Spontaneous firing and evoked pauses in the tonically active cholinergic interneurons of the striatum

Neuroscience 198 (2011) 27– 43

REVIEW SPONTANEOUS FIRING AND EVOKED PAUSES IN THE TONICALLY ACTIVE CHOLINERGIC INTERNEURONS OF THE STRIATUM J. A. GOLDBERGa* AND J. N. J. REYNOLDSb

Key words: TAN, autonomous discharge, acetylcholine dopamine balance, synaptic plasticity, neuromodulation, reward signaling.

a

Department of Physiology, Feinberg School of Medicine, Northwestern University, Chicago, IL 60611, USA b Department of Anatomy and the Brain Health Research Centre, University of Otago, PO Box 913, Dunedin 9054, New Zealand

Contents Tonically active neurons of the striatum in behaving primates The role of dopamine in the conditioned pause response Spontaneous firing patterns and synchronization of TANs TANs are mostly cholinergic interneurons Cellular mechanisms of spontaneous discharge Rhythmic single spiking Rhythmic bursting Irregular single spiking Synchronization of background discharge and the pause response Dopaminergic inputs to cholinergic interneurons Cortical inputs to cholinergic interneurons Thalamic inputs to cholinergic interneurons Striatal interneuronal inputs to cholinergic interneurons Which input is responsible for the baseline synchronization? So. . . what causes the pauses in cholinergic interneurons? Dopamine sensitivity of the pause response Plasticity of the pause response Synaptic mechanisms Postsynaptic excitability Role of cholinergic interneurons and their pause in striatal function Nicotinic ACh receptors Muscarinic ACh receptors affect striatal circuitry at multiple time scales Functional implications of the pause response Dynamics of pause generation Conclusions Acknowledgments References

Abstract—The tonically active neurons (TANs) are a population of neurons scattered sparsely throughout the striatum that show intriguing patterns of firing activity during reinforcement learning. Following repeated pairings of a neutral stimulus with a primary reward, TANs develop a transient cessation of firing activity in response to the stimulus, termed the “conditioned pause response.” In tasks where specific cues are arranged to signal the probability of particular outcomes, the pause response to both cue and outcome may differ in ways that suggest the involvement of different inputs to the same neuron. Here we review the cellular properties of cholinergic interneurons and describe the response to their afferents in terms of inducing TAN-like pauses in tonic firing. Recent work has shown that thalamostriatal inputs to cholinergic neurons transiently suppress firing activity via dopamine release. Because these pauses are initiated by subcortical pathways with limited sensory processing abilities, we propose that they are an ideal correlate for the pauses observed in TANs in response to cues signaling trial initiation. On the other hand, pauses that accompany outcome presentation contain higher-level information, including an apparent sensitivity to reward prediction error. Thus, these pauses may be mediated by cortical inputs to cholinergic interneurons. Although there is evidence linking cholinergic pauses to synaptic plasticity, much remains to be discovered about the effect of this relatively sparse but influential population on the striatal learning system. This article is part of a Special Issue entitled: Function and Dysfunction of the Basal Ganglia. © 2011 IBRO. Published by Elsevier Ltd. All rights reserved.

27 28 28 29 29 29 31 31 31 31 31 31 32 32 32 33 35 35 35 35 35 36 37 39 39 40 40

TONICALLY ACTIVE NEURONS OF THE STRIATUM IN BEHAVING PRIMATES Extracellular recording in the striatum of awake behaving primates reveals the presence of tonically active neurons (TANs) that possess particularly broad action potentials (Anderson, 1978; Kimura et al., 1984; Aosaki et al., 1994b). TANs were initially shown not to be responsive to movement per se (Crutcher and DeLong, 1984). Instead, upon presentation of primary reward (e.g. a drop of liquid) to a primate, TANs respond with a pause in their tonic firing that lasts a few hundred milliseconds (Kimura et al., 1984; Apicella et al., 1997). The pause response is often preceded and/or followed by an excitation. A subsequent study showed that these neurons acquire this pause in

*Corresponding author. Tel: ⫹1-312-503-1139. E-mail address: [email protected] (J. A. Goldberg). Abbreviations: ACh, acetylcholine; AHP, afterhyperpolarization; BK, big conductance calcium and voltage activated potassium; ChAT, choline acetyltransferase; CICR, calcium induced calcium release; CM-PF, centromedian-parafascicular; DRT, dopamine replacement therapy; GABA, ␥-aminobutyric acid; HCN, hyperpolarization-activated, cyclic nucleotide-gated cation; IPSC, inhibitory postsynaptic current; ISI, interspike interval; LTD, long-term depression; LTP, longterm potentiation; mAChR, muscarinic ACh receptor; mAHP, medium AHP; MPTP, 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine; nAChR, nicotinic acetylcholine receptor; PF, parafascicular; sAHP, slow AHP; SC, superior colliculus; SK, small conductance calcium-activated potassium; SNc, substantia nigra pars compacta; SPN, spiny projections neuron; TAN, tonically active neuron.

0306-4522/11 $ - see front matter © 2011 IBRO. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.neuroscience.2011.08.067

27

28

J. A. Goldberg and J. N. J. Reynolds / Neuroscience 198 (2011) 27– 43

shows positive and negative responses to stimulus delivery and omission, respectively, while the cholinergic neuron response is uniformly a pause in tonic firing at the time of expected reward (or its omission) (Fiorillo et al., 2003; Morris et al., 2004). In addition, the pause response depends to a large extent on normal striatal dopamine innervation. Direct application of dopamine receptor antagonists to TANs suppresses pause responses that have previously been acquired through reward-related learning (Watanabe and Kimura, 1998). Depletion of striatal dopamine by lesioning the dopamine projection from the SNc with 1methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) treatment causes primates to exhibit the cardinal symptoms of Parkinson’s disease (PD), including akinesia, bradykinesia, cogwheel rigidity, and tremor (Burns et al., 1983). This treatment results in a loss of learned pause responses that can be restored by exogenous dopamine agonists (Aosaki et al., 1994a). However, even an almost complete degeneration of nigrostriatal dopamine terminals fails to obliterate established pauses in all TANs, indicating that dopamine release is not necessary for the pauses, or at least that some pauses are not dopamine dependent. Instead, dopamine depletion returns the percentage of responding TANs to the same level as before conditioning (approximately 10 –20%) (Aosaki et al., 1994a).

18

Spk/s

Reward Neutral Aversive

10

2 -0.5

0

Time (s)

2

Fig. 1. Conditioned pause response of a single striatal TAN in a well-trained behaving monkey to trigger stimuli signaling different outcomes. Spike raster (upper panels) and poststimulus time histogram (PSTH, lower panel) aligned to the timing of distinct visual cues that predicted the subsequent delivery of a tone accompanied by a food reward (blue), a tone followed by neither a rewarding nor an aversive outcome (green) or a tone accompanying an aversive air puff (red). Note the similarity in response to the different visual cues, namely, a suppression in firing lasting up to 200 ms without a preceding activation, followed by a prolonged excitatory phase (Figure kindly provided by Avital Adler and Hagai Bergman).

response to sensory stimuli that become associated with the reward, and then lose the pause when this association is extinguished (Aosaki et al., 1994b). Therefore, the pause response is a neural correlate of classical conditioning, and has come to be known as the “conditioned pause response” (Fig. 1). It is now widely accepted that the pause response of the TANs is related to the detection of the motivational significance of an external stimulus, both rewarding and aversive (Ravel et al., 1999; Apicella, 2002; Morris et al., 2004; Joshua et al., 2008). The role of dopamine in the conditioned pause response Simultaneous extracellular recordings from the dopaminergic neurons of the substantia nigra pars compacta (SNc) and TANs of behaving primates have shown that the pause response is coincident with an increase in discharge of the SNc dopaminergic neurons (Morris et al., 2004). Both of these neurons respond to unexpected salient stimuli during reinforcement learning: the dopaminergic neuron

Spontaneous firing patterns and synchronization of TANs Aosaki and colleagues also described the statistics of the TANs’ spontaneous firing patterns in primates (Aosaki et al., 1994b). Their analysis of the TANs’ interspike interval (ISI) histograms demonstrated the existence of irregular tonic firing patterns, characterized by a unimodal ISI histogram, and burst firing, characterized by a bimodal distribution (the second modal representing the intraburst periods). It has been suggested that these two firing patterns represent two different classes of TANs (Apicella, 2002); however, this is unlikely, as individual TANs alternate between these two firing patterns over time (Aosaki et al., 1994b). The uniformity and temporal alignment of the pause response among TANs that are distributed widely within the striatum led Graybiel and colleagues to suggest that these neurons form synchronous cell assemblies. These assemblies presumably modulate in a spatially uniform fashion the activity of the phasically active spiny projections neurons (SPNs) that give rise to the GABAergic (GABA, ␥-aminobutyric acid) output of the striatum (Graybiel et al., 1994). Bergman and coworkers tested this idea by recording the simultaneous discharge of several TANs that were millimeters apart. They found that 90% of the cross-correlograms of pairs of TANs that paused in response to reward presentation had significant zero-lag peaks, implying synchronous discharge. Extracellular recording of the spontaneous firing of TANs in MPTP-treated primates demonstrates that the cells remain synchronized, but their firing rates become oscillatory in the 10 –20 Hz range (Raz et al., 1996; Goldberg et al., 2004). In conclusion, the baseline discharge of TANs is synchronized to a certain extent under all conditions. More importantly, a reward or stimuli associated with it cause numerous TANs to

J. A. Goldberg and J. N. J. Reynolds / Neuroscience 198 (2011) 27– 43

pause their ongoing discharge synchronously. The finding that the majority of cells discharge synchronously supports the notion that these neurons can form a single functional unit to synchronously modulate large portions of the striatal network.

TANs ARE MOSTLY CHOLINERGIC INTERNEURONS TANs recorded in primates were shown to be spatially distributed within the striatum in a fashion that most closely resembles the distribution of choline acetyltransferase (ChAT)-immunoreactive neurons in these animals (Aosaki et al., 1995), indicating that the population of TANs is composed primarily of the cholinergic interneurons of the striatum. Although it is likely that tonically active striatal neurons are mostly cholinergic interneurons, throughout this review we will continue to refer to unidentified neurons that show this firing pattern as “TANs” and reserve “cholinergic interneurons” for studies where cell identity was verified. Cholinergic interneurons were first identified as giant interneurons by Kölliker (1896). They are rare, accounting for only about 2% of the neurons in the mammalian striatum (Kemp and Powell, 1971; Bolam et al., 1984; Phelps et al., 1985). Therefore, intracellular recording studies of these neurons in vivo are few and rely on a handful of neurons that are encountered fortuitously over many years (Bishop et al., 1982; Wilson et al., 1990; Reynolds et al., 2004; Reynolds and Wickens, 2004). The earlier studies demonstrated that these cells discharge tonically in vivo at a rate of 3–10 spikes/s and continue to fire spontaneously in ipsilaterally decorticate animals. In addition, they were shown to exhibit a broad action potential followed by a deep afterhyperpolarization (AHP). The cells were tentatively identified as cholinergic using intracellular staining methods, which demonstrated that they have the same cytological features as well as synaptic distribution and morphology as the ChAT-expressing cells of the striatum (Wilson et al., 1990). Subsequent studies identified them conclusively by double labeling (Kawaguchi, 1992). The neurons possess- a large soma, two to four large primary tapering relatively aspiny dendrites that are riddled with small spine-like appendages [see Fig. 4 in Kemp and Powell (1971), and Fig. 2 in Wilson et al. (1990)]. These dendrites bifurcate repeatedly but infrequently and can extend over a range of 1 mm. The axon that arises from one of the primary dendrites branches densely and profusely over a large portion of the striatum (Fig. 2A) (Chang and Kitai, 1982; DiFiglia and Carey, 1986; DiFiglia, 1987; Kawaguchi, 1992, 1993). In the first brain slice recording study of these cells, only 11 cells from over 350 striatal cells recorded were found to be cholinergic interneurons (Jiang and North, 1991). That study described a prominent sag in voltage in response to hyperpolarizing current injections that was shown to result from the hyperpolarization-activated, cyclic nucleotide-gated cation (HCN) inward current. This study also reported that 40% of the cells were spontaneously

29

active in vitro, when recorded using conventional sharp electrodes. Thanks to their large somata, cholinergic interneurons are easily discernable against the background of the other smaller spiny neurons using infrared differential interference contrast microscopy in live striatal slices. Numerous studies have since contributed to the systematic characterization of the physiology of these cells. The first of these studies, which identified the cholinergic cells with ChAT immunoreactivity, was published by Kawaguchi (1993). In that study he described most of the salient features of these cells’ responses to current injections. The cells respond to suprathreshold depolarizing pulses with increased firing that undergoes spike frequency adaptation (Fig. 2B), and sometimes with spike threshold accommodation during the pulse. The action potentials are very broad and are followed by an AHP, called the medium AHP (mAHP), that lasts 100 –200 ms (Fig. 2C). At the end of a long depolarizing pulse, the membrane potential undergoes a long-lasting AHP, called the slow AHP (sAHP), lasting several seconds. When the cells are hyperpolarized with a current pulse they exhibit a rapid hyperpolarization followed by the above-mentioned sag due to HCN currents. After the hyperpolarizing pulse there is a rebound of increased firing, but then the cell relaxes back to spontaneous discharge (Fig. 2B) (Bennett and Wilson, 1999; Reynolds et al., 2004; Wilson, 2005).

CELLULAR MECHANISMS OF SPONTANEOUS DISCHARGE The cellular mechanism underlying the rich repertoire of spontaneous firing patterns has been the subject of intense research in the past decade. Wilson and coworkers described three firing patterns in cholinergic interneurons: irregular discharge and two periodic firing patterns, namely, rhythmic single spiking and rhythmic burst firing (Fig. 2D) (Bennett and Wilson, 1999). Two lines of evidence demonstrate that these discharge patterns are autonomously generated, that is, independent of synaptic input. The first is their insensitivity to blockade of a wide variety of synaptic transmitter receptors (Bennett and Wilson, 1999). The second is the mechanistic description of the membrane currents and calcium dynamics that give rise to autonomous activity (Bennett et al., 2000). Following is a brief overview of these mechanisms. A detailed review of these issues appears elsewhere (Goldberg and Wilson, 2010). Rhythmic single spiking The cycle of the rhythmic single spiking is generated as follows. Persistent sodium currents via NaV1.6 channels (Maurice et al., 2004) prevent a stable resting potential and drive the membrane potential to the action potential threshold. A sodium action current then depolarizes the cell, leading to an influx of calcium via high voltage-activated CaV2 calcium channels. The CaV2.1 (primarily of the Qtype) current activates big conductance calcium- and volt-

30

J. A. Goldberg and J. N. J. Reynolds / Neuroscience 198 (2011) 27– 43

C 40 mV

B

A

20 mV

-60 mV

2 ms 1s

+40 pA -40 pA

20 µm

E

D

400 ms 50 pA 0 mV

-50 mV -60 mV -70 mV pA 0

-70

-60

-50

-40

mV

-20

20 mV 5 s

-40 -60

Fig. 2. Morphology and electrophysiological characteristics of cholinergic interneurons. (A) A Z-axis projection of an Alexa 568 (Invitrogen, Carlsbad, CA, USA) filled mouse cholinergic interneuron imaged with two-photon laser scanning microscopy. Note the spine-like appendages that cover the dendrites of these aspiny interneurons. (B) Response to intracellular current injection of a different cholinergic interneuron recorded using whole-cell patch-clamp technique. Hyperpolarizing current injection results in a rapid hyperpolarization due to the action of the inward rectifying potassium current and a depolarizing “sag” followed by rebound spike firing due to the action of the HCN current (red trace). Suprathreshold depolarizing current injection (1 s) induces prolonged firing with spike frequency adaptation (blue trace). Termination of the burst is followed by a modest sAHP, which lasts more than a second before returning to baseline tonic firing. (C) The action potential of cholinergic interneurons is very broad and shown here in comparison to an action potential of a GABAergic interneuron. (D) Examples of slightly irregular single spiking and rhythmic bursting patterns in cholinergic interneurons. (E) Steady-state currents recorded in voltage clamp in response to voltage pulses to ⫺70 mV (blue), ⫺60 mV (green), and ⫺50 mV (red) reveal a nonmonotonic current voltage relation. The current-voltage (IV) curve for the range of ⫺70 mV to ⫺50 mV reveals a negative conductance region. This IV curve belongs to the cell whose single spiking is depicted in the left side of panel (D).

age-activated potassium (BK) channels, which participate, alongside delayed-rectifier potassium currents, in repolarizing the cell, and in determining action potential width. The CaV2.2 (N-type) current activates the small conductance calcium-activated potassium (SK) current that gives rise to

the mAHP (Bennett et al., 2000; Goldberg and Wilson, 2005). KV4 channels are activated during the mAHP and also affect its trajectory (Song et al., 1998; Hattori et al., 2003). Finally, HCN currents depolarize the cell to approximately ⫺60 mV, above which the persistent sodium cur-

J. A. Goldberg and J. N. J. Reynolds / Neuroscience 198 (2011) 27– 43

rent takes over (Bennett et al., 2000; Deng et al., 2007; Oswald et al., 2009). Rhythmic bursting A large SK current prevents the appearance of bursting in cholinergic interneurons because the mAHP it generates after each action potential hyperpolarizes the cell sufficiently to (a) de-inactivate the persistent sodium current, and (b) prevent a sustained depolarization, which is required to activate the IsAHP, the yet unidentified current that underlies the sAHP, that is necessary to terminate the burst. Thus a prerequisite for bursting in these cells— either endogenous or induced by the SK channel blocker, apamin—is a significantly reduced mAHP, which allows the cell to produce a rapid succession of action potentials that make up the burst (compare the sizes of the mAHPs shown in Fig. 2D). Hence, the two rhythmic firing patterns, single-spiking and bursting, are mutually exclusive. Which pattern is exhibited at any given moment depends on the size of the SK current. The cycle of intrinsic bursting is as follows: (a) during the burst and before it there is a subthreshold buildup in the cell of calcium entering through CaV1 (L-type) channels; (b) the CaV1 current induces calcium-induced calcium release (CICR) from intracellular stores that activates IsAHP; (c) IsAHP terminates the burst and triggers an sAHP; (d) inward-rectifying potassium currents amplify the sAHP into a rapid and prolonged hyperpolarization, during which calcium concentrations drop; (e) HCN currents, followed by persistent sodium currents, depolarize the cell so that it approaches action potential threshold. Another rapid succession of action potentials ensues (again unhampered due to the diminished mAHPs), which makes up the next burst and initiates the next cycle of calcium buildup in the cell (Bennett et al., 2000; Goldberg and Wilson, 2005; Wilson, 2005; Wilson and Goldberg, 2006; Goldberg et al., 2009). Irregular single spiking Two mechanisms for irregular firing have been proposed (Goldberg and Wilson, 2010). The first is sodium dependent: the tetrodotoxin (TTX)-sensitive persistent sodium current creates a negative conductance region in the steady-state current voltage (IV) curve, which passes a few picoamperes (or less) negative to the zero current line (Fig. 2E) (Bennett et al., 2000). As the cell depolarizes to the vicinity of ⫺60 mV, because of the proximity to zero current line, the total ionic current available to depolarize the cell is small, leading potentially to lingering of the voltage around ⫺60 mV. During the lingering, the voltage trajectory becomes more sensitive to noise from which irregularity can arise. The calcium-dependent mechanism is attributable to the subthreshold calcium entry via CaV1 channels, which can also trigger CICR (Goldberg and Wilson, 2005). During tonic firing, many cells that fire irregularly exhibit subthreshold calcium entry, whereas regular ones do not. This subthreshold calcium entry interferes with the mechanism of rhythmic single spiking and introduces irregularity (Goldberg et al., 2009). These sodium-

31

and calcium-dependent mechanisms contribute to the irregularity observed in the leftmost trace of Fig. 2D. Just as observed in extracellular TAN recordings (Aosaki et al., 1994b), cholinergic interneurons recorded in vitro wander among the various firing patterns in the absence of synaptic input (Wilson, 2005). Why this happens is still unknown, but could arise from changes in neuromodulatory tone or intrinsic fluctuations in intracellular signaling proteins.

SYNCHRONIZATION OF BACKGROUND DISCHARGE AND THE PAUSE RESPONSE In order to speculate about the mechanism that synchronizes the background discharge of TANs, as well as their concerted pauses, we need to consider the possibility of correlated inputs, and the possibility of effective recurrent connections among the cholinergic interneurons. Let us begin by briefly reviewing the afferent inputs to cholinergic interneurons. Dopaminergic inputs to cholinergic interneurons The cell bodies of dopamine neurons that project to the dorsal and ventral striatum are situated primarily in the SNc and ventral tegmental area (VTA). Single dopamine axons arborize extensively throughout a large volume of the striatum (Prensa and Parent, 2001). Their axonal branches surround cholinergic interneurons in a distribution suitable for volume transmission of dopamine, as well as forming infrequent synaptic connections on their somata and proximal dendrites (Dimova et al., 1993). Thus, the dopamine system is appropriately situated to play a direct role in pause generation in widely distributed cholinergic interneurons. Cortical inputs to cholinergic interneurons Cortical inputs have been difficult to demonstrate anatomically, probably because their synaptic connections are located in a dendritic region that is difficult to highlight using standard immunohistochemical procedures (Lapper and Bolam, 1992). Using antibodies against M2 muscarinic receptors in conjunction with a cortical lesion, Thomas et al. (2000) demonstrated that degenerating cortical terminals made synaptic contact with the distal extremities of the dendrites of large M2-immunoreactive striatal neurons, which have been shown to be cholinergic interneurons (Alcantara et al., 2001). Although cortical synaptic contacts have now been demonstrated, they are still relatively rare in comparison with the numerous thalamic synapses located more proximally on the dendrites of cholinergic interneurons (Lapper and Bolam, 1992; Dimova et al., 1993; Thomas et al., 2000). Thus, thalamic inputs, to which we turn next, presumably exert a greater excitatory influence on cholinergic interneurons in total than do cortical inputs (Sidibé and Smith, 1999). Thalamic inputs to cholinergic interneurons Cholinergic interneurons receive rich synaptic input from the intralaminar nucleus of the thalamus. In primates, this

32

J. A. Goldberg and J. N. J. Reynolds / Neuroscience 198 (2011) 27– 43

innervation originates in the centromedian-parafascicular (CM-PF) nuclear complex, homologous to the lateral and medial parafascicular (PF) nuclei in the rat (Smith et al., 2004). These areas project in a topographic manner to the striatum, such that in the rat, the lateral PF nucleus projects to lateral sensorimotor areas of the dorsal striatum (Berendse and Groenewegen, 1990). Single thalamostriatal neurons from the PF nucleus arborize extensively within the striatum, innervating a considerable territory (Deschênes et al., 1996). Terminals of thalamic neurons form abundant asymmetric synapses, which are glutamatergic and excitatory on the somata and on proximal and distal dendrites of cholinergic interneurons (Lapper and Bolam, 1992; Mouroux and Féger, 1993; Bennett and Wilson, 1998). Striatal interneuronal inputs to cholinergic interneurons Cholinergic interneurons receive GABAergic input mediated by GABAA receptors (Chang and Kitai, 1982; DiFiglia and Carey, 1986; DiFiglia, 1987; Bennett and Wilson, 1998) from the GABAergic striatal interneurons (Sullivan et al., 2008) and possibly from axon collaterals of SPNs as well. The GABAergic interneurons also provide two mechanisms of lateral inhibition among cholinergic interneurons. Firstly, action potentials in one cholinergic interneuron create polysynaptic inhibitory postsynaptic currents (IPSCs) in itself and in other cholinergic interneurons, and transiently suppress ongoing discharge in these neurons. The IPSCs are polysynaptic and are nicotinic acetylcholine receptor (nAChR) dependent. This latter dependency is conferred upon them because the GABAergic interneurons that generate the IPSCs express nACHRs (Sullivan et al., 2008). Secondly, application of muscarinic agonists can silence cholinergic interneurons, and focal stimulation of the slice can induce what have been described as muscarinic IPSPs in these neurons. Both these effects are mediated by postsynaptic M2 receptors (Calabresi et al., 1998b; Bonsi et al., 2008). These findings imply that cholinergic interneurons possess a negative feedback mechanism to control their own discharge rate, both individually and collectively as a network (see section “Muscarinic ACh receptors affect striatal circuitry at multiple time scales”). Which input is responsible for the baseline synchronization? The cholinergic interneurons clearly receive inputs from multiple regions and may also be reciprocally coupled via muscarinic autoreceptors, or disynaptically via GABAergic interneurons. This means that their synchrony could arise from multiple sources. We currently do not know how the baseline synchrony comes about, and why dopamine depletion increases the degree of synchrony. However, the activity of TANs is synchronized to a certain degree to pallidal border cells (Raz et al., 2001) (which may be part of the basal forebrain cholinergic group). In addition, spiking activity in some cholinergic interneurons shows variable coherence with local field potential activity in the cortex (Schulz et al., 2011). In MPTP-induced parkinson-

ism, TANs become significantly synchronized to global oscillations present in the local field potential in cortex and the basal ganglia (Goldberg et al., 2004). These findings suggest that dopamine depletion may render extrinsic inputs more effective at driving membrane potential fluctuations of distributed cholinergic interneurons in a coordinated fashion, but this issue awaits further investigation.

SO. . . WHAT CAUSES THE PAUSES IN CHOLINERGIC INTERNEURONS? Inactivation of CM-PF with muscimol markedly attenuates the pause response and the ensuing excitation in TANs in response to a trigger stimulus in a classical conditioning task (Matsumoto et al., 2001), indicating that the thalamic input is required for the full expression of the pause response under these circumstances. Because thalamic inputs to striatal neurons are glutamatergic, it follows that these excitatory inputs must be translated into inhibitory effects on the discharge of cholinergic interneurons. Given the various synaptic inputs on to cholinergic interneurons, several mechanisms can be envisioned as described below: Firstly, GABAergic interneurons driven by thalamic excitation may mediate feedforward inhibition of cholinergic interneurons. This is supported by the finding that intense stimulation of excitatory fibers close to striatum produces a disynaptic inhibitory postsynaptic potential (PSP) lasting between 100 and 300 ms in cholinergic interneurons in vitro (Suzuki et al., 2001). Similarly, strong intrastriatal stimulation of cholinergic interneurons themselves results in multiple IPSPs of around 100 ms duration in cholinergic interneurons that are mediated by GABAergic interneurons (Sullivan et al., 2008). More modest synaptic activation of excitatory afferent fibers outside striatum in vivo activates GABAergic mechanisms that induce brief (20 –30 ms) periods of inhibition in cholinergic interneurons (Kita, 1993; Reynolds and Wickens, 2004). Thus, there is clear evidence to support the existence of feedforward inhibition of cholinergic interneuron firing via GABAergic interneurons. However, the intensity of activation that would be required to elicit a pause of the duration reported in behaving animals, coupled with the persistence of pauses under blockade of GABAergic transmission in some preparations (Ding et al., 2010; Schulz et al., 2011), suggests that GABAergic inhibition is unlikely to either act alone or to contribute to pauses in all situations. Secondly, activation of thalamostriatal inputs may lead to dopamine release via local glutamatergic activation of dopamine terminals, leading to a direct inhibition of cholinergic interneuron firing. This would be an attractive mechanism to explain the requirement for both dopamine and thalamostriatal inputs in the expression of pause responses. Consistent with this, there is evidence that glutamate spillover may induce dopamine release in the striatum (Ochi et al., 1995; Kulagina et al., 2001); however, this seems to be primarily by activation of fibers of cortical origin. The lack of spatial convergence of thalamic and dopaminergic synapses at the level of axon terminals, or

J. A. Goldberg and J. N. J. Reynolds / Neuroscience 198 (2011) 27– 43

onto the dendrites of individual cholinergic interneurons, would make direct thalamically driven dopamine release seem less likely (Smith et al., 1994). Furthermore, Aosaki et al. (1994a) showed that dopamine replacement therapy (DRT) with apomorphine, a mixed D1 and D2 dopamine receptor agonist, which would provide tonic activation of these receptors, restored pause responses in MPTPtreated animals, suggesting that stimulus-evoked transient increases in dopamine release are unlikely to be the main trigger for pause responses. Thirdly, glutamatergic input may trigger a pause directly through a mechanism intrinsic to the cholinergic interneuron. Two major scenarios have been suggested. According to scenario 1 an AMPA receptor-mediated depolarization leads to a slight deactivation of HCN currents and/or to a slight reduction in the availability of the NaV1.6 channels underlying the persistent sodium currents. Either of these effects can prolong the mAHP and slow the ramp up to action potential threshold, thereby delaying the next spike (Maurice et al., 2004; Deng et al., 2007; Oswald et al., 2009). Scenario 1 is depicted schematically in Fig. 3A in which a small depolarization results in a decrease in HCN activation or NaV1.6 availability, which would slow down the ramp to action potential threshold as shown in the deviation of the black voltage trace from the gray one. According to scenario 2, the depolarization activates IsAHP, which generates an sAHP that is similar to the one generated by a long (even subthreshold) depolarization with a somatic current pulse (Reynolds et al., 2004; Wilson and Goldberg, 2006). The inward rectifying potassium channel may also contribute by amplifying and prolonging the (after)hyperpolarization initiated by IsAHP (Wilson, 2005). In scenario 2, the HCN and persistent sodium currents are still responsible for the depolarizing portion of the sAHP trajectory. In fact, a recent paper suggests that the main determinant of the length of the sAHP is the HCN current (Oswald et al., 2009). A recent study by Surmeier and coworkers (Ding et al., 2010) sheds new light on the role of the thalamic input through the development of an in vitro assay that recapitulates a pause in cholinergic interneurons in response to synaptic activation. Using a parahorizontal slice that maintains both corticostriatal and thalamostriatal projections intact (Arbuthnott et al., 1985; Kawaguchi et al., 1989; Smeal et al., 2007; Ding et al., 2008), this group demonstrated that only stimulation of thalamic axons, but not cortical ones, led to a burst–pause response in the tonic firing of cholinergic interneurons. The pause persisted in the presence of blockers of GABA receptors and was abolished with AMPA receptor antagonists. Whole cell voltage recordings during thalamic stimulation seem to suggest that the pause usually does not correspond to an sAHP of scenario 2 described above. Instead the ramp up to the next action potential is prolonged, in a fashion that is consistent with the scenario 1, without any visible AHP (left trace in Fig. 3B) (Ding et al., 2010). However, every once in a while thalamic stimulation can result in voltage trajectories that are indistinguishable from an sAHP (middle and right trace in Fig. 3B).

33

In vivo recordings from cholinergic interneurons show that electrical stimulation delivered by an electrode placed in the rat cortex gives rise to a fast and powerful excitatory postsynaptic potential (EPSP) (Wilson et al., 1990; Reynolds et al., 2004; Reynolds and Wickens, 2004). A fast conduction system could be the means by which the initial excitatory component of the pause response is distributed nearly simultaneously to widely spaced interneurons. However, the CM-PF inactivation studies suggest that it is the thalamic input that primarily generates the pause. Consistent with this premise, pause responses can be induced by light flash stimuli that access the striatum primarily via subcortical pathways activating thalamostriatal inputs (Fig. 3C) (Coizet et al., 2007; Schulz et al., 2009, 2011). Dopamine sensitivity of the pause response The pause is abolished in roughly 80% of the TANs after dopamine depletion, and DRT with apomorphine reinstates the pause (Aosaki et al., 1994a). We therefore need to account for dopamine sensitivity of the pause according to both scenarios. Dopamine sensitivity could arise because both HCN and persistent sodium currents are downregulated by D2 dopamine receptor activation (Maurice et al., 2004; Deng et al., 2007). Maurice et al. (2004) reported both a reduction in the size of the NaV1.6 current and a leftward shift in the inactivation curve of NaV1.6 in the presence of D2 receptor agonist. Deng et al. (2007) showed that an sAHP induced with current injections could be prolonged by D2 dopamine receptor activation and that this effect is mimicked and occluded by pharmacological block of HCN currents. Because D2 dopamine receptors down-regulate the HCN and NaV1.6 channels, elevating the ambient level of dopamine (e.g. by application of cocaine that blocks dopamine uptake) should down-regulate these channels, by activating D2 receptors. Because the length of the sAHP increases as the HCN current is reduced (Deng et al., 2007), it follows that elevating ambient levels of dopamine should prolong the sAHP and the pause it creates. Indeed, Ding et al. (2010) demonstrated that the thalamically induced pause was prolonged by cocaine. This implies that an ambient level of dopamine is present in the slices and influences the length of the pause. Moreover, it is through D2 dopamine receptors that dopamine affects the synaptically activated pause (but not the sAHP induced with current injections because sulpiride, a D2 receptor antagonist, shortens the pause) (Deng et al., 2007; Ding et al., 2010). Fig. 3A gives a schematic description of how dopamine sensitivity is conferred upon the pause response according to scenario 1. Based on the data of Maurice et al. (2004), we depict in red the presumed shift in the NaV1.6 inactivation curve or HCN activation curve after dopamine depletion. This shift predicts an elevated availability of HCN or NaV1.6 channels in this condition. Similarly, the shifted curve predicts that a small depolarization will result in a smaller reduction in availability of these channels. A smaller reduction means less of a prolongation of the rise time to action potential threshold (red voltage trace), resulting in a diminished pause. This explains the dopamine

34

J. A. Goldberg and J. N. J. Reynolds / Neuroscience 198 (2011) 27– 43

A model of EPSP induced pause (scenario 1) HCN Activation OR NaV1.6 Availability [%]

100

Dopamine Depletion 80

Normal

60 40 20 0 -100 -90

EPSP

-80

-70

-60

-50

-40

-30

Membrane Potential [mV]

B

In vitro vit 10 mV

-45 mV

↑ 300 pA 1s

10 s

Thalamic stimulation

C Time (min)

In vivo vi 4 3 2 1 8 Spk/s

6 4 2 -0.5

0

0.5

1

Time (s)

2

3

Fig. 3. Pauses in tonic firing of cholinergic interneurons induced by activation of the thalamostriatal pathway. (A) Hypothetical model of scenario 1 (see text) explaining how a pause is induced by an AMPA-mediated EPSP. In the normal condition, a small depolarization due to an EPSP deactivates HCN or reduces availability of NaV1.6 channels (blue arrows on sigmoidal curve). This results in a slowing down of the ramp up to action potential threshold (compare black voltage trace to gray one). After dopamine depletion the sigmoidal curve is shifted to the right [inspired by Fig. 6C of Maurice et al. (2004)]. This results in a higher availability of channels to start off with, but the depolarization generates a more modest reduction in availability (magenta arrows), and a reduced effect on the ramp up to threshold, hence a shortening of the pause (red traces). (B) Graded pauses induced with thalamic stimulation in mouse parahorizontal slices, as seen in whole cell recording. Left: in most cases the pause is manifested as a prolongation of the interspike interval (indicated with an arrow, and compare to the two flanking interspike intervals), by reducing the slope of the rise to the next action potential, presumably due to deactivation of HCN or interference with recovery from inactivation of persistent sodium currents. Middle: in some cases a clear hyperpolarization is visible following the stimulation, presumably due to activation of IsAHP. Right: in rare cases a prominent deep and long-lasting AHP is observed that is indistinguishable from the sAHP that follows a strong depolarizing current injection (300 pA, 200 ms). (C) Thalamostriatal inputs activated physiologically in an anesthetized rat by a light flash induce a strong pause in a TAN, followed by a rebound excitation (unpublished data from Zhang YF, Oswald MJ, and Reynolds JNJ). The response was present only after the light was rendered “salient” under anesthesia by disinhibition of the superior colliculus (SC), thereby enabling subcortical visual pathways (Schulz et al., 2009). The time following the onset (arrow) of collicular disinhibition following local ejection of the GABAA antagonist bicuculline (BIC) injection is indicated on the raster plot. Note the progressive reduction in pause latency as the bicuculline takes effect and the “wear off” commencing after 4 min.

J. A. Goldberg and J. N. J. Reynolds / Neuroscience 198 (2011) 27– 43

sensitivity of the pause according to scenario 1 without invoking IsAHP. It is more difficult to envision how an elevated availability of HCN or NaV1.6 channels can confer dopamine sensitivity to the pause according to scenario 2. Even if the recovery from hyperpolarization is more rapid when these channels are effectively up-regulated, IsAHP would presumably still trigger an sAHP that is longer than the mean ISI of the neuron, and so a pause should be visible. A possible solution to this problem is that the pauses that correspond to sAHPs are indeed not susceptible to dopamine depletion and contribute to the 20% of TANs that do not lose their ability to pause after depletion (Aosaki et al., 1994a).

PLASTICITY OF THE PAUSE RESPONSE The ability of the pause to increase or decrease with conditioning raises the question of the loci of this plasticity. The two (nonexclusive) possibilities are: (a) plasticity of the synapses that trigger the pause; and (b) changes in the postsynaptic excitability of cholinergic interneurons, which affect the intrinsically generated component of the pause (e.g. an up-regulation with conditioning of HCN channels or calcium channels that activate potassium channels that underlie the pause response). Synaptic mechanisms A number of studies addressed the question of plasticity at the corticostriatal synapses onto cholinergic interneurons. The first of these demonstrated that tetanic stimulation of slices induced a potentiation of both an early depolarizing and a later inhibitory component of the synaptic response. The first component, of either cortical or thalamic origin, is mediated by AMPA, whereas the later component is GABAergic arising from intrastriatal sources, presumably via GABAergic interneurons (Suzuki et al., 2001). A later study suggested that the potentiated depolarizing component is actually mediated by GABAA receptors (Bonsi et al., 2004), in difference to the findings of Suzuki et al. (2001). The potentiation of these synapses depended on D5 dopamine receptor activation (Suzuki et al., 2001; Bonsi et al., 2004). The finding that the plasticity has a D5 dopamine receptor sensitivity is compelling in light of indirect evidence that the plasticity of the pause depends on an intact dopaminergic innervation. Firstly, haloperidol, a dopamine receptor antagonist, abolishes pause responses, presumably transiently, during acquisition. Secondly, only a minority of cells expresses the unconditioned response, which sometimes does not even include a pause. Therefore, the fact that the conditioned pause is sensitive to dopamine depletion suggests that the plasticity by which it is created is also sensitive to this depletion (Aosaki et al., 1994a,b; Watanabe and Kimura, 1998). Another important property of the plasticity of the pause response is that it is bidirectional. This could presumably reflect a capacity for bidirectional plasticity at these cortico-/thalamostriatal synapses onto cholinergic interneurons. The study of Fino et al. (2008) provides some evidence for an asymmetric spike timing-dependent plasticity (STDP) at these synapses. In-

35

terestingly, they claim that the asymmetry is partially reversed: when presynaptic activation preceded postsynaptic activation the synapses undergo long-term depression (LTD), but when the postsynaptic activation preceded the presynaptic one both LTD and long-term potentiation (LTP) were observed. While the evidence favors a thalamic (rather than cortical) origin to the pause, this issue is still debated. Unfortunately, these studies of nominally cortical synapses on cholinergic interneurons did not entirely rule out a thalamic source. This may be the source of some of the differences between those studies. The study of plasticity in identified thalamic vs. cortical synapses onto cholinergic interneurons remains a topic for future research. Postsynaptic excitability The most direct demonstration of the plasticity of the pause response under controlled conditions was provided by Reynolds and co-workers. They showed that in three cells recorded intracellularly in vivo, application of tetanic stimulation in the SNc induced an sAHP in these neurons that did not exist before the execution of the protocol. This is also evidence that postsynaptic excitability can play a large part in enhancing an sAHP-based pause response (Reynolds and Wickens, 2004). In general, the mechanisms proposed above that could confer dopamine sensitivity to the pause response itself, would by extension also impact the ability of any intrinsic component of the pause response to display plasticity in face of dopamine depletion.

ROLE OF CHOLINERGIC INTERNEURONS AND THEIR PAUSE IN STRIATAL FUNCTION The cholinergic interneurons are the sole source of acetylcholine (ACh) to the striatum. Despite their small numbers they are responsible for levels of acetylcholine, ChAT, and choline esterase that are among the highest in the brain (Mesulam et al., 1992; Contant et al., 1996). The spontaneous discharge of these neurons is what is responsible for an ambient ACh level in the striatum. ACh affects multiple striatal targets across multiple time scales. A full review of cholinergic signaling in the striatum is well beyond the scope of the present exposition and has been reviewed elsewhere (Goldberg et al., in press). Instead we will mention some general principles of the role of ACh in the striatum. Nicotinic ACh receptors At the shortest time scale of tens of milliseconds, ACh activates nAChRs that are located in the striatum primarily on GABAergic interneurons (location 1 in Fig. 4) and on the axons of dopaminergic neurons (location 2). nAChRs drive firing in striatal GABAergic interneurons and vicariously through them inhibit the SPNs as well as the cholinergic interneurons themselves (Koos and Tepper, 2002; Zhou et al., 2002; Sullivan et al., 2008; Witten et al., 2010). Activation of nAChRs on dopaminergic axons allows the cholinergic interneuron to dynamically gate the release of dopamine in the striatum (Wilson et al., 1990; Zhou et al.,

36

J. A. Goldberg and J. N. J. Reynolds / Neuroscience 198 (2011) 27– 43

Cholinergic interneuron 7

cortex/CM-PF

3

glu

4

2 DA

GABA

1

SNc

5 6 GABA

GABAergic interneuron

EC

SPN Legend: x nAChR y M1 mAChR z M2 mAChR EC endocannabinoids

GPe/GPi/SNr Fig. 4. Role of cholinergic interneurons within the basic striatal circuit. The receptors by which ACh exerts its influence on neighboring neurons are indicated by the italicized numerals in conjunction with the legend. SPN, spiny neuron; GPe/i, globus pallidus externus/internus; SNr/c, substantia nigra pars reticulata/compacta; CM-PF, centromedian parafasicularis.

2001, 2002; Rice and Cragg, 2004; Zhang and Sulzer, 2004; Exley and Cragg, 2008; Ding et al., 2010; Witten et al., 2010), which is important for maintaining the so-called dopamine-ACh balance that is disrupted in movement disorders such as Parkinson’s disease (McGeer et al., 1961; Barbeau, 1962; Lehmann and Langer, 1983; DeBoer et al., 1996). Muscarinic ACh receptors affect striatal circuitry at multiple time scales In contrast to the relatively restricted distribution and function of nAChRs, muscarinic receptors (mAChRs) are expressed postsynaptically on all neuronal types in the striatum, as well as presynaptically on terminals of afferent fibers. These receptors can be divided into two classes on the basis of their coupling to G proteins: M1 class (M1, M3, M5) and M2 class (M2, M4). M1-class receptors couple to Gq/11 G␣ proteins that activate phosphlipase C (PLC) isoforms resulting in phosphatidylinositol 4,5-bisphosphate (PIP2) hydrolysis to inositol 1,4,5-triphosphate (IP3) and diacyl glycerol (DAG). M2 class couple to Gi/o G proteins that inhibit adenylyl cyclase (AC) through Gi␣ subunits and close CaV2 channels and open Kir3 channels through associated G␤␥ subunits (Wess, 1996; Wess et al., 2007). The physiological role of mAChRs spans the time scales of tens of milliseconds to minutes and hours. At the shorter time scales, ACh (released even by a single action potential emitted by a single cell) can powerfully reduce afferent glutamatergic transmission via the activation of presynaptic mAChRs (location 3 in Fig. 4) and can aug-

ment synaptic integration in the postsynaptic SPNs via postsynaptic mAChRs (location 4) (Akaike et al., 1988; Calabresi et al., 1998a; Barral et al., 1999; Alcantara et al., 2001; Pakhotin and Bracci, 2007; Higley et al., 2009; Ding et al., 2010). The recent finding that cholinergic interneurons can potently inhibit SPNs (Witten et al., 2010) synergizes with the presynaptic reduction of fast synaptic transmission via mAChR activation at glutamatergic synapses in the striatum (location 3). At a scale of seconds to minutes, we find the neuromodulatory roles of mAChRs to down-regulate a slew of somatodendritic potassium and high voltage-activated (HVA) calcium channels that alter the excitability of SPNs (location 4 in Fig. 4). Prominent among these are the Kv4, Kv7, and Kir2 channels giving rise to A-type and M-type and inward-rectifying potassium currents, respectively, and the CaV1 and CaV2 families of calcium channels (Dodt and Misgeld, 1986; Akins et al., 1990; Howe and Surmeier, 1995; Hsu et al., 1996; Galarraga et al., 1999; Tkatch et al., 2000; Vilchis et al., 2000; Figueroa et al., 2002; Shen et al., 2004, 2005, 2007; Olson et al., 2005; Perez-Rosello et al., 2005; Day et al., 2008; Perez-Burgos et al., 2008). An important theme emerging from recent studies is that the two classes of SPNs—the striatonigral (direct pathway, D1 DA receptor expressing) dSPNs and the striatopallidal (indirect pathway, D2 DA receptor expressing) iSPNs— exhibit a different degree of intrinsic excitability (Day et al., 2008; Gertler et al., 2008). In the same vein, muscarinic modulation has a differential effect on these two classes (Shen et al., 2007; Day et al., 2008). The details of the dichotomy between the two classes of SPNs and the impact of mAChRs on them are beyond the scope of this review and are reviewed elsewhere (Surmeier et al., 2010). In summary, to a first approximation activation of postsynaptic mAChRs, which are primarily of the M1 type, increases the excitability of SPN. Another disinhibitory route by which M1-type mAChRs increase the excitability of SPNs is by enhancing release of endocannabinoids postsynaptically from SPNs, which activate CB1 receptors presynaptically on GABAergic terminals thereby suppressing inhibition of SPNs (location 5 in Fig. 4) (Narushima et al., 2007). In contrast, activation of presynaptic mAChRs, primarily of the M2 type, reduces synaptic transmission both at glutamatergic synapses (location 3) (Akaike et al., 1988; Malenka and Kocsis, 1988; Higley et al., 2009) and at GABAergic synapses (location 6) (Marchi et al., 1990; Koos and Tepper, 2002). Of special interest is the effect of mAChRs on the cholinergic interneurons themselves via autoreceptors (location 7 in Fig. 4). Application of muscarinic agonists can silence cholinergic interneurons, and focal stimulation of the slice can induce what have been described as muscarinic IPSPs in these neurons. Both these effects are mediated by postsynaptic M2 receptors (Calabresi et al., 1998b; Bonsi et al., 2008). Activation of the M2-class receptors down-regulates CaV2.1 and CaV2.2 channels in cholinergic interneurons (Yan and Surmeier, 1996). Because CaV2.2 channels activate the SK channels that determine the size of the mAHP in these neurons (Goldberg and Wilson, 2005), activation of mAChRs in these neurons

J. A. Goldberg and J. N. J. Reynolds / Neuroscience 198 (2011) 27– 43

reduces the mAHPs and induces irregular discharge (Ding et al., 2006). These findings together with the finding that cholinergic interneurons inhibit each other’s discharge (Sullivan et al., 2008) imply that cholinergic interneurons possess a negative feedback mechanism to control their own discharge rate, both individually and collectively as a network. This negative feedback, mediated by cholinergic autoreceptors, inhibits the various above-mentioned mechanisms by which activation of M1 receptors increase the excitability of SPNs (location 4 in Fig. 4). Thus activation of M2-class receptors antagonizes the effect of M1 receptors on SPNs by activating autoreceptors on cholinergic interneurons (location 7) (Galarraga et al., 1999). At the longest time scales, mAChRs (in conjunction with nAChRs and DA receptors) play a prominent role in the induction of long-term plasticity in corticostriatal synapses (Calabresi et al., 2000; Wang et al., 2006; Pisani et al., 2007; Kreitzer and Malenka, 2008). This is an intricate topic that is reviewed elsewhere (Surmeier et al., 2007). To a first approximation, M1-class receptors are the primary players in conveying the influence of cholinergic interneurons on synaptic plasticity in glutamatergic striatal synapses: enhancement of M1-class receptor activation favors LTP, and inhibition of these receptors favors LTD (Centonze et al., 1999; Pisani et al., 2007). However, M2-class receptors also have an indirect influence on LTD. In their capacity as autoreceptors on cholinergic interneurons, they lead to self-inhibition of ACh release (Calabresi et al., 1998b), which leads presumably to a reduced activity of M1-class receptors and hence to LTD (Bonsi et al., 2008). Functional implications of the pause response Because ACh is hydrolyzed rapidly in the extracellular space, the occurrence of synchronous pauses in the tonic activity of cholinergic interneurons presumably leads to a significant drop in ACh levels for the duration of the pause. In light of the potent modulatory effect of ACh on multiple targets across multiple time scales in striatal circuitry, the concerted pauses should have a dramatic effect on striatal function due to the reduction in ACh levels that it creates. Cholinergic pauses have been proposed to act as a temporal window for the induction of synaptic plasticity in SPNs (Morris et al., 2004). In agreement with this, recent studies using optogenetics have demonstrated that these pauses remove the powerful inhibition on spiny neuron discharge conferred by cholinergic interneuron firing, resulting in a permissive effect on spiny neurons (Witten et al., 2010). In addition, the reduction in ACh levels accompanying a pause would reduce the low-pass filtering effect of nAChRs tonic activation on dopamine terminals (location 2 in Fig. 3) (Rice and Cragg, 2004; Zhang and Sulzer, 2004; Cragg, 2006; Exley and Cragg, 2008), thereby facilitating phasic dopamine release at higher burst frequencies associated with reward-relevant activity (Hyland et al., 2002). In association with corticostriatal motor or sensory activity preceding a pause-inducing stimulus, these three factors (namely, pre- and postsynaptic activity and phasic dopamine) would be brought together around the time of the pause, a conjunction that favors the induction of corti-

37

costriatal LTP (Reynolds and Wickens, 2002). However, we have already heard that LTD and not LTP is likely to be favored during the pause itself, due to reduced tonic activation of M1 receptors (Centonze et al., 1999; Pisani et al., 2007). How, then, is the requirement for M1 receptor activation for LTP to be met? The possible answer to this requires both a closer consideration of the excitatory phases flanking the pause response and a broad view regarding the temporal requirements for neuromodulation in the striatum (Redgrave et al., 2008). Perhaps M1 activation could be facilitated by the TAN burst firing that accompanies the initial facilitation [as shown by Ding et al. (2010)], or, indeed, by the postpause rebound. The latter is observed more reliably in TAN studies in response to trial outcome (Matsumoto et al., 2001; Apicella et al., 2011) and overlaps dopamine phasic responses to some degree (Morris et al., 2004; Joshua et al., 2008). Moreover, it may reflect more than simply a membrane-intrinsic rebound response and appears to signal information akin to a true reward prediction error (Apicella et al., 2009, 2011) at a time scale more appropriate for this purpose than the short-latency response of dopamine neurons (see Redgrave, Vautrelle and Reynolds, this issue). Thus, mismatch between reward expectation and delivery activates a late cholinergic burst that might be used as a global teaching signal for striatal plasticity. It is tempting to further speculate that a determinant of the direction of resulting synaptic change may be the size of the postpause burst. Thus, an unexpected outcome that induces the largest burst favors LTP induction, compared to a wholly predicted outcome, where the effect of reduced M1 tone would dominate. The utility of the excitatory phases of the conditioned pause response to satisfy the need for M1 activation in corticostriatal LTP will, however, depend on the timing requirements for neuromodulators in synaptic plasticity induction, a topic for which there are few empirical data at present (Pawlak et al., 2010). It is clear that a better understanding of the functional significance of the conditioned pause response requires more knowledge about the contributions of afferent pathways and conditions that elicit each phase of the response. In recent in vivo work Reynolds and coworkers (Schulz et al., 2011; Zhang YM and Reynolds JNJ, unpublished observations) have found that pauses in cholinergic interneurons initiated by thalamostriatal inputs in the rat are invariably preceded by an excitatory burst of variable duration (see Fig. 3C). The ACh transient so-induced would be necessary to induce the burst–pause mechanism described by Ding et al. (2010), which relies on nicotinicreceptor driven dopamine release. In contrast, stimulation applied to the contralateral cortex (therefore specifically activating corticostriatal inputs) induced pauses in most of the same striatal cholinergic interneurons without a preceding excitation. This may be difficult to resolve in the context of monkey behavioral experiments because little is currently understood about the specific contributions of cortical and thalamic inputs to particular phases of the pause response in TANs. However, it does suggest that different sources of drive converging on cholinergic in-

38

J. A. Goldberg and J. N. J. Reynolds / Neuroscience 198 (2011) 27– 43

?

Orbitofrontal cortex Motor and Value of sensory Reward reward cortical expectation outcome areas

{

Visual cortical areas

Trigger

?Reward prediction error

?Context LGN Thalamus

Outcome Probability predicted by trigger

CM-PF Thalamus

Superior Colliculus

Corticostriatal synapse Low DA frequency firing

CIN

High frequency burst firing ACh

DA

ACh

SPN Substantia nigra pars compacta

Striatum

Response to trigger stimulus

Initial excitation +/-

latency if present: 50 to 80 ms

Pause +++ short latency: 100 to 180 ms

Reward probability modulation weak positive

Response to outcome stimulus

Post pause excitation

Reward probability modulation

++

No

latency: 220 to 320 ms

Initial excitation +/-

Pause

Reward probability modulation

+++

negative

longer latency: 210 to 330 ms

Post pause excitation +++ longer latency: 310 to 490 ms

Reward probability modulation strong negative

Reward probability 0.25 0.5 0.75

Fig. 5. Proposed cholinergic interneuron circuitry underlying the two types of conditioned pause responses observed in TANs during probabilistic tasks. The effect of the neurotransmitter in the depicted circumstance is indicated by the postsynaptic bar (white: excitatory, black: inhibitory, gray: dependent on receptor subtype on the target neuron). The attributes of the pause responses elicited by a trigger stimulus (in this case a visual cue) and an outcome stimulus (in this case limited to reward) are tabulated below the circuit (data combined from Apicella et al., 2009, 2011; Joshua

J. A. Goldberg and J. N. J. Reynolds / Neuroscience 198 (2011) 27– 43

terneurons can induce pauses with varying phase relationships. A major challenge to cellular physiologists is to determine the input requirements for the induction of pauses in different behavioral situations requiring varying degrees of cortical resources, for example, classical conditioning in contrast to instrumental learning tasks (Apicella et al., 2009, 2011). Dynamics of pause generation Recent TAN studies that have separated the timing between the conditioned cue (a trigger stimulus that signals the start of a trial) and the outcome (reward, aversive stimulus, or no outcome) have begun to distinguish differences in characteristics between the pause responses elicited in each behavioral context. These studies are probabilistic, in that the trigger stimulus (usually visual) conveys information about the likelihood of the predicted outcome actually being delivered. Knowledge of the anatomy of the striatal circuit and the timing of the components of these conditioned responses allows us to speculate here on mechanisms in cholinergic interneurons that might drive pause responses seen in TANs following a trigger stimulus and following the outcome of that trial (see Fig. 5). The response to the trigger stimulus is more stereotypical, and more importantly shows minimal sensitivity to the probability of ensuing reward, unlike the outcome stimulus (Morris et al., 2004; Joshua et al., 2008; Apicella et al., 2009, 2011). This suggests that only relatively rudimentary preprocessing has been performed by the time the trigger stimulus is presented to the cholinergic interneuron. Inputs to the cholinergic interneurons from the CM-PF are driven by areas, such as the superior colliculus (SC), which are specialized for preattentive processing. An unexpected cue with some behavioral significance (e.g. trial initiation) is “seen” by this subcortical circuit and induces a pause in cholinergic interneuron firing by ACh-burst-driven dopamine release. Note that in addition to this local nAChRdriven release, the SC directly activates the dopamine cells in the SNc (Redgrave and Gurney, 2006). At low (e.g. tonic) firing frequencies of the dopamine cells, ACh release will enhance terminal dopamine release (Rice and Cragg, 2004) and drive the pause. This pause will arrest ongoing behavior [through a mechanism proposed by Ding et al. (2010) involving enhancement of indirect pathway “NOGO” activity], and focus behavioral resources on the task at hand (e.g. moving the hand to quickly press a key). Thus, the trigger stimulus drives attentional shifts required for focusing motor output either for reward consummation (classical conditioning) or immediate task performance (instrumental conditioning).

39

In contrast, the pause response elicited by the ensuing outcome (rewarding or not) may be driven by additional cortical inputs that converge with thalamostriatal inputs onto cholinergic interneurons. Indeed, the latencies to all phases of the outcome response are invariably longer than those observed for the trigger response (Joshua et al., 2008, Apicella et al., 2009, 2011), suggesting that outcome-generated pauses could be initiated by inputs emanating from the cortex, after reward processing. Following trigger presentation and on a slower time scale, neurons in the orbitofrontal cortex will calculate the expected value of the outcome to follow (in the case depicted in Fig. 5, a food reward) based on prior cue association, and modify motivational state. When the reward arrives, a sudden change in luminance in the visual field may activate the thalamostriatal inputs and initiate the pause response as before. However, this time the cholinergic response may be modified by dendritic inputs from cortex that have become active following the cue, such as those coding for motivational state and the preliminary calculation of the ensuing reward value. These could alter the time-course and magnitude of the early phases of thalamically driven pauses through dendritic interactions. Alternatively, cortical inputs responding to reward presentation may trigger a pause directly through intrinsic membrane mechanisms, assuming such information is made available to cholinergic interneurons. Notably the orbitofrontal cortex can now calculate the reward-prediction error (e.g. a large error would be reported if a trial of low-reward probability, say 25%, did result in delivery of a reward) and report this to the neurons via a subsequent activation forming the rebound phase (Apicella et al., 2011). The burst in cholinergic interneuron firing that accompanies the outcome stimulus or the subsequent activation may provide the necessary M1 receptor activation to induce LTP. Clearly, the pause response cannot be considered one entity because of its variable attributes under different behavioral conditions. Functionally, the control of the generation, timing, and synchronization of the pause response by inputs from the CM-PF and cortex provides a mechanism by which these structures can direct attention and control affective learning. These possibilities await future study.

CONCLUSIONS In this review, we have attempted to correlate the expression of the conditioned pause response of TANs with known properties and mechanisms of activation of striatal cholinergic interneurons and their afferents. Because di-

et al., 2008). The likelihood of observing a particular component of the pause response is indicated in the table by “⫹” or “⫺”. The component of the response is said to show “positive reward modulation” if the size of the response increases with increasing probability of reward, and vice versa. For instance, the size of the postpause excitation decreases as the trigger stimulus predicts reward delivery with greater certainty (compare green and red traces in cartoon below the table). The circuits that are proposed to principally contribute to components of the pause responses elicited by trigger and outcome stimulus are color coded into cortical (yellow) and subcortical (green) pathways (see section “Dynamics of pause generation” for further explanation). Note some known connections are omitted for clarity. CIN, cholinergic interneuron; SPN, spiny neuron; LGN, lateral geniculate nucleus; CM-PF, centromedian parafasicularis; ACh, acetylcholine; DA, dopamine. For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.

40

J. A. Goldberg and J. N. J. Reynolds / Neuroscience 198 (2011) 27– 43

rect GABAergic and dopaminergic cell activation seems less likely to underlie the predominant pause mechanism, we presented two scenarios for how glutamatergic inputs to cholinergic interneurons can transiently engage intrinsic membrane mechanisms that suppress tonic firing. In the first, deactivation of HCN channels and/or reduction in the availability of NaV1.6 channels prolongs the mAHP and delays the following spike (Deng et al., 2007; Oswald et al., 2009). In the second, synaptic depolarization activates a hyperpolarization mediated by the as yet unidentified IsAHP (Wilson and Goldberg, 2006). Only the first scenario is sensitive to tonic dopamine levels and therefore represents a plausible mechanism to underlie TAN pause responses in the majority of situations. A mechanism recently proposed that invokes pauses through extrinsic neuromodulator mechanisms involves burst activation of cholinergic interneurons by thalamostriatal inputs (Ding et al., 2010). These bursts have been shown to elicit nAChRdriven dopamine release, thereby suppressing cholinergic firing via dopamine D2 receptor activation, enhancing activity through the indirect pathway and arresting ongoing behavior. Both glutamatergic initiation and dopamine receptor dependence of the pause are accounted for by this mechanism. However, considering the full behavioral expression of the pause response, it is apparent that other afferent pathways must be engaged under different circumstances. We account for the apparent reward prediction error sensitivity of the late phase of the outcomeinduced pause by considering the effect of cortical afferents that converge onto cholinergic interneurons. Such information is unlikely to be available to thalamostriatal inputs driven by preattentive processing mechanisms. Clearly, much work needs to be done to understand the effect of dendritic convergence of these pathways likely carrying quite different information to cholinergic interneurons. Investigations also need to focus on elucidating the proposed temporal effect that cholinergic pauses likely have on the induction of synaptic plasticity in the corticostriatal pathway. Understanding the role of dopamine in the acquisition and maintenance of the pause response is a further challenge, with the translational aim of assisting our understanding of pathological conditions that disturb the dopamine-ACh balance. Acknowledgments—We thank Jim Surmeier, Charlie Wilson, Avital Adler, and Hagai Bergman for contributing data for the figures. During the preparation of this article JAG was supported by the IDP Foundation, Chicago, IL and JR received support from a Rutherford Discovery Fellowship from the Royal Society of New Zealand.

REFERENCES Akaike A, Sasa M, Takaori S (1988) Muscarinic inhibition as a dominant role in cholinergic regulation of transmission in the caudate nucleus. J Pharmacol Exp Ther 246:1129 –1136. Akins PT, Surmeier DJ, Kitai ST (1990) Muscarinic modulation of a transient K⫹ conductance in rat neostriatal neurons. Nature 344:240 –242. Alcantara AA, Mrzljak L, Jakab RL, Levey AI, Hersch SM, GoldmanRakic PS (2001) Muscarinic m1 and m2 receptor proteins in local

circuit and projection neurons of the primate striatum: anatomical evidence for cholinergic modulation of glutamatergic prefronto-striatal pathways. J Comp Neurol 434:445– 460. Anderson ME (1978) Discharge patterns of basal ganglia neurons during active maintenance of postural stability and adjustment to chair tilt. Brain Res 143:325–338. Aosaki T, Graybiel AM, Kimura M (1994a) Effect of the nigrostriatal dopamine system on acquired neural responses in the striatum of behaving monkeys. Science 265:412– 415. Aosaki T, Kimura M, Graybiel AM (1995) Temporal and spatial characteristics of tonically active neurons of the primate’s striatum. J Neurophysiol 73:1234 –1252. Aosaki T, Tsubokawa H, Ishida A, Watanabe K, Graybiel AM, Kimura M (1994b) Responses of tonically active neurons in the primate’s striatum undergo systematic changes during behavioral sensorimotor conditioning. J Neurosci 14:3969 –3984. Apicella P (2002) Tonically active neurons in the primate striatum and their role in the processing of information about motivationally relevant events. Eur J Neurosci 16:2017–2026. Apicella P, Deffains M, Ravel S, Legallet E (2009) Tonically active neurons in the striatum differentiate between delivery and omission of expected reward in a probabilistic task context. Eur J Neurosci 30:515–526. Apicella P, Legallet E, Trouche E (1997) Responses of tonically discharging neurons in the monkey striatum to primary rewards delivered during different behavioral states. Exp Brain Res 116:456 – 466. Apicella P, Ravel S, Deffains M, Legallet E (2011) The role of striatal tonically active neurons in reward prediction error signaling during instrumental task performance. J Neurosci 31:1507–1515. Arbuthnott GW, MacLeod NK, Rutherfod A (1985) The rate corticostriatal pathway in vitro. J Physiol 367:102. Barbeau A (1962) The pathogenesis of Parkinson’s disease: a new hypothesis. Can Med Assoc J 87:802– 807. Barral J, Galarraga E, Bargas J (1999) Muscarinic presynaptic inhibition of neostriatal glutamatergic afferents is mediated by Q-type Ca2⫹ channels. Brain Res Bull 49:285–289. Bennett BD, Callaway JC, Wilson CJ (2000) Intrinsic membrane properties underlying spontaneous tonic firing in neostriatal cholinergic interneurons. J Neurosci 20:8493– 8503. Bennett BD, Wilson CJ (1998) Synaptic regulation of action potential timing in neostriatal cholinergic interneurons. J Neurosci 18:8539 – 8549. Bennett BD, Wilson CJ (1999) Spontaneous activity of neostriatal cholinergic interneurons in vitro. J Neurosci 19:5586 –5596. Berendse HW, Groenewegen HJ (1990) Organization of the thalamostriatal projections in the rat, with special emphasis on the ventral striatum. J Comp Neurol 299:187–228. Bishop GA, Chang HT, Kitai ST (1982) Morphological and physiological properties of neostriatal neurons: an intracellular horseradish peroxidase study in the rat. Neuroscience 7:179 –191. Bolam JP, Wainer BH, Smith AD (1984) Characterization of cholinergic neurons in the rat neostriatum. A combination of choline acetyltransferase immunocytochemistry, Golgi-impregnation and electron microscopy. Neuroscience 12:711–718. Bonsi P, De Persis C, Calabresi P, Bernardi G, Pisani A (2004) Coordinate high-frequency pattern of stimulation and calcium levels control the induction of LTP in striatal cholinergic interneurons. Learn Mem 11:755–760. Bonsi P, Martella G, Cuomo D, Platania P, Sciamanna G, Bernardi G, Wess J, Pisani A (2008) Loss of muscarinic autoreceptor function impairs long-term depression but not long-term potentiation in the striatum. J Neurosci 28:6258 – 6263. Burns RS, Chiueh CC, Markey SP, Ebert MH, Jacobowitz DM, Kopin IJ (1983) A primate model of parkinsonism: selective destruction of dopaminergic neurons in the pars compacta of the substantia nigra by N-methyl-4-phenyl-1,2,3,6-tetrahydropyridine. Proc Natl Acad Sci U S A 80:4546 – 4550.

J. A. Goldberg and J. N. J. Reynolds / Neuroscience 198 (2011) 27– 43 Calabresi P, Centonze D, Gubellini P, Pisani A, Bernardi G (1998a) Endogenous ACh enhances striatal NMDA-responses via M1-like muscarinic receptors and PKC activation. Eur J Neurosci 10:2887– 2895. Calabresi P, Centonze D, Gubellini P, Pisani A, Bernardi G (2000) Acetylcholine-mediated modulation of striatal function. Trends Neurosci 23:120 –126. Calabresi P, Centonze D, Pisani A, Sancesario G, North RA, Bernardi G (1998b) Muscarinic IPSPs in rat striatal cholinergic interneurones. J Physiol 510 (Pt 2):421– 427. Centonze D, Gubellini P, Bernardi G, Calabresi P (1999) Permissive role of interneurons in corticostriatal synaptic plasticity. Brain Res Brain Res Rev 31:1–5. Chang HT, Kitai ST (1982) Large neostriatal neurons in the rat: an electron microscopic study of gold-toned Golgi-stained cells. Brain Res Bull 8:631– 643. Coizet V, Overton PG, Redgrave P (2007) Collateralization of the tectonigral projection with other major output pathways of superior colliculus in the rat. J Comp Neurol 500:1034 –1049. Contant C, Umbriaco D, Garcia S, Watkins KC, Descarries L (1996) Ultrastructural characterization of the acetylcholine innervation in adult rat neostriatum. Neuroscience 71:937–947. Cragg SJ (2006) Meaningful silences: how dopamine listens to the ACh pause. Trends Neurosci 29:125–131. Crutcher MD, DeLong MR (1984) Single cell studies of the primate putamen. II. Relations to direction of movement and pattern of muscular activity. Exp Brain Res 53:244 –258. Day M, Wokosin D, Plotkin JL, Tian X, Surmeier DJ (2008) Differential excitability and modulation of striatal medium spiny neuron dendrites. J Neurosci 28:11603–11614. DeBoer P, Heeringa MJ, Abercrombie ED (1996) Spontaneous release of acetylcholine in striatum is preferentially regulated by inhibitory dopamine D2 receptors. Eur J Pharmacol 317:257–262. Deng P, Zhang Y, Xu ZC (2007) Involvement of I(h) in dopamine modulation of tonic firing in striatal cholinergic interneurons. J Neurosci 27:3148 –3156. Deschênes M, Bourassa J, Doan VD, Parent A (1996) A single-cell study of the axonal projections arising from the posterior intralaminar thalamic nuclei in the rat. Eur J Neurosci 8:329 –343. DiFiglia M (1987) Synaptic organization of cholinergic neurons in the monkey neostriatum. J Comp Neurol 255:245–258. DiFiglia M, Carey J (1986) Large neurons in the primate neostriatum examined with the combined Golgi-electron microscopic method. J Comp Neurol 244:36 –52. Dimova R, Vuillet J, Nieoullon A, Kerkerian-Le Goff L (1993) Ultrastructural features of the choline acetyltransferase-containing neurons and relationships with nigral dopaminergic and cortical afferent pathways in the rat striatum. Neuroscience 53:1059 –1071. Ding J, Guzman JN, Tkatch T, Chen S, Goldberg JA, Ebert PJ, Levitt P, Wilson CJ, Hamm HE, Surmeier DJ (2006) RGS4-dependent attenuation of M4 autoreceptor function in striatal cholinergic interneurons following dopamine depletion. Nat Neurosci 9:832– 842. Ding J, Peterson JD, Surmeier DJ (2008) Corticostriatal and thalamostriatal synapses have distinctive properties. J Neurosci 28:6483– 6492. Ding JB, Guzman JN, Peterson JD, Goldberg JA, Surmeier DJ (2010) Thalamic gating of corticostriatal signaling by cholinergic interneurons. Neuron 67:294 –307. Dodt HU, Misgeld U (1986) Muscarinic slow excitation and muscarinic inhibition of synaptic transmission in the rat neostriatum. J Physiol 380:593– 608. Exley R, Cragg SJ (2008) Presynaptic nicotinic receptors: a dynamic and diverse cholinergic filter of striatal dopamine neurotransmission. Br J Pharmacol 153 (Suppl 1):S283–S297. Figueroa A, Galarraga E, Bargas J (2002) Muscarinic receptors involved in the subthreshold cholinergic actions of neostriatal spiny neurons. Synapse 46:215–223.

41

Fino E, Deniau JM, Venance L (2008) Cell-specific spike-timing-dependent plasticity in GABAergic and cholinergic interneurons in corticostriatal rat brain slices. J Physiol 586:265–282. Fiorillo CD, Tobler PN, Schultz W (2003) Discrete coding of reward probability and uncertainty by dopamine neurons. Science 299: 1898 –1902. Galarraga E, Hernández-López S, Reyes A, Miranda I, BermudezRattoni F, Vilchis C, Bargas J (1999) Cholinergic modulation of neostriatal output: a functional antagonism between different types of muscarinic receptors. J Neurosci 19:3629 –3638. Gertler TS, Chan CS, Surmeier DJ (2008) Dichotomous anatomical properties of adult striatal medium spiny neurons. J Neurosci 28:10814 –10824. Goldberg JA, Ding JB, Surmeier DJ (in press) Muscarinic modulation of striatal function and circuitry. In: Handbook of experimental pharmacology (Fryer AD et al., eds). Heidelberg: Springer-Verlag. Goldberg JA, Rokni U, Boraud T, Vaadia E, Bergman H (2004) Spike synchronization in the cortex/basal-ganglia networks of Parkinsonian primates reflects global dynamics of the local field potentials. J Neurosci 24:6003– 6010. Goldberg JA, Teagarden MA, Foehring RC, Wilson CJ (2009) Nonequilibrium calcium dynamics regulate the autonomous firing pattern of rat striatal cholinergic interneurons. J Neurosci 29:8396 – 8407. Goldberg JA, Wilson CJ (2005) Control of spontaneous firing patterns by the selective coupling of calcium currents to calcium-activated potassium currents in striatal cholinergic interneurons. J Neurosci 25:10230 –10238. Goldberg JA, Wilson CJ (2010) The cholinergic interneurons of the striatum: intrinsic properties underlie multiple discharge patterns. In: Handbook of basal ganglia structure and function (Steiner H, Tseng K, eds), pp 133–149. London: Academic Press. Graybiel AM, Aosaki T, Flaherty AW, Kimura M (1994) The basal ganglia and adaptive motor control. Science 265:1826 –1831. Hattori S, Murakami F, Song WJ (2003) Quantitative relationship between Kv4. 2 mRNA and A-type K⫹ current in rat striatal cholinergic interneurons during development. J Neurophysiol 90:175– 183. Higley MJ, Soler-Llavina GJ, Sabatini BL (2009) Cholinergic modulation of multivesicular release regulates striatal synaptic potency and integration. Nat Neurosci 12:1121–1128. Howe AR, Surmeier DJ (1995) Muscarinic receptors modulate N-, P-, and L-type Ca2⫹ currents in rat striatal neurons through parallel pathways. J Neurosci 15:458 – 469. Hsu KS, Yang CH, Huang CC, Gean PW (1996) Carbachol induces inward current in neostriatal neurons through M1-like muscarinic receptors. Neuroscience 73:751–760. Hyland BI, Reynolds JNJ, Hay J, Perk CG, Miller R (2002) Firing modes of midbrain dopamine cells in the freely moving rat. Neuroscience 114:475– 492. Jiang ZG, North RA (1991) Membrane properties and synaptic responses of rat striatal neurones in vitro. J Physiol 443:533–553. Joshua M, Adler A, Mitelman R, Vaadia E, Bergman H (2008) Midbrain dopaminergic neurons and striatal cholinergic interneurons encode the difference between reward and aversive events at different epochs of probabilistic classical conditioning trials. J Neurosci 28:11673–11684. Kawaguchi Y (1992) Large aspiny cells in the matrix of the rat neostriatum in vitro: physiological identification, relation to the compartments and excitatory postsynaptic currents. J Neurophysiol 67: 1669 –1682. Kawaguchi Y (1993) Physiological, morphological, and histochemical characterization of three classes of interneurons in rat neostriatum. J Neurosci 13:4908 – 4923. Kawaguchi Y, Wilson CJ, Emson PC (1989) Intracellular recording of identified neostriatal patch and matrix spiny cells in a slice preparation preserving cortical inputs. J Neurophysiol 62:1052–1068.

42

J. A. Goldberg and J. N. J. Reynolds / Neuroscience 198 (2011) 27– 43

Kemp JM, Powell TP (1971) The synaptic organization of the caudate nucleus. Philos Trans R Soc Lond B Biol Sci 262:403– 412. Kimura M, Rajkowski J, Evarts E (1984) Tonically discharging putamen neurons exhibit set-dependent responses. Proc Natl Acad Sci U S A 81:4998 –5001. Kita H (1993) GABAergic circuits of the striatum. Prog Brain Res 99:51–72. Kölliker A (1896) Handbuch der Gewebelehre des Menchen, Bd. II. Leipzig: Kengelman. Koos T, Tepper JM (2002) Dual cholinergic control of fast-spiking interneurons in the neostriatum. J Neurosci 22:529 –535. Kreitzer AC, Malenka RC (2008) Striatal plasticity and basal ganglia circuit function. Neuron 60:543–554. Kulagina NV, Zigmond MJ, Michael AC (2001) Glutamate regulates the spontaneous and evoked release of dopamine in the rat striatum. Neuroscience 102:121–128. Lapper SR, Bolam JP (1992) Input from the frontal cortex and the parafascicular nucleus to cholinergic interneurons in the dorsal striatum of the rat. Neuroscience 51:533–545. Lehmann J, Langer SZ (1983) The striatal cholinergic interneuron: synaptic target of dopaminergic terminals? Neuroscience 10:1105– 1120. Malenka RC, Kocsis JD (1988) Presynaptic actions of carbachol and adenosine on corticostriatal synaptic transmission studied in vitro. J Neurosci 8:3750 –3756. Marchi M, Sanguineti P, Raiteri M (1990) Muscarinic receptors mediate direct inhibition of GABA release from rat striatal nerve terminals. Neurosci Lett 116:347–351. Matsumoto N, Minamimoto T, Graybiel AM, Kimura M (2001) Neurons in the thalamic CM-Pf complex supply striatal neurons with information about behaviorally significant sensory events. J Neurophysiol 85:960 –976. Maurice N, Mercer J, Chan CS, Hernandez-Lopez S, Held J, Tkatch T, Surmeier DJ (2004) D2 dopamine receptor-mediated modulation of voltage-dependent Na⫹ channels reduces autonomous activity in striatal cholinergic interneurons. J Neurosci 24:10289 –10301. McGeer PL, Boulding JE, Gibson WC, Foulkes RG (1961) Druginduced extrapyramidal reactions. Treatment with diphenhydramine hydrochloride and dihydroxyphenylalanine. JAMA 177:665– 670. Mesulam MM, Mash D, Hersh L, Bothwell M, Geula C (1992) Cholinergic innervation of the human striatum, globus pallidus, subthalamic nucleus, substantia nigra, and red nucleus. J Comp Neurol 323:252–268. Morris G, Arkadir D, Nevet A, Vaadia E, Bergman H (2004) Coincident but distinct messages of midbrain dopamine and striatal tonically active neurons. Neuron 43:133–143. Mouroux M, Féger J (1993) Evidence that the parafascicular projection to the subthalamic nucleus is glutamatergic. Neuroreport 4:613– 615. Narushima M, Uchigashima M, Fukaya M, Matsui M, Manabe T, Hashimoto K, Watanabe M, Kano M (2007) Tonic enhancement of endocannabinoid-mediated retrograde suppression of inhibition by cholinergic interneuron activity in the striatum. J Neurosci 27: 496 –506. Ochi M, Inoue H, Koizumi S, Shibata S, Watanabe S (1995) Long-term enhancement of dopamine release by high frequency tetanic stimulation via a N-methyl-D-aspartate-receptor-mediated pathway in rat striatum. Neuroscience 66:29 –36. Olson PA, Tkatch T, Hernandez-Lopez S, Ulrich S, Ilijic E, Mugnaini E, Zhang H, Bezprozvanny I, Surmeier DJ (2005) G-protein-coupled receptor modulation of striatal CaV1.3 L-type Ca2⫹ channels is dependent on a Shank-binding domain. J Neurosci 25:1050 –1062. Oswald MJ, Oorschot DE, Schulz JM, Lipski J, Reynolds JN (2009) IH current generates the afterhyperpolarisation following activation of subthreshold cortical synaptic inputs to striatal cholinergic interneurons. J Physiol 587:5879 –5897. Pakhotin P, Bracci E (2007) Cholinergic interneurons control the excitatory input to the striatum. J Neurosci 27:391– 400.

Pawlak V, Wickens JR, Kirkwood A, Kerr JN (2010) Timing is not everything: neuromodulation opens the STDP gate. Front Synaptic Neurosci 2:146. Perez-Burgos A, Perez-Rosello T, Salgado H, Flores-Barrera E, Prieto GA, Figueroa A, Galarraga E, Bargas J (2008) Muscarinic M(1) modulation of N and L types of calcium channels is mediated by protein kinase C in neostriatal neurons. Neuroscience 155:1079 – 1097. Perez-Rosello T, Figueroa A, Salgado H, Vilchis C, Tecuapetla F, Guzman JN, Galarraga E, Bargas J (2005) Cholinergic control of firing pattern and neurotransmission in rat neostriatal projection neurons: role of CaV2.1 and CaV2.2 Ca2⫹ channels. J Neurophysiol 93:2507–2519. Phelps PE, Houser CR, Vaughn JE (1985) Immunocytochemical localization of choline acetyltransferase within the rat neostriatum: a correlated light and electron microscopic study of cholinergic neurons and synapses. J Comp Neurol 238:286 –307. Pisani A, Bernardi G, Ding J, Surmeier DJ (2007) Re-emergence of striatal cholinergic interneurons in movement disorders. Trends Neurosci 30:545–553. Prensa L, Parent A (2001) The nigrostriatal pathway in the rat: a single-axon study of the relationship between dorsal and ventral tier nigral neurons and the striosome/matrix striatal compartments. J Neurosci 21:7247–7260. Ravel S, Legallet E, Apicella P (1999) Tonically active neurons in the monkey striatum do not preferentially respond to appetitive stimuli. Exp Brain Res 128:531–534. Raz A, Feingold A, Zelanskaya V, Vaadia E, Bergman H (1996) Neuronal synchronization of tonically active neurons in the striatum of normal and parkinsonian primates. J Neurophysiol 76:2083– 2088. Raz A, Frechter-Mazar V, Feingold A, Abeles M, Vaadia E, Bergman H (2001) Activity of pallidal and striatal tonically active neurons is correlated in mptp-treated monkeys but not in normal monkeys. J Neurosci 21:RC128. Redgrave P, Gurney K (2006) The short-latency dopamine signal: a role in discovering novel actions? Nat Rev Neurosci 7:967–975. Redgrave P, Gurney K, Reynolds J (2008) What is reinforced by phasic dopamine signals? Brain Res Rev 58:322–339. Reynolds JN, Hyland BI, Wickens JR (2004) Modulation of an afterhyperpolarization by the substantia nigra induces pauses in the tonic firing of striatal cholinergic interneurons. J Neurosci 24: 9870 –9877. Reynolds JN, Wickens JR (2002) Dopamine-dependent plasticity of corticostriatal synapses. Neural Netw 15:507–521. Reynolds JN, Wickens JR (2004) The corticostriatal input to giant aspiny interneurons in the rat: a candidate pathway for synchronising the response to reward-related cues. Brain Res 1011:115– 128. Rice ME, Cragg SJ (2004) Nicotine amplifies reward-related dopamine signals in striatum. Nat Neurosci 7:583–584. Schulz JM, Oswald MJ, Reynolds JN (2011) Visual-induced excitation leads to firing pauses in striatal cholinergic interneurons. J Neurosci 31:11133–11143. Schulz JM, Redgrave P, Mehring C, Aertsen A, Clements KM, Wickens JR, Reynolds JN (2009) Short-latency activation of striatal spiny neurons via subcortical visual pathways. J Neurosci 29:6336 – 6347. Shen W, Hamilton SE, Nathanson NM, Surmeier DJ (2005) Cholinergic suppression of KCNQ channel currents enhances excitability of striatal medium spiny neurons. J Neurosci 25:7449 – 7458. Shen W, Hernandez-Lopez S, Tkatch T, Held JE, Surmeier DJ (2004) Kv1. 2-containing K⫹ channels regulate subthreshold excitability of striatal medium spiny neurons. J Neurophysiol 91:1337–1349. Shen W, Tian X, Day M, Ulrich S, Tkatch T, Nathanson NM, Surmeier DJ (2007) Cholinergic modulation of Kir2 channels selectively el-

J. A. Goldberg and J. N. J. Reynolds / Neuroscience 198 (2011) 27– 43 evates dendritic excitability in striatopallidal neurons. Nat Neurosci 10:1458 –1466. Sidibé M, Smith Y (1999) Thalamic inputs to striatal interneurons in monkeys: synaptic organization and co-localization of calcium binding proteins. Neuroscience 89:1189 –1208. Smeal RM, Gaspar RC, Keefe KA, Wilcox KS (2007) A rat brain slice preparation for characterizing both thalamostriatal and corticostriatal afferents. J Neurosci Methods 159:224 –235. Smith Y, Bennett BD, Bolam JP, Parent A, Sadikot AF (1994) Synaptic relationships between dopaminergic afferents and cortical or thalamic input in the sensorimotor territory of the striatum in monkey. J Comp Neurol 344:1–19. Smith Y, Raju DV, Pare JF, Sidibe M (2004) The thalamostriatal system: a highly specific network of the basal ganglia circuitry. Trends Neurosci 27:520 –527. Song WJ, Tkatch T, Baranauskas G, Ichinohe N, Kitai ST, Surmeier DJ (1998) Somatodendritic depolarization-activated potassium currents in rat neostriatal cholinergic interneurons are predominantly of the A type and attributable to coexpression of Kv4.2 and Kv4.1 subunits. J Neurosci 18:3124 –3137. Sullivan MA, Chen H, Morikawa H (2008) Recurrent inhibitory network among striatal cholinergic interneurons. J Neurosci 28:8682– 8690. Surmeier DJ, Ding J, Day M, Wang Z, Shen W (2007) D1 and D2 dopamine-receptor modulation of striatal glutamatergic signaling in striatal medium spiny neurons. Trends Neurosci 30:228 –235. Surmeier DJ, Shen W, Day M, Gertler T, Chan S, Tian X, Plotkin JL (2010) The role of dopamine in modulating the structure and function of striatal circuits. Prog Brain Res 183:149 –167. Suzuki T, Miura M, Nishimura K, Aosaki T (2001) Dopamine-dependent synaptic plasticity in the striatal cholinergic interneurons. J Neurosci 21:6492– 6501. Thomas TM, Smith Y, Levey AI, Hersch SM (2000) Cortical inputs to m2-immunoreactive striatal interneurons in rat and monkey. Synapse 37:252–261. Tkatch T, Baranauskas G, Surmeier DJ (2000) Kv4.2 mRNA abundance and A-type K(⫹) current amplitude are linearly related in basal ganglia and basal forebrain neurons. J Neurosci 20:579 – 588.

43

Vilchis C, Bargas J, Ayala GX, Galván E, Galarraga E (2000) Ca2⫹ channels that activate Ca2⫹-dependent K⫹ currents in neostriatal neurons. Neuroscience 95:745–752. Wang Z, Kai L, Day M, Ronesi J, Yin HH, Ding J, Tkatch T, Lovinger DM, Surmeier DJ (2006) Dopaminergic control of corticostriatal long-term synaptic depression in medium spiny neurons is mediated by cholinergic interneurons. Neuron 50:443– 452. Watanabe K, Kimura M (1998) Dopamine receptor-mediated mechanisms involved in the expression of learned activity of primate striatal neurons. J Neurophysiol 79:2568 –2580. Wess J (1996) Molecular biology of muscarinic acetylcholine receptors. Crit Rev Neurobiol 10:69 –99. Wess J, Eglen RM, Gautam D (2007) Muscarinic acetylcholine receptors: mutant mice provide new insights for drug development. Nat Rev Drug Discov 6:721–733. Wilson CJ (2005) The mechanism of intrinsic amplification of hyperpolarizations and spontaneous bursting in striatal cholinergic interneurons. Neuron 45:575–585. Wilson CJ, Chang HT, Kitai ST (1990) Firing patterns and synaptic potentials of identified giant aspiny interneurons in the rat neostriatum. J Neurosci 10:508 –519. Wilson CJ, Goldberg JA (2006) Origin of the slow afterhyperpolarization and slow rhythmic bursting in striatal cholinergic interneurons. J Neurophysiol 95:196 –204. Witten IB, Lin SC, Brodsky M, Prakash R, Diester I, Anikeeva P, Gradinaru V, Ramakrishnan C, Deisseroth K (2010) Cholinergic interneurons control local circuit activity and cocaine conditioning. Science 330:1677–1681. Yan Z, Surmeier DJ (1996) Muscarinic (m2/m4) receptors reduce Nand P-type Ca2⫹ currents in rat neostriatal cholinergic interneurons through a fast, membrane-delimited, G-protein pathway. J Neurosci 16:2592–2604. Zhang H, Sulzer D (2004) Frequency-dependent modulation of dopamine release by nicotine. Nat Neurosci 7:581–582. Zhou FM, Liang Y, Dani JA (2001) Endogenous nicotinic cholinergic activity regulates dopamine release in the striatum. Nat Neurosci 4:1224 –1229. Zhou FM, Wilson CJ, Dani JA (2002) Cholinergic interneuron characteristics and nicotinic properties in the striatum. J Neurobiol 53: 590 – 605.

(Accepted 30 August 2011) (Available online 8 September 2011)