groenewegen and Berendse 1990; Kita et al. 1983; Moriizumi and Hattori 1992). How STN neurons respond to these inputs depends on their membrane properties. In slice studies, STN neurons fire regularly, increasing firing frequencies linearly with the magnitude of injected currents (Bevan and Wilson 1999; Nakanishi et al. 1987). This would suggest that the STN works as a linear transformer relaying excitatory inputs to its targets. Indeed, a recent study in anesthetized rat suggests that STN neurons follow the activity of cortical neurons (Magill et al. 2000). STN neurons, however, have intrinsic membrane properties, which can significantly change neuronal firing pattern. The generation of a plateau potential, a long-lasting depolarizing potential, for example, has been described in a subset of STN neurons (Beurrier et al. 1999; Nakanishi et al. 1987; Otsuka et al. 1998, 1999; Overton and Greenfield 1995; Song et al. 1998). Because of its slow decay kinetics, the plateau potential would lead to a longlasting, high-frequency discharge in the absence of synaptic inputs. In behaving animals, burst activity lasting several hundred milliseconds has been observed in STN neurons (Matsumura et al. 1992; Wichmann et al. 1994a). The generation of a plateau potential may be one possible underlying mechanism for such long-lasting burst. Although the plateau potential can profoundly alter the output properties of STN neurons, no work has focused on how the ability to generate plateau potentials affects synaptic inte- Address for reprint requests: W.-J. Song, Dept. of Electronic Engineering, Graduate School of Engineering, Osaka University, 2-1 Yamadaoka, Suita 565-0871, Japan (E-mail: [email protected]). The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked ‘‘advertisement’’ in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. INTRODUCTION 1816 0022-3077/01 $5.00 Copyright © 2001 The American Physiological Society www.jn.org PLATEAU POTENTIALS OF SUBTHALAMIC NEURONS gration. Beurrier et al. (1999) reported a plateau potential in STN neurons that in part supports recurrent membrane oscillations, leading to rhythmic burst firing of STN neurons. Others have observed plateau potentials evoked by current injections (Nakanishi et al. 1987; Overton and Greenfield 1995). But it is not known whether plateau potentials can be triggered by synaptic potentials. In the present study, we therefore tested whether activation of excitatory synaptic inputs to STN neurons can trigger a plateau potential and if so, how the plateau potential would interact with synaptic potentials. We found that activation of glutamate receptor-mediated synaptic potentials triggered a plateau potential in about half of STN neurons, only when the cells were hyperpolarized. Our results thus suggest that about half of STN neurons transform excitatory synaptic inputs into either a single spike or a train of spikes, depending on membrane potential. Part of the results has been published in abstract form (Otsuka et al. 1998, 1999; Song et al. 1998). METHODS Slice preparation Slice preparation including the STN was obtained from SpragueDawley rats (14 –20 postnatal days). All experiments were conducted in compliance with the Guidelines for Use of Laboratory Animals of Osaka University. Rats were anesthetized with ether and perfused with a high-sucrose solution containing (in mM): 200 sucrose, 2.5 KCl, 0.5 CaCl2, 10 MgSO4, 1.25 KH2PO4, 26 NaHCO3, 10 glucose, 0.2 ascorbic acid, and 1 pyruvic acid (300 ⫾ 5 mOsm/l; pH, 7.4). The brain was then removed, iced, and blocked for slicing. In the medium described in the preceding text, 250- to 300-m (mostly 300 m)thick horizontal slices were prepared using a Microslicer (Dosaka EM, Japan). The slices were then incubated at room temperature (20°C) in oxygenated Kreb’s solution containing (in mM): 126 NaCl, 2.5 KCl, 1.25 KH2PO4, 1 MgSO4, 2 CaCl2, 26 NaHCO3, and 10 glucose (300 ⫾ 5 mOsm/l, pH, 7.4; bubbled with 95% O2-5% CO2). Ascorbic acid (0.2) and 1 pyruvic acid (in mM) were added to the holding solutions for improving tissue viability. After 1-h recovery period, a slice was transferred to a recording chamber mounted on the stage of an upright microscope (Olympus, Tokyo). The recording chamber was continuously superfused with the oxygenated Kreb’s solution. Recordings and data analyses Whole cell recordings of STN neurons employed standard techniques (Edwards et al. 1989; Stuart et al. 1993). Electrodes were pulled from glass capillary tubes (Narishige, Tokyo) and fire-polished. In most experiments, the recording pipettes were filled with a solution containing (in mM): 120 KCl, 3 MgCl2, 10 HEPES, 0.2 EGTA, 2 Na2ATP, 0.2 Li2GTP, 12 phosphocreatine, and 0.1 leupeptin (pH, 7.2; 270 ⫾ 5 mOsm/l) and had resistances of 5–7 M⍀ in the bath. The liquid-junction potential was estimated to be 4.3 mV. All potentials reported in this paper were corrected for this potential. In experiments chelating internal Ca2⫹, the internal solution consisted of (in mM) 100 KCl, 3 MgCl2, 10 HEPES, 20 BAPTA, 2 Na2ATP, 0.2 Li2GTP, 12 phosphocreatine, 0.1 leupeptin (pH, 7.2; 270 ⫾ 5 mOsm/l), and the concentration of KCl of the external solution was changed to 1.5 mM to adjust K⫹ equilibrium potential. We also recorded with a low-Cl⫺ internal solution containing (in mM) 70 K2SO4, 2.5 MgCl2, 1.0 EGTA, 0.1 CaCl2, 35 HEPES, 30 N-methyl-D-glucamine (NMG), 2 Na2ATP, 0.2 Li2GTP, and 0.1 leupeptin (pH 7.2, 270 ⫾ 5 mOsm/l). Recordings were obtained with an EPC-7 amplifier (List-MedicalElectronics, Darmstadt, Germany) controlled with a Pentium PC running pCLAMP (version 6.0) or Axoscope (Axon Instruments, J Neurophysiol • VOL 1817 Foster City, CA). Synaptic responses were evoked by electrical stimulation applied through a bipolar tungsten electrode using a Master-8 stimulator (AMPI, Jerusalem, Israel). The electrode was placed in a region rostral to the STN. A single current pulse of 0.1– 0.5 mA in amplitude and 100 s in duration was used for stimulation. Data analyses were performed with AxoGraph (Axon Instruments) and Kaleidagraph (Albeck Software, Reading, PA). Data are represented as means ⫾ standard deviation (SD) and statistical difference between samples was tested using Mann-Whitney U test, unless mentioned otherwise. Significance was accepted when P ⬍ 0.05. Drugs All reagents were obtained from Sigma Chemical (St. Louis, MO) except nifedipine and 6,7-dinitroquinoxaline-2,3-dione (DNQX), which were obtained from RBI (Natick, MA). Biocytin was diluted in the internal solution at the concentration of 5 mg/ml. Other drugs were diluted in oxygenated Kreb’s solution and applied to the slice. Nifedipine was made up as concentrated stocks in 95% ethanol and diluted immediately before use. DNQX was dissolved in dimethylsulfoxide. When using these drugs, equal concentrations of the solvent were added to all control solutions. For the tetraethylammonium chloride (TEA) experiment, the Kreb’s solution was modified by replacing NaCl (50 mM) with NMG (50 mM) and was used as the control solution; TEA solutions were prepared by replacing NMG with an equimolar concentration of TEA. Solutions containing bicuculline or nifedipine were protected from ambient light. Histology To verify that recorded cells were in fact STN neurons, biocytin was always included in the internal solution (Horikawa and Armstrong 1988). The avidin-biotin-horseradish peroxidase reaction was used to visualize recorded neurons. After recording, slices were fixed with a solution containing 4% paraformaldehyde in phosphate buffer (0.1 M; pH, 7.4). After rinsing in Tris-buffered saline (TBS; 50 mM; pH, 7.4), the slices were treated with a mixture of 10% methanol and 0.3% H2O2 for 20 –30 min, rinsed again in TBS, and treated with TBS containing 0.5% Triton x-100. After washes with TBS, the slices were incubated for 2 h in the avidin-biotin-horseradish peroxidase complex (1%; Vector Laboratories, Burlingame, CA) at room temperature, washed in TBS, and reacted with a mixture of 3,3⬘-diaminobenzidine tetrahydrochloride (0.05%) and H2O2 (0.003%) in TBS for 5–10 min. The slices were then rinsed several times in TBS and mounted on gelatin-coated glass slides. The mounted slices were stained with methylene blue for identification of the nucleus, dehydrated in graded ethanol series, cleared in xylene, and coverslipped with Entellan New (Merck, Darmstadt, Germany) for observation with a light microscope. RESULTS Recordings were obtained from 327 neurons. Each recorded neuron was labeled intracellularly with biocytin, and all were found within the STN (Fig. 1A). The stained STN neurons had soma diameters of 15–30 m and two to six primary dendrites. The resting membrane potential was ⫺62.65 ⫾ 5.10 mV and the input resistance at rest, estimated with a negative current pulse (⫺10 pA), was 0.66 ⫾ 0.22 G⍀. Induction of plateau potentials by synaptic potentials To examine the response of STN neurons to excitatory synaptic inputs, recordings were obtained while stimulation at a location rostral to the STN was performed to evoke synaptic 86 • OCTOBER 2001 • www.jn.org 1818 T. OTSUKA, F. MURAKAMI, AND W.-J. SONG FIG. 1. Responses to synaptic stimulation in subthalamic nucleus (STN) neurons. A: a photomicrograph of a 300-m-thick slice stained with methylene blue. Arrowhead indicates the recorded cell labeled intracellularly with biocytin. Scale bar ⫽ 50 m. B: a synaptic potential evoked by stimulation of a site rostral to the STN. Coapplication of 6,7-dinitroquinoxaline-2,3-dione (DNQX) and 2-amino-5-phosphonovalerate (APV) abolished the potential. C: in the same cell, increasing the stimulus strength evoked a single action potential at the resting membrane potential. D: at a hyperpolarized membrane potential, however, stimulation with the same strength as in C evoked a long-lasting potential and a short train of spikes. Membrane potentials were hyperpolarized by injection of a constant current. potentials. Bicuculline (50 M) was included in the external solution to block inhibitory synaptic potentials. In response to stimulation, depolarizing potentials or inward currents were observed in most of the neurons examined (n ⫽ 32/41). Stimulation at sites rostrolateral to the STN failed to evoke a response (n ⫽ 7). Coapplication of DNQX (10 M), a non-Nmethyl-D-aspartate (non-NMDA) receptor blocker, and 2-amino-5-phosphonovalerate (APV; 50 M), an NMDA receptor blocker, abolished the potentials (n ⫽ 4, Fig. 1B), indicating that these potentials are excitatory postsynaptic potentials (EPSPs) mediated by glutamate receptors. Application of either DNQX or APV alone revealed that 54 –71% of the EPSP amplitude was mediated by non-NMDA receptors at ⫺80 mV (n ⫽ 3). At 20°C, the latency of the EPSPs was 4.4 –7.2 ms and changed ⬍0.5 ms with varying stimulus strength. Assuming a Q10 value of 3.5 (Hirano et al. 1986), the latencies correspond to latencies of 0.5– 0.9 ms and the latency change corresponds to a value ⬍0.1 ms, at 37°C, suggesting that the EPSPs are of monosynaptic nature. With increased stimulus strength, the EPSPs always triggered a single action potential in STN neurons at resting membrane potentials (n ⫽ 22; Fig. 1C). When the membrane potential was hyperpolarized to about ⫺75 mV, however, stimulation with the same strength evoked a long-lasting potential, or a plateau potential, in a subpopulation of STN neurons (n ⫽ 13 of 22, Fig. 1D). The holding current required to hyperpolarize the membrane potential to ⫺75 to ⫺85 mV ranged from ⫺10 to ⫺30 pA. On the rising phase of the plateau potential, a short train of action potentials (1–5 spikes) was evoked (Fig. 1D). When the stimulus intensity was decreased, only EPSPs were observed. Gradually changing the stimulus strength revealed that the amplitude of the EPSPs right before the occurrence of plateau potentials ranged from ⫺61 to ⫺67 mV (n ⫽ 5). Neurons in which plateau potentials were not observed always discharged a single action potential at the peak of the EPSPs either at rest or at hyperpolarized states, for stimulus intensities up to 0.5 mA (n ⫽ 9). These results suggest that EPSPs are capable of triggering a plateau potential in a subpopulation of STN neurons, in a voltage-dependent manner. J Neurophysiol • VOL Induction of plateau potentials by injected current pulses To examine how plateau potentials affect synaptic integration in STN neurons, we next studied the properties of plateau potentials in STN neurons. For this purpose, plateau potentials were evoked with current injection instead of synaptic activation for the ease of experimentation. At resting membrane potentials, injection of depolarizing current pulses evoked action potentials either during the initial phase of the current (70%, n ⫽ 205/294; Fig. 2A, top) or during the entire period of current injection (n ⫽ 89; Fig. 2C, top). The membrane potential returned to rest in an exponential manner after termination of the current pulse (Fig. 2, A and C, top). At hyperpolarized potentials, however, a long-lasting potential was induced that far outlasted current injection (Fig. 2C, bottom), again in a subpopulation of STN neurons. Other cells exhibited an exponential relaxation of membrane potential after termination of current injection (Fig. 2A, bottom). For a quantitative description of the membrane potentials after current termination, we measured the half-decay time defined as the time interval from the pulse end to the time when the potential had decayed to half-amplitude at the current pulse end. STN neurons could easily be divided into a population having short half-decay times (30 ⫾ 0.13 ms, n ⫽ 146) and a population having much longer half-decay times (0.66 ⫾ 0.49 s, range 0.20 –3.0 s; n ⫽ 148). We defined a potential with a half-decay time ⱖ0.2 s as a plateau potential. About half of STN neurons tested (n ⫽ 148/294) generated a plateau potential at hyperpolarized state. Plateau potentials were also observed with an internal solution of low Cl⫺ concentration (see METHODS) at 30°C, again in a subset of neurons (n ⫽ 14/33). Under this recording condition, some STN neurons showed spontaneous activity at resting membrane potentials (2–10 Hz, n ⫽ 9/33), although these activities disappeared with membrane hyperpolarization; all STN neurons discharged during the entire period of current injection. Plateau-generating neurons fired 5.1 ⫾ 1.6 (n ⫽ 14) spikes during current injection (100 ms in duration, 50 pA in 86 • OCTOBER 2001 • www.jn.org PLATEAU POTENTIALS OF SUBTHALAMIC NEURONS 1819 FIG. 2. Induction of plateau potentials with current pulses. A: recordings from a nonplateau-generating cell. Top: injection of a depolarizing pulse current (50 pA), shown as the bottom trace, at the resting membrane potential, depolarized the membrane potential and evoked action potentials. The membrane potential returned to the resting level immediately after the termination of the current in an exponential manner. Middle: injection of the same current at a hyperpolarized state evoked a similar response, though an increase of membrane resistance is obvious. B: a recording from the same neurons as in A. A rebound action potential was evoked on termination of a hyperpolarizing current (⫺50 pA). During injection of the current, the membrane potential exhibited a tendency to return to the resting level, indicating the presence of an H current in this neuron. C: recordings from a plateau-generating neuron. Top: injection of a depolarizing pulse current (50 pA, bottom), at the resting membrane potential, depolarized the membrane potential and evoked action potentials. The membrane potential relaxed to the resting level immediately after the termination of the current, as in A. Middle: injection of the same current at a hyperpolarized state, however, evoked a plateau potential that outlasted current injection. D: a recording from the same neuron as in C. A plateau potential was induced as a rebound potential on termination of a negative current (⫺50 pA). amplitude), which is not significantly different from that of nonplateau-generating neurons (4.5 ⫾ 2.1, n ⫽ 19; P ⬎ 0.05). A plateau potential was also generated after termination of a hyperpolarizing current pulse (n ⫽ 19; Fig. 2D). Such a potential was observed only in neurons in which a plateau potential could be evoked by a depolarizing current pulse, at hyperpolarized states (compare Fig. 2, B–D). We next tested to what membrane potentials the cell has to be hyperpolarized for a depolarizing pulse to evoke a plateau potential. Membrane potentials were gradually hyperpolarized by changing the amount of injected constant current, while a depolarizing current pulse (50 ms) was injected to test if a plateau potential could be induced. As a result, a plateau potential was induced in an all-or-none manner, across certain membrane potentials (Fig. 3A). The membrane potential at which a plateau potential was first induced was referred to as threshold potential. Shown in Fig. 3B is a histogram of the threshold potentials from 19 neurons. The threshold potential was ⫺74.98 ⫾ 1.96 mV. From a hyperpolarized state, the membrane potential had to be depolarized to a certain potential for plateau potential to occur. This depolarizing threshold for the plateau potential was difficult to determine from the rising phase of the potential, but with gradual change of injected current, the potentials immediately before the occurrence of plateau potentials ranged from ⫺62 to ⫺67 mV, corresponding to an injected current of ⬃10 pA. The amplitude of the plateau potential was independent of the suprathreshold current amplitude, ranging from 10 to 100 pA. To estimate the distribution of plateau-generating neurons in the STN, we divided the nucleus into three equal parts by length, either along the rostrocaudal or lateromedial axis. Along the rostrocaudal axis, the ratio of plateau-generating neurons to nonplateau-generating neurons was 46.9% (n ⫽ FIG. 3. The threshold hyperpolarization level for the plateau potentials to be induced. A: the cell was held at different membrane potentials by injecting constant currents, while a short depolarizing pulse was applied to test if plateau potential can be induced. The plateau potential was induced only when the membrane potential was kept at ⫺74 mV or more negative. We call ⫺74 mV the threshold potential for this case. B: the histogram of threshold potentials determined in 19 STN neurons. The distribution of threshold potentials had a peak at about ⫺74 mV. J Neurophysiol • VOL 86 • OCTOBER 2001 • www.jn.org 1820 T. OTSUKA, F. MURAKAMI, AND W.-J. SONG 15/32) in the rostral part, 47.6% (n ⫽ 32/68) in the middle part, and 51.2% (n ⫽ 15/32) in the caudal part. Along the lateromedial axis, however, the ratio in the lateral, middle, and medial part of the STN was 60.3% (n ⫽ 38/63), 44.9% (n ⫽ 22/49), and 29.0% (n ⫽ 9/31), respectively. Thus plateaugenerating neurons tend to be located in the lateral part of the nucleus. Although the morphology of plateau-generating neurons did not appear to differ from that of nonplateau-generating neurons, the input resistance at resting membrane potentials of plateau-generating neurons was found to be significantly larger than that of nonplateau-generating neurons (0.813 ⫾ 0.07 vs. 0.524 ⫾ 0.05 G⍀, n ⫽ 148; Student’s t-test, P ⬍ 0.005). To test whether plateau potentials evoked by synaptic potentials and those evoked by current injection are of the same nature, we tried to evoke plateau potentials in the same neuron both by synaptic activation and by current injection. In such experiments, EPSPs evoked plateau potentials only in neurons in which a plateau potential could be evoked by injection of current pulses at hyperpolarized states (n ⫽ 13), suggesting that plateau potentials evoked by EPSPs and those evoked by current injections are attributable to a common membrane mechanism. Effect of temperature It was puzzling that action potentials were not evoked during the plateau phase of the plateau potential. We suspected that this might be because the experiment was carried out at room temperature. To verify this hypothesis, we raised the temperature from 20 to 25°C. As shown in Fig. 4, although a plateau potential did not evoke action potentials at 20°C (Fig. 4A), it evoked action potentials even at its late phase at 25°C (Fig. 4B; n ⫽ 6). Raising the temperature also appeared to increase the duration of the plateau potential (compare Fig. 4, A to B). Stability of plateau To study how the plateau potentials interact with synaptic inputs, we tested the stability of the plateau potentials. Because plateau potentials were not terminated by action potentials (Fig. 4B), they appear to be resistant to perturbations of depolarizing potentials. To test the effect of inhibitory synaptic potentials, a negative current pulse, used to represent an inhibitory synaptic current, was injected at successive timings during the course of the plateau potential induced with a short pulse (50 ms; Fig. 5). The negative current had an amplitude of ⫺60, ⫺80, or ⫺90 pA. The duration of the current was determined in reference to inhibitory synaptic currents in STN neurons (Shen and Johnson 2000), and a 20-ms duration was used. The stability of the plateau potential was evaluated by the ratio of the peak potential after the current pulse to the potential immediately before the current, referred to as stability index here; a stable potential would give rise to a stability index of one. Shown in Fig. 5B is the stability index estimated at successive time intervals during the course of plateau potentials. The stability index was close to one during the early phase of the potential and then fell gradually afterwards. During the initial 10% of the plateau, however, the stability index was always ⬎1 (1.18 ⫾ 0.06; n ⫽ 4). Toward the end of the potential, the stability index fell abruptly. To compare the effect of the amplitude of the negative current, data between 30 and 70% of the plateau duration was fitted with an exponential function in the form Stability index ⫽ 1⫺ exp共共T ⫺ A兲ⲐB兲 where T is percent plateau duration, B is time constant, and A represents delay (Fig. 5B, inset). The stability index decreased to 0.7 in 69 ⫾ 8% of plateau duration when the current amplitude was ⫺60 pA. This percentage was 62 ⫾ 8% for a current of ⫺80 pA and 49 ⫾ 8% for a current of ⫺90 pA. Thus the early part of the plateau potential was resistant to inhibitory perturbations and would be able to support a train of spikes, although the plateau potential may be terminated by inhibitory synaptic inputs toward the end of the potential. Subcellular origin for the generation of plateau potentials FIG. 4. Temperature dependence of the firing properties of a plateaugenerating neuron. A: a recording at 20°C. No action potential was observed during the plateau phase of the plateau potential. B: a recording from the same neuron as in A at 25°C. Action potentials were evoked during the plateau phase. The injected current was 50 pA for both A and B. J Neurophysiol • VOL Whether plateau potentials in STN neurons are generated in the soma or in the dendrites is of obvious significance for synaptic integration. To address this question, we tested the effect of voltage-clamping somatic membrane on the ability of EPSPs in evoking plateau potentials (currents). Anatomical evidence suggests that excitatory synaptic inputs end on distal dendrites of STN neurons (Bevan et al. 1995). Thus EPSPs in STN neuron, as those shown in Fig. 1, are expected to be induced from distal dendrites. If a plateau potential occurs at the soma, voltage-clamping the soma would prevent the induction of the plateau by EPSPs; in contrast, considering the filtering effect of dendrites, a plateau potential occurring at distal dendrites may not be blocked by voltage-clamping the soma and would thus give rise to a plateau current at the soma. To verify that excitatory synaptic inputs occur at sites electrotonically distant from the soma, we tried to determine the reversal potential of the EPSPs, which is calculated to be ⫹6.4 mV for an appropriate voltage control, assuming equal permeability of Na⫹ and K⫹ and a Ca2⫹ permeability of 1.17 relative 86 • OCTOBER 2001 • www.jn.org PLATEAU POTENTIALS OF SUBTHALAMIC NEURONS FIG. 5. The stability of the plateau potential tested with hyperpolarizing currents. A: plateau potentials were evoked from a hyperpolarized state with a current pulse (100 pA, 50 ms). A negative pulse current (⫺60 pA, 20 ms) was injected at successive time intervals after the initiation of the potential. B: the ratio of the peak value of the potential after the negative current pulse to the potential immediately before the current injection, for three current amplitudes (⫺60, ⫺80, and ⫺90 pA), was plotted against plateau potential duration normalized to one. The ratio is close to one during the early phase of the plateau potential. Toward the end of the potential, the ratio fell abruptly. Inset: data between 30 and 70% of the plateau duration was fitted with exponential curves. With increasing current amplitude, the potential was of shorter duration. to Na⫹; intracellular Ca2⫹ concentration was assumed to be 100 nM (Götz et al. 1997; Mayer and Westbrook 1987). EPSPs were evoked in the same way as described in the preceding text. As a result, the reversal potential was ⫹21.0 ⫾ 10.8 mV (n ⫽ 4). In three other tested cells, the EPSPs did not reverse 1821 for potentials up to ⫹20 mV. These results suggest that excitatory synaptic inputs to STN neurons are electrotonically distant from the soma and are consistent with previous anatomical observations (Bevan et al. 1995). To test the effect of voltage-clamping somatic membrane on the ability of EPSPs in evoking plateau potentials, a plateau potential was first elicited in current-clamp mode by EPSPs (Fig. 6A). In voltage-clamp mode, however, only a transient inward current was evoked in the same neuron, in response to the same stimulation (Fig. 6B). To study the nature of the current, we compared its time course to that of currents recorded in the same way from nonplateau-generating neurons. Shown in Fig. 6, C and D, is an example of recordings from a nonplateau-generating neuron. In both plateau- and nonplateau-generating cells, the current recorded in voltage-clamp mode decayed in an exponential manner. The time constant was 12.0 ⫾ 2.0 ms (n ⫽ 7) in plateau-generating neurons and 14.5 ⫾ 3.9 ms (n ⫽ 6) in nonplateau-generating neurons, which is not significantly different from that of plateau-generating cells (P ⬎ 0.05). This result suggests that the current recorded in plateau generating cells was not related to the plateau potential but was likely a pure synaptic current. In plateau-generating cells, doubling the stimulus intensity increased the amplitude of the current, but did not change its kinetics (decay time constant, 12.7 ⫾ 2.2 ms, n ⫽ 7; P ⬎ 0.05). Taken together, these results suggest the possibility that plateau potentials in STN neurons occur at a region where membrane potential is controlled by an electrode in the soma. This region is likely to be the soma and/or proximal dendrites. Ionic mechanisms of a plateau potential The observation that plateau potentials can be induced only at hyperpolarized membrane potentials suggests that voltagedependent conductances are involved in the generation of the plateau potentials. Because a plateau potential could be induced only from a hyperpolarized state, it is natural to consider channels deinactivated at hyperpolarized potentials to be involved in the generation of plateau potentials. An obvious FIG. 6. Subcellular origin of the plateau potential. A: in current-clamp mode, stimulation of a site rostral to STN evoked a plateau potential. Subthreshold membrane oscillations (arrowhead) during the plateau phase was observed in this and 33 other STN neurons. The oscillating frequency was 19.9 ⫾ 7.7 Hz. Application of TTX (1 M) abolished the membrane oscillation (n ⫽ 5). Inset: autocorrelation function of the potential indicated by the arrowhead. B: in voltage-clamp mode, stimulation at the same strength as in A evoked a transient current. Voltage was clamped to ⫺80 mV. C: current-clamp recording from a nonplateaugenerating cell. Only a synaptic potential was evoked on top of which an action potential was triggered. D: a voltage-clamp recording from the same neuron with the same stimulus strength as in C. Holding potential was ⫺80 mV. Shaded curves in B and D are exponential fit to the current. J Neurophysiol • VOL 86 • OCTOBER 2001 • www.jn.org 1822 T. OTSUKA, F. MURAKAMI, AND W.-J. SONG candidate is the low-threshold Ca2⫹ channel (or T channel). The existence of T channels in STN neurons has been documented before (Beurrier et al. 1999; Nakanishi et al. 1987; Song et al. 2000), and its possible involvement in the generation of plateau potentials in STN neurons has been suggested (Beurrier et al. 1999). Because of the transient nature of the T channels (Carbone and Lux 1984), it is unlikely that a plateau potential is maintained by T channels. To search for conductances involved in the maintenance of plateau potentials, we tested the effect of blocking channels that inactivate or deactivate slowly. Previous studies have suggested that Ca2⫹ (Nakanishi et al. 1987) and L-type Ca2⫹ channels (Beurrier et al. 1999) are required for the generation of plateau potentials in STN neurons. These suggestions were confirmed in our study: the generation of a plateau potential was blocked by removing Ca2⫹ from the external solution (n ⫽ 3; data not shown), and nifedipine (10 M), an L-type channel blocker, strongly reduced the duration of plateau potentials (Fig. 7A; n ⫽ 5), shortening the half decay time from 0.808 ⫾ 0.185 s in control condition to 0.135 ⫾ 0.026 s (P ⬍ 0.05). It is conceivable that the current through L-type channels is involved in the maintenance of the plateau potential, but it is also possible that it is the increase in intracellular Ca2⫹ concentration that is important for the maintenance of plateau potentials because intracellular Ca2⫹ activates a number of ion channels of slow kinetics. Actually, a Ca2⫹-dependent cation conductance has been suggested to be involved in the plateau potential (Beurrier et al. 1999). We examined the effect of intracellular Ca2⫹ by including a high concentration (20 mM) of the Ca2⫹ chelating reagent bis-(o-aminophenoxy)N,N,N⬘,N⬘-tetraacetic acid (BAPTA) in the internal recording solution. Immediately after going from the cell-attached to the whole cell configuration, plateau potentials elicited with a depolarizing pulse had half decay times of 0.34 ⫾ 0.01 s (Fig. 7B, top). Over time, the duration of plateau potentials gradually increased, reaching a stable value at ⬃10 min of whole cell recording (Fig. 7B, bottom; n ⫽ 5). The half decay time at 10 min of whole cell recording was 0.65 ⫾ 0.01 s, which is significantly longer than the value on establishment of the whole cell configuration (P ⬍ 0.05). In addition to the elongation of the duration of the plateau potential, a reduction in plateau amplitude was also noticed (Fig. 7B). These results suggest that a Ca2⫹-dependent outward current and possibly inward currents are involved in the maintenance of the plateau potential. To search for conductances involved in the repolarization of plateau potentials, we tested the effect of TEA. As shown in Fig. 7C (top), a plateau potential was induced in the presence of TTX (1 M), a result in agreement with a previous finding (Beurrier et al. 1999); addition of TEA (10 mM) to the bath solution increased the duration of the plateau potential (Fig. 7C, bottom). The half decay time was significantly increased by TEA from 0.49 ⫾ 0.05 to 0.79 ⫾ 0.08 s (P ⬍ 0.05, n ⫽ 5). Taken together with the effects of intracellular Ca2⫹ chelation, these results suggest that the repolarization of plateau potentials is mediated by both Ca2⫹-dependent K⫹ channels and TEA-sensitive K⫹ channels. DISCUSSION In the present study, we investigated how intrinsic membrane properties of STN neurons may interact with synaptic inputs. We have found that activation of excitatory synaptic inputs evoked a plateau potential, in a voltage-dependent manner. Plateau potentials, evoked with a depolarizing pulse, were observed in about half of STN neurons. L-type Ca2⫹ channels, Ca2⫹-dependent K⫹ channels, and TEA-sensitive K⫹ channels were found to be involved in the generation of the potential. We conclude that EPSPs, evoked at a hyperpolarized state, activate L-type Ca2⫹ channels and other channels, leading to the generation of a plateau potential. Thus about half of STN neurons can transform short-lasting synaptic excitation into a long train of output spikes, in a voltage-dependent manner. FIG. 7. Ionic mechanisms of the plateau potential. A: current pulse (50pA; bottom)-evoked plateau potentials (top) were suppressed by application of nifedipine (10 M; middle). B: chelation of intracellular Ca2⫹ with bis-(o-aminophenoxy)-N,N,N⬘,N⬘tetraacetic acid (BAPTA) increased the duration of a plateau potential. Top: a plateau potential evoked with a current pulse (bottom) immediately after establishment of the whole cell recording configuration. The pipette solution contained 20 mM BAPTA. Middle: the recording 10 min after establishment of the whole cell recording configuration in the same neuron. C: plateau potentials were observed in the presence of TTX (1 M). Application of TEA (50 mM) greatly enhanced the duration of the plateau potential. J Neurophysiol • VOL 86 • OCTOBER 2001 • www.jn.org PLATEAU POTENTIALS OF SUBTHALAMIC NEURONS Ionic mechanisms Although the occurrence of a plateau potential in STN neurons has been reported before (Beurrier et al. 1999; Nakanishi et al. 1987; Overton and Greenfield 1995), we have shown for the first time that a plateau potential in STN neurons can be induced by activation of synaptic inputs from hyperpolarized membrane potentials. The EPSPs recorded here were evoked by stimulation of a site rostral to the STN. Because axon collaterals of fibers descending in the cerebral peduncle enter the STN from the rostral aspect (Iwahori 1978), the EPSPs are possibly of cortical origin, although we cannot exclude the contribution of parafascicular thalamic origin. Previous in vivo studies have found that STN neurons exhibit prolonged depolarizations in response to stimulation of the cortex (Fujimoto and Kita 1993; Kitai and Deniau 1981). These depolarizations, however, appear different from plateau potentials in nature, because the duration of the depolarizations is only ⬃20 ms. We have reported in abstract form that plateau potentials in STN neurons can be evoked only when the membrane potential is hyperpolarized (Otsuka et al. 1998, 1999; Song et al. 1998). Similar results were reported by Beurrier et al. (1999), although there was a quantitative difference: in Beurrier et al. (1999) plateau potentials were observed in a voltage range from ⫺50 to ⫺70 mV, while in the present study, STN neurons had to be hyperpolarized to about ⫺75 mV for the plateau potential to occur. Why then does the membrane potential have to be hyperpolarized for plateau potentials to occur? One possibility is that hyperpolarization deinactivates voltage-dependent channels that are involved in the generation of plateau potentials. T-type Ca2⫹ channels, which are present in STN neurons (Beurrier et al. 1999, 2000; Nakanishi et al. 1987; Song et al. 2000), are well known to be inactivated at resting membrane potentials and deinactivated at hyperpolarized potentials (Carbone and Lux 1984; Mouginot et al. 1997). However, because of the transient nature of T-type current, it is expected that T-type channels may play a role in the induction, but not the maintenance, of plateau potential. What channels are then activated to produce a plateau potential? In principle, the occurrence of a plateau potential requires the steady-state current-voltage (I-V) curve to cross zero current with a negative slope. Thus in STN neurons, synaptic potentials have to activate an inward current that decays only slowly for plateau potentials to occur. The channel giving rise to this current has to be again inactivated (or partially inactivated) at the resting membrane potential and deinactivated at hyperpolarized potentials. Possible candidates for such channels include highthreshold Ca2⫹ channels and probably noninactivating Na⫹ channels. Ca2⫹-dependent cation channels may also be involved if Ca2⫹ channels are activated. Although a persistent, TTX-sensitive Na⫹ channel has been reported in STN neurons (Beurrier et al. 2000; Bevan and Wilson 1999), it does not appear to be essential for plateau potentials to occur because plateau potentials in STN neurons could be evoked in the presence of TTX (see Fig. 7) (Beurrier et al. 1999). Plateau potentials have also been found and well studied in motoneurons and interneurons of the spinal cord (reviewed in Kiehn 1991). In the spinal cord, L-type Ca2⫹ channels can impart a negative slope region to the steady-state I-V curve (Hounsgaard and Kiehn 1989; Svirskis and Hounsgaard 1997). Nakanishi et J Neurophysiol • VOL 1823 al. first showed that plateau potentials in STN neurons were Ca2⫹ dependent (Nakanishi et al. 1987). After that, Beurrier et al. identified the Ca2⫹ channel to be of the L type (Beurrier et al. 1999), and this was confirmed in the present study. In a computer simulation study, we found that the voltage dependence of plateau potential induction can be solely attributed to voltage-dependent inactivation of L-type Ca2⫹ channel (Otsuka et al. 2000). We, therefore propose that the voltage dependence of L-type channels plays an important role in the voltage-dependent induction of plateau potentials in STN neurons. Besides carrying an inward current, Ca2⫹ influx may also cause activation of Ca2⫹-dependent inward currents. Actually, Beurrier et al. have found that chelating intracellular Ca2⫹ greatly shortened the duration of STN plateau potentials and concluded that a Ca2⫹-dependent cation channel is activated and is essential for the maintenance of plateau potentials (Beurrier et al. 1999). Chelating intracellular Ca2⫹, however, enhanced the duration of plateaus in the present study (see Fig. 7). Because a Ca2⫹-dependent K⫹ current is obvious in STN neurons (e.g., Fig. 1), enhancement of plateau duration by chelating intracellular Ca2⫹ is expected. The reduction in plateau amplitude by intracellular BAPTA observed in the present study, however, might be attributable to the blocking of Ca2⫹dependent cation channels. Subcellular origin of the plateau potential Although plateau potentials have been reported in STN neurons for a long time, it is not known whether the potential is generated in the dendrites or soma. If dendrites are the origin of plateau potentials, then plateau potentials of different durations may be generated in several dendrites of the same neuron. The integrated potential at the soma could thus be a step-wise potential (but this may not occur in slice preparation, where dendrites are not all preserved) (Reuveni et al. 1993) and would produce complex spiking behavior. In contrast, if the soma is the origin, STN neurons would show simpler discharge patterns. Thus the site of origin of plateau potentials has significance in synaptic integration and in regulating firing behavior. Our strategy for addressing this issue took advantage of the fact that excitatory synaptic inputs to STN neurons end on distal dendrites. Thus if a plateau potential occurs in dendrites, voltage-clamping the soma would still allow the plateau potential to be evoked in dendrites by the synaptic input. In this case, an inward current similar in shape to the plateau is expected to be recorded at the soma. Because such currents were never observed, plateau potentials in STN neurons appear to be generated in the soma. This inference is consistent with previous findings that L-type Ca2⫹ channels, which are essential for plateau potentials, tend to be localized in the soma (Hell et al. 1993). Nevertheless other possibilities cannot be excluded at this time because our current argument is based on negative findings. Functional significance The STN receives excitatory inputs from the cortex (Fujimoto and Kita 1993; Hartmann-von Monakow et al. 1978; Kitai and Deniau 1981; Nambu et al. 1996). Because of the 86 • OCTOBER 2001 • www.jn.org 1824 T. OTSUKA, F. MURAKAMI, AND W.-J. SONG voltage dependence of the generation of plateau potential, cortical input to the STN will be transformed into either a single or a train of action potentials, depending on the membrane potential of STN neurons at the time of the input. What may hyperpolarize the cell for a plateau potential to occur? Opening of K⫹ channels by metabolic signaling pathways is one possibility. High-frequency inhibitory input from, for example, the globus pallidus, would also help hyperpolarize the cell. It is an important task to identify the pathway that hyperpolarizes STN neurons in future studies. Because a plateau potential is more or less a long-lasting, stereotyped potential, generation of a plateau potential would lead to a reduction in spatiotemporal integrative ability of the cell. The benefit of using a plateau potential, in addition to transforming a single input volley into a train of output spikes, may lie in the generation of rhythmic bursting in STN-related neuronal networks. A previous study has shown that STN neurons “switch from single-spike activity to burst-firing mode” and suggested plateau potentials are involved in the generation of rhythmic burst activity (Beurrier et al. 1999). In our experiments, rhythmic burst activity was seldom observed. This is consistent with the recent observation in explant culture that STN neurons exhibit pacemaker activity when STN is in the STN-GP network but not when STN is alone (Plenz and Kitai 1999). Our observation that plateau potentials were observed in about half of STN neurons is in coincidence with the observation by Plenz and Kitai that about half of STN neurons exhibits pacemaker activity (Plenz and Kitai 1999). This coincidence suggests the possibility that plateau potentials in STN neurons are involved in the generation of oscillatory bursting in the STN-GP network. Two features of STN plateau potentials may be relevant. First, because a plateau potential can be evoked as a rebound potential (see Fig. 2) (Beurrier et al. 1999; Overton and Greenfield 1995), a short train of spikes in GP neurons would hyperpolarize STN neurons and a plateau potential would then occur as a rebound potential, evoking a train of spikes in STN neurons. Second, STN activity would cause immediate feedback inhibition from the GP, but this inhibition might not immediately terminate STN spiking activity because the early part of plateau potentials appears to be resistant to inhibitory perturbations. We thank S. Maeda for technical assistance in data analysis. This work was supported by the Uehara Foundation, Grants-in-Aid for Scientific Research on Priority Areas (Grants 12210099, 12053247, and 11170232) from the Ministry of Education, Science, and Culture, Japan to W.-J. Song, and a grant from Core Research for Evolutional Science and Technology, Japan Science and Technology Corporation to F. Murakami. REFERENCES BEURRIER C, CONGAR P, BIOULAC B, AND HAMMOND C. Subthalamic nucleus neurons switch from single-spike activity to burst-firing mode. J Neurosci 19: 599 – 609, 1999. BEURRIER C, CONGAR P, BIOULAC B, AND HAMMOND C. Slowly inactivating sodium current (INaP) underlies single-spike activity in rat subthalamic neurons. J Neurophysiol 83: 1951–1957, 2000. BEVAN MD, FRANCIS CM, AND BOLAM JP. The glutamate-enriched cortical and thalamic input to neurons in subthalamic nucleus of rat: convergence with GABA-positive terminals. J Comp Neurol 361: 491–511, 1995. BEVAN MD AND WILSON CJ. Mechanisms underlying spontaneous oscillation and rhythmic firing in rat subthalamic neurons. J Neurosci 19: 7617–7628, 1999. CARBONE E AND LUX HD. A low voltage-activated calcium conductance in embryonic chick sensory neurons. Biophys J 46: 413– 418, 1984. J Neurophysiol • VOL EDWARDS FA, KONNERTH A, SAKMANN B, AND TAKAHASHI T. A thin slice preparation for patch clamp recordings from neurones of the mammalian central nervous system. Pflügers Arch Arch 414: 600 – 612, 1989. FÉGER J, BEVAN M, AND CROSSMAN AR. The projections from the parafasicular thalamic nucleus to the subthalamic nucleus and the striatum arose from separate neuronal populations: a comparison with the corticostriatal and corticosubthalamic efferents in a retrograde fluorescent double-labeling study. Neuroscience 60: 125–132, 1994. FUJIMOTO K AND KITA H. Response characteristics of subthalamic neurons to the stimulation of the sensorimotor cortex in the rat. Brain Res 609: 185–192, 1993. GÖTZ T, KRAUSHAAR U, GEIGER J, LUBKE J, BERGER T, AND JONAS P. Functional properties of AMPA and NMDA receptors expressed in identified types of basal ganglia neurons. J Neurosci 17: 204 –215, 1997. GROENEWEGEN H AND BERENDSE HW. Connections of the subthalamic nucleus with ventral striatopallidal parts of the basal ganglia in the rat. J Comp Neurol 294: 607– 622, 1990. HAMADA I AND HASEGAWA N. Disturbance in task performance after inhibition of subthalamic nucleus neurons. In: The Basal Ganglia V, edited by Ohye C, Kimura M, and McKenzie JS. New York: Plenum, 1996, p. 225–229. HARTMANN-VON MONAKOW K, AKERT K, AND KUNZLE H. Projections of the precentral motor cortex and other cortical areas of the frontal lobe to the subthalamic nucleus in the monkey. Exp Brain Res 33: 395– 403, 1978. HELL JW, WESTENBROEK RE, WARNER C, AHLIJANIAN MK, PRYSTAY W, GILBERT MM, SNUTCH TP, AND CATTERALL WA. Identification and differential subcellular localization of the neuronal class C and class D L-type calcium channel alpha 1 subunits. J Cell Biol 123: 949 –962, 1993. HIRANO T, KUBO Y, AND WU MM. Cerebellar granule cells in culture: monosynaptic connections with Purkinje cells and ionic currents. Proc Natl Acad Sci USA 83: 4957– 4961, 1986. HORIKAWA H AND ARMSTRONG WE. A versatile means of intracellular labeling: injection of biocytin and its detection with avidin conjugates. J Neurosci Methods 25: 1–11, 1988. HOUNSGAARD J AND KIEHN O. Serotonin-induced bistability of turtle motoneurons induced by a nifedipine-sensitive calcium plateau potential. J Neurophysiol 414: 265–282, 1989. IWAHORI N. A Golgi study on the subthalamic nucleus of the cat. J Comp Neurol 182: 383–398, 1978. KIEHN O. Plateau potentials and active integration in the “final common pathway” for motor behavior. Trends Neurosci 14: 68 –73, 1991. KITA H, CHANG HT, AND KITAI ST. Pallidal inputs to subthalamic: intracellular analysis. Brain Res 264: 255–265, 1983. KITAI ST AND DENIAU JM. Cortical inputs to the subthalamus: intracellular analysis. Brain Res 214: 411– 415, 1981. KITAI ST AND KITA H. Anatomy and physiology of the subthalamic nucleus: a driving force of the basal ganglia. In: The Basal Ganglia II, edited by Carpenter MB and Jayaraman A. New York: Plenum, 1987, p. 357–373. MAGILL PJ, BOLAM JP, AND BEVAN MD. Relationship of activity in the subthalamic nucleus-globus pallidus network to cortical electroencephalogram. J Neurosci 20: 820 – 833, 2000. MATSUMURA M, KOJIMA J, GARDINER TW, AND HIKOSAKA O. Visual and oculomotor function of monkey subthalamic nucleus. J Neurophysiol 67: 1615–1632, 1992. MAYER ML AND WESTBROOK GL. Permeation and block of N-methyl-Daspartic acid receptor channels by divalent cations in mouse cultured central neurons. J Physiol (Lond) 394: 501–527, 1987. MORIIZUMI T AND HATTORI T. Separate neuronal populations of the rat globus pallidus projecting to the subthalamic nucleus, auditory cortex and pedunculopontine tegmental area. Neuroscience 46: 701–710, 1992. MOUGINOT D, BOSSU J-L, AND GAHWILER BH. Low-threshold Ca2⫹ currents in dendritic recordings from Purkinje cells in rat cerebellar slice cultures. J Neurosci 17: 160 –170, 1997. MOUROUX M AND FÉGER J. Evidence that the parafascicular projection to the subthalamic nucleus is glutamatergic. Neuroreport 4: 613– 615, 1993. NAKANISHI H, KITA H, AND KITAI ST. Electrical membrane properties of rat subthalamic neurons in an in vitro slice preparation. Brain Res 437: 35– 44, 1987. NAMBU A, TAKADA M, INASE M, AND TOKUNO H. Dual somatotopical representations in the primate subthalamic nucleus: evidence for ordered but reversed body-map transformations from the primary motor cortex and supplementary motor area. J Neurosci 16: 2671–2683, 1996. OTSUKA T, ABE T, TSUKAGAWA T, AND SONG W-J. Single compartment model of the voltage-dependent generation of a plateau potential in subthalamic neurons (Abstract). Neurosci Res Suppl 24: S81, 2000. 86 • OCTOBER 2001 • www.jn.org PLATEAU POTENTIALS OF SUBTHALAMIC NEURONS OTSUKA T, SONG W-J, AND MURAKAMI F. Voltage-dependent generation of a plateau potential in rat subthalamic nucleus neurons (Abstract). Neurosci Res Suppl 23: S44, 1999. OTSUKA T, SONG W-J, AND MURAKAMI F. Generation of a plateau potential at hyperpolarized potentials in rat subthalamic nucleus neurons (Abstract). Neurosci Res Suppl 22: 175, 1998. OVERTON PG AND GREENFIELD SA. Determinants of neuronal firing pattern in the guinea-pig subthalamic nucleus: an in vivo and in vitro comparison. J Neural Transm 10: 41–54, 1995. PLENZ D AND KITAI ST. A basal ganglia pacemaker formed by the subthalamic nucleus and external globus pallidus. Nature 400: 677– 682, 1999. REUVENI I, FRIEDMAN A, AMITAI Y, AND GUTNICK MJ. Stepwise repolarization from Ca2⫹ plateaus in neocortical pyramidal cells: evidence for nonhomogeneous distribution of HVA Ca2⫹ channels in dendrites. J Neurosci 13: 4609 – 4621, 1993. SHEN K-Z AND JOHNSON ST. Presynaptic dopamine D2 and muscarine M3 receptors inhibit excitatory and inhibitory transmission to rat subthalamic neurons in vitro. J Physiol (Lond) 525: 331–341, 2000. SONG W-J, BABA Y, OTSUKA T, AND MURAKAMI F. Characterization of Ca2⫹ J Neurophysiol • VOL 1825 channels in rat subthalamic nucleus neurons. J Neurophysiol 84: 2630 – 2637, 2000. SONG W-J, OTSUKA T, AND MURAKAMI F. Hyperpolarization changes the firing mode of rat subthalamic nucleus neurons by enabling generation of a plateau potential. Soc Neurosci Abstr 24: 1642, 1998. STUART GJ, DODT HU, AND SAKMANN B. Patch-clamp recordings from the soma and dendrites of neurons in brain slices using infrared video microscopy. Pflügers Arch Arch 423: 511–518, 1993. SVIRSKIS G AND HOUNSGAAED J. Depolarization-induced facilitation of a plateau-generating current in ventral horn neurons in the turtle spinal cord. J Neurophysiol 78: 1740 –1742, 1997. WHITTIER JR. Ballism and the subthalamic nucleus (nucleus hypothalamicus; corpus luysi). Arch Neurol Psychiatry 58: 672– 692, 1947. WICHMANN T, BERGMAN H, AND DELONG MR. The primate subthalamic nucleus. I. Functional properties in intact animals. J Neurophysiol 72: 494 –506, 1994a. WICHMANN T, BERGMAN H, AND DELONG MR. The primate subthalamic nucleus. III. Changes in motor behavior and neuronal activity in the internal pallidum induced by subthalamic inactivation in the MPTP model of Parkinsonism. J Neurophysiol 72: 521–530, 1994b. 86 • OCTOBER 2001 • www.jn.org