| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Department of Neurobiology and Behavior, Cornell University, Ithaca, New York
Submitted 19 January 2006; accepted in final form 3 April 2006
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
; Jurgens 1994; Kennedy 1975; Seller 1981; Wild 1997). Although inputs from forebrain limbic structures, as demonstrated in mammals, support the PAG's role in coupling motivational state to vocal behavior (Bandler and Shipley 1994; Behbehani 1995; Holstege 1998; Jurgens 2002; Sewards and Sewards 2003), few studies show how PAG neuronal activity shapes the temporal parameters of vocal motor output. PAG activity is clearly necessary for call initiation (Behbehani 1995; Jurgens 2002). Single-unit recordings in vocalizing macaques show PAG activity correlated with temporal features (e.g., duration) of vocalizations (Larson 1991; Larson and Kistler 1986), implying a more direct role in vocal patterning than classically assumed (Davis et al. 1996; Jordan 1998; Jurgens 1994, 2002; Swanson 2000). However, similar recordings in squirrel monkeys do not reveal such correlations (Dusterhoft et al. 2000, 2004). Although these findings may represent species differences, the interpretation is currently complicated by an incomplete understanding of the relationship between ensemble muscle activity and the acoustic properties of primate vocalizations. This makes it difficult to establish the PAG's exact role in vocal patterning. We therefore chose to examine this question in a species with a simple vocal repertoire and a more direct translation from neural activity pattern to muscle action and vocal output.
Midshipman fish (Porichthys notatus) depend on vocal communication for successful courtship and reproduction. Territorial males use sonic swim bladder muscles (Fig. 1A) to produce several call types, differing primarily in duration (Brantley and Bass 1994) (Fig. 1B). The sonic muscles are innervated by a hindbrain–spinal sonic motor nucleus (SMN) that receives input from nearby pacemaker neurons (PN) (Bass and Baker 1990) (Fig. 1C). The rhythmic output of the PN–SMN circuit, referred to as fictive vocalization, directly establishes muscle contraction rate and, in turn, call fundamental frequency (Bass and Baker 1990). A ventral medullary nucleus (VM) links the PN–SMN circuitry to a midbrain region similar in location to the mammalian PAG (Fig. 1C) (Bass et al. 1994; Goodson and Bass 2002). Electrical stimulation of this midbrain region elicits vocalization (Demski and Gerald 1972, 1974; Fine 1979; Goodson and Bass 2002). The simplicity of vocal motor production, specifically a one-to-one translation from the SMN spike train to acoustic properties of the call, make this an ideal system for identifying the role of PAG neurons in vocal initiation and/or patterning.
|
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Midshipman fish (Porichthys notatus) have two male reproductive morphs that differ in vocal and spawning behaviors (Bass 1996). All experiments were performed on adult type I (territorial) males, which have the most dynamic vocal repertoire (Bass et al. 1999). Fish were collected from tidal pool nesting sites or by offshore trawls in northern California and Washington, shipped to Cornell University, and maintained in artificial seawater (ASW) tanks at about 15°C. All experimental procedures were approved by the Cornell University Institutional Animal Care and Use Committee.
Surgeries
Surgical procedures were similar to those described previously (Bass and Baker 1990; Goodson and Bass 2000c). Fish were anesthetized by immersion in 0.025% benzocaine (ethyl p-amino benzoate, Sigma, St. Louis, MO). Local anesthetic [0.2 ml of 0.25% bupivacaine (Abbott Labs, North Chicago, IL) with 0.01 mg/ml epinephrine (International Medication Systems, South El Monte, CA)] was then injected subdermally to the top of the head. The hypothalamus, midbrain, and hindbrain were exposed by dorsal craniotomy. Before transfer to the experimental apparatus, fish were immobilized with an intramuscular injection of pancuronium bromide (about 5 mg/kg, Baxter Healthcare, Deerfield, IL). For all experiments, fish were suspended in a parafilm sling in a Plexiglas tank with head stabilized, and ASW at about 15°C was perfused continuously across the gills. Exposed portions of the brain were kept covered with an inert, electrically conductive fluorocarbon (Fluorinert, 3M, St. Paul, MN). Fish were left to rest for about 1 h before starting electrophysiology experiments to allow all residual benzocaine to wash out of their system.
Extracellular recordings
FICTIVE VOCALIZATIONS.
The vocal–motor output of the hindbrain vocal pattern generator—the "fictive vocalization"—was monitored with an extracellular electrode [75-µm diameter Teflon-coated silver wire (A-M Systems, Sequim, WA) with an exposed ball tip, 125–200 µm in diameter] placed on an occipital nerve root that carries the axons of motor neurons from the ipsilateral sonic motor nucleus to the ipsilateral sonic muscle; both sonic motor nuclei fire in phase (Bass and Baker 1990). These nerve root recordings were amplified 1,000x and band-pass filtered from 1 Hz to 20 kHz with an A-M Systems differential AC amplifier (Model 1700).
PAG NEURONS.
Extracellular recordings from PAG neurons were obtained using glass microelectrodes (6010, A-M Systems) pulled to a tip resistance of 10–15 M
on a Flaming/Brown micropipette puller (Model P-97, Sutter Instruments, Novato, CA) filled with 2 M NaCl. The electrode tip solution also contained 5% dextran tetramethylrhodamine (10,000 MW, D-1868, Molecular Probes, Eugene, OR) to label each recording site. Preamplified signals were amplified 1,000x total (Model NB-100 amplifier, Biomedical Engineering, Thornwood, NY; and an A-M Systems Model 1700 differential AC amplifier) and band-pass filtered from 0.3 to 10 kHz.
Forebrain stimulation of PAG neurons and fictive vocalizations
RATIONALE.
Previous studies established the medial longitudinal fasciculus (MLF) of the midbrain as a site where low-amplitude stimulation evokes naturalistic vocalizations (Demski and Gerald 1972; Fine and Perini 1994). Thus our preliminary experiments focused on the MLF and showed that stimulation here elicited reliable vocal responses that did not fatigue with repeated stimulation. These experiments facilitated the reliable localization and isolation of single PAG neurons with stimulus-modulated activity concurrent with stimulus-evoked fictive vocalizations. However, our ongoing anatomical studies showed the MLF to be the main descending pathway for PAG axons (see last section of RESULTS). Thus changes in PAG neuron firing evoked by MLF stimulation were likely antidromically mediated and therefore would not resemble the endogenous pattern of vocal-related PAG activity during spontaneous vocalization. Neuroanatomical and brain stimulation studies suggested that the ventral tuberal hypothalamus (vT) was a likely candidate for a source of descending input to the PAG (Goodson and Bass 2000a, 2002), as we later confirmed in our own anatomical studies (see RESULTS). Stimulation at sites in and around vT evoked a less-reliable vocal response, at longer and more variable latency, that tended to fatigue with repeated stimulation. Although these features made it more difficult to isolate vocal-related PAG neurons, all of the available data indicated that PAG activity evoked by stimulating vT should be more reflective of endogenous PAG vocal activity patterns. Therefore, for the purposes of this paper, we present data from 48 PAG neurons recorded during vT stimulation. The results of the MLF stimulation experiments were used to analyze how descending midbrain output, inclusive of PAG axons, affects features of the vocal response (see below).
PROCEDURES.
Surface landmarks and micromanipulator coordinates were used to guide insulated tungsten stimulating electrodes (125-µm diameter, 8 ° tip angle, 5-M impedance; A-M Systems) to sites in or near either vT or the MLF. Brief trains of stimuli (3–15 pulses, 1-ms pulse duration, 333-Hz repetition rate, 50–75 µA) were delivered using a WPI stimulus isolation unit (Model 850A, World Precision Instruments, Sarasota, FL), with stimulus timing parameters driven by Tucker-Davis System II software and hardware (Alachua, FL). The stimulating electrode was lowered into the brain until a stable reliable vocal output was obtained. The extracellular electrode was then positioned over the PAG and advanced into the brain using a Burleigh microdrive (Model LSS-1000, Burleigh Instruments, Fishers, NY). Once the electrode was roughly at the correct depth, we began searching for units whose activity was clearly modulated by the stimulus. Because the vocal response elicited by vT stimulation tended to fatigue with repeated stimulation, we typically repeated only about 40 stimulus sweeps (1 every 2 s) at a time, followed by a 5- to 10-min rest interval. This stimulation pattern was followed both while searching for PAG units and while acquiring data once a stable unit was isolated. Because PAG neurons tended to have either low or no spontaneous activity (see RESULTS), units were often detected based only on the presence of stimulus-evoked action potentials. Once a stable unit was isolated, spike and vocal output data were recorded while the stimulus duration was varied, resulting in changes in the latency, duration, and probability of the vocal response. Spontaneous spike data were also recorded. After recording, the recording sites were marked by injecting pulsed positive current (+3 µA, 50% duty cycle, about 10 min) to iontophorese the fluorescent dye out of the electrode tip. In addition, stimulation sites were confirmed after each experiment by the small electrolytic lesion left at each site.
Data analysis and statistics of PAG neuronal activity
Both PAG and nerve root signals, digitized at 50 kHz, were acquired using a Tucker-Davis System II data acquisition system with BrainWare v6.3 software (Tucker-Davis). Initial post hoc analyses of the extracellular and vocal nerve recordings were performed with Brainware v6.3 software (Tucker-Davis). The shapes of all action potentials were examined, and multiple units, when present, were defined based on differences in various parameters of spike shape (amplitude of each peak, interpeak interval). When all spikes were plotted in parameter space, distinct clusters were readily apparent. Visual inspection of voltage versus time plots of each cluster confirmed that all spikes within a cluster were from the same unit (e.g., Fig. 1D). The few spikes that did not conform to cluster boundaries were discarded. Of the 143 total neurons recorded, the majority of all recording sites were single units (112 of 127); of the 15 multiunit sites, 14 revealed two clearly distinguishable units each, and one revealed three clear units. The 48 units recorded during vT stimulation represent 42 separate recording sites: 36 sites with only a single unit plus six sites each yielding two clearly distinguishable single units. The spike times for each unit, and for the coincident nerve recording of the vocal response pulses, were exported for further analysis.
Custom Matlab scripts [version 7.0.1 (R14), The MathWorks, Natick, MA] were used to compile, annotate, and analyze the spike records. For each unit confirmed by histology to be in the PAG (see following text), we calculated (Table 1): 1) the mean spontaneous firing rate (from trials with no stimulus), 2) the net mean number of stimulus-evoked spikes (over the first 800 ms poststimulus, reduced by the number of expected spontaneous spikes), 3) the mode of the interspike interval distribution, 4) the latency to the first poststimulus spike, and 5) the mean lag time between the first poststimulus unit spike and the first pulse of the vocal response. In addition, for each of these fish, we determined various parameters of the vocal response, including (Table 1): 1) the total mean number of pulses (over the first 800 ms poststimulus), 2) the mean number of pulses in the first vocal burst (a burst was defined as a series of pulses, each separated from the next by no more than 50 ms), 3) the latency to the first pulse of the vocal response, 4) the mean interpulse interval, and 5) the mean interburst interval. Poststimulus time histograms (PSTHs) for each unit and vocal response, perievent time histograms (PETHs) for each unit (centered on the time of the first vocal response pulse), and frequency distributions of interspike intervals (for each unit and vocal response) were plotted using Origin version 6.1 software (Origin Lab, Northampton, MA).
|
It seemed likely that variation in the number of stimulus pulses from trial to trial could affect both the unit activity and the vocal response separately. We therefore wanted to be sure that any detected correlations between the unit activity and the vocal response were not being driven by changes in the stimulus duration. Thus for all four correlational analyses, we ran multivariate regressions with both stimulus pulse number and unit activity (either latency or net spikes) as independent variables. In addition, we tested for correlations between the vocal response and unit activity on the subset of trials in which stimulus duration was kept constant. Because for some neurons there were few data at particular stimulus durations, the multivariate regression analysis was used as the criterion for whether there was a significant correlation between unit activity and vocal response for each cell. For clarity, however, the illustrative data presented in the figures is only for trials in which stimulus duration was held constant. Finally, for each cell, we tested whether the net number of unit spikes was different on trials on which a vocal response occurred than on trials on which the vocal response failed (while holding stimulus duration constant).
To compare how tightly each unit's spike pattern was timed relative to the stimulus onset versus the vocal onset, we calculated the width of each PSTH and PETH distribution peak at half-maximal height. These distributions were noisy: the spikes per time bin often did not decay smoothly away from the peak. We calculated the width at the first time point greater than and less than the peak time where the number of spikes per bin fell below half the number of spikes in the peak bin, and again at the last time points (i.e., furthest from the peak) where the spikes per bin fell below this half-peak value. The former number will tend to underestimate the half-height width of the distribution, whereas the latter will overestimate this width. We averaged these minimum and maximum values to arrive at a final estimate of the width of both the PSTH and PETH spike distributions for each unit. A comparison of these two numbers was used to judge whether spike timing for each cell was more tightly locked to the stimulus onset or to the vocal onset.
MLF stimulation of vocal output
We wanted to directly test whether altering the ensemble output of PAG neurons influenced specific parameters of the vocal response, including temporal features such as call duration and interpulse interval. To do so, we analyzed the properties of fictive vocal responses elicited by stimulation of the MLF (as described above). All stimulation sites were confirmed histologically to be in the MLF, through which PAG axons descend to connect with the hindbrain vocal circuit. We systematically altered the number of stimulation pulses (stimulus duration) and analyzed four features of the vocal response that might change in correlation with stimulus duration: 1) the duration of the stimulus-evoked vocal burst (number of vocal pulses), 2) the probability of eliciting a vocal response, 3) the latency to the vocal response, and 4) the mean interval between pulses within the vocal response. All four features were quantified at each stimulus duration. Using SAS version 9.1.3 statistical software (SAS Institute), we ran repeated-measures ANOVAs to determine whether there was a significant effect of stimulus duration on each of these four vocal parameters across experiments.
Lidocaine inactivation experiments
In these experiments, we sought to determine whether reversible inactivation of the PAG affected the stimulus-evoked vocal output. Surgeries, vT stimulation, and vocal nerve recording were as described above. Once a stable and reliable vocal response was obtained, we recorded the baseline response for 
40 min (16–24 stimulus trials every 10 min). About 6–7 min after the final baseline recording, a glass micropipette containing 4% lidocaine hydrochloride (Sigma) and 4% fluorescent dextran conjugate (either fluorescein or tetramethylrhodamine, Molecular Probes) in 0.1 M phosphate-buffered saline (PBS) was stereotactically guided to PAG. Micropipettes were made as described above for extracellular recording electrodes, but the tips were broken back to an inner diameter of about 5–10 µm. Either lidocaine or control (vehicle plus dye only) solution was pressure-ejected from the pipette using a picospritzer (Biomedical Engineering) set to deliver 1 pulse/s, 10- to 50-ms duration each, at 25–30 psi, for 30 s, for a total injection volume of about 0.5 nl. We continued to record the stimulus-evoked vocal responses for up to 1 h postinjection. PAG injections were bilateral, ipsilateral to the stimulus, or contralateral. In some fish, control injections were targeted to midbrain areas outside the PAG, including the torus and ventral tectum. In some cases post hoc analysis revealed that injections intended for the PAG had in fact missed; these injections were also treated as controls. Vehicle (PBS) only or sham injections to the PAG were used as additional controls. Typically, one experimental and one control injection were made in each fish, with 60 min between the two injections, allowing for complete recovery from any effect of the first injection on the vocal response. The locations of the injection sites were confirmed post hoc (see following text).
Post hoc analyses were similar to the extracellular single-unit analysis described above. Vocal response pulses were discriminated using Brainware and pulse times were exported for further analysis. Matlab scripts were used to compile and annotate the data and to calculate the mean number of vocal response pulses over the first 800 ms poststimulus for each time point pre- and postinjection. Data were graphed in Origin and statistical analysis was performed using JMP. Within each experiment, an injection was determined to have had a significant effect on the vocal response if: 1) there was a significant overall effect of time on the mean vocal response over the duration of the experiment and 2) a Tukey–Kramer test indicated that the vocal response at a minimum of at least one time point within the first 20 min postinjection was significantly less than at a minimum of at least two time points during the preinjection baseline. To analyze the data by treatment type, we first normalized the postinjection vocal response data to the mean preinjection baseline response for each experiment and then pooled these normalized data by treatment type. Within each treatment type, we performed a repeated-measures ANOVA across the experimental time points. Because the data were normalized to the preinjection baselines, a significant effect of time indicates a consistent postinjection effect within the group. In such cases we also determined which postinjection time points were different from preinjection using a Tukey–Kramer test, with a significance level of 0.05.
Tract-tracing experiments
To characterize the connectivity, both antero- and retrograde, of neurons in the PAG, we made focal iontophoretic injections of 5% neurobiotin (Vector Labs, Burlingame, CA). Surgeries were as described above, except that only a small craniotomy was made over the midbrain on one side. Fish were transferred to the recording rig and glass micropipettes were stereotactically guided to the PAG. Micropipettes were the same as those used as extracellular recording electrodes. Neurobiotin (in 2 M KCl) was iontophoresed for 5–10 min using +3-µA pulsed current (15 s on/15 s off). Only a single injection was made in each fish. After a 10-h survival time to allow transport of the tracer, fish were perfused and brains processed as described below.
Histology
At the end of all experiments, fish were deeply anesthetized (0.025% benzocaine) and perfused with ice-cold, teleost Ringer solution with 10 units/ml of heparin (Elkins-Sinn, Cherry Hill, NJ), followed by 4% paraformaldehyde in 0.1 M phosphate buffer (PB). Brains were removed, postfixed overnight, and transferred to 0.1 M PB (pH 7.2) for storage. Brains were equilibrated in 30% sucrose for 24 h and then sectioned transversely at 50 µm on a freezing microtome. For brains with fluorescent dye injections (fluorescein- or rhodamine-dextran, as in the recording and lidocaine inactivation experiments), sections were collected in 0.1 M PB, mounted immediately on gelatin-subbed slides, dried overnight, and coverslipped using a fluorescent mounting medium (Vectashield, Vector Labs). Neurobiotin injections, as used in the tract-tracing experiments, were visualized using a standard avidin–biotin–peroxidase protocol. Briefly, sections were collected in PBS, permeabilized for 30 min in PBS with 0.04% Triton-X100 (Sigma), incubated for 3 h at room temperature in avidin–biotin–peroxidase in PBS (Vectastain ABC Elite, Vector Labs), rinsed twice in PB, reacted with 0.05% diaminobenzidine (Sigma) and 0.0072% H2O2 in PB, rinsed twice more in PB, and mounted on gelatin-subbed slides. Alternate sections were mounted on separate slides, dried overnight, cleared in xylene, and coverslipped. Later, one set of sections was stained with 0.5% cresyl violet (Sigma) to delineate the borders of the different cell clusters. Sections were examined and photomicrographs taken on a Nikon Eclipse E-800 microscope.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Extracellular recordings were made from a total of 143 isolated units in the midbrains of 53 type I male midshipman fish, while stimulating different sites to elicit vocal output (see METHODS). Of these, a total of 48 neurons were recorded from 23 fish while vocal responses were being elicited by microstimulation of vT, a known vocal structure providing descending afferent input to the PAG (see METHODS and the last section of RESULTS). Twenty-four of these 48 neurons, in 12 fish, were confirmed histologically post hoc to be in the PAG (e.g., Fig. 6C), a compact cell layer along the periventricular surface of the midbrain (Fig. 6, A, B, and K). Recording sites were often (21/24) localized to the PAG axons just ventral to the main PAG cell body layer (e.g., Fig. 6C), but all 24 of these sites had clearly back-filled neurons within the PAG proper. Of the 24 neurons outside the PAG whose activity was also modulated by vT stimulation, 11 were localized to the ventral tectum, nine were in the midbrain tegmentum, and four were rostral of the PAG, in the dorsal thalamus; there were no back-filled neurons in the PAG in any of these recording sites. The mean and median values, including range and SDs, for spontaneous and stimulus-evoked firing properties of PAG neurons are summarized in Table 1. All cells fired spontaneous action potentials at either a low rate (
8 Hz; 17 cells) or not at all (seven cells). Ventral tuberal stimulation induced an excitatory response in all PAG neurons recorded. From cell to cell, this response ranged from a single, short-latency spike (nine cells), to either a discrete burst of spikes or a broad, longer-lasting increase in firing that decayed back to the spontaneous baseline over several hundred milliseconds. In 18 of the 24 cells, the mean latency to the first poststimulus spike was <50 ms.
|
|
In nine of the 13 "vocal" neurons, the net number of stimulus-evoked spikes (i.e., total spike count minus spontaneous activity) was significantly correlated with the total number of vocal response pulses (Table 2, Fig. 2, A5 and B5), suggesting that spike number is correlated with either the duration (i.e., number of vocal pulses) of each vocal burst or with the number of vocal bursts. Alternatively, these data could reflect either increased or decreased neuronal firing when a vocal response occurred, with no correlation to the duration of the response itself. This latter possibility would support the hypothesis that PAG neuronal activity contributes to initiation of the vocal response, but has no clear role in the patterning of response duration. To directly test whether spike number in any of the vocal neurons predicted the duration of the subsequent vocal burst, we reran the correlation analyses excluding all trials when the vocal response failed, considering only the duration of the first vocal burst on each trial and considering spikes occurring only before the end of this first vocal burst. In five of the nine neurons that initially showed significant correlations between net spike number and total vocal pulses, significant correlations (three positive, two negative) persisted between net spike number and vocal burst duration (Fig. 2A6). In the remaining four neurons, this second test revealed no significant relation between spike count and vocal duration, implying that the initial correlation between spike count and total vocal pulses in these four neurons was attributed to an underlying relationship between spike count and the binary presence or absence of a vocal response or between spike count and the number of vocal bursts. Indeed, in seven of the 13 vocal neurons, including these four, there was a significant relationship between the net stimulus-evoked spike number and the simple presence or absence of a vocal response (Fig. 2, A4 and B4), evidence that further supports a role in initiation (but not excluding a role in patterning). Only a minority of the 13 vocal neurons revealed significant correlations either between spike number and vocal latency (three cells, Fig. 2B6) or between mean unit latency and the presence of a vocal response (four cells). Finally, in none of the 24 confirmed PAG neurons was there any significant relationship between the firing frequency of the unit spike train and the discharge frequency of the vocal response (i.e., rate of fictive sound pulses; see Table 1) or between the latency of the first unit spike and the latency of the vocal response. Histograms of interspike interval (ISI) distributions for each neuron never showed a clear peak at or near 10 ms, the characteristic interpulse interval of both stimulus-evoked fictive vocalizations (see Vocal response data in Table 1) and natural calls (Fig. 1B). In summary, approximately half of all histologically confirmed PAG neurons recorded were significantly correlated with the initiation and/or duration of the vocal output.
|
Another approach to examining the relationship between spike trains and vocal output was to replot each unit's PSTH as a perievent time histogram (PETH) relative to the onset of the vocal response on each trial. By comparing the PSTH and PETH for each unit, it is possible to quantify how tightly each unit's spiking is distributed relative to the stimulus versus to the onset of the vocal response (Fig. 3). We quantified this difference by measuring the width of the peak of each distribution at half of the maximum height (see METHODS) and comparing the widths of the two distributions for each unit. Across all PAG neurons the PETH width was, on average, 119 ± 29 ms (SE) greater than the PSTH width, indicating that spiking was more tightly time locked to the stimulus than to the vocal response. However, when cells were subdivided into "vocal" and "nonvocal" categories, based on correlations with the vocal response as described above, the vocal cells had a significantly smaller difference between the PETH and PSTH distribution widths (Fig. 3, top left inset). In seven of the 13 cells classified as vocal (and only one of the nonvocal cells), the PETH width was within ±50 ms of the PSTH width. Thus in these cells, spiking was nearly as well time locked to the vocal response as to the stimulus itself. Although the difference between the PETH and PSTH half-height widths was significantly less in vocal cells than in nonvocal, the absolute values of these widths for either PETH or PSTH distributions were no different between the two groups (P > 0.2 and P > 0.9, respectively, Wilcoxon test). This analysis validates the separation of PAG neurons into vocal and nonvocal subgroups and shows that the spike trains of vocal neurons were not only significantly correlated with the vocal response, but were also more tightly distributed relative to the onset of the fictive vocalization.
|
If PAG activity does contribute to both initiation of the vocal response and establishing vocal duration, one prediction is that increasing or decreasing the duration of PAG ensemble spiking should cause changes in both the probability and the duration of the resultant vocal output. Our preliminary recording experiments involved stimulation in the MLF (see METHODS), which is the sole pathway by which PAG axons connect to the hindbrain vocal circuit (see following text). Thus MLF stimulation presumably elicits vocal responses by excitation of the descending PAG axons, and changes in the duration of such stimulation will produce longer or shorter trains of action potentials in these descending inputs to the hindbrain. We examined the relationship between stimulus duration and vocal response properties in 11 fish in which the stimulation site was histologically confirmed post hoc to be in the MLF and in which vocal responses had been recorded at multiple stimulus durations. Repeated-measures ANOVAs revealed a significant positive relationship across experiments between stimulus duration and both the duration of each vocal burst (Fig. 4A) and the probability of a vocal response (Fig. 4B). In contrast, stimulus duration did not significantly affect either the latency of the vocal response (Fig. 4C) or the interval between vocal pulses within the response (Fig. 4D). Thus not only is PAG neuronal activity correlated with the initiation and duration of the vocal output, but directly altering the output of these neurons also causes changes in these same vocal parameters.
|
To directly test whether PAG activity was necessary for the stimulus-evoked vocalizations to occur, focal injections of the reversible sodium channel blocker lidocaine were made into the PAG and surrounding midbrain structures after recording baseline vocal responses to vT stimulation. After injection, we continued recording to monitor any effects of the injection and recovery from these effects. Fluorescent dyes included with the lidocaine allowed us to examine post hoc the location of each injection and to estimate the relative size and spread away from the injection site.
Representative results of two lidocaine inactivation experiments are shown in Fig. 5, A–D. The vocal response was completely suppressed immediately after all lidocaine injections to the PAG ipsilateral to the site of vT stimulation (n = 6) and nine of ten lidocaine injections to the PAG bilaterally. In contrast, none of the control injections, including sham (n = 3) and vehicle (n = 4) injections to ipsilateral PAG and lidocaine injections to either the lateral midbrain tectum (n = 3) or the underlying torus semicircularis (n = 2), resulted in the complete and immediate blockade of the vocal response (see summary in Fig. 5E).
|
10 min postinjection and, in two of these five cases, the vocal response was not completely blocked at any postinjection time point. The delayed effect of these injections was consistent with the diffusion of lidocaine from the injection site to the PAG. Indeed, in all but one of these injections, we confirmed the spread of fluorescent dye from the injection site to structures surrounding the ventricle, including the PAG (see example in Fig. 5B'). In two of four cases, vehicle injections to the PAG caused a decrease of >70%, but not a complete suppression, in the vocal response immediately after injection; in both cases the response returned to baseline levels within 10 min. Although this shows an apparent influence of the vehicle alone on the results, we emphasize that these results sharply contrast with the lidocaine injections in the ipsilateral PAG that show an immediate and complete cessation of vocal activity (see above and statistical tests below). Sham injections to the PAG never resulted in any detectable change in the vocal response (see METHODS for how we defined an effect of treatments within each experiment). The effects of lidocaine injections to only the PAG contralateral to the stimulation were inconsistent. Of the four injections, three caused an immediate decrease in the vocal response, with two of these completely blocking the vocal response. The latter two injection sites both showed clear diffusion of dye across the midline to the ipsilateral PAG, but the immediacy of the effects nonetheless suggests the involvement of the contralateral PAG in vocalizations evoked by unilateral vT stimulation. That the results of these experiments were inconsistent may imply relatively weak contralateral connectivity (see anatomical experiments below for confirmation of bilateral PAG to PAG connectivity). Nevertheless, only lidocaine injections directly in the ipsilateral PAG blocked the vocal response consistently, immediately, and completely, demonstrating that neuronal activity in the PAG, but not in the surrounding midbrain structures, is necessary for stimulation-evoked vocal responses.
These results were confirmed by statistical analyses of the different treatment types. Repeated-measures ANOVAs confirmed that there was a significant overall effect across postinjection time of both bilateral and ipsilateral lidocaine injections to the PAG (P < 0.0001, both groups). Toral and tectal lidocaine injections also had significant effects on the vocal response (P < 0.0001 and P < 0.001, respectively), but the timing of these effects was different. Tukey–Kramer tests indicated that lidocaine injections to the PAG, both ipsi- and bilateral, caused a significant decrease in the vocal response, relative to baseline, at the 0, 10, and 20 min postinjection time points. In contrast, lidocaine injections to the torus and tectum caused decreases at the 10, 20, and 30, but not 0, min time points. This confirms that the effects of lidocaine injections outside the PAG were indeed delayed relative to injections that directly hit the PAG, supporting the hypothesis that activity only in the PAG, and not in neighboring midbrain structures, is essential for vocal initiation. By this analysis, there were no significant overall effects of vehicle or sham injections to the PAG or of lidocaine injections to the contralateral PAG (P > 0.6, P > 0.35, and P > 0.06, respectively).
Anatomical connectivity of the PAG
Two main types of neurophysiological data support the hypothesis that vT provides descending afferent input to the PAG. First, the latency of vocal responses elicited by stimulation of vT is much longer than responses evoked by MLF stimulation (roughly 200 vs. 15–30 ms). Second, PAG inactivation blocks vocal responses evoked by vT stimulation. To more clearly define the connectivity of the PAG, both anterograde and retrograde, we made focal, unilateral injections of neurobiotin targeted to the vocal PAG regions in four type I male and in one female midshipman (e.g., Fig. 6, A, B, and K). Our two goals were 1) to determine whether there were direct connections from our stimulation site in the anterior hypothalamus (vT) to the PAG and from the PAG to the known vocal motor structures in the hindbrain and 2) to be able to compare the connectivity of the presumptive teleost PAG with that of the mammalian PAG. Photomicrographs of labels resulting from two such injections are shown in Fig. 6, A, B, and K. We found strong retrograde labeling of cell bodies in vT (Fig. 6, D and E), as well as in other vocally active portions of the hypothalamus (the anterior tuberal nucleus) and preoptic area (the anterior and posterior parvocellular preoptic nuclei; not shown, but see Goodson and Bass 2002). In the hindbrain, we found labeled fibers with swellings indicative of presynaptic boutons in VM (Fig. 6, F, G, L, and M) and in an area ventral and medial of VM (Fig. 6, F and H) where commissural axons from VM neurons cross the midline to the contralateral VM (Bass et al. 1994). The projection from the PAG to VM was highly specific because terminals were not found in surrounding structures. In no case did we find fibers or terminals in the PN or SMN, the two other components of the hindbrain vocal motor circuit (Fig. 1C). Labeled axons terminating in VM traveled caudally from the PAG within the MLF. Other labeled axons crossed the midline at the level of the PAG (Fig. 6I) and small numbers of retrogradely labeled neurons and anterogradely labeled terminals were observed in the contralateral PAG (Fig. 6J). This reciprocal connectivity provides an anatomical basis for the immediate effects on the vocal response observed after some contralateral PAG lidocaine injections (see previous section of RESULTS). In sum, these experiments confirm a descending pathway from the anterior hypothalamus (and, more specifically, from vT) to the PAG and then to VM, but neither to the motoneurons innervating the sonic muscles nor to the premotor pacemaker neurons (see Fig. 1C for overview).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
).
Single neurons in the midshipman PAG, excited by forebrain hypothalamic (vT) stimulation, showed neuronal responses preceding the evoked fictive vocal response, with significant correlations between spike number and the presence, duration, and/or latency of fictive calls. Such "vocal" neurons were more tightly timed with regard to the vocal onset than were other "nonvocal" PAG neurons. However, the frequency of neuronal firing did not correlate with the discharge frequency of the fictive call. This is consistent with other studies showing that this trait, which establishes the fundamental frequency of natural calls, is primarily, if not solely, determined by the hindbrain–spinal vocal circuitry (Bass and Baker 1990). Stimulation of the MLF, the sole descending pathway for PAG axons connecting to the hindbrain vocal circuitry, affected the probability and duration of vocal bursts, but not their latency or discharge frequency. These data strongly support the hypothesis that the correlations between individual PAG neuron spike activity and vocal output are relevant for vocal initiation and duration patterning, but not for establishing the fine temporal structure (in this case, fundamental frequency) of a vocalization. Consistent with these findings, reversible lidocaine inactivation experiments of the PAG showed that its activity is necessary for the production of vocal responses that are elicited by forebrain stimulation. Last, the above neurophysiological experiments, together with the demonstration of a direct anatomical pathway from vT to the PAG, and then from the PAG to VM, an integral component of the hindbrain vocal motor complex (Fig. 1C), provide strong support for the general hypothesis that the vT–PAG–VM circuit is the major descending pathway for vocal motor commands originating in the forebrain.
Comparisons with mammals
These findings also provide critical empirical support for proposals that this portion of the teleost brain is similar both structurally and functionally to the mammalian, and more generally tetrapod, PAG (Goodson and Bass 2002). Neurons in the vocal portion of the mammalian PAG do not project directly to the motoneuron pools involved in vocalization, but rather to the nucleus retroambiguus (NRA), a primary premotor structure in the caudal medulla (Holstege 1989; Jurgens 2002; Vanderhorst et al. 2000). The NRA then connects to the various motoneuron pools innervating the vocal musculature (e.g., muscles of the larynx, pharynx, tongue, and jaw; see Jurgens 2002). Similarly, we find that the teleost PAG projects to VM, which in turn connects to the PN–SMN circuit that includes sonic motoneurons. As in mammals, we find no evidence of direct connections from the PAG to vocal motoneurons. The mammalian PAG receives dense inputs from a variety of structures in the limbic system, including hypothalamic and preoptic nuclei (Dujardin and Jurgens 2005; Jurgens 2002), inputs that are also present in midshipman (see RESULTS and Goodson and Bass 2002).
An important distinction between our results and those in mammals is that the mammalian PAG appears to be topographically organized. The vocal region itself occupies a discrete portion of the PAG (Holstege 1989; Larson and Kistler 1986; Vanderhorst et al. 2000). Furthermore, there is evidence of topographic mapping of call types across the vocal area of the PAG (Dujardin and Jurgens 2005; Zhang et al. 1994). In contrast, we found no obvious spatial clustering of vocal neurons within the teleost PAG.
Functionally, either lesions or pharmacological inactivation of the mammalian PAG blocks both spontaneous vocalizations and those elicited by stimulation of PAG afferents (Jurgens 1994; Siebert and Jurgens 2003), just as lidocaine inactivation of the teleost PAG prevented vocal responses to vT stimulation. Single-unit PAG recordings in bats (Suga and Yajima 1988), squirrel monkeys (Dusterhoft et al. 2000, 2004), and macaques (Larson 1991; Larson and Kistler 1986) demonstrate that neuronal activity correlates with vocal output, as we describe here. Generally, PAG neurons in midshipman show similar patterns of vocal-related activity to those described in these other systems. Specifically, both here (see above) and in mammals: 1) only a small subset of PAG neurons has vocal-related activity; 2) vocal neurons usually begin firing before vocal onset; and 3) vocal responses are heterogeneous, sometimes ceasing to fire before vocal onset and sometimes continuing to fire during vocalization. Furthermore, different neurons are correlated with different aspects of the vocal output. 4) From trial to trial, there is significant variability in the lag time between the onset of neuronal spiking and the vocal onset. Thus the variability in vocal-related spike activity we found in neurons in the midshipman PAG is, in fact, typical of that described previously in the mammalian PAG (Larson 1991; Larson and Kistler 1986). One minor difference between our results and those in mammals is that we found a small subset (2/24) of "vocal" PAG neurons with an apparent net inhibitory relationship to the vocal response. Such neurons have not been described in the mammalian PAG (Dusterhoft et al. 2000, 2004; Larson 1991; Larson and Kistler 1986; Suga and Yajima 1988). Despite the presence of these inhibitory neurons, the summed effect of PAG activity is still excitatory vis-à-vis vocalization, as demonstrated by the fact that PAG inactivation completely blocks vocalization and that MLF stimulation excites vocal responses. Taken together, the structural and functional similarities between the mammalian and teleost PAG suggest that our findings regarding the role of PAG neurons in the initiation and temporal patterning of vocalization are likely to be generally applicable across sonic vertebrates.
Temporal patterning of vocalization
A role for the PAG in the direct patterning of the temporal and motor structure of vocalization remains unresolved. In macaques, the activity of single PAG neurons is correlated both with the activation of single or functionally related groups of vocal muscles and with various acoustic properties of vocalizations, including duration and fundamental frequency (Larson 1991; Larson and Kistler 1986). These data have been interpreted to imply a role for the PAG in vocal patterning. However, other data imply that the PAG initiates each call type, but that detailed patterning of the temporal and acoustic features of calls occurs in premotor and motor circuitry of the hindbrain (Jurgens 1994). Consistent with this, the activity of single PAG neurons in squirrel monkeys correlates with call type but does not predict the gross temporal structure of amplitude and frequency modulations (Dusterhoft et al. 2000, 2004). However, analysis of the temporal details of the unit spike trains in these experiments may yet reveal temporal correlations between unit activity and acoustic features of the vocalizations. Although abundant experimental evidence exists for supra-hindbrain temporal patterning of learned vocalizations in songbirds (Ashmore et al. 2005; Hahnloser et al. 2002; Kittelberger and Mooney 2005; Leonardo and Fee 2005; Vu et al. 1994; Yu and Margoliash 1996), none of these studies addresses the role of the vocal midbrain.
In midshipman fish, the direct translation of SMN activity to vocal traits simplifies the analysis of whether PAG spike trains predict the fine temporal details of vocalization. Natural calls in the midshipman have either a fundamental frequency (for harmonic "hums") or pulse repetition rate (for nonharmonic "grunts") close to 100 Hz (Fig. 1B). Stimulus-evoked fictive vocalizations invariably have a similar frequency (Bass and Baker 1990). We find no indication that PAG neuronal spike trains establish this characteristic frequency. However, the number of spikes scaled with the duration of vocal bursts in a number of neurons that could participate in establishing call duration, the primary feature distinguishing call types (Fig. 1).
Although the activity of single PAG neurons in the midshipman was significantly correlated with various parameters of the vocal output, including duration, no single neuron's activity explained a high percentage of the variance in any one parameter (as indicated by the generally low r values of the linear regressions such as those shown in Fig. 2). This may reflect the fact that initiation and patterning of the vocal output depend on a population of neurons, both within the PAG and in other portions of the vocal circuit (see Fig. 1C). Thus the activity of no single neuron within this population will likely be sufficient to account for the entire variance in any vocal parameter, such as initiation or duration, even when that neuron is functionally involved in shaping that parameter.
In support of the conclusion that PAG neuronal activity shapes vocal duration, changing the duration of electrical stimuli delivered to the MLF consistently altered fictive call duration, with concurrent effects on the probability of eliciting a vocal response, but not on discharge frequency. MLF stimulation likely excites descending axons from mid- and forebrain sources other than the PAG. Thus, taken alone, these experiments demonstrate only that call duration can be influenced by activity manipulation above the level of the hindbrain. However, several pieces of evidence suggest the most parsimonious interpretation is that the effects of MLF stimulation on vocalization are attributable to effects specifically on descending PAG axons. First, injections of biotin tracers into the hindbrain region of VM that innervates the PN–SMN circuit (Fig. 1) mainly label PAG neurons (Goodson and Bass 2002). The tract-tracing experiments reported here confirm the specificity of this projection. Consistent with this circuitry, the current study also demonstrates that PAG activity is predictive of a vocal response, and lidocaine experiments show that activity in the PAG, but not in midbrain structures surrounding the PAG, is necessary for forebrain-evoked fictive vocalizations. Previously published reports also support a role for the midshipman PAG in patterning vocal duration. First, neuropeptide injections near the PAG induce dose-dependent changes in fictive call duration but not in discharge frequency (Goodson and Bass 2000b). Furthermore, certain effects of androgenic steroids on vocal duration have been localized to the midbrain (Remage-Healey and Bass 2004); a particularly high density of androgen receptors in the PAG (Forlano et al. 2005) makes it a likely locus for these effects.
The PAG is a critical site in the initiation of a variety of "emotional" behaviors that involve vocal communication signals, from defense and escape to courtship and reproduction (Behbehani 1995; Holstege 1998). Similarly, the multiple midbrain regions that constitute the mesencephalic locomotor center initiate locomotion by projections to reticulospinal neurons driving spinal motoneuron pools (Grillner 2003; Jordan 1998). In all of these contexts, midbrain neuronal activity has been thought to play only a minor role in directly patterning the motor output. Here we show that PAG neuronal activity correlates with discrete aspects of vocal initiation. Furthermore, we demonstrate that PAG activity correlates with the duration of the vocal output, but clearly not with other fine temporal features, such as fundamental frequency. In general, the results of this study provide a compelling example of how distantly related groups of vertebrates have adopted similar mechanisms to solve common problems in vocal–acoustic communication, in this case the patterning of brain stem vocal motor output. To the extent that the specifics of these findings prove to be generally applicable across motor systems and groups of vertebrates, initiation and patterning functions may not be localized in distinct, separate nodes of the descending motor pathway, but rather may be distributed properties of the motor circuit considered in its entirety.
|
GRANTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
FOOTNOTES |
|---|
Address for reprint requests and other correspondence: J. M. Kittelberger, Dept. of Neurobiology and Behavior, Seeley G. Mudd Hall, Cornell University, Ithaca, NY 14853 (E-mail: mk348{at}cornell.edu)
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Bandler R and Shipley MT. Columnar organization in the midbrain periaqueductal gray: modules for emotional expression? Trends Neurosci 17: 379–389, 1994.[CrossRef][Web of Science][Medline]
Bass AH. Shaping brain sexuality. Am Scientist 84: 352–363, 1996.
Bass AH and Baker R. Sexual dimorphisms in the vocal control system of a teleost fish: morphology of physiologically identified neurons. J Neurobiol 21: 1155–1168, 1990.[CrossRef][Web of Science][Medline]
Bass AH, Bodnar D, and Marchaterre MA. Complementary explanations for existing phenotypes in an acoustic communication system. In: The Design of Animal Communication, edited by Hauser MD and Konishi M. Cambridge, MA: MIT Press, 1999, p. 493–514.
Bass AH, Marchaterre MA, and Baker R. Vocal-acoustic pathways in a teleost fish. J Neurosci 14: 4025–4039, 1994.[Abstract]
Behbehani MM. Functional characteristics of the midbrain periaqueductal gray. Prog Neurobiol 46: 575–605, 1995.[CrossRef][Web of Science][Medline]
Brantley RK and Bass AH. Alternative male spawning tactics and acoustic signals in the plainfin midshipman fish Porichthys notatus girard (Teleostei, Batrachoididae). Ethology 96: 213–232, 1994.[Web of Science]
Davis PJ, Zhang SP, and Bandler R. Midbrain and medullary regulation of respiration and vocalization. Prog Brain Res 107: 315–325, 1996.[Web of Science][Medline]
Demski LS and Gerald JW. Sound production evoked by electrical stimulation of the brain in toadfish (Opsanus beta). Anim Behav 20: 507–513, 1972.[CrossRef][Web of Science][Medline]
Demski LS and Gerald JW. Sound production and other behavioral effects of midbrain stimulation in free-swimming toadfish, Opsanus beta. Brain Behav Evol 9: 41–59, 1974.[Web of Science][Medline]
Dujardin E and Jurgens U. Afferents of vocalization-controlling periaqueductal regions in the squirrel monkey. Brain Res 1034: 114–131, 2005.[Medline]
Dusterhoft F, Hausler U, and Jurgens U. On the search for the vocal pattern generator. A single-unit recording study. Neuroreport 11: 2031–2034, 2000.[Medline]
Dusterhoft F, Hausler U, and Jurgens U. Neuronal activity in the periaqueductal gray and bordering structures during vocal communication in the squirrel monkey. J Neurosci Methods 123: 53–60, 2004.[CrossRef]
Esposito A, Demeurisse G, Alberti B, and Fabbro F. Complete mutism after midbrain periaqueductal gray lesion. Neuroreport 10: 681–685, 1999.[Medline]
Fine ML. Sounds evoked by brain stimulation in the oyster toadfish Opsanus tau L. Exp Brain Res 35: 197–212, 1979.[Web of Science][Medline]
Fine ML and Perini MA. Sound production evoked by electrical stimulation of the forebrain in the oyster toadfish. J Comp Phys A Sens Neural Behav Physiol 174: 173–185, 1994.
Forlano PM, Marchaterre MA, Deitcher DL, and Bass AH. Distribution of androgen receptor mRNA in vocal and nonvocal circuitry of a teleost fish. Soc Neurosci Abstr 1001.1006, 2005.
Goodson JL and Bass AH. Forebrain peptides modulate sexually polymorphic vocal circuitry. Nature 403: 769–772, 2000a.[CrossRef][Medline]
Goodson JL and Bass AH. Rhythmic midbrain-evoked vocalization is inhibited by vasoactive intestinal polypeptide in the teleost Porichthys notatus. Brain Res 865: 107–111, 2000b.[CrossRef][Medline]
Goodson JL and Bass AH. Vasotocin innervation and modulation of vocal-acoustic circuitry in the teleost Porichthys notatus. J Comp Neurol 422: 363–379, 2000c.[CrossRef][Medline]
Goodson JL and Bass AH. Vocal-acoustic circuitry and descending vocal pathways in teleost fish: convergence with terrestrial vertebrates reveals conserved traits. J Comp Neurol 448: 298–322, 2002.[CrossRef][Web of Science][Medline]
Grillner S. The motor infrastructure: from ion channels to neuronal networks. Nat Rev Neurosci 4: 573–586, 2003.[Web of Science][Medline]
Hahnloser RHR, Kozhevnikov A, and Fee MS. An ultra-sparse code underlies the generation of neural sequences in a songbird. Nature 419: 65–70, 2002.[CrossRef][Medline]
Holstege G. Anatomical study of the final common pathway for vocalization in the cat. J Comp Neurol 284: 242–252, 1989.[CrossRef][Web of Science][Medline]
Holstege G. The emotional motor system in relation to the supraspinal control of micturition and mating behavior. Behav Brain Res 92: 103–109, 1998.[Medline]
Jordan LM. Initiation of locomotion in mammals. Ann NY Acad Sci 860: 83–93, 1998.[CrossRef][Web of Science][Medline]
Jurgens U. The role of the periaqueductal grey in vocal behaviour. Behav Brain Res 62: 107–117, 1994.[CrossRef][Web of Science][Medline]
Jurgens U. Neural pathways underlying vocal control. Neurosci Biobehav Rev 26: 235–258, 2002.[CrossRef][Web of Science][Medline]
Kennedy MC. Vocalization elicited by electrical stimulation of the midbrain. Brain Res 91: 321–325, 1975.[Medline]
Kittelberger JM and Mooney R. Acute injections of brain-derived neurotrophic factor in a vocal premotor nucleus reversibly disrupt adult birdsong stability and trigger syllable deletion. J Neurobiol 62: 406–424, 2005.[CrossRef][Web of Science][Medline]
Larson CR. On the relation of PAG neurons to laryngeal and respiratory muscles during vocalization in the monkey. Brain Res 552: 77–86, 1991.[CrossRef][Web of Science][Medline]
Larson CR and Kistler MK. The relationship of periaqueductal gray neurons to vocalization and laryngeal EMG in the behaving monkey. Exp Brain Res 63: 596–606, 1986.[CrossRef][Web of Science][Medline]
Leonardo A and Fee MS. Ensemble coding of vocal control in birdsong. J Neurosci 25: 652–661, 2005.
Remage-Healey L and Bass AH. Rapid, hierarchical modulation of vocal patterning by steroid hormones. J Neurosci 24: 5892–5900, 2004.
Seller TJ. Midbrain vocalization centres in birds. Trends Neurosci 4: 301–303, 1981.
Sewards TV and Sewards MA. Representations of motivational drives in mesial cortex, medial thalamus, hypothalamus and midbrain. Brain Res Bull 61: 25–49, 2003.[CrossRef][Web of Science][Medline]
Siebert S and Jurgens U. Vocalization after periaqueductal grey inactivation with the GABA agonist muscimol in the squirrel monkey. Neurosci Lett 340: 111–114, 2003.[Medline]
Suga N and Yajima Y. Auditory-vocal integration in the midbrain of the mustached bat: periaqueductal gray and reticular formation. In: The Physiological Control of Mammalian Vocalization, edited by Newman JD. New York: Plenum, 1988, p. 87–107.
Swanson LW. Cerebral hemisphere regulation of motivated behavior. Brain Res 886: 113–164, 2000.[CrossRef][Web of Science][Medline]
Vanderhorst VGJM, Terasawa E, Ralston HJ, and Holstege G. Monosynaptic projections from the lateral periaqueductal gray to the nucleus retroambiguus in the rhesus monkey: implications for vocalization and reproductive behavior. J Comp Neurol 424: 251–268, 2000.[CrossRef][Web of Science][Medline]
Vu ET, Mazurek ME, and Kuo YC. Identification of a forebrain motor programming network for the learned song of zebra finches. J Neurosci 13: 6924–6934, 1994.
Wild JM. Functional anatomy of neural pathways contributing to the control of song production in birds. Eur J Morphol 35: 303–325, 1997.[CrossRef][Medline]
Yu AC and Margoliash D. Temporal hierarchical control of singing in birds. Science 273: 1871–1875, 1996.
Zhang SP, Davis PJ, Bandler R, and Carrive P. Brain stem integration of vocalization: role of the midbrain periaqueductal gray. J Neurophysiol 72: 1337–1356, 1994.
This article has been cited by other articles:
|
|
 |
|
  L. Remage-Healey and A. H. Bass Plasticity in Brain Sexuality Is Revealed by the Rapid Actions of Steroid Hormones J. Neurosci., January 31, 2007; 27(5): 1114 - 1122. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
