| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Section of Oral Biology, The Ohio State University, Columbus, Ohio 43210
 |
ABSTRACT |
|---|
 Abstract
 Introduction
Methods
Results
Discussion
References
|
|---|
Halsell, Christopher B. and Susan P. Travers. Anterior and posterior oral cavity responsive neurons are differentially distributed among parabrachial subnuclei in rat. J. Neurophysiol. 78: 920-938, 1997. The responses of single parabrachial nucleus (PBN) neurons were recorded extracellularly to characterize their sensitivity to stimulation of individual gustatory receptor subpopulations (G neurons, n = 75) or mechanical stimulation of defined oral regions (M neurons, n = 54) then localized to morphologically defined PBN subdivisions. Convergence from separate oral regions onto single neurons occurred frequently for both G and M neurons, but converging influences were more potent when they arose from nearby locations confined to the anterior (AO) or posterior oral cavity (PO). A greater number of G neurons responded optimally to stimulation of AO than to PO receptor subpopulations, and these AO-best G neurons had higher spontaneous and evoked response rates but were less likely to receive convergent input than PO-best G neurons. In contrast, proportions, response rates, and convergence patterns of AO- and PO-best M neurons were more comparable. The differential sensitivity of taste receptor subpopulations was reflected in PBN responses. AO stimulation with NaCl elicited larger responses than PO stimulation; the converse was true for QHCl stimulation. Within the AO, NaCl elicited a larger response when applied to the anterior tongue than to the nasoincisor duct. Hierarchical cluster analysis of chemosensitive response profiles suggested two groups of PBN G neurons. One group was composed of neurons optimally responsive to NaCl (N cluster); the other to HCl (H cluster). Most N- and H-cluster neurons were AO-best. Although they were more heterogenous, all but one of the remaining G neurons were unique in responding best or second-best to quinine and so were designated as quinine sensitive (Q+). Twice as many Q+ neurons were PO- compared with AO-best. M neurons were scattered across PBN subdivisions, but G neurons were concentrated in two pairs of subdivisions. The central medial and ventral lateral subdivisions contained both G and M neurons but were dominated by AO-best N-cluster G neurons. The distribution of G neurons in these subdivisions appeared similar to distributions in most previous studies of PBN gustatory neurons. In contrast to earlier studies, however, the external medial and external lateral-inner subdivisions also contained G neurons, intermingled with a comparable population of M neurons. Unlike cells in the central medial and ventral lateral subnuclei, nearly every neuron in the external subnuclei was PO best, and only one was an N-cluster cell. In conclusion, the present study supports a functional distinction between sensory input from the AO and PO at the pontine level, which may represent an organizing principle throughout the gustatory neuraxis. Furthermore, two morphologically distinct pontine regions containing orosensory neurons are described.
Primary gustatory and somatosensory afferent fibers in the facial, glossopharyngeal, trigeminal, and vagal cranial nerves terminate in an orderly sequence within the rostral division of the nucleus of the solitary tract (NST) (e.g., Hamilton and Norgren 1984 Subjects and preparation
Eighty-two adult male Sprague Dawley rats ranging in weight from 270 to 460 g were used for recording. The animals were anesthetized with ethyl carbamate (urethan, 1.0 g/kg ip) and pentobarbital sodium (Nembutal, 25 mg/kg ip). Supplemental doses of Nembutal (10 mg/kg ip) were administered as needed to maintain surgical levels of anesthesia, i.e., areflexia of hindlimb withdrawal and corneal blink, for the duration of the recording session. A bolt and headholder assembly was attached to the skull to provide a secure attachment to the stereotaxic frame and clear access to the oral cavity (Halsell et al. 1993 Recording
Glass-coated tungsten microelectrodes (Z = 1.0-3.9 M Stimulation
GUSTATORY STIMULI.
All taste stimuli were presented at room temperature and were preceded by an identical stimulation with water to control for the somatosensory aspects of stimulus application. The taste stimuli were chosen to represent four standard taste qualities (sweet, salty, sour, and bitter) and consisted of (in M) 0.01 Na saccharin, 0.3 NaCl, 0.01 HCl, 0.03 HCl, 0.003 quinine hydrochloride, and 0.01 quinine hydrochloride. Two mixtures of the four representative stimuli also were used and consisted of the above Na saccharin and NaCl concentrations and either the higher or lower concentrations of both HCl and quinine. No major differences were noted in the response properties of neurons stimulated with the higher or lower concentrations of HCl and quinine. Na saccharin was substituted for the more typical stimulus, sucrose, because the viscosity of the latter solution made it inconvenient to flow through the small pipette. Na saccharin was predicted to be a reasonable alternative, due to its effectiveness in other neurophysiological studies (Frank 1991 MECHANICAL STIMULATION.
Discrete oral regions also were stimulated to assess the neuron's mechanosensitivity. Tactile stimulation consisted of stroking the area of interest with a blunt glass rod using a series of discrete strokes that lasted for ~5 s. Areas routinely tested included the AT, intermolar eminence (IME), hard palate including the incisal papilla and rugae (HP), buccal wall (BW), upper and lower molars (UM, LM) and incisors (UI, LI), FOL, CV, posterior tongue between and behind these papillae (PT), retromolar mucosa (RM), and SP. The force produced was not quantified but was firm enough to indent the epithelium without producing any tissue damage. A subset of neurons also was tested with a camel's hair brush to simulate a low-threshold stimulus. Stimulus onset and offset was indicated by voice and recorded on video tape.
Neurophysiological data analysis
Neurophysiological responses were analyzed and quantified off-line. Gustatory response measures were the number of action potentials occurring during the 10 s of stimulus presentation minus the number of spontaneous action potentials occurring immediately preceding the initial water stimulation (Fig. 1C). Although the tactile and thermal components of manual tastant application occasionally elicited small transient responses (Travers and Norgren 1991
Histological reconstruction
Electrolytic lesions (anodal current, 3 µA, 3 s) were used to mark recording sites (Fig. 1A). Lesions were made either at the recording site (36% of the cells), near the recording site on the same track (191 ± 16.7 µm distant; 43%) or on an adjacent track (21%). The mean diameter of the lesions was 138 ± 3.24 µm. At the end of the recording session, the animal was given a lethal dose of Nembutal (150 mg/kg) and perfused transcardially with physiological saline and phosphate-buffered formalin. The brains were removed and cryoprotected in 20% sucrose phosphate-buffered formalin and frozen sectioned in the transverse plane at 40 µm. Alternate sections were stained for Nissl substance with cresyl violet and for myelin with the Weil technique (Fig. 1, A and B).
General response properties
One-hundred ninety-three single neurons were recorded from PBN, and 142 of these were isolated for a period of time sufficient to provide meaningful data. These cells provide the basis for the present results. The additional neurons were discarded for technical reasons, e.g., cell isolation was lost before complete testing was finished. Seventy-five of the 142 responded to stimulation with either the taste mixture or at least one of the individual tastants and so were classified as gustatory neurons (G neurons) (Ogawa et al. 1990
Response properties: G neurons
GUSTATORY RECEPTIVE FIELDS.
Sixty-seven of 75 G neurons were characterized for their responsiveness to stimulation of individual taste receptor subpopulations. Two cells responded when a tastant was applied to the entire oral cavity but not to any of the individual receptor subpopulations. For 48 G neurons, the most vigorous taste response arose from stimulating the anterior oral cavity (AO: AT or NID), and for 17 G neurons, from the posterior oral cavity (PO: CV, FOL, or SP). Figure 2 depicts the taste responses elicited by stimulating individual receptor subpopulations. Responses are separated into four groups defined by whether the neuron responded to stimulation of a single taste receptor subpopulation (specific neurons) or to multiple receptor subpopulations (convergent neurons) (Travers and Norgren 1995
GUSTATORY AND MECHANICAL PROPERTIES FOR MULTIMODAL (GM) CELLS.
GM neurons responded to both taste and mechanical stimulation. Interestingly, about half of these neurons exhibited transient (~1-s duration) suprathreshold responses to the onset of the water flow that immediately preceded stimulus application. Complete information on the location of the mechanical receptive field was obtained for 45/55 GM neurons. The remaining 10 cells were not used for this analysis. For 38 of the 45 cells, the optimal receptive fields for the two modalities were located in the same, nearby, or apposed oral loci confined to the AO or PO. AO-best cells, however, were more likely to receive optimal multimodal input from precisely the same oral locations (21/23) than PO-best cells (6/15; P = 0.001, Fisher's Exact).
GUSTATORY CHEMOSENSITIVITY.
Neural response profiles. For 58 neurons, individual taste stimuli were applied to each taste receptor subpopulation effective in activating the cell. Chemosensitive response profiles or spectra (Frank 1991
Response properties: M neurons
Sensitivity to the different oral regions was determined for 47/54 M neurons. For the other six M neurons, receptive field data were incomplete so were not used in the following analysis. Twenty-nine M-neurons responded best to AO (IME, IP, HP, BW, UM, LM, UI, LI) stimulation and 18 to PO (FOL, CV, PT, RM, SP) stimulation. In contrast to G neurons, there was no significant difference between the magnitude of evoked responses in AO-best versus PO-best neurons (F = 0.877, df = 7, P = 0.536). Sixteen M neuronsresponded to stimulation of a single oral region, seven to PO sites (2 PT, 2 SP, 3 CV/FOL) and nine to AO sites (5 HP, 2 LI, 1 NID, 1 BW). Most M neurons (31/47), however, responded to input from multiple oral regions and are referred to as M-composite neurons. The term composite was chosen instead of convergent because the inference of central convergence is more uncertain for M than G neurons due to the continuous distribution of oral mechanoreceptors compared with the discrete locations of different taste receptor subpopulations. For example, although central convergence is the most likely explanation for a neuron that responds to mechanical stimulation of both the tongue and palate, either convergence or peripheral branching is a plausible mechanism for a neuron that responds to stimulation of both the hard and soft palate, because these oral loci are directly adjacent to one another (see also Travers and Norgren 1995
Topographic organization
ORGANIZATION BY MODALITY.
Recording site locations were reconstructed on one of four standard sections in the transverse plane (Fig. 7). A differential distribution was evident along the rostrocaudal axis for the cells that were or were not responsive to the stimuli used in this study. More unresponsive neurons were located in the rostral half (Fig. 7, levels C and D) of the PBN region we recorded from, and more G and M neurons were in the caudal half (Fig, 7, levels A and B;
OROTOPIC ORGANIZATION.
A differential distribution along the rostrocaudal axis was also apparent for receptive field (Fig. 7, Table 5). Collapsing across G and M neurons, most AO-best neurons were found caudally compared with just under half of the PO-best neurons ( CHEMOTOPIC ORGANIZATION.
Because more than two-thirds of the neurons that we characterized for chemosensitivity were N-cluster neurons, definitive conclusions regarding the topographic arrangement of chemical sensitivity in PBN cannot be drawn from the present data. Although not statistically significant, a greater number of each neuron class were located caudally than rostrally. On the other hand, more N- and H-cluster neurons were located ventrally and within the core, whereas Q+ neurons were about evenly distributed dorsoventrally and between the core and shell.
SUBNUCLEAR ORGANIZATION.
One goal of the present study was to determine if there was a relationship between morphologically defined PBN subdivisions and orosensory response characteristics. The evidence presented above suggests that, although orosensory neurons were not confined to specific PBN subdivisions, there were two separate foci of orosensory cells. One was centered in the CM/VL subdivisions and the other in the EM/ELi subdivisions, and these subnuclei were dominated by distinctive cell types (Table 6). CM/VL was populated mostly by G neurons (both GM and G-only neurons), which were AO-best, with 76% belonging to the N cluster. In contrast, EM/ELi was populated by both M and G PO-best neurons, with all the G neurons being GM neurons. Only four neurons in this subdivision were characterized for chemosensitivity, but only one was a N-cluster cell. Fig. 1 depicts recording sites and response properties of two PBN G neurons (cells 212 and 213) that typify some of these relationships. Both cells were recorded at the same rostrocaudal PBN level (about level C) but neuron 213 was located in EM and neuron 212 was in VL (Fig. 1, A and B). Correspondingly, neuron 213 was PO-best (specifically CV-best, Fig. 1, D). Although this neuron was classified as a N-cluster neuron, it was the last neuron amalgamated in this cluster and not typical of cells in this group. It responded optimally and equally to NaCl and HCl, but it also responded quite vigorously to quinine (Fig. 1, E). As was typical of PO-best G neurons, this cell could be further classified as a GM neuron, and it responded equivalently or more robustly to mechanical stimulation of several PO sites (Fig. 1, F). In contrast, neuron 212 was an AO(AT)-best G-only neuron that responded well to NaCl (Fig. 1, D).
The results of the present study support a functional distinction between sensory input arising from the anterior and posterior oral cavities within PBN. Significant differences were noted in the spontaneous and evoked response rates and chemosensitive profiles of neurons receiving optimal input from AO or PO receptor subpopulations, in convergence patterns across receptor fields and modalities and in the topographic organization of AO- and PO-best neurons within PBN subdivisions. The AO/PO distinction appears to be conserved generally from the periphery through the NST and PBN and possibly at the thalamic and cortical levels (vide infra) and may represent an underlying organizing principle throughout the gustatory pathway.
Receptive field properties
The substantial convergence of input from different taste receptor subpopulations and oral regions onto both G and M neurons reported herein supports previous reports of spatial convergence at several levels of the central gustatory pathway, from NST to cortex (Hayama et al. 1987 Response properties
GUSTATORY CHEMOSENSITIVITY.
In general, the chemosensitivity of different groups of PBN neurons was consistent with predictions based on their receptive field characteristics and known sensitivities of peripheral taste fibers. The majority of N neurons received their strongest input from the AO, and most of these responded only to AT stimulation. NaCl also elicited a significantly larger response from the AT than the other taste receptor subpopulations. These relationships mirror the peripheral responsiveness of gustatory afferent nerves (Boudreau et al. 1987 TACTILE RESPONSES.
The responses of the mechanical neurons recorded in the present study were generally similar to those reported by Ogawa and colleagues at the pontine, thalamic, and cortical levels (Nomura and Ogawa 1985 Topographic and subnuclear organization in PBN
When sensitivities of neurons were compared with their locations, a clear anatomic organization by modality and receptive field emerged (Table 5). Thus the same response properties appear reflected in the topography of orosensory PBN as in NST (Halsell et al. 1993 ORGANIZATION BY MODALITY.
The significant differences in G and M neuron distribution observed along the rostrocaudal (levels A and B vs. levels C and D) and mediolateral (core vs. lateral shell) axes (Table 5) were reflected in the differential distribution of G and M neurons within CM/VL and EM/ELi (Table 6), because these subnuclei are discrete along the same anatomic axes. Histological reconstructions from a previous study by Ogawa's group were also suggestive of some differential distribution of G and M neurons, but the separation was less striking and not noted by the authors (Ogawa et al. 1987 OROTOPY.
Collapsed across G and M neurons, orotopy also occurred along the rostrocaudal and mediolateral axes (Table 5). Although CM/VL and EM/ELi were dominated by different neuronal populations (Table 6), it should be emphasized that there was not a complete segregation of neurons according to receptive field in PBN. Indeed, although AO-best G neurons dominated the CM/VL and were more numerous in this region, PO-best neurons were just as frequently encountered here as in EM/ELi, though they were the only type of G neuron encountered in the latter area.
ARRANGEMENT OF CHEMOSENSITIVITY.
A previous neurophysiological study in rat suggested a chemotopic organization in PBN, such that NaCl-best cells were most frequently located caudally and ventrally and HCl-best cells rostrally and dorsally (Ogawa et al. 1987 TASTE FUNCTION IN EM/ELI.
The literature on the gustatory function of PBN has been based primarily on CM/VL (Norgren 1995
INTRODUCTION
Abstract
Introduction
Methods
Results
Discussion
References
). Neurophysiological recordings demonstrate that the organization of NST reflects its orotopic input (Dickman and Smith 1989; McPheeters et al. 1990; Ogawa et al. 1984; Sweazey and Bradley 1989; Travers and Norgren 1995; see also Travers 1993), with neurons responsive to anterior oral cavity (AO) stimulation lying anterior and lateral to those responsive to posterior oral cavity (PO) stimulation. The receptive fields of single neurons show a similar orderly organization. Convergence of sensory inputs provides evidence for central integration but this integration is not random; instead it occurs preferentially between receptors confined to the AO or PO (Travers and Norgren 1995; Travers et al. 1986).
; McPheeters et al. 1990; Mistretta 1988; Scott et al. 1986). Finally, the NST is organized by modality, with neurons responsive solely to intraoral mechanical stimulation extending lateral to those cells responsive to gustatory stimulation (Halsell et al. 1993; Ogawa and Hayama 1984; Ogawa et al. 1984; Travers and Norgren 1995). The topographic organization of NST may serve as an anatomic substrate for differential control of ingestive and rejection responses by different peripheral nerves, modalities, and gustatory stimuli (Frank 1991; Grill et al. 1992; Harrer and Travers 1996; Zeigler et al. 1985; see also Travers 1993, Travers et al. 1987). An even more striking segregation between peripheral afferent inputs occurs in the teleost medulla (e.g., catfish: Finger 1976; Hayama and Caprio 1989; Kanwal and Caprio 1987), a segregation that mirrors the specific effects of facial versus vagal input in controlling food selection versus swallowing in these species (Atema 1971).
; Ogawa et al. 1987). In one study, it was suggested that the anterior tongue was represented ventral to the posterior tongue (Norgren and Pfaffmann 1975). This conclusion, however, was based on specifically stimulating the anterior tongue by placing it in a chamber but attempting to stimulate the posterior tongue by flowing tastants over the rest of the oral cavity, a technique that probably failed to optimally stimulate buried circumvallate and foliate taste buds while unintentionally activating other taste buds including some in the AO (e.g., nasoincisor duct). Another investigation using more specific stimulation procedures supported the contention that anterior tongue neurons are located ventrally but suggested that dorsal neurons receive input, not from the posterior tongue, but rather from the palate, or palate and tongue, although only a few neurons optimally responsive to posterior tongue stimulation were recorded (Ogawa et al. 1987). Further, despite the specificity of stimulation, there was no evidence that PBN neurons had orderly receptive fields like those in NST. In sum, it remains unclear whether the differential medullary representation of the AO and PO persists in the pons.
). This investigation provided provocative evidence that the taste-responsive region of PBN may extend beyond the boundaries usually recognized (Yamamoto et al. 1993). Previous neurophysiological studies have focused on a PBN region located caudally and centered on the middle third of the superior cerebellar peduncle and, although not typically related to specific subnuclei, available histological reconstructions suggest taste-responsive neurons are located mainly in the central medial and ventral lateral subnuclei (Hill 1987; Norgren and Pfaffmann 1975; Ogawa et al. 1987; Schwartzbaum 1983; Scott and Perrotto 1980; Travers and Smith 1984; Van Buskirk and Smith 1981). Indeed, the one study that localized taste responses within PBN subnuclei found them confined to these subnuclei in hamster (Halsell and Frank 1991). The Fos results in rat suggest, however, that more rostral and lateral neurons in the external medial and external lateral subnuclei are also responsive to taste stimuli and, interestingly, that they respond preferentially to behaviorally aversive stimuli such as quinine and HCl. The Fos data therefore suggest that neurophysiological studies may have missed a second taste-responsive PBN region. Indeed, recent anterograde tracing studies in rat demonstrate that the external medial and external lateral subnuclei are labeled heavily with fibers and terminal-like varicosities after rostral NST injections (Herbert et al. 1990), particularly when the injections are centered into the PO-responsive NST region (Becker 1992; H. C. Hu and S. P. Travers, unpublished observations).
METHODS
Abstract
Introduction
Methods
Results
Discussion
References
; Travers and Norgren 1991, 1995; Travers et al. 1986). The hypoglossal and superior laryngeal nerves were sectioned bilaterally to prevent tongue movement and reflex swallowing. A tracheal cannula was inserted and a drain was passed into the mouth from the pharynx. The clinical crowns of the upper incisors were amputated. Sutures were placed in the corners of the mouth and around the lower incisors. When the sutures were retracted, the entire oral cavity was visible. Sutures also were placed through the tongue between the intermolar eminence and the circumvallate papilla. These sutures were used to position the tongue to maximize exposure of the foliate or circumvallate papillae. The portions of the parietal and interparietal cranial plates above the transverse sinus and lateral to lambda were trephined. The dura was reflected around the transverse sinus and the sinus was ligated and resected.
) were used to record extracellular action potentials arising from single neurons, recognized by their consistent waveform and amplitude. Neural activity was amplified, monitored, and stored on video format tape using conventional techniques. Voice commentary and stimulus markers (vide infra) were stored concurrently to facilitate off-line analysis. The electrode was positioned perpendicular to the dorsal surface of the skull with the skull flat, i.e., positioned so bregma and lambda were on the same horizontal plane.
; Giza et al. 1996; Nejad 1986; Travers and Norgren 1991). To maximize the specificity of the "sweet" component of this stimulus, however, we used a saccharin concentration near the behavioral preference peak for rats (Ogawa 1972), which is somewhat less than the 0.02-0.03 M range employed in these other neurophysiological studies.
; Halsell et al. 1993). If the neuron was responsive to whole-mouth taste stimulation, the other taste receptor subpopulations [anterior tongue (AT), nasoincisor duct (NID), foliate papillae (FOL), and soft palate (SP)] were tested specifically using a soft nylon brush. If the cell was still viable, it was further tested by stimulating the responsive receptor subpopulation with individual taste stimuli via a pipette. Approximately 2 ml of stimuli were presented to the whole mouth via the syringe, and ~1 ml was presented to the individual receptor subpopulations via the micropipette. The pipette was positioned to specifically stimulate the individual taste receptor subpopulations on the ipsilateral side of the mouth (Halsell et al. 1993). Stimulation of the CV was optimized by placing the tip of the pipette into the trench that surrounds this papilla and contains all CV taste buds. Because FOL taste buds are distributed among several adjacent folds, the pipette was positioned near the surface of the tongue so that all folds were contacted by the tastant and stimulus access to the folds was facilitated by stretching them open. To stimulate the SP taste buds, the pipette was positioned in front of the anterior SP so the fluid would flow down the SP behind the tongue and into the drain tube. The AT was stimulated by gently pulling the tongue forward and out of the mouth with the sutures. The pipette was positioned above the AT so the fluid would drip off the end of the tongue. The specificity of receptor stimulation was verified in initial experiments by flowing dye through the pipette. Proper positioning of the tongue with the sutures and constant suction through the drain tube prevented the taste fluids from flowing to adjacent taste receptor subpopulations. During data collection, the flow of the taste fluids was monitored visually to verify that the fluid was not contacting fields other than the one being targeted.
). The timing of fluid flow was controlled using a programmable spreadsheet and a digital output interface (Modular Instruments) that triggered opening and closing of solenoid valves (General Valve). The triggering TTL pulses were recorded on one channel of a video tape (Vetter) to be used as stimulus onset/offset markers. The stimulus sequence consisted of 10 s of no stimulus, and then a continuous flow through the same pipette consisting of 10 s of water, 10 s of stimulus, and 20 s of water rinse. The interstimulus rest period totalled 1 min. The stimuli were applied in a varied order.
, 1995; Travers et al. 1986), there were no differences in the classification of neurons when taste responses were corrected with spontaneous rate or the activity occurring during water flow. Minor differences in response magnitude were noted, but for consistency, the taste response was corrected with spontaneous rate for both manual and automated application of taste stimuli. The automated system in the present study provided for a continuous flow of water followed by tastant so that any tactile and thermal responses were adapted during the gustatory stimulation period. Indeed, subtracting the water response would have underestimated the gustatory contribution when the automated flow system was used.
View larger version (105K):
[in a new window]
FIG. 1.
A: photomicrograph of a Weil-stained section through parabrachial nucleus (PBN) at rostrocaudal level C (see Fig. 7). 
, lesion in external medial (EM), that corresponds to recording site for cell 213. 
, lesion in ventral lateral (VL), that corresponds to recording site for cell 212. B: photomicrograph of adjacent cresyl violet stained section. , same lesion as in A; *, lesions centers. Note that a second lesion was made ~200 µm ventral to lesion indicated by in A and B. Scale bar is 200 µm for both A and B. C: record of action potentials for cell 213 during QHCL stimulation of circumvallate papilla (CV). Small , time of water flow on/off; Large , time of QHCL flow on/off. Down is flow on and up is flow off. Scale bar is 5 s. Artifacts apparent in record occur during solenoid switching. D: bar graph showing response of cells 212 (
) and 213 (
) in spikes/second to taste mixture stimulation of different receptive fields. Cell 212 is anterior tongue (AT)-best but also has a suprathreshold response to foliate papillae (FOL). Cell 213 is CV-best but also has suprathreshold responses to nasoincisor duct (NID) and FOL. E: bar graph showing response to different taste stimuli on CV for cell 213 in spikes/second. - - -, criteria level for this cell. F: peristimulus time histogram of responses to mechanical stimulation of different receptive fields for cell 213. Bins are 500 ms. B, buccal wall; CV, I, intermolar eminence; PT, posterior tongue between CV and FOL; SP, soft palate; S, Na sacccharin; N, NaCl; H, HCl; Q, quinine hydrochloride.
10 spikes per 10 s that was 2.5 SD of the spontaneous rate (Travers and Norgren 1995). Spontaneous rate was measured for 10 s preceding each stimulation, and the SD was calculated across these trials. The across-stimulus correlations between gustatory neurons were analyzed using multivariate techniques (Systat, SPSS). For hierarchical cluster analysis, the Pearson product-moment correlation coefficients and average-linkage method were used to calculate taste profile similarities. For multidimensional scaling, the Pearson product-moment correlation coefficients were used to derive and plot interneuronal distances in two dimensions. A two-dimensional space was chosen because the calculated stress levels were very similar for more than two dimensions.
2.5 SD of the spontaneous rate. For both gustatory and mechanical responses, neurons without spontaneous rate met a second criterion of a change of firing rate at least 1 spikes/s. The level of significance for all statistical analyses was set at P 
0.05 and variancesare SE.
and Fulwiler and Saper (1984). To ascertain whether there was a topographic organization of orosensory cells according to their response properties, the positions of cells were analyzed in relation to their relative positions along the three anatomic axes and in relation to their subnuclear locations. Because the number of neurons in some subnuclei was quite small, an inclusive and quantitative analysis was feasible only using the former method. To this end, we compared neurons recorded within the rostral versus caudal halves of the PBN region we recorded from and neurons recorded dorsal versus ventral to a line bisecting the brachium conjunctivum. For the mediolateral axis, we divided the PBN into equal thirds instead of halves, because dividing the PBN into mediolateral halves would have split the largest group of orosensory neurons. The middle versus the medial and lateral thirds were referred to as the "core" and "shell" regions, respectively. Although the mediolateral parcellation was not based on morphological criteria, there was overlap with morphologically defined subdivisions. The core region generally corresponds to the lateral half of the central medial subdivision (CM), the ventral lateral subdivision (VL), as well as portions of the ventral medial and central lateral subdivisions. The shell corresponds to the dorsal medial, external medial (EM), external lateral -inner and -outer (ELi, ELo) subdivisions, as well as the remaining portions of the central lateral and ventral medial subdivisions.
RESULTS
Abstract
Introduction
Methods
Results
Discussion
References
; Travers and Norgren 1995). There were two subsets of G neurons: one responded to both gustatory and intraoral mechanical stimulation (GM neurons, n = 55), whereas the other was not mechanically responsive (G-only neurons, n = 20). Of the 67 nongustatory neurons, 54 responded to intraoral mechanical stimulation so were designated as M neurons, whereas the remaining 13 cells were unresponsive to both gustatory and mechanical stimulation. Unless otherwise specified, the properties of the evoked taste responses described below refer to responses elicited by whole-mouth stimulation with the taste mixture, and mechanical response properties refer to responses elicited by stroking the most effective region of the oral cavity.
1 spike/s)for a smaller proportion of G than M neurons (
2 = 14.4, P < 0.005; Table 1). Even among cells with spontaneous firing, G neurons had significantly higher rates of unstimulated activity than M neurons (t = 3.94, df = 64, P < 0.001; Table 1). Differences between G-only versus GM neurons followed this trend. The mean spontaneous rate for spontaneously active cells was higher for G-only (means = 10.9 ± 2.94 spikes/s) than GM neurons (means = 6.0 ± 0.77 spikes/s; t = 2.11, df = 47, P < 0.05). In addition to higher rates of resting activity, G neurons had significantly higher rates of evoked activity than M neurons (t = 5.28, df = 88, P < 0.001; Table 1). There was, however, no significant difference between the mean taste responses for GM versus G-only neurons (t = 0.26, df = 66, P = 0.798).
View this table:
TABLE 1.
Response properties of G versus M neurons
; Travers et al. 1986) and whether the most vigorous, "best" taste response arose from stimulating the AO or PO. Within groups, AO-best neurons are ordered by their responses to the AT and PO-best neurons by their responses to the FOL (Fig. 2, ). Compared with PO-best neurons, AO-best neurons had significantly higher spontaneous (t = 4.13, df = 73, P < 0.001) and evoked (t = 3.70, df = 73, P < 0.001) response rates (Table 2). The vast majority of both AO and PO responses were excitatory. Taste stimulation elicited decrements in firing rate in only two neurons. Interestingly, both inhibitory responses were evoked by PO stimulation in neurons that responded best and in an excitatory manner to AO stimulation (Fig. 2, downward-oriented filled bars). The magnitudes of these inhibitory responses, however, were very small.
View larger version (19K):
[in a new window]
FIG. 2.
Responses of 60 neurons to stimulation of individual taste receptor subpopulations. Two whole-mouth-only cells are not shown because their receptive fields were not localized. Five of 26 cells that only responded to AT stimulation are also not shown because responses to stimulation of individual taste receptor subpopulations were not taped during testing. Responses are in spikes/second and responses for a given cell are aligned on vertical axis. Neurons are grouped by whether individual cells had suprathreshold responses to a single (SPECIFIC; 2 groups on left) or to multiple (CONVERGENT; 2 groups on right) receptor subpopulations. Neurons are further grouped by whether most effective receptor subpopulation was in anterior (AO; AT or NID) or posterior (PO; FOL, CV or SP) oral cavity. Within the 2 AO-best groups, neurons are ordered by magnitude of responses to AT stimulation, and within 2 PO-best groups, neurons are ordered by magnitude of responses to FOL stimulation. , suprathreshold responses; 
, responses below threshold.
View this table:
TABLE 2.
Response properties of AO- versus PO-best taste neurons
2 = 13.74, df = 1, P < 0.005).
2 = 3.08, df = 1, P = 0.079; Table 2). For neurons that responded to stimulation of more than one taste receptor subpopulation, there was a tendency for an orderly pattern of convergence. Almost twice as many convergent neurons received their two most effective inputs from within the AO or PO than from both the AO and PO (2 = 3.62, df = 1, P = 0.057; Table 2). When all suprathreshold responses were taken into account, however, notable convergence between the AO and PO was evident. About equal numbers of convergent neurons responded to stimulation of taste receptor subpopulations confined to the AO or PO as responded to stimulation of fields in both these regions.
View larger version (34K):
[in a new window]
FIG. 3.
Suprathreshold evoked response (spikes/second above spontaneous) to taste ( ) and mechanical (
) stimulation of best receptive field for 43 individual gustatory and intraoral mechanical responsive (GM) neurons. Neurons are grouped by whether the cell receives both taste and mechanical input from only anterior oral cavity fields (AO only), both AO and PO fields (AO & PO) or only posterior oral cavity fields (PO only). Within the groups, the neurons are ordered by the response to taste. Two GM neurons that responded to mechanical stimulation of fields in either the AO or PO and to taste stimulation of only whole mouth are not included because locations of their taste receptor subpopulations are unknown.
) for these neurons were derived by using the most vigorous response to a given stimulus, regardless of the receptive field location (Travers et al. 1986). For 82% of the neurons, each response used in the profile was evoked by stimulation of the same receptor subpopulation; for the remaining 18%, responses used for different stimuli arose from different receptor subpopulations. For example, cell 191 received convergent input from the NID, CV, and FOL and responded to all four stimuli, but responses varied depending on which receptor subpopulation was stimulated. The optimal responses evoked by QHCl and NaCl arose from the CV, whereas those evoked by HCl and Na saccharin arose from the NID and FOL, respectively. Consequently, the chemosensitive response profile for cell 191 included responses from three different receptor subpopulations.
0.82). In one cluster, 38/39 neurons responded most vigorously to NaCl (N cluster) and in the other cluster, 9/9 responded most vigorously to HCl (H cluster). The remaining 10 neurons did not separate into major clusters, although three neuron pairs were intercorrelated as highly as the N and H clusters. On the other hand, nine of the cells could be distinguished from neurons in the N and H clusters on the basis of their greater responsiveness to quinine. All responded best (n = 5) or second-best (n = 4) to quinine. In contrast, quinine was never the second-best stimulus for N- or H-cluster neurons. Thus although the nine neurons with quinine sensitivity were not highly intercorrelated, they were considered a functional group and, for brevity, designated as the quinine-sensitive group (Q+ group). The last neuron amalgamated in the dendrogram was distinct and actually was correlated negatively with the other 57 (correlation distance >1.0). This was the only cell that responded optimally to Na saccharin, which was generally a poor stimulus in the present study.
View larger version (27K):
[in a new window]
FIG. 4.
Dendrogram derived from hierarchical cluster analysis on the response spectra of 58 G neurons. Cells are labeled with taste qualities eliciting suprathreshold responses (ordered from most-effective to least-effective quality: N, NaCl; H, HCl; S, Na saccharin; Q, quinine), by whether they responded optimally to anterior (A) or posterior (P) oral cavity stimulation and by cell number. 
, Pearson interneuron correlations > 0.82 (distance <0.18). Two main clusters of neurons at this correlation level are labeled: N, NaCl cluster; H, HCL cluster. Inset: map derived from multidimensional scaling of interneuron correlations. Cells that are members of N cluster and H clusters are outlined. Two N-cluster neurons that are not NaCl-best are labeled with their best stimulus designation [H is cell 81; N is cell 213 (see Fig. 1)]. Ten neurons not grouped into true clusters are labeled with their best stimulus designation and include 9 Q+ neurons and the lone S neuron.
View larger version (23K):
[in a new window]
FIG. 5.
Mean evoked responses to each of 4 tastants for each cluster or group classification (N, NaCl cluster; H, HCl cluster, Q+, quinine sensitive group; S, Na saccharin-best cell). Responses are in spikes/second and error bars are means ± SE. Number of neurons comprising each cluster or group with best receptive fields in either AO or PO is indicated.
. An analysis of variance (ANOVA) revealed a significant main effect for cluster type (F = 17.39, df = 2, P < 0.001; Table 3). Q+ neurons were tuned more broadly than N- or H-cluster neurons, and H-cluster neurons were more broadly tuned than N-cluster neurons. Finally, neurons in the N and H clusters were differentiable from the Q+ group on the basis of their receptive field designation. Significantly more N- and H-cluster neurons were AO-best whereas more Q+ neurons were PO-best (P < 0.005, Fisher's Exact; Table 3).
View this table:
TABLE 3.
Properties of neuron classes
View this table:
TABLE 4.
Mean responses to stimulation of receptor subpopulations with standard taste stimuli
).
). Despite their complexity, the configurations of these receptive fields were orderly. Many more composite neurons received their two most effective inputs from within the AO or PO (n = 24) compared with both the AO and PO (n = 7; P < 0.05, Fisher's Exact), although about equal numbers of M-composite neurons exhibited suprathreshold responses to stimulation of both the AO and PO.
View larger version (59K):
[in a new window]
FIG. 6.
Responses of 31 M-composite neurons (suprathreshold responses to stimulation of multiple individual oral receptive fields). Responses are given as a ratio of suprathreshold responses to each field for a given cell with the most vigorous response set as maximum y-axis value and responses below threshold set at 0. Individual neurons are aligned along rows and are labeled with cell number. Individual receptive fields are aligned along columns. Neurons are grouped by whether the cell receives input only from anterior oral cavity receptive fields (5 fields on right, AO COMPOSITE), only from posterior oral cavity (4 fields on left, PO COMPOSITE), from both anterior and posterior oral cavities but with most effective receptive field in AO (BOTH COMPOSITE: AO BEST) or PO (BOTH COMPOSITE: PO BEST). , excitatory responses; , inhibitory responses. AI, anterior tongue and incisal papilla; I, lower incisors; HP, hard palate; M, upper and lower molars; A+, other anterior oral cavity regions (e.g., intermolar eminence, buccal wall); CF, circumvallate and foliate papillae; RM, retromolar mucosa; P+, other posterior oral cavity fields (e.g., posterior tongue).
2 = 12.67, P < 0.001; Table 5). Although as a group, orosensory neurons were more numerous in caudal PBN, this depended on modality (Fig. 7, Table 5). G neurons were located predominantly caudally but M neurons were distributed evenly between the rostral and caudal halves of PBN (2 = 14.59, P < 0.001). Indeed, at the most rostral level (Fig. 7, level D), the only orosensory neurons were M neurons.
View larger version (31K):
[in a new window]
FIG. 7.
Location of neurons plotted online drawings of 4 standard transverse levels through PBN based on modality and best receptive field. Level A is most caudal and level D, most rostral. Distance between levels A and B, B and C is ~200-250 µm each and between levels C and D is 300-350 µm. Filled symbols are AO-best and open symbols are PO-best. 
and 
, gustatory neurons (G neurons); and 
, oral somatosensory neurons (M neurons). Unresponsive cells are indicated by small . - - -, division of PBN into 3 equal mediolateral segments (core and surrounding shell). Dorsal is toward top, and lateral is to right. Inset: PBN subdivisions. CL, central lateral; CM, central medial; DM, dorsal medial; EL, external lateral; ELi, external lateral-inner; ELo, external lateral-outer; EM, external medial; IL, internal lateral; VL, ventral lateral; VM, ventral medial; KF, Kolliker-Fuse nucleus; LC/Me5, Locus Coeruleus and Mesencephalic trigeminal nucleus. Scale bar is 0.5 mm.
View this table:
TABLE 5.
Topographic organization summary
2 = 16.99, P < 0.001). Nearly one-third of M neurons were clustered in the lateral shell at level C, within the EM and Eli subdivisions. In contrast, the majority of G neurons in the core were located as a tight cluster in CM and VL and in the cell bridges that span the superior cerebellar peduncle between these two subdivisions ("waist" region) at level B. G neurons in the shell were somewhat more scattered, occupying the dorsal medial, CM, VL, central lateral, and EM and ELi subdivisions, although almost half were concentrated in EM and Eli. The G neurons in EM and ELi were GM neurons.
2 = 2.712,P = 0.1; Fig. 7, Table 5).
2 = 11.2, df = 1, P = 0.001). When considered alone, G neurons exhibited a similar orotopic arrangement modified by the overriding tendency for G neurons to be located caudally. More than six times as many AO-best G neurons were located caudally versus rostrally, compared with only 1.7 times as many PO-best neurons (P = 0.006, Fisher's exact). Although PO-best G neurons in caudal PBN outnumbered those in rostral PBN, the proportion of G neurons that were PO-best at a given level actually increased at successively more rostral locations: level A, 8.3%; level B, 24.3%; and level C, 50%. The trend for a reversed orotopy was also evident for M neurons, with three times as many PO-best M neurons being located rostrally as caudally whereas slightly more AO-best neurons were located caudally compared with rostrally (2 = 3.82, df = 1, P = 0.051).
2 = 17.50, df = 1, P < 0.001). This mediolateral orotopy, however, only reflected the distribution of G neurons:44/46 AO-best G neurons versus 5/16 PO-best G neurons in the core (P < 0.001, Fisher's Exact). Less than half of the M neurons were found in the core, regardless of whether they were AO- or PO-best (2 = 0.001, df = 1, P = 0.97).
View this table:
TABLE 6.
Numbers of different neuron types within CM/VL versus EM/ELi
DISCUSSION
Abstract
Introduction
Methods
Results
Discussion
References
; Frank 1991; Frank et al. 1988). That larger taste responses arise from AO-best than PO-best neurons is consistent with a previous report of an unquantified "smaller" taste response within the PBN arising from stimulating the AT versus the rest of the mouth (Norgren and Pfaffmann 1975) and is similar to our previous observations in the NST (Travers and Norgren 1995).
; Norgren and Pfaffmann 1975; Ogawa and Hayama 1984; Ogawa and Nomura 1988; Ogawa et al. 1982, 1992a,b; Sweazey and Bradley 1989, 1993; Travers and Norgren 1995). Numerous cells responsive to both gustatory and mechanical stimulation similar to those in the present study also have been reported in these earlier studies and may represent another aspect of convergence. However, some of this dual responsiveness may merely reflect the multimodal responses reported in peripheral fibers (discussed in Travers 1993).
; Norgren and Pfaffmann 1975; Ogawa et al. 1982), a range that may be attributed partially to differences in methods of stimulation and data analysis. For example, responses arising from the CV and FOL often are stimulated or analyzed as a single field, thus underestimating the amount of convergence between these two lingual fields. In the present study, a significant proportion (15%) of G neurons received both CV and FOL input, a clear instance of central processing, because single glossopharyngeal nerve fibers do not respond to stimulation of both papillae (Frank 1991).
; Travers et al. 1986), but except for one report of mechanically sensitive neurons in the sheep NST (Sweazey and Bradley 1989), other reports of central convergence in gustatory relays have not noted this pattern (Hayama et al. 1985, 1987; Ogawa et al. 1984). The lack of systematic convergence in other studies may be mainly a matter of definition. Receptive fields typically are defined simply by the presence or absence of suprathreshold responses. The orderliness observed in PBN was dependent on comparing the relative magnitudes of responses arising from different receptive fields. When all suprathreshold responses were treated equivalently, no orderly pattern emerged in the present data either. Orderliness also was noted for across-modality convergence. The taste and tactile receptive fields of most GM neurons were both restricted to the AO or PO. A similar organization was found for dually responsive NST neurons, such that gustatory and mechanical receptive fields showed substantial spatial overlap (Travers and Norgren 1995).
; Travers et al. 1986), a trend for increased integration was suggested. Although proportions of convergent neurons were similar at the two levels (53% in NST; 48% in PBN), differences emerged when the details of receptive field configuration were assessed. If receptive fields were defined using all suprathreshold inputs, 69% of medullary neurons had receptive fields within the AO or PO, but in PBN, only 41% of the cells exhibited these confined receptive fields (2 = 5.41, df = 1, P = 0.015). Similarly, there was a trend for more medullary (83%) than pontine (65%) neurons to receive their two most effective inputs from the same half of the oral cavity (2 = 3.48, df = 1, P = 0.062). On the other hand, no significant difference was apparent in patterns of convergence for NST (Travers and Norgren 1995) versus PBN M neurons, though data from fewer NST cells were available for comparison. Almost identical proportions of NST and PBN M neurons (45.7 and 46.9%, respectively, 2 = 0.009, df = 1, P = 0.924) had confined suprathreshold receptive fields. If there was any difference in the two levels, it was a slight trend in the direction opposite to that for G neurons when inputs from the two most effective receptive fields were analyzed; 66% of NST versus 80% of PBN neurons had confined optimal receptive fields (2 = 1.758,df = 1, P = 0.185). The lack of increased integration for PBN M neurons is consistent with a more important role for stimulus localization for the somatosensory compared with the gustatory system.
, 1991; Sweazey and Bradley 1989; Travers 1993). In addition, the limited convergence between certain sets of taste receptor subpopulations, e.g., the AT and CV, may be necessary for maintaining central separation in the processing of tastants, such as salt and quinine, that are differentially effective for these groups of receptors.
). Indeed, similar to taste mixture interactions (Travers and Smith 1984; Vogt and Smith 1993a,b), linear summation was rarely observed. Further experiments are required to gain insight into this question. In particular, it will be interesting to analyze differences in integration between different taste bud groups, because some findings have suggested inhibitory interactions between the AO and PO (Halpern and Nelson 1965; Lehman et al. 1995; Miller and Bartoshuk 1991, but see Dinkins and Travers 1996). A hint of such an interaction was evident in a few neurons in the present study in which antagonistic responses were evoked by AO versus PO stimulation.
; Frank 1991; Frank et al. 1983; Nejad 1986). The highly intercorrelated, relatively narrowly tuned AT-best N neurons we observed were similar to types previously reported in hamster PBN (Smith et al. 1979, 1983), NST-PBN projection neurons (Monroe and Di Lorenzo 1995; Ogawa et al. 1984), and rodent chorda tympani fibers and geniculate ganglion cells (Boudreau et al. 1983; Frank et al. 1983; Smith et al. 1979, 1983).
; Formaker and Hill 1991; Giza and Scott 1991; Hettinger and Frank 1990; Ninomiya and Funakoshi 1988; Scott and Giza 1990). This pathway has been suggested to be necessary for the discrimination of sodium and involved in sodium appetite and is probably modifiable based on the homeostatic state of the organism (Contreras and Frank 1979; Jacobs et al. 1988; Spector and Grill 1992; Sollars and Bernstein 1992). Interestingly, PBN lesions that abolish the adaptive change after sodium depletion appear to be effective only when the lesions are centered in the CM/VL PBN subnuclei (Flynn et al. 1991; Hill and Almli 1983; see also Spector 1995), the location where N neurons were most numerous in the present study. This is also the PBN location where expression of the immediate early gene c-fos is observed when rats ingest 0.2 M NaCl (Yamamoto et al. 1993). When NaCl is presented with amiloride, however, c-fos is not expressed in this area. Preliminary electrophysiological evidence from our lab supports the preferential amiloride-induced suppression of NaCl-evoked responses in N-cluster neurons in PBN. Using a modification of the procedure of Hettinger and Frank (1990), we found that the responses of N-cluster, AT-best neurons that were evoked by NaCl were reduced by 49% after amiloride application, but the comparable reduction in H-cluster neurons was only 27% (unpublished observations). This reduction is similar to that found in NST (Giza and Scott 1991), although attenuated compared with chorda tympani fibers (Hettinger and Frank 1990).
). In the present study, we used Na saccharin instead of sucrose and obtained poor responses to this sweet tastant (vide infra). Thus it seems possible that some of the apparent AT-only NaCl-best neurons observed here might have exhibited a sensitivity to palatal stimulation with sweet stimuli had a more effective representative tastant been employed. Indeed, in chronic recording studies, many sodium-best PBN neurons exhibit significant sideband responses to sweet stimuli (and vice versa), although narrowly tuned NaCl-best neurons are also evident (Nishijo and Norgren 1990).
; Di Lorenzo and Schwartzbaum 1982; Giza and Scott 1991; Ogawa et al. 1987; Travers and Smith 1979 1984). The small proportion of PO-best units classified in the H cluster (Table 3) is somewhat surprising in light of the fact that the glossopharyngeal nerve contains a high proportion of fibers highly responsive to HCl (Frank 1991). However, HCl-best cells are also common in the chorda tympani nerve and the broadly tuned average profile for H neurons in the present study bears a greater resemblance to their chorda tympani than glossopharyngeal counterparts, given that chorda tympani H fibers are tuned more broadly (Boudreau et al. 1983; Frank 1991; Frank et al. 1983, 1988). In fact, in contrast to the average response profile, the two H-best neurons that received optimal input from the PO exhibited relatively specific responses to HCl.
; Frank et al. 1983). On the other hand, in contrast to the narrowly tuned quinine-best fibers observed in the glossopharyngeal nerve, the Q+ neurons that we recorded were the most broadly tuned of all the PBN groups. If these Q+ PBN neurons are central recipients of information from glossopharyngeal Q units, the responses are transformed dramatically at the second-order relay. The apparent transformation in neural responsiveness to quinine could be due partially to convergence, because, for most of the Q+ neurons, responses from two or three receptor subpopulations contributed to the chemosensitive profiles. Convergence is probably an insufficient explanation for the attenuation of quinine specificity, however, because most of Q+ unit profiles were composites of similar profiles from the CV and FOL. It is somewhat difficult to compare the Q+ neurons in the present study with quinine-responsive neurons in other central recording studies. Although quinine-best neurons have been reported at all levels of the rodent central taste neuraxis, they tend to comprise a small percentage of recorded cells, and their properties are variable or incompletely described. Indeed, in many investigations, even when whole-mouth stimulation or chronic recording techniques were used, such cells constituted <5% of the population (e.g., Giza and Scott 1991; Nishijo and Norgren 1991; Nomura and Ogawa 1985; Norgren and Pfaffmann 1975; Ogawa et al. 1987; Ogawa and Nomura 1988; Ogawa et al. 1982, 1987; Travers et al. 1986; Yamada et al. 1990). Two studies of rat cortex, however, reported proportions of quinine-best neurons comparable with other best-stimulus classes, though absolute numbers of cells were small. In one study, quinine-best neurons were similar to the cells we observed in exhibiting very broad tuning and relatively low response rates (Ogawa et al. 1992a), whereas in the other investigation, they were relatively narrowly tuned and had comparable response rates (Yamamoto et al. 1989). Clearly, much remains to be learned about the central representation of bitter-tasting stimuli. The lack of more robust central responses to this class of stimuli is somewhat puzzling (Norgren et al. 1989) but observations in this lab suggest that PO-responsive neurons are more difficult to isolate than AO-responsive cells (Halsell et al. 1993; Travers and Norgren 1995; M. Dinkins and S. P. Travers, unpublished observations), which may make it problematic to obtain a representative sample of PO cells, including quinine-best neurons.
), this lone S-best neuron responded most vigorously to stimulation of SP taste receptors on the Geschmacksstreifen. Aside from this match in sensitivities, the generally poor saccharin response in the present study is perplexing. We routinely stimulated each of the rat taste receptor subpopulations exhibiting good sensitivity to sweeteners; i.e., the hard and SP (Nejad 1986; Travers and Norgren 1991; Travers et al. 1986) and posterior tongue (Frank 1991). It is possible that a bias against isolating neurons responsive to these tastants may have existed due to mixture suppression effects when we employed our quaternary taste mixture to "search" for responsive neurons. There is evidence for suppression between qualitatively different stimuli in binary mixtures, specifically suppression of sugar-evoked responses by quinine or acid (Travers and Smith 1984; Vogt and Smith 1993a,b). Although our previous studies in NST revealed a sizable population of cells responsive to sweeteners despite using mixtures as a search stimuli, those mixtures contained 0.3-1.0 M sucrose: concentrations which are probably more effective than the Na saccharin concentration (0.01 M) in the present study (e.g., see Travers and Norgren 1991). Thus although it may not completely explain our results, a bias against sampling cells optimally responsive to sweeteners may have existed due to the additive effects of mixture suppression and a suboptimal representative stimulus. These same factors could have contributed to the relatively poor palatal response. Only 9% of the cells in the current investigation responded best to palatal stimulation, compared with 27-31% in previous NST investigations in which 1.0 M sucrose was used instead of saccharin (Travers and Norgren 1995; Travers et al. 1986).
; Ogawa et al. 1982, 1984, 1987). The percentage of M neurons recorded in the present study (42%) was also similarto that reported by Ogawa and colleagues in the PBN (mean =32%) (Ogawa et al. 1982, 1987) using similar stimulating techniques. Because the NST also contains M neurons and is the source of a robust afferent PBN projection, a logical assumption would be that PBN M neurons simply receive input from the analogous neuronal population in NST. Several experiments, however, raise questions regarding the connectivity of PBN M neurons. An antidromic stimulation study demonstrated that only 13% of NST-PBN projection neurons were M neurons; the remaining majority were taste-responsive cells (Ogawa et al. 1984). This physiological evidence suggesting only a sparse NST-PBN M neuron projection is supported by both anterograde and retrograde tracing data. After injections of biocytin into the lateral, mechanically responsive NST, there was only sparse PBN terminal labeling, compared with much denser labeling after similar injections into the more medial gustatory-responsive NST (Becker 1992). Complimentary results were obtained with Fluorogold injections into PBN (Halsell et al. 1996). Many more NST neurons were labeled retrogradely in the more medially located rostral central subnucleus compared with the more laterally placed rostral lateral subnucleus. Thus it would appear that a direct NST-PBN circuit may not fully account for PBN M neurons. Indeed, although the response properties of PBN M neurons generally resembled NST M neurons (Travers and Norgren 1995), there were some differences in receptive field location, e.g., an increased number of cells responsive to periodontal stimulation, implying input from a source outside NST. One candidate for this additional input is the spinal trigeminal nuclei, especially the paratrigeminal islands, as suggested by anterograde tract-tracing studies (Cechetto et al. 1985; Feil and Herbert 1995). It is also possible, of course, that divergence of M neuron projections from NST contributes to the discrepancy between the sizeable population of PBN M neurons and the apparently small population of mechanically sensitive NST-PBN projection neurons.
). In agreement with this hypothesis, recent observations in our lab have demonstrated a projection from orosensory PBN to the region of the parvocellular reticular formation implicated in coordinating oromotor behavior (Karimnamazi et al. 1996).
; Travers and Norgren 1995). However, it is difficult to compare the relative degree of anatomic segregation in the two relays quantitatively.
; Halsell and Frank 1991). This corresponds to the location of the classically described pontine taste area (Norgren and Leonard 1973) where gustatory neurons appear to be located in most neurophysiological studies of PBN (Hill 1987; Norgren and Pfaffmann 1975; Ogawa et al. 1987; Schwartzbaum 1983; Scott and Perrotto 1980; Travers and Smith 1984; Van Buskirk and Smith 1981). The second group of G and M neurons was located further rostrally and laterally within the coextensive EM and ELi subdivisions (Fulwiler and Saper 1984; Herbert et al. 1990). Neurophysiological studies have not previously linked this region with gustatory responsiveness, but Yamamoto's group recently suggested a taste function based on c-fos immunohistochemistry (Yamamoto et al. 1994). This region also has been suggested to be part of a visceral and somatosensory-pain "relay" to dorsal thalamus based on anatomic and physiological studies (Bernard and Besson 1988, 1990; Bernard et al. 1989; Feil and Herbert 1995; Fulwiler and Saper 1984). The results of the present study suggest that an orosensory function be added to the multimodal nature of EM/ELi.
).
; Cechetto 1987; Chamberlin and Saper 1992; Han et al. 1991) or to input from higher brain regions (for review, see Travers 1993).
; Yamamoto and Kawamura 1975; Yamamoto and Kitamura 1990; Yamamoto et al. 1980). A robust AO/PO distinction occurs in the teleost medulla and possibly in the PBN homolog (Finger 1976; Hayama and Caprio 1989; Kanwal and Caprio 1987; Lamb and Caprio 1992; Morita and Finger 1985), thus suggesting a phylogenetically conserved organization. Indeed, in the fish, anatomic studies suggest an arrangement of the facial and vagal nerve representation in the pontine superior secondary gustatory nucleus (Morita and Finger 1985) that is strikingly similar to what we have observed in the mammalian PBN. On the other hand, the available neurophysiological data in the fish suggest extensive intermingling and convergence of extra- and intraoral input within this nucleus (Lamb and Caprio 1992).
). Data from the present study provide a hint that this is the case. Four of six G neurons in the medial shell, including the three most medially segregated neurons, were PO-best (level B, , Fig. 7). In addition, multiunit responses to posterior oral cavity stimulation were noted routinely in this area (unpublished observations). Consistent with these neurophysiological observations, a previous anatomic study from this lab demonstrated that biocytin injections into the PO-responsive NST produced terminal labeling in CM and VL in PBN that was focused medial to that observed after similar injections into the AO-responsive NST (Becker 1992). Further data are clearly necessary to determine whether the medial half of CM/VL is similar to EM/ELi in responding preferentially to posterior oral stimulation.
; Ogawa et al. 1987), the locations of single-unit recording sites in the present study did not suggest strong orotopy along the dorsoventral axis, either with respect to the representation of AO and PO or tongue and palate. On the other hand, multiunit responses obtained while attempting to isolate single units suggested that optimal AT responses were located ventral to optimal PO lingual and palatal responses (unpublished observations). Because the taste-responsive PBN is relatively small dorsoventrally, our technique of marking recording sites with lesions (sometimes made below the recording site) may not have been sensitive enough to reveal this organization. A similar discrepancy between single- and multiunit results was evident in Norgren and Pfaffmann (1975). Further work is required to clarify the nature of orotopic organization along the dorsoventral axis.
). The present data similarly suggest a caudal and ventral clustering of NaCl-best cells. The consistent finding of a caudal representation of salt responsiveness is further supported by c-fos immunohistochemistry in rat (Yamamoto et al. 1994) and by neurophysiological findings in another species, the hamster (Van Buskirk and Smith 1981). HCl-best neurons in the current investigation were most frequent caudally, in contrast to their rostral location in the previous study (Ogawa et al. 1987). The reason for this discrepancy is unclear.
). Although both types of data suggest an important role for EM/ELi in processing aversive gustatory stimuli, our data further suggest it is unlikely that EM/ELi are unique in processing these tastants. Yamamoto and colleagues (Yamamoto et al. 1994) make no mention of quinine or acid-elicited Fos expression in CM/VL but the present study clearly demonstrates that responses to these stimuli occur in the other subnuclei. Indeed, about half of both the H cluster and Q+ occurred in CM/VL (Table 6).
; Spector 1995; Travers 1993). Recently, however, on the basis of a shift in gustatory-elicited Fos after conditioned taste aversion or satiety, Yamamoto and colleagues (Yamamoto et al. 1994) suggested the interesting hypothesis that a subset of neurons in EL represent the negative hedonic properties of gustatory stimuli rather than taste quality, per se. The present data can provide no explicit support for such a suggestion, especially in light of the poor saccharin responses in our population of cells. However, the fact that each neuron in EM/ELi responded well to aversive taste stimuli does not conflict with this interpretation. Data from anatomic studies provide further support that EM/ELi may be a region distinct from CM/VL. Although both sets of PBN subnuclei project to the same taste-related forebrain nuclei, there are notable differences in the topography of their projections within these regions, for example, to the central nucleus of the amygdala (Bernard et al. 1993). Further, in contrast to CM/VL, neurons in EM/ELi do not project to medullary oromotor reflex circuits (Karimnamazi et al. 1996). Likewise, immunohistochemical studies suggest differences between these regions. A higher proportion of neurons in EM/ELi are immunoreactive for calcitonin-gene-related peptide, compared with those in CM/VL (Yasui et al. 1989; unpublished observations). Interestingly, the regions of EM/ELi that we recorded from were positioned adjacent and possibly within the PBN region involved in the spinoreticular autonomic and pain arousal system (Saper 1995). It seems possible that the robust arousal elicited by aversive taste stimuli (Grill and Norgren 1978) could use this same circuitry.
| |
ACKNOWLEDGEMENTS |
|---|
The authors thank N. Burkhardt, T. Copelin, and L. McConnell for assistance with the animal preparation and histology and L. DiNardo, Dr. Joseph Travers, and Dr. Hamid Karimnamazi for valuable comments on an earlier version of this manuscript.
This work was supported by the National Institute of Deafness and Other Communication Disorders Grant DC-00416 to S. P. Travers.
| |
FOOTNOTES |
|---|
Address for reprint requests: C. Halsell, Section of Oral Biology, College of Dentistry, The Ohio State University, 305 West 12th Ave., Columbus, OH 43210. E-mail: halsell.1{at}osu.edu
Received 30 September 1996; accepted in final form 6 May 1997.
| |
REFERENCES |
|---|
|
Abstract
Introduction
Methods
Results
Discussion
References
|
|---|
This article has been cited by other articles:
|
|
 |
|
  T. Suwabe and R. M. Bradley Characteristics of Rostral Solitary Tract Nucleus Neurons With Identified Afferent Connections That Project to the Parabrachial Nucleus in Rats J Neurophysiol, July 1, 2009; 102(1): 546 - 555. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. C. Geran and S. P. Travers Bitter-Responsive Gustatory Neurons in the Rat Parabrachial Nucleus J Neurophysiol, March 1, 2009; 101(3): 1598 - 1612. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. W. Biondolillo, L. A. Williams, and M. S. King Blocking Glutamate Receptors in the Waist Area of the Parabrachial Nucleus Decreases Taste Reactivity Behaviors in Conscious Rats Chem Senses, March 1, 2009; 34(3): 221 - 230. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Hashimoto, K. Obata, and H. Ogawa Characterization of Parabrachial Subnuclei in Mice with Regard to Salt Tastants: Possible Independence of Taste Relay from Visceral Processing Chem Senses, March 1, 2009; 34(3): 253 - 267. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. R. Kinzeler and S. P. Travers Licking and gaping elicited by microstimulation of the nucleus of the solitary tract Am J Physiol Regulatory Integrative Comp Physiol, August 1, 2008; 295(2): R436 - R448. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. T. King, M. Garcea, D. S. Stolzenberg, and A. C. Spector Experimentally cross-wired lingual taste nerves can restore normal unconditioned gaping behavior in response to quinine stimulation Am J Physiol Regulatory Integrative Comp Physiol, March 1, 2008; 294(3): R738 - R747. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. M. Morganti, A. K. Odegard, and M. S. King The Number and Location of Fos-like Immunoreactive Neurons in the Central Gustatory System Following Electrical Stimulation of the Parabrachial Nucleus in Conscious Rats Chem Senses, July 1, 2007; 32(6): 543 - 555. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. C. Geran and S. P. Travers Single Neurons in the Nucleus of the Solitary Tract Respond Selectively to Bitter Taste Stimuli J Neurophysiol, November 1, 2006; 96(5): 2513 - 2527. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. C. Spector and S. P. Travers The representation of taste quality in the Mammalian nervous system. Behav Cogn Neurosci Rev, September 1, 2005; 4(3): 143 - 191. [Abstract] [PDF]
|
|
|
|
|
|
J. V. Verhagen, B. K. Giza, and T. R. Scott Responses to Taste Stimulation in the Ventroposteromedial Nucleus of the Thalamus in Rats J Neurophysiol, January 1, 2003; 89(1): 265 - 275. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. P. Travers Quinine and citric acid elicit distinctive Fos-like immunoreactivity in the rat nucleus of the solitary tract Am J Physiol Regulatory Integrative Comp Physiol, June 1, 2002; 282(6): R1798 - R1810. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J.-P. Baird, J. B. Travers, and S. P. Travers Parametric analysis of gastric distension responses in the parabrachial nucleus Am J Physiol Regulatory Integrative Comp Physiol, November 1, 2001; 281(5): R1568 - R1580. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J.-P. Baird, S. P. Travers, and J. B. Travers Integration of gastric distension and gustatory responses in the parabrachial nucleus Am J Physiol Regulatory Integrative Comp Physiol, November 1, 2001; 281(5): R1581 - R1593. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. Cerf-Ducastel, P.-F. Van de Moortele, P. MacLeod, D. Le Bihan, and A. Faurion Interaction of Gustatory and Lingual Somatosensory Perceptions at the Cortical Level in the Human: a Functional Magnetic Resonance Imaging Study Chem Senses, May 1, 2001; 26(4): 371 - 383. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Z. Chen, S. P. Travers, and J. B. Travers Muscimol infusions in the brain stem reticular formation reversibly block ingestion in the awake rat Am J Physiol Regulatory Integrative Comp Physiol, April 1, 2001; 280(4): R1085 - R1094. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Blomqvist, E.-T. Zhang, and A. D. Craig Cytoarchitectonic and immunohistochemical characterization of a specific pain and temperature relay, the posterior portion of the ventral medial nucleus, in the human thalamus Brain, March 1, 2000; 123(3): 601 - 619. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. B. Travers, K. Urbanek, and H. J. Grill Fos-like immunoreactivity in the brain stem following oral quinine stimulation in decerebrate rats Am J Physiol Regulatory Integrative Comp Physiol, August 1, 1999; 277(2): R384 - R394. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. J. St. John and A. C. Spector Behavioral Discrimination between Quinine and KCl Is Dependent on Input from the Seventh Cranial Nerve: Implications for the Functional Roles of the Gustatory Nerves in Rats J. Neurosci., June 1, 1998; 18(11): 4353 - 4362. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
