INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

THE PARABRACHIAL NUCLEUS of the pons (PBN) has gained prominence recently as a site of potential importance in the control of feeding behavior (see, e.g., Refs. 29, 30, 33, 58). Signals related to feeding, including taste, gastric, duodenal, hepatic, and osmotic stimuli, have PBN representation (33, 36, 37, 61, 67, 69, 71). Although several studies have systematically investigated the representation of taste in the PBN (34, 49, 50, 53, 69), fewer studies (67) have sought to systematically characterize the electrophysiological response characteristics of PBN cells to gastric distension, an important feeding-related gut signal.

Gastric projections to the PBN originate in the greater splanchnic and vagal afferents (3, 15, 72). Gastric splanchnic afferents terminate in dorsal horn laminae I and V, layers that give rise to PBN projections (18, 26, 48, 72). Vagal gastric projections relay to the PBN through the caudal region of the nucleus of the solitary tract (NST; see Refs. 1, 54, and 66). Earlier reports of PBN visceral representation suggested that visceral projections terminate mainly in the external and lateral subnuclei in the rostral portion of the PBN (51, 60). However, recent studies reveal a more widespread distribution for vagal-visceral projections from the caudal NST (41, 62). For example, although terminal fields were densest in the external lateral subnucleus, biotinylated dextran injections in caudal NST sites electrophysiologically responsive to gastric distension resulted in terminal labeling throughout the PBN, including the well-characterized taste-responsive "waist region" (41). These results suggest that gastric distension signals likely have widespread influence over PBN areas that are involved in a variety of functions, including respiratory, cardiac, and taste processing (see, e.g., Refs. 13, 34, 45, 53, 62, 69).

Gastric distension responses in PBN have been only partially characterized. Suemori et al. (67) described PBN cells with excitatory and inhibitory responses to gastric distension limited principally to the dorsal and lateral regions of the PBN. However, it is not clear whether more caudal sites, traditionally associated with ingestive function, were sampled. Because previous studies strongly implicate the PBN as a locus for integration of taste and visceral signals (33, 41, 58), cells responsive to gastric distension may also be found more caudal in this nucleus. The results of Suemori et al. (67) also suggested that some responses tended to outlast distension, an effect not seen in the periphery. These prolonged responses may reflect a central code for distension magnitude or possibly satiation level. However, a lack of precise control over the timing and inflation rate did not permit a detailed quantitative analysis of distension response parameters, including response duration. Thus, to provide a more comprehensive characterization of the response properties of gastric distension-sensitive PBN neurons, we used a parametric, repeated-measures design and a computer-controlled pump to inflate and deflate the stomach at precisely controlled rates, testing a range of distension volumes from subthreshold to those that approached but did not elicit tissue damage.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

General Note

Subjects

Surgery

Abdominal and oropharyngeal surgery. A laparotomy was made, and the antrum was exposed. Gastric contents were removed via a small incision, and a gastric balloon (see below) was gently inserted in the corpus; the wound was ligated around the balloon's shaft. The stomach was then viewed via a surgical microscope over a range of balloon inflation volumes to ensure proper function and to evaluate possible tissue damage. If small tears (typically ~1 mm, perpendicular to the longitudinal axis) were observed, inflation was immediately halted, and greater inflations were not tested. The laparotomy was then closed, a tracheotomy was performed, and an oral drain tube was inserted along with retractable mouth sutures to expose the oral cavity for taste stimulation (34).

Cranial surgery. The rat was placed in a stereotaxic frame, and the skull was exposed and leveled. The left hemisphere was trephined 1.7 mm lateral and 0.4 mm anterior to lambda to expose ~15 mm² of brain surface, allowing full access to the PBN.

Apparatus and materials. The gastric balloon was made from a latex glove finger and was fastened via Teflon tape to one end of a Tygon tube (1/32 in. ID × 3/32 in. OD). The gastric balloon was connected to a 140-ml syringe, and a programmable infusion/withdrawal pump (KD Scientific 210P) was used to deliver gastric stimuli. Oral stimuli tested were dH₂O and a "taste mixture" containing 0.1 M NaCl, 0.3 M sucrose, 0.05 M citric acid, and 0.003 M QHCl. This mixture was used as the "probe" taste stimulus (see below) and was applied to the entire oral cavity.

General Procedure

Gastric stimulation. The balloon was inflated for 10 s, held fully inflated for 10 s, and then deflated for 10 s. Thus, as distension volume was increased, the inflation rate was more rapid. These inflation rates span most of the range of distension rates reported in rat recording studies (5, 10, 21, 42, 55, 65, 68), although they remain 12 times (3 ml) to 100 times (18 ml) faster than experienced during a meal (40; cf. 68). Gastric stimulation was preceded by 30 s of baseline recording and was followed by a 90-s pause. If a cell continued to respond into the pause period, the next stimulation was delayed until the response returned to baseline.

Recording. Glass and parylene-coated tungsten microelectrodes [impedance (Z) = 0.5-2.5 M] were used to record extracellular activity of single neurons in the PBN. Neural activity was amplified 10,000 times with filter settings at 300-10,000 Hz. Activity was monitored on a storage oscilloscope and was stored on magnetic tape and a computer hard drive. Stimulus markers were recorded on separate channels. Data were collected using the MII (Modular Instruments Systems) or Spike2 (Cambridge Electronic Design) systems.

Search procedure. The electrode was inclined 20° posterior to prevent transverse sinus rupture and was advanced with a piezoelectric microdrive. The general strategy was to first locate the gustatory-responsive region in the caudal-medial PBN and then probe in a rostrolateral direction with later tracks. Tracks were separated by 100-200 µm. Initial coordinates for targeting the gustatory PBN were +0.4 mm rostral and +1.7 mm lateral to lambda. Testing began ~5 mm below the cortical surface, at the transition to the pontine surface. The taste mixture was applied at 25-µm intervals, regardless of whether spontaneous activity was present. Water rinses were applied after each stimulation. Once gustatory responses were identified, gastric probe stimulations (6 ml) were also introduced and tested at each 25-µm interval. The intertest interval for gastric probes was ~180 s. Cells with higher thresholds were detected by testing unresponsive isolated cells at larger volumes. On all tracks, after the identification of the taste region, both gastric and gustatory probe stimuli were tested to identify responsive cells; again, probe stimuli were tested every 25 µm after encountering the pontine surface, regardless of whether spontaneous activity was evident. Cells with obvious rhythmic spontaneous activity, such as those with a respiratory rhythm in Kolliker-Fuse, were occasionally observed but rarely responded to taste or gastric stimuli. Responses to jaw stretch (indicating the mesencephalic trigeminal nucleus) were often tested to determine the ventral limit of the PBN. If a jaw stretch response was observed, the electrode was withdrawn, and a new track was begun.

Testing procedure. If a single neuron responsive to a gastric probe stimulus was detected, formal testing proceeded. All cells were tested for gustatory responsiveness (6), and, if the cell showed a response to taste stimulation, it was not used in this study. In addition, we usually tested neurons for responsiveness to nociceptive stimulation by pinching the cheek, ear, and contralateral hindlimb. Cells with nociceptive responses were also discarded. Three gastric "anchor" volumes (6, 12, and 18 ml, if permissible by stomach stretch standards) were tested at least two times each. If the cell was lost before completion of these tests, the data were discarded. If this protocol was completed, 3-, 9-, and 15-ml (if possible) volumes were then tested. When possible, four tests at each volume were conducted.

Histological reconstruction. A lesion (anodal current: 3 µA × 3 s) was made at the recording site, subjacent to it, or at the microdrive reading of the recording site after the track had been completed and the electrode was being removed. Concluding testing, the rat was given a lethal dose of anesthetic (ip) and was perfused with isotonic saline followed by 10% buffered formalin. To aid histological reconstruction, the brain was blocked in the recording plane by remounting the head in the stereotax (see Fig. 7, inset, and Ref. 41). Alternate sections were stained with Weil and cresyl violet to distinguish myelinated fibers and somatic components. Lesions made at the recording site allowed localization within morphologically distinct PBN subnuclei (28, 34). If the lesion was made below the recording site, or was made upon removal of the electrode, the precise location of the recorded cell in the dorsoventral axis could not be specified. Tracks and marker lesions, however, still provided useful information about cell location in the rostrocaudal and mediolateral axes when the lesions were relatively close to the recording site. Thus tracks were plotted when lesions were made within 250 µm of the recording site and included a clear length of track (>500 µm) on the same section.

Neurophysiological data analysis. Electrophysiological data were analyzed off-line using customized computer analysis programs. The present study used a quantitative approach to evaluate central responses to gastric distension. Initially, we used a criterion based on the summed response over the entire stimulation period, similar to criteria commonly applied in taste and other sensory neurophysiological paradigms (see, e.g., Refs. 22 and 34). Thus, to quantify distension responses, the spike total for a 10-s spontaneous period preceding gastric inflation was subtracted from the spike counts for each of the three 10-s distension periods (inflation, hold, and deflation). To be rated significant, the difference score in at least one of the three distension periods had to exceed 2 SD of the spontaneous period spike count. In addition, responses had to vary by at least one spike per second.

View larger version (21K): [in this window] [in a new window]   Fig. 1.   Example of data transform. In A-C, histograms represent neural firing; the solid line represents the onset and offset of gastric distension. Shown is a cell that is completely shut off during distension and then has an "off" response when the stomach is deflating (akin, but inverse to those noted in Refs. 21 and 55). A: raw data accumulated in 1-s bins. The spike total during the spontaneous period (10th to 20th s) was 15 spikes. The average spontaneous rate was therefore 1.5 spikes/s, and this was subtracted from each 1-sec bin to yield the net response (B). The 3-s moving average of this net response profile (C) was then used for analysis.

Statistical analyses. Most response measures were derived from the second-by-second analysis. Response magnitude was expressed as average frequency (net spike count during the response divided by duration). Average responses are reported as means ± SE. Cells were categorized on the basis of response direction (inhibition or excitation), duration, and threshold. Frequency analysis among these categories was conducted using the ² method. Categorical variables (e.g., response direction, volume) were used as grouping or repeated-measures factors in one- or two-way ANOVAs. The mean time course of gastric distension responses was analyzed by averaging appropriate groups of cells at particular volumes. For the mean time course, response offsets and onsets were defined using t-tests comparing the last second of the spontaneous period with successive seconds during and after inflation, respectively. The time to peak was defined as the time relative to stimulus onset during which the response attained 90% of its maximum value. Recording sites were plotted on one of four representative serial sections of PBN, with each section ~200 µm apart. Only a minority of the cells was definitively localized to subnuclei; therefore, these histological data were treated qualitatively. However, each identified recording site was categorized by the serial section it belonged in, and ² analysis was used to determine whether there was a preferential distribution of gastric cells along the rostrocaudal axis.

Control study: Blood pressure and cardiac effects of distension. The PBN contains neurons responsive to a variety of visceral stimuli. Therefore, we determined whether distension induced changes in heart rate or blood pressure. Because the recording preparation was quite involved, this was done in a second group of six rats fitted with a gastric balloon (see above) and jugular (PE-50 tubing; right side) and carotid (PE-50 tubing; left side-recording side) catheters. The carotid catheter was connected to a bridge force transducer (Columbus Instruments). EKG was simultaneously monitored in four rats using subcutaneous electrodes. EKG and pressure responses were analyzed with the Spike2 system. The same volumes (0, 3, 6, 9, 12, 15, and 18 ml, 3-4 trials each) as in the recording experiments were tested. In two rats, 18 ml were damaging; therefore, only one 18-ml test was performed as the last test in these rats. After testing, epinephrine (0.1 ml, 0.01%) and sterile isotonic saline (0.1 ml) were injected via the jugular catheter to confirm measurement sensitivity. Repeated-measures ANOVAs were used to analyze heart rate for the 30-s distension period and for the average arterial pressure (100 Hz sample rate) for the 30-s distension period and the 3-s period surrounding the peak pressure response during distension.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Fifty-two cells showed a significant response to gastric inflation with no response to gustatory or cutaneous pain tests. Ninety-two cells were gustatory or gustatory- and distension-responsive and were assigned to another study (6).

Response Direction

Threshold

View larger version (10K): [in this window] [in a new window]   Fig. 2.   Distribution of response thresholds for cells tested at all volume levels (n = 31 neurons).

Volume-Response Functions

Population effects. Sixteen neurons were tested at all six gastric distension volumes. The second-by-second analysis of responses from these cells indicated that volume affected both response rate and duration. Figure 3A shows that response duration increased considerably as a function of volume, particularly when volumes >9 ml were used [Fig. 3A; F(5,75) = 7.54, P < 0.0001]. Post hoc tests indicated that responses at the two largest volumes were significantly longer than those at 9 ml or less (P < 0.05). Figure 3B further shows a commensurate increase/decrease in the response rate. The change in rate for excitatory responses was significant [F(5,50) = 6.15, P < 0.0001] but only approached significance for the smaller inhibitory population [F(5,20) = 2.01, P = 0.12]. Post hoc comparisons support the interpretation that, for excitatory responses, the response rate systematically increased with volume (significant pair-wise comparisons: 3 ml < 12, 15, and 18 ml; 6 ml < 15 ml; 9 ml < 15 and 18 ml; P < 0.05).

View larger version (17K): [in this window] [in a new window]   Fig. 3.   Increasing distension volume systematically prolonged mean response duration (mean ± SE; A) and enhanced the mean spike rate (mean ± SE; B).

View larger version (31K): [in this window] [in a new window]   Fig. 4.   Average peristimulus time histogram (3-s moving average; mean + SE) for cells tested at anchor volumes of 6 (A), 12 (B), and 18 (C) ml, which had a significant excitatory response for 2 of the 3 volumes tested (n = 17).

Time course. Although the preceding analysis suggests that PBN responses to gastric distension generally outlasted distension, inspection of data from individual neurons revealed variation among cells. To analyze this variation, we defined three categories related to stimulus parameters. "Prolonged" responses began during distension but continued to respond for at least 10 s after deflation ended (e.g., Fig. 5C), "brief" responses lasted 10 s or less, possibly indicating a sensitivity to a particular phase of distension (e.g., Fig. 5A), and "sustained" responses fell between these boundaries (e.g., Fig. 5B). Because volume affected time course, initial analysis of these categories was confined to a single volume, 12 ml, and to excitatory responses. At this volume, brief responses were the least frequent (n = 6/27), small (1.1 ± 0.1 net spikes/s during distension) and highly variable, with latencies ranging throughout the distension period (range = 10-28 s). Thus, contrary to expectation, brief responses did not appear to reflect any particular phase of distension, e.g., inflation. In contrast, sustained (n = 8) and prolonged (n = 13) responses were more numerous, larger (sustained = 2.0 ± 0.2, prolonged = 3.3 ± 0.2 net spikes/s during distension), and homogenous. Mean time courses for these responses (Fig. 6) indicated that differences between them extended beyond duration. Prolonged responses had a long latency (10 s) and time to peak (22 s) and exhibited only a negligible decline during deflation (slope = 0.04 spikes/s). By contrast, sustained responses began (6 s) and peaked (10 s) earlier, showed a decline during deflation (slope = 0.17 spikes/s), and were no longer statistically different from baseline by the last second of deflation [t(7) = 1.79, P > 0.11]. Thus sustained responses more closely tracked distension dynamics.

View larger version (23K): [in this window] [in a new window]   Fig. 5.   Typical examples of the three time courses observed. A: a brief response was one that began during stimulation and lasted 4-10 s. B: a sustained response also began during stimulation, lasted 11 s or longer, and did not continue to respond >10 s after deflation offset. C: a prolonged response began during stimulation but continued to respond >10 s after deflation offset. Despite the average effect of volume on duration, all three time courses were observed across a wide range of volumes. For example, the brief response in A is from an 18-ml volume test, the sustained response in B is from a 12-ml volume test, and the prolonged response in C is from a 6-ml test.

View larger version (22K): [in this window] [in a new window]   Fig. 6.   Mean time course (3-s moving average; mean + SE) for sustained excitatory responses (n = 8; A) and prolonged excitatory responses (n = 13; B) at 12 ml.

Other features associated with time course. We extended the analysis by time course type to the entire population. Because some cells changed time course across volumes, cells were grouped based on the time course expressed over the majority of tests. Cells with mostly brief (n = 11) or sustained (n = 20) responses did not exhibit a preferential response direction when grouped, but the majority of cells with mostly prolonged responses tended to be excitatory (17/21). We also evaluated whether there was a relationship between threshold and time course (Table 2). Of the 31 cells in which a threshold was determined, 28 had responses at more than one volume. Thus time course type at threshold and at the largest volume tested for each cell was evaluated separately using a two-way ². As shown in Table 2, all three time courses were evident at threshold and did not vary according to whether a neuron was high or low threshold [²(2) = 0.46, NS]. At the largest volume (Table 2), more of these same 28 cells exhibited longer response types. However, this shift occurred in both threshold groups. For example, the number of prolonged responses doubled from threshold to the largest effective volume in both the low (from 3 to 6 cells)- and the high (from 2 to 4 cells)-threshold groups. Thus, at the largest volume, there was also no relationship between cell threshold and its characteristic time course [²(2) = 3.63, NS].

Topography

View larger version (24K): [in this window] [in a new window]   Fig. 7.   Schematic representation of the locations of gastric distension-responsive neurons in the parabrachial nucleus (PBN). Serial sections are arranged caudal ("A") to rostral ("D"), ~200 µm apart. Solid symbols: a lesion was made at recording site without moving electrode. Open symbols with vertical lines: visible tracks with accompanying marker lesions made along the track but not at the recording site. Open symbols were placed at the top of track without depicting actual loci to avoid confusion with the more precisely localized recording sites (solid symbols). The tip of each track is depicted to end at the site of the marker lesion closest to the recording site, although each track continued ventral from this point. Tracks with 2 open symbols indicate 2 cells were found on that track. Complex cells were usually inhibitory at small volumes and excitatory at large volumes. 7n, Exit of VIIth nerve; bc, brachium conjunctivum; cl, central lateral subnucleus; dl, dorsal lateral subnucleus; el, external lateral subnucleus; em, external medial subnucleus; i, internal lateral subnucleus; LC, locus ceruleus; Me5, mesencephalic trigeminal nucleus; s, superior lateral subnucleus; vl, ventral lateral subnucleus; Mo5, motor trigeminal nucleus; Pr5, primary sensory trigeminal nucleus; vsc, ventral spinocerebellar tract; cm, central medial.

More cells appeared to be inhibitory or complex at successively rostral levels; 80% of cells were excitatory at the most caudal level A, 57% at level B, 50% at level C, and only 25% at level D. However, ² analysis of the response direction and topographic level only approached significance [²(1) = 2.48, P = 0.11]. There was no significant relationship between the response time course (brief, sustained, or prolonged) and caudal vs. rostral locations [²(2) = 1.28, P = 0.53]. There was also no systematic relationship between cell location and other response characteristics, including threshold, response rate, and spontaneous rate.

Blood Pressure

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The present study utilized a quantitative, parametric approach to analyze PBN responses to gastric distension over a range of volumes. Various criteria have been used previously to distinguish central gastric distension responses (see, e.g., Refs. 5, 11, 25, 27, 74); however, most studies describe gastric responses qualitatively (see, e.g., Refs. 2, 7, 8, 17, 19, 35, 38, 39, 57, 67), and statistical analysis of response features (e.g., latency, duration, and rate) is rare. Overall, gastric responses differed from those in the NST and vagus in four respects. First, a subset of cells had thresholds higher than observed in the vagus. Second, the second-by-second analysis introduced in this study revealed a systematic relationship between distension volume and response duration. In some cells, the response noticeably outlasted distension, an effect not seen in the vagus or NST. Third, more PBN distension responses were inhibitory compared with those in the vagus. Finally, PBN gastric cells were found in caudal regions of PBN that anatomically overlap with taste cells. In NST and periphery, gastric and taste-responsive fibers/cells are distinctly separate.

Response Threshold

Stimulus Intensity Coding in PBN

Response Time Course

Previous descriptions of the time course of gastric distension responses have distinguished between "fast-adapting" or "phasic" vs. "slowly adapting" or "tonic" responses (4, 21, 35, 57, 67, 73). Fast-adapting vagal responses occur only during changes in distension (4, 35, 63, 64) and are thought to reflect the response of mechanoreceptors in the higher-tension antrum (4) or other accommodative processes (55). The brief responses in the present study were not similar to fast-adapting peripheral responses. They were not more prevalent at larger volumes, and they appeared throughout the distension period, during inflation, the "hold" period (e.g., Fig. 5A), or deflation in some cases. Furthermore, many cells with brief responses at smaller volumes exhibited sustained or prolonged responses at larger volumes, and, notably, these longer responses typically lacked phasic components. The lack of a phasic component in the responses we observed is likely the result of inflation dynamics. We inflated the stomach over a 10-s period. A recent study noted that step gastric inflation produced responses with a clear dynamic component, whereas more gradual distension at rates roughly comparable to those in the present study produced slowly rising responses that resemble those reported here (55).

Sustained responses generally tracked the temporal dynamics of the inflation-deflation curve, and similar responses have been noted at all levels of the central nervous system (CNS; see, e.g., Refs. 2, 11, 27, 35, 38, 57, 67, 73). Sustained responses appear to resemble slowly adapting responses in vagal afferents, which are thought to reflect activity of intramural mechanoreceptors in the gastric corpus (reservoir; see Refs. 4, 14, and 15).

The third time course was a prolonged response that outlasted distension by at least 10 s, continuing in many cases for up to 2 min. Although prolonged gastric distension responses have not been reported in NST (19, 57, 73) or vagal afferents with normal-range distension values (see, e.g., Refs. 14, 16, 32, 63, 64, 65), there is one recent report of prolonged vagal afferent responses elicited by damaging distension (40-100 mmHg fluid; see Ref. 55). By comparison, prolonged responses to noxious stimuli are commonly seen throughout the CNS (see, e.g., Refs. 12, 44, 46) and even in response to "normal" distension volumes in higher-order sites such as the PBN (67) and hypothalamus (38, 39). The function of prolonged gastric responses is unknown. In the pain system, responses that outlast the stimulus have been associated with the affective component of pain (23, 44). Although some cells exhibited prolonged responses at low volume (e.g., Fig. 5C), it is interesting to note that the longest responses usually occurred at the largest volumes. Indeed, Figure 3A suggests a notable "jump" in duration for responses at 12 ml and greater, volumes at which the stomach no longer had visible slack. Prolonged gastric responses could play a role in the PBN's contribution to satiation, possibly reflecting amplification of a feedback signal that serves to inhibit feeding as food accumulates in the stomach (20). It is interesting to note that preloads up to 5 ml fluid confined to the stomach have only a small impact on ingestion, but when the total gastric load reaches ~10 ml rats do stop eating (56). Moreover, c-fos expression in caudal NST after a meal is lost when rats are permitted to ingest only one-half of the meal, which suggests enhanced CNS activity over time with greater gastric fills (24).

Response Direction: Functional Implications

It is tempting to consider the possibility that the bidirectional nature of PBN distension responses reflects spinal (noxious) vs. vagal (digestive) influences. Vagal and splanchnic pathways have opposing influences on duodenal distension responses in the dorsal motor nucleus of the vagus (59), and splanchnic stimulation inhibits vagally elicited excitatory responses in both the NST and PBN (9, 72). However, these opposing responses need not reflect differential origins. PBN responses to noxious colorectal distension, which were conveyed by the splanchnic pathway, were both excitatory and inhibitory (12). Nevertheless, it seems unlikely that the inhibitory and excitatory responses in the present study were splanchnic in origin. PBN neurons with splanchnic input have high-threshold cutaneous receptive fields (12), and we not only rejected neurons with nociceptive cutaneous receptive fields from the present study, but most of our cells had gastric distension thresholds well below the noxious range.

A vagal origin for the present PBN distension responses is more consistent with proposed roles for the targets of PBN projections. Excitatory and inhibitory responses to gastric distension occur in the lateral and ventromedial hypothalamus, regions long associated with feeding (2, 38, 39, 42). Water deprivation diminishes distension responses in the lateral hypothalamic and preoptic regions, supporting an ingestive function (10). However, the principal PBN projection to the lateral and ventromedial areas is from cholecystokinin-containing cells in the superior lateral subnucleus (28, 62), an area where we sampled but did not detect distension responses. On the other hand, there are PBN projections to other feeding-related hypothalamic regions. The central lateral subnucleus projects to the paraventricular nucleus (reviewed in Ref. 62), where many inhibitory responses to distension are observed (70). A number of tracks with distension-responsive cells passed through this region (levels B and C in Fig. 7), and two neurons were definitively localized there. In addition, projections of the external lateral subnucleus, where several distension-responsive neurons were recorded, include the medial preoptic area, insular cortex, and amygdala (62), regions that include feeding among their varied functions. Finally, evidence that PBN gastric distension responses, including those that are inhibitory, have a vagal origin and are related to ingestive behavior is provided by a subpopulation of neurons we encountered but did not include in the present analysis (6). These cells responded in an excitatory manner to gustatory stimuli, but nearly all were inhibited by distension.

Specialization of Function Within PBN

The more surprising result of the present study is the observation of distension-responsive units in PBN areas better known to be taste responsive. The classic PBN gustatory-responsive region encompasses approximately the medial and lateral PBN areas that abut the brachium conjunctivum in the medial two-thirds of the caudal three levels represented in Fig. 7 (see Refs. 34 and 53). Four of the 12 lesions placed at the recording site fell clearly within these bounds (Fig. 7, A and C). In addition, several other tracks passed through this region. The present data therefore support the findings of Hermann and Rogers (37), who showed that cervical vagal stimulation activated taste-responsive neurons in PBN taste regions, suggesting functional gustatory-visceral overlap in the PBN. This notion is buttressed by two additional observations. First, in several instances, both taste and gastric-responsive neurons were located on the same track (6). Second, we found a subset of gastric-responsive units (n = 15) that also responded to taste stimulation, which were assigned to another study. These "coactivated" cells were located within the caudal taste-associated areas of the PBN (6).

The extension of the PBN visceral "boundary" suggested by the present data parallels recent studies that report taste responses in subnuclei of the external region (34, 41, 71). The overlap of gastric and taste functions in PBN underscores the interpretation of PBN as a locus for the integration and coordination of complex autonomic reflexes (12, 62) and suggests that simple dichotomies of the PBN into "visceral" vs. "taste" zones should be considered carefully.

Caveats: Blood Pressure and Other Indirect Effects

Perspectives