| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
The Journal of Neurophysiology Vol. 80 No. 4 October 1998, pp. 1828-1838
Copyright ©1998 by the American Physiological Society
Department of Neuroscience, University of Pennsylvania, Philadelphia, Pennsylvania 19104-6074
 |
ABSTRACT |
|---|
 Abstract
 Introduction
Methods
Results
Discussion
References
|
|---|
Dutar, Patrick, Huan M. Vu, and David J. Perkel. Multiple cell types distinguished by physiological, pharmacological, and anatomic properties in nucleus HVc of the adult zebra finch. J. Neurophysiol. 80: 1828-1838, 1998. Nucleus HVc of the songbird is a distinct forebrain region that is essential for song production and shows selective responses to complex auditory stimuli. Two neuronal populations within HVc give rise to its efferent projections. One projection, to the robust nucleus of the archistriatum (RA), serves as the primary motor pathway for song production, and can also carry auditory information to RA. The other projection of HVc begins a pathway through the anterior forebrain, (area X 
medial portion of the dorsolateral nucleus of the thalamus (DLM) lateral portion of the magnocellular nucleus of the anterior neostriatum (L-MAN) RA) that is crucial for song learning but, although active during singing, is not essential for adult song production. To test whether these different projection neuron classes have different functional properties, we recorded intracellularly from neurons in nucleus HVc in brain slices. We observed at least three classes of neuron based on intrinsic physiological and pharmacological properties as well as on synaptic inputs. We also examined the morphological properties of the cells by filling recorded neurons with neurobiotin. The different physiological cell types correspond to separate populations based on their soma size, dendritic extent, and axonal projection. Thus HVc neurons projecting to area X have large somata, show little spike-frequency adaptation, a hyperpolarizing response to the metabotropic glutamate receptor (mGluR) agonist (1S,3R)-trans-1-aminocyclopentane-1,3-dicarboxylic acid (ACPD), and exhibit a slow inhibitory postsynaptic potential (IPSP) following tetanic stimulation. Those HVc neurons projecting to motor nucleus RA have smaller somata, show strong accommodation, are not consistently hyperpolarized by ACPD, and exhibit no slow IPSP. A third, rarely recorded class of neurons fire in a sustained fashion at very high-frequency and may be interneurons. Thus the neuronal classes within HVc have different functional properties, which may be important for carrying specific information to their postsynaptic targets.
Nucleus HVc [used here as the full name of the nucleus, following Fortune and Margoliash (1995) Preparation of slices and electrophysiological recording
Adult male zebra finches ( Visualization of the neurons
In some experiments, recording electrodes were filled with 2-4% neurobiotin (Vector Laboratories, Burlingame, CA) dissolved in the recording solution. At the end of the recording, neurobiotin was injected into the neuron by passing depolarizing pulses (1 Hz, 600- to 700-ms duration, 0.3- to 1-nA intensity). We then allowed a survival time for the neuron and slice of 30 min to 1 h to permit the diffusion of neurobiotin into the axon and dendrites. The slice was then fixed in paraformaldehyde (4% in 0.1 M phosphate buffer) and kept at 4°C for Electrophysiological studies with the use of intracellular recordings from HVc neurons led to the identification of three main neuronal types having different physiological and pharmacological properties. We first determined that the recorded neurons fell into categories based on physiological properties. We tentatively called these type I, type II, and type III HVc neurons. We then demonstrated that these neurons have different pharmacological properties. Finally, we observed evidence for different anatomic properties. These data are based on stable intracellular recordings lasting up to 6 h from 75 neurons from 75 slices from 57 birds.
Physiological properties
The neurons that we recorded in HVc belong to two main types (and occasionally to a third type) having very different physiological properties.
TYPE I NEURONS.
The first type of neuron is illustrated in Fig. 1A. This neuronal population had a low spike threshold, a resting membrane potential of
TYPE II NEURONS.
A type II neuron is illustrated in Fig. 1B. Compared with the first type, it had a very hyperpolarized resting membrane potential (
TYPE III NEURONS.
The third type of neuron was encountered less frequently (n = 7, Fig. 1C). They were able to produce very high-frequency firing ( Pharmacological properties
The pharmacological properties of HVc neurons were studied with the use of two approaches. First, we studied the synaptic pharmacology of neurons following electrical stimulation within HVc. Second, we studied the neuronal effects of exogenously applied neurotransmitter agonists.
Synaptic responses
TYPE I NEURONS.
In response to high-frequency stimulation of afferents located within HVc (Schmidt and Perkel 1998
TYPE II NEURONS.
The synaptic response of these neurons was very different from that of the type I neurons in that tetanic stimulation induced a very large EPSP (amplitude 24.6 ± 2.1 mV and duration 7.72 ± 0.77 s, n = 13), which was not followed by an IPSP (Fig. 4B1). The lack of IPSP could not be explained by the small driving force on potassium because, even during depolarization by steady current injection, there was no slow IPSP (average depolarization 16 mV; n = 10/10). The amplitude of the slow EPSP was reduced to 44.1 ± 5% (n = 4) of control in the presence of CNQX (20 µM) and APV (50 µM). A component of the slow EPSP (12.8 ± 2.3 mV; n = 4) was resistant to APV and CNQX, and its nature has not yet been identified (Fig. 4B2).
TYPE III NEURONS.
The number of recorded neurons is small. However, these neurons exhibited an EPSP/IPSP sequence in response to stimulation within HVc. In contrast with type I neurons, the IPSP was totally blocked by application of a cocktail of GABAA and GABAB antagonists (not illustrated).
Response to drug application
Our results demonstrate that the different neuronal types differ in their sensitivity to exogenously applied neurotransmitter agonists. We applied agonists of GABAB receptors and metabotropic glutamate receptors. GABAB receptors are known to cause a dramatic hyperpolarization in HVc neurons (Perkel and Schmidt 1998) and in many other vertebrate neurons (Bowery 1993 TYPE I AND TYPE II NEURONS.
Both cell types were reliably hyperpolarized by superfusion of the GABAB receptor agonist baclofen (30 µM). The hyperpolarization was associated with a decrease in input resistance. The baclofen-induced hyperpolarization was stronger in type I neurons (
TYPE III.
Because of the small sample, our results must be interpreted cautiously. The type III neurons tested were hyperpolarized by baclofen ( Anatomy
Most of the neurons were recorded with the use of neurobiotin filled electrodes (see METHODS). This technique allowed us to stain the neurons and to characterize the morphology of the neurons and the direction of their axonal projections.
TYPE I NEURONS.
Of 30 recorded neurons having the type I physiological properties, we could find 27 cell bodies labeled. Cell somata were mostly round in shape, and their mean surface area was 145.1 ± 7.7 µm2 (n = 26; 1 cell was not adequately labeled for size measurement). They were widely distributed throughout nucleus HVc (Fig. 7A). The neurons had a large, extensively arborized dendritic field with no apparent polarization (Fig. 8). The maximal dendritic extension averaged 253.3 ± 6.4 µm (n = 24). In addition, the dendrites of these neurons had a great number of spines. We found the axons of the neurons leaving nucleus HVc in 15 neurons (Fig. 7B). In each case, the axon left in an anteroventral direction and appeared to be headed toward area X. In one case (Fig. 8A-C), after a single recording, we observed two filled somata and two axons. A more typical case is shown in Fig. 8D, in which a single cell was labeled. No axon of a type I neuron was found heading posteroventrally, i.e., toward nucleus RA. This finding indicates that the physiologically defined type I neurons are in fact the HVc neurons projecting to area X.
TYPE II NEURONS.
We labeled 13 neurons having type II physiology. Their somatic surface area averaged 106.9 ± 7.8 µm2. Thus these neurons were smaller than the type I neurons (P < 0.001). They were widely distributed throughout the nucleus (Fig. 7A). The maximal extension of the dendrites averaged 201.0 ± 9.4 µm, less than for type I neurons (P = 0.005). Like those of type I cells, these dendrites had numerous spines.
TYPE III NEURONS.
We filled only two neurons having type III physiology. These neurons resembled those of type I except that the dendrites were apparently devoid of spines. In addition, numerous fine processes were present throughout HVc, possibly representing extensive axonal divergence within HVc. The mean surface area of the soma was 135.5 ± 12.8 µm2. In addition, although we found a well-filled axon within HVc, we were unable to see it leaving HVc, suggesting that this neuron could be an interneuron.
We have described at least three types of neurons recorded in nucleus HVc. Although the initial classification was made by intrinsic physiological properties, the presence of the slow IPSP, sensitivity to mGluR activation, and location of axonal projection could each serve to distinguish these cell classes. Thus we have shown that neurons with type I physiological properties project to area X, whereas type II neurons project to RA. X-projecting neurons have a resting potential around
INTRODUCTION
Abstract
Introduction
Methods
Results
Discussion
References
] of the songbird is essential for singing and forms a part of the forebrain central pattern generator for song production (Nottebohm et al. 1976; Vu et al. 1994). HVc neurons show premotor firing before singing behavior (McCasland 1987; McCasland and Konishi 1981; Yu and Margoliash 1996), and disruption of HVc firing during singing can reset the song rhythm (Vu et al. 1994). In addition, HVc neurons have highly selective auditory properties, responding strongly to complex acoustic stimuli, particularly the bird's own song (Margoliash 1986). HVc is thus a critically important station for sensorimotor integration during song learning and production.
) and in zebra finch (Katz and Gurney 1981; Lewicki and Konishi 1996; Lewicki 1995) indicate that there are several morphologically distinguishable neuron types. Different subpopulations of neurons project to HVc's two targets, the anterior forebrain auditory region called area X, and the premotor robust nucleus of the archistriatum (RA) (Fortune and Margoliash 1995; Nottebohm et al. 1976). It is thus reasonable to hypothesize that different HVc cell types participate in different functions. Indeed, Katz and Gurney (1981) suggested that HVc neurons that project to area X (X-projecting cells) were more responsive to auditory stimuli than cells projecting to the robust nucleus of the archistriatum (RA-projecting cells), and the same trend was reported by Lewicki (1996). Recent immunohistochemical staining for Fos, the protein product of the immediate early gene c-fos, suggests that the RA-projecting cells are active while birds sing and thus that the two populations of HVc projection neurons have different functional properties (Kimpo and Doupe 1997). However, expression of the immediate early gene zenk is elevated in area X following singing, even in deafened birds, suggesting that a motor or efference copy signal is sent to area X as well (Jarvis and Nottebohm 1997).
; Lewicki 1996; Lewicki and Konishi 1995), and in vitro studies identified slow hyperpolarizing potentials (Kubota and Saito 1991; Schmidt and Perkel 1998). However, the basic cellular properties of the different HVc neuron types remain unclear. We used intracellular recording to examine the intrinsic firing properties, synaptic responses, and pharmacological sensitivity of these cells in brain slices. In addition, we asked whether neurons projecting to different targets have different functional properties. Indeed, the two populations of projection neurons differ strikingly in their characteristics. We report here three cell types distinguished by a number of physiological, pharmacological, and anatomic criteria. Of particular note, X-projecting neurons support sustained firing, exhibit a slow inhibitory postsynaptic potential (IPSP), and have an unusual hyperpolarizing response to metabotropic glutamate receptor (mGluR) activation. In contrast, RA-projecting neurons show strong accommodation, lack the slow IPSP, and are not hyperpolarized by mGluR agonists. These distinct functional properties may be important for the different functions performed by subcircuits within HVc and the rest of the song system.
METHODS
Abstract
Introduction
Methods
Results
Discussion
References
120 days old) were obtained from a local supplier or were bred in our colony at the University of Pennsylvania. Animals were housed three or four per cage in a room on a 13:11 h, light:dark cycle. It is likely that they were in breeding condition and singing before slicing, but their behavioral state was not systematically verified by audio recordings. Slices were prepared as described by Schmidt and Perkel (1998), and the procedures were approved by the Institutional Animal Care and Use Committee at the University of Pennsylvania. Briefly, a bird was anesthetized with halothane and decapitated. The brain was rapidly removed and placed in ice-cold artificial cerebrospinal fluid (ACSF) that was pregassed with 95% O2-5% CO2. The composition of the ACSF was (in mM) 119 NaCl, 2.5 KCl, 1 NaH2PO4, 26.2 NaHCO3, 11 D-glucose, 2.5 CaCl2, and 1.3 MgSO4. Parasagittal slices (300- to 400-µm thick) were prepared with a vibrating microtome and stored submerged in bubbled ACSF. For recording, a slice was transferred to a chamber where it was submerged and superfused at 1-2 ml/min with pregassed ACSF.
4 h. It was then placed in a cold sucrose solution (30% in 0.1 M phosphate buffer). After a few hours or days, the slice was resectioned at 40- to 60-µm thickness with a freezing microtome. Sections were then processed for visualization of injected neurons with the use of avidin and biotinylated horseradish peroxidase complex (ABC Elite Kit, Vector Laboratories) followed by reaction with diaminobenzidine. Drawings were made with a camera lucida. Soma area was calculated by multiplying one-half the major axis by one-half the minor axis by 
.
RESULTS
Abstract
Introduction
Methods
Results
Discussion
References
68.7 ± 1.2 (SE) mV (n = 42), and a high input resistance (161.8 ± 8.8 M
, n = 44). This cell type was able to fire several action potentials in response to a depolarizing current pulse with little accommodation of the spike discharge (Fig. 1A1, pulse duration 300 ms). The fast AHP following each spike had a large amplitude (15.2 ± 0.6 mV, n = 36), a slow time-to-peak (12.4 ± 1.6 ms, n = 32), and an especially long duration at half-amplitude (66.8 ± 4.7 ms, n = 32; Figs. 1A and 3B). Most of these neurons (28/37) also showed time-dependent inward rectification in response to hyperpolarizing pulses (Fig. 1A2). This is characterized by a sag in the hyperpolarizing response. The mean value of the sag for this neuronal population (measured as the difference between the peak and asymptotic values divided by the peak value) was 13.1 ± 1.7% (n = 28). Spontaneous synaptic events were rarely encountered.
View larger version (18K):
[in a new window]
FIG. 1.
Firing properties of the different types of HVc neurons. A: type I neuron. A1: 300-ms depolarizing current pulse (0.2 nA) applied to the neuron induced repetitive firing of action potentials. Each action potential was followed by a large, long-lasting afterhyperpolarization (AHP). These AHPs prevented high-frequency firing. A2: hyperpolarizing current pulse ( 0.2 nA) elicited a voltage deflection having a pronounced time-dependent inward rectification characterized by a sag in the voltage response and a rebound overshoot (average of 4 consecutive sweeps). Resting membrane potential: 72 mV. B: type II neuron. B1: 300-ms depolarizing current pulse (0.2 nA) induced an action potential followed by a short-duration AHP. It was common to elicit a few action potentials per pulse immediately following penetration of the cell, but after a few minutes the neuron reached a very negative resting membrane potential and gave this typical response with usually a single action potential. Increasing the amplitude of the depolarizing pulse did not increase the number of spikes to >2. B2: applying a negative constant current pulse (0.2 nA) induced a hyperpolarizing response with no apparent membrane rectification (4 averaged sweeps). Resting membrane potential: 82 mV. C: type III neuron. C1: this type of neuron, more rarely encountered, was characterized by its ability to discharge at very high frequency in response to a 300-ms depolarizing current pulse. In this example, the discharge frequency was 83 Hz. C2: hyperpolarizing pulse induced a voltage deflection exhibiting time-dependent inward rectification. Resting membrane potential: 65 mV
View larger version (25K):
[in a new window]
FIG. 3.
Physiological distinction between type I and type III neurons. A: number of action potentials elicited in type I and type III neurons during a 300-ms depolarizing current pulse as a function of current intensity. We observed that the type I neurons ( 
, n = 16) fire a lower number of spikes at higher stimulation intensity, possibly because the long-duration AHP prevented high-frequency firing. In contrast, type III neurons (
, n = 4) were able to fire action potentials at very high frequency. The curve reaches almost 50 spikes/300 ms for the higher current intensities, giving a mean maximum discharge rate of about 165 Hz. B: examples of the firing of type I (left) and III (right) neurons in response to depolarizing pulses applied with increasing intensities: bottom, 0.1 nA; middle, 0.2 nA; top, 0.4 nA. Resting membrane potential for both neurons: 70 mV.
80.9 ± 1.5 mV, n = 27) and a low input resistance (87.1 ± 8.2 M, n = 30) and was able to fire only one or two action potentials in response to a depolarizing pulse (Fig. 1B1, pulse duration 300 ms). These values significantly differed from those of type I cells (P < 10
4 and P < 104, respectively). These neurons were not simply unable to fire repetitively because they fired at 10 Hz when given brief (50 ms) depolarizing pulses at that frequency (n = 4/4). Each spike was followed by a fast AHP that had kinetics different from those of the other neuronal types. Although the amplitude of the AHP in type II cells (12.8 ± 1.5 mV, n = 22) was comparable with that in other neuronal types, in contrast, the time-to-peak (2.6 ± 0.3 ms, n = 21) and the duration at half-amplitude (10.4 ± 2.1 ms, n = 21) of the AHP were much shorter in these neurons. The response to hyperpolarizing current pulses exhibited no trace of time-dependent inward rectification or sag (Fig. 1B2). Spontaneous depolarizing synaptic potentials were frequently observed, but they were not examined further. When we plotted one physiological parameter against another (e.g., number of action potentials induced by a 0.4-nA, 300-ms depolarizing current pulse vs. AHP duration, Fig. 2), we observed that the neurons we tentatively called type I and type II form two separate populations (see also Fig. 6, B and C).
View larger version (14K):
[in a new window]
FIG. 2.
Types I and II neurons have distinct physiological properties. Scatter plot of the number of action potentials induced by a 300-ms long, 0.4-nA depolarizing current pulse against the duration of the fast AHP, measured at half-amplitude. The neurons of type I ( 
) and type II (
) fell into two separate populations. Note the different scales for the fast AHP duration for the 2 cell types.
View larger version (16K):
[in a new window]
FIG. 6.
Pharmacology of type I and type II neurons. A: average changes in membrane potential of type I ( 
) and type II () neurons in response to drug application (mean ± SE). Note the substantially smaller responses of type II neurons to baclofen and serotonin (5-HT) and their small depolarization in response to ACPD or insensitivity to (2S,1'S,2'S)-2-(carboxycyclopropyl)glycine (L-CCG-I). Numbers indicate the number of neurons tested. B: scatter plot, for each type I (n = 24) and type II neuron (n = 19) tested of the response to metabotropic glutamate receptor (mGluR) agonist application against the AHP duration, which was a key distinguishing intrinsic property. C: scatter plot, for each type I (n = 17) and type II neuron (n = 12) tested of the response to baclofen against the response to mGluR agonist. Type I neurons showed a correlation of the 2 responses, whereas the type II neurons did not (see text).
150 spikes/s) in response to a depolarizing pulse (see Figs. 1C1 and 3). To test whether they really fell into a different category from the type I cells, we examined their firing in response to depolarizing stimuli. The number of action potentials elicited in response to a depolarizing pulse (0.2 nA, 300 ms) was 16.5 ± 2.9 (n = 4) for type III neurons compared with 3.6 ± 0.5 (n = 21) for type I cells. This was the primary feature that distinguished type III from type I cells. These neurons had a mean resting membrane potential of 67.6 ± 2.0 mV (n = 7), a membrane input resistance of 107.7 ± 13.2 M 12 (n = 7), and a pronounced time-dependent inward rectification (sag 22.3% ± 4.3, Fig. 1C2). The amplitude, time-to-peak, and duration at half-amplitude of AHPs were respectively 13.2 ± 1.4, 5.1 ± 2.0, and 28.4 ± 7.3 ms (n = 7). Of these seven neurons, three showed some phases of bursting activity.
), type I neurons exhibited a slow excitatory postsynaptic potential (EPSP; amplitude 19.3 ± 1.3 mV and duration 2.78 ± 0.19 s; n = 19), followed by a slow IPSP (Fig. 4A1). The IPSP following the EPSP had an amplitude of 6.1 ± 0.5 mV (n = 19).
View larger version (11K):
[in a new window]
FIG. 4.
Absence of a slow hyperpolarization in type I neurons. A: in a type I neuron, high-frequency electrical stimulation of nucleus HVc (50 pulses delivered at 100 Hz, 
) induced a synaptic response recorded intracellularly having a depolarizing and a hyperpolarizing component. In control conditions (A1), a slow depolarizing phase (duration measured from end of stimulation until return to baseline potential) was followed by a long-lasting hyperpolarizing response. A2: in the presence of the ionotropic glutamate receptor antagonists 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX; 20 µM) and D,L-2-amino-5-phosphonovaleric acid (APV; 50 µM) and of the 
-aminobutyric acid-A (GABAA) and GABAB receptor antagonists bicuculline methiodide (BMI; 40 µM) and CGP35348 (500 µM), the depolarizing potential and a large part of the hyperpolarization were blocked. However, a hyperpolarizing response persisted in this neuronal population. This phase, defined as a slow hyperpolarization by Schmidt and Perkel (1998), was blocked by the sodium channel blocker tetrodotoxin (TTX; 2 µM). A3: resting membrane potential: 66 mV. B: same protocol was applied to a type II neuron. We observed in the control medium (B1) the presence of the depolarizing potential but no hyperpolarizing response, even when the neuron was held depolarized. The absence in this neuronal population of the slow hyperpolarization is more obvious during bath application of CNQX, APV, and bicuculline (B2). Resting membrane potential: 86 mV.
-aminobutyric acid-A (GABAA) antagonists picrotoxin (50 µM) or bicuculline methiodide (40 µM) and the -aminobutyric acid-B (GABAB) antagonist CGP35348 (500 µM). As reported previously (Schmidt and Perkel 1998), a portion of this hyperpolarization was sensitive to bicuculline and CGP35348 and thus was mediated by activation of GABAA and GABAB receptors. A part of the IPSP was resistant to the GABA antagonists (Fig. 4A2) but was blocked by tetrodotoxin (TTX; Fig. 4A3). Its pharmacological nature is still unknown (Schmidt and Perkel 1998).
). Metabotropic glutamate receptors are widely distributed throughout the vertebrate nervous system and affect most neuronal types tested (Pin and Duvoisin 1995; Schoepp and Conn 1993).
14.0 ± 1.4 mV; n = 14; Fig. 5, top panel) than in type II neurons (6.6 ± 0.7 mV; n = 17; Fig. 5, middle panel). The difference between type I and type II neurons was statistically significant (P < 0.005). This difference can be explained, at least in part, by the lower resting membrane potential and input resistance of type II neurons.
View larger version (16K):
[in a new window]
FIG. 5.
Different neuronal types have different pharmacological properties. Top panel: response of a type I neuron to bath application of the GABAB receptor agonist baclofen (30 µM) and the metabotropic glutamate receptor agonist (1S,3R)-trans-1-aminocyclopentane1,3-dicarboxylic acid (ACPD; 100 µM). Baclofen (left) always induced a large membrane hyperpolarization associated with a decrease in input resistance (not shown), and ACPD, applied to the same neuron (right) caused a hyperpolarization associated with a decrease in input resistance (not shown). Duration of agonist application is indicated by the horizontal bars. Resting membrane potential: 78 mV. Middle panel: in this type II neuron, baclofen (30 µM, left) caused a hyperpolarization. In contrast with the type I neuron, ACPD (100 µM, right) caused a reversible depolarization of the cell membrane. Resting membrane potential: 73 mV. Bottom panel: in a type III neuron, baclofen induced a large hyperpolarization, but ACPD had no consistent effect.
10.8 ± 1.6 mV; n = 10), ACPD (12.7 ± 0.9 mV; n = 24), or L-CCG-I (14.7 ± 1.3 mV; n = 13). Type II neurons were less sensitive to 5-HT (2.5 ± 0.7 mV, n = 6; P < 0.02) and gave variable responses to mGluR agonists (Fig. 6A). They were slightly hyperpolarized, insensitive, or even depolarized by L-CCG-I or ACPD (Fig. 5, middle panel). On average, ACPD depolarized type II neurons by 1.9 ± 0.7 mV (n = 17), and L-CCG-I depolarized these cells by 0.14 ± 1.1 mV (n = 14). In addition, both types of neurons were not consistently affected by the muscarinic cholinergic agonist carbachol (type I, n = 5; type II, n = 3), dopamine (n = 2; n = 2), or norepinephrine (n = 5; n = 3).
6.4 ± 1.1 mV, n = 6) but were insensitive to mGluR agonists ACPD (0 mV, n = 4) and L-CCG-I (0.6 ± 0.6 mV, n = 5; Fig. 5, bottom panel). In two cases, 5-HT depolarized the neuron and induced bursting activity.
View larger version (20K):
[in a new window]
FIG. 7.
Distribution of labeled cell bodies and axons. A: different neuronal types have no topographic organization. This schematic representation shows the location of the recorded neurons superimposed onto a canonical drawing of HVc. This distribution was drawn after observation of the cell bodies filled with neurobiotin and reconstruction with a camera lucida. Cell bodies of physiologically defined types I (open triangles) and II (filled circles) neurons were found distributed across the whole nucleus. Three pairs of adjacent circles represent locations where 2 cell bodies were found filled with neurobiotin after a single type II recording was obtained. Only 2 type III neurons (filled squares) were filled with neurobiotin. Neuronal types were intermingled and widely distributed throughout the entire HVc nucleus. B: distribution of labeled axons leaving HVc. Camera lucida drawing of the course of the axons of type I and type II neurons at the boundary of nucleus HVc. Axons of type I neurons were all found anterior to the dotted line and coursed in the anterior or anteroventral direction (toward area X). Axons of type II neurons were found posterior to the dotted line and are directed in the ventroposterior direction (toward nucleus RA). Axonal course is illustrated only in the region of the HVc boundary. Ordinarily we could follow the processes for some distance following their exit from the nucleus. Only 13 type I axons are illustrated here because, in 1 case, coronal sections were made, and we could not superimpose the results on this summary graph, but the axon clearly left HVc in the anteroventral direction (toward area X).
View larger version (22K):
[in a new window]
FIG. 8.
Partial reconstruction of a type I neuron intracellularly filled with neurobiotin. A: camera lucida drawing of a parasagittal view of the cell body, the dendritic arborization, and the axonal trajectory. This is a 2-D projection of a partial 3-D reconstruction of the neuron generated from examination of multiple sections. B and C: bright field photomicrograph of parts of this neuron showing, at high-magnification, the cell body (B), and a part of the axon outside HVc (C) heading in the direction of area X. In this case, 2 cell bodies within HVc and 2 axons were filled following a single recording. Other parts of the axon used for the reconstruction illustrated in A are on different sections indicating a lateromedial direction for the HVc X projection. D: camera lucida drawing of a parasagittal view of another type I neuron. Orientation and scaling are the same as in A.
; Lewicki 1996). Second, in three cases we were able to find two cell bodies labeled following a single somatic injection (Fig. 9), suggesting a high incidence of coupling in these neurons. In two cases of coupled neurons, it was possible to follow two independent axons, both projecting toward RA.
View larger version (18K):
[in a new window]
FIG. 9.
Partial reconstruction of a type II neuron intracellularly filled with neurobiotin. A: camera lucida drawing of a parasagittal view of the cell body, the dendritic arborization, and the axonal trajectory. Note the presence of 2 labeled cell bodies for a single neurobiotin injection, suggesting the presence of coupling in this neuronal population. Finally, it was possible to follow the axon going toward the nucleus RA in a single slice, indicating that the HVc-RA projection is organized in the parasagittal plane. Insets are bright field photomicrographs of parts of this neuron showing the cell bodies (B) and a part of the axon leaving HVc (C). Note also the different directions of axons in type I (Fig. 8) and type II (this figure) neurons.
DISCUSSION
Abstract
Introduction
Methods
Results
Discussion
References
68 mV, show sustained firing with a large AHP following each action potential, exhibit a slow IPSP in response to repetitive synaptic stimulation, hyperpolarize in response to ACPD, and in a few cases show dye coupling. RA-projecting neurons have an apparently quite hyperpolarized resting potential (approximately 80 mV), show dramatic spike-frequency accommodation, have no slow IPSP, are not hyperpolarized by mGluR agonists, and can show dye coupling with other HVc neurons. A third cell type was also observed, which could sustain very rapid firing rates, did not respond to mGluR agonists, and whose axon was not observed to leave HVc.
identified in HVc slices neurons that appear to correspond to the type I (X-projecting) cell reported here. They have similar resting potential, input resistance, and time-dependent inward rectification; in addition, both fire multiple action potentials in response to depolarizing current pulses, and both have a prominent slow AHP following prolonged depolarization (Schmidt and Perkel 1998). The large fast AHP following individual action potentials resembles those of auditory neurons illustrated by Lewicki and Konishi (1995) and Lewicki (1996).
) as well as for bushy cells of the mammalian anteroventral cochlear nucleus (Manis and Marx 1991; Oertel 1983; Wu and Oertel 1994) and for their putative avian homologues (Reyes et al. 1994). A rapidly activated potassium conductance may prevent repetitive firing in response to a sustained current pulse (Manis and Marx 1991; Reyes et al. 1994). These neuronal types are specialized for carrying temporal information with great specificity. To explain repetitive firing in response to rapid, brief depolarizing stimuli, such a rapidly activated potassium conductance would also need to be rapidly deactivated upon hyperpolarization. It is possible that the RA-projecting HVc neurons use a similar strategy, perhaps even a similar potassium conductance, to help generate a temporally precise rhythm.
). An intriguing speculation is that these premotor neurons must undergo some switch of their electrophysiological properties that permits rapid firing before and during singing. We did not observe such an action by the cholinergic agonist carbachol, 5-HT, norepinephrine, or dopamine, but many other candidate modulators were not tested.
. Apparently, this cell type does not send an axon outside HVc. It is possible that this cell type represents an interneuron (Nixdorf et al. 1989). We might have missed other classes of cells, either because we did not record from them in this study or because we did not sufficiently characterize the cells reported here to discover latent subtypes.
) in X-projecting but not RA-projecting neurons. Although the small potassium driving force might interfere with the ability to record the slow IPSP in RA-projecting cells, it is unlikely that we missed the slow IPSP in those cells because the response remained depolarizing even when the membrane potential was held depolarized by intracellular current injection. A hypothetical role for the slow hyperpolarization is to shape auditory responses during and following singing (Schmidt and Perkel 1998). The presence of a slow hyperpolarization selectively in X-projecting neurons is consistent with a role for this response in reducing auditory responsiveness for several seconds following singing (McCasland and Konishi 1981). Indeed, a long-lasting hyperpolarization can occur in response to song presentation (Lewicki 1996). Of course, other roles of the slow hyperpolarization specific to the X-projecting cells remain possible.
). In RA-projecting neurons, these agonists cause a smaller hyperpolarization, at least in part caused by the smaller potassium driving force and lower input resistance. Nonetheless, these hyperpolarizations are reliable, indicating that there is sufficient driving force to elicit this response.
). In contrast, RA-projecting neurons are not consistently hyperpolarized by ACPD, suggesting that mGluRs are not linked to potassium channels in this cell type. However, RA-projecting cells are hyperpolarized by baclofen or 5-HT, indicating that some neurotransmitter-activated potassium channels are present. Thus, a key distinction between the X- and RA-projecting populations is in the effector to which mGluRs are coupled. This distinction might have practical value for song-system research in that it might serve to help identify X- or RA-projecting cells during single-unit recording in vivo. The differential sensitivity to different neurotransmitters suggests there could be selective modulation of auditory and motor function by neuromodulators.
reported auditory responses in some neurons projecting to area X but not in RA-projecting cells. Lewicki and Konishi (1995) and Lewicki (1996) noted a similar trend. Kimpo and Doupe (1997) found FOS expression in RA-projecting but not X-projecting neurons following singing, the clearest functional difference reported to date. It is tempting to speculate that the auditory neurons in HVc are those that project to area X, whereas the RA-projecting cells do not respond to auditory stimuli. However, Doupe and Konishi (1991) showed that RA receives HVc-dependent auditory responses in the absence of the lateral portion of the magnocellular nucleus of the anterior neostriatum (L-MAN). This raises the possibility that there may be two types of RA-projecting neuron, auditory and nonauditory. Our results do not support the presence of physiologically distinct populations of RA-projecting neurons; we never recorded RA-projecting neurons with type I properties. Thus if auditory HVc neurons project to RA, then either we did not record from them in this study or auditory neurons do not have a distinct profile of physiological characteristics.
and of Fortune and Margoliash (1995). However, both types of cell possess dendritic spines. The X-projecting cells we recorded appear to correspond to the thick dendrite class of HVc neuron described by Nixdorf et al. (1989) and filled by Benton et al. (1998) in the canary. Fortune and Margoliash (1995) reported a similar cell class retrogradely labeled from area X and also observed a class of smaller neurons. It is possible we did not record from this smaller class of X-projecting neurons.
. However, we did not observe the heterogeneity described by Fortune and Margoliash (1995) in RA-projecting cells, again perhaps reflecting sampling bias caused by intracellular recording. Our type III neurons may represent their aspiny, putative interneuron class.
reported electron microscopic evidence for the presence of gap junctions in female canary HVc. Testosterone induced increases in the number of gap junctions in HVc as well as increased singing. Our observation that RA-projecting cells (and less commonly X-projecting cells) are dye coupled to other HVc cells could be explained by gap junctions within HVc of male zebra finches.
| |
ACKNOWLEDGEMENTS |
|---|
We are grateful to members of the Perkel laboratory for valuable discussions and to Dr. Marc F. Schmidt for helpful comments on the manuscript.
This work was supported by National Institute of Deafness and Other Communication Disorders Grant R03 DC-02477 to D. J. Perkel and a grant from the Fondation Pour la Recherche Médicale to P. Dutar.
| |
FOOTNOTES |
|---|
Address for reprint requests: D. J. Perkel, Dept. of Neuroscience, 215 Stemmler Hall, University of Pennsylvania, Philadelphia, PA 19104-6074.
Received 3 April 1998; accepted in final form 1 June 1998.
| |
REFERENCES |
|---|
|
Abstract
Introduction
Methods
Results
Discussion
References
|
|---|
This article has been cited by other articles:
|
|
 |
|
  J. A. Thompson, W. Wu, R. Bertram, and F. Johnson Auditory-Dependent Vocal Recovery in Adult Male Zebra Finches Is Facilitated by Lesion of a Forebrain Pathway That Includes the Basal Ganglia J. Neurosci., November 7, 2007; 27(45): 12308 - 12320. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Meitzen, I. T. Moore, K. Lent, E. A. Brenowitz, and D. J. Perkel Steroid Hormones Act Transsynaptically within the Forebrain to Regulate Neuronal Phenotype and Song Stereotypy J. Neurosci., October 31, 2007; 27(44): 12045 - 12057. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. J. Coleman, A. Roy, J. M. Wild, and R. Mooney Thalamic Gating of Auditory Responses in Telencephalic Song Control Nuclei J. Neurosci., September 12, 2007; 27(37): 10024 - 10036. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. H. R. Hahnloser, A. A. Kozhevnikov, and M. S. Fee Sleep-Related Neural Activity in a Premotor and a Basal-Ganglia Pathway of the Songbird J Neurophysiol, August 1, 2006; 96(2): 794 - 812. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. J. Rosen and R. Mooney Synaptic Interactions Underlying Song-Selectivity in the Avian Nucleus HVC Revealed by Dual Intracellular Recordings J Neurophysiol, February 1, 2006; 95(2): 1158 - 1175. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. M. Solis and D. J. Perkel Rhythmic Activity in a Forebrain Vocal Control Nucleus In Vitro J. Neurosci., March 16, 2005; 25(11): 2811 - 2822. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Mooney and J. F. Prather The HVC Microcircuit: The Synaptic Basis for Interactions between Song Motor and Vocal Plasticity Pathways J. Neurosci., February 23, 2005; 25(8): 1952 - 1964. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. J. Coleman and R. Mooney Synaptic Transformations Underlying Highly Selective Auditory Representations of Learned Birdsong J. Neurosci., August 18, 2004; 24(33): 7251 - 7265. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. D. I. Abarbanel, L. Gibb, G. B. Mindlin, and S. Talathi Mapping Neural Architectures Onto Acoustic Features of Birdsong J Neurophysiol, July 1, 2004; 92(1): 96 - 110. [Abstract] [Full Text] [PDF]
|
|
|
|
