THE JOURNAL OF COMPARATIVE NEUROLOGY 303:355474 (1991)

Autonomic Responses andEf€erent Pathwap From the InsularCortex in the Rat YUKIHlKO YMUI, CHRISTOPHER D. BREDER, CLIFFORD B. SAPER, AM) DAVID F. C E C H E W Department of Pharmacology and Physiology W.Y., C.D.B., C.B.S., D.F.C.) and Department of Neurology (C.B.S.), University of Chicago and John P. Robarts Research Institute and Department of Physiology, University of Western Ontario (D.F.C.), London, Ontario, Canada N6A 5K8

ABSTRACT The anatomical distribution of autonomic, particularly cardiovascular, responses originating in the insular cortex was examined by using systematic electrical microstimulation. The localization of these responses to cell bodies in the insular cortex was demonstrated by using microinjection of the excitatory amino acid, D,L-homocysteic acid. The efferents from the cardiovascularresponsive sites were traced by iontophoretic injection of the anterograde axonal tracer Phaseoleus vulgaris leucoagglutinin (PHA-L). Two distinct patterns of cardiovascular response were elicited from the insular cortex: an increase in arterial pressure accompanied by tachycardia or a decrease in arterial pressure with bradycardia. The pressor responses were obtained by stimulation of the rostral half of the posterior insular cortex while depressor sites were located in the caudal part of the posterior insular area. Both types of site were primarily located in the dysgranular and agranular insular cortex. Gastric motility changes originated from a separate but adjacent region immediately rostral to the cardiovascular responsive sites in the anterior insular cortex. Tracing of efferents with PHA-L indicated a number of differences in connectivitybetween the pressor and depressor sites. Pressor sites had substantially more intense connections with other limbic regions including the infralimbic cortex, the amygdala, the bed nucleus of the stria terminalis and the medial dorsal and intralaminar nuclei of the thalamus. Alternatively, the depressor region of the insular cortex more heavily innervated sensory areas of the brain including layer I of the primary somatosensory cortex, a peripheral region of the sensory relay nuclei of the thalamus and the caudal spinal trigeminal nucleus. In addition, there were topographical differences in the projection to the lateral hypothalamic area, the primary site of autonomic outflow for these responses from the insular cortex. These differences in connectivity may provide the anatomic substrate for the specific cardiovascular responses and behaviors integrated in the insular cortex. Key words: microstimulation,gastric and cardiovascular responses

A variety of autonomic responses, including changes in blood pressure, heart rate, respiration, piloerection, pupillary dilatation, gastric motility, peristaltic activity, salivation, and adrenaline secretion can be elicited by electrical stimulation of the orbito-insular cortex, in a wide crosssection of mammalian species including rats, cats, dogs, monkeys, and humans (Delgado, '60; Hall et al., '77; Hoff et al., '63; Hoffman and Rasmussen, '53;Kaada, '51; Penfield and Faulk, '55; Ruggiero et al., '87; von Euler and Folkow, '58; Wall and Davis, '51). The results of many early investigations are difficult to interpret; as the sites of stimulation were not always precisely localized, large stimo 1991 WILEY-LISS, INC.

ulation voltages resulted in considerable current spread, and there were differences in anesthesia, species and protocol (for review see Cechetto and Saper, '90). A recent microstimulation study of the insular cortex of the rat more precisely localized arterial blood pressure responses to a region of the insular cortex (Ruggiero et al., '87) that correlates well with a visceral sensory region identified by Cechetto and Saper ('87a). Accepted September 11,1990. Dr. Yasui is now at The Department of Anatomy (1st Division), School of Medicine, Mie University, Tsu,Mie 514,Japan.

Y. YASUI ET AL.

356 Within the insular cortex, neurons responding to different visceral sensations are organized in a viscerotopic map. Gustatory neurons are located most rostrally in the dysgranular area, gastric mechanoreceptive neurons are found in the adjacent rostra1 granular insular field, and cardiorespiratory baro- and chemo-receptorresponsive neurons are located in the caudal granular area. It would be of considerable interest to explore the correspondence of cardiovascular and other autonomic responses to this map. Therefore, in the first series of experiments we systematically explored the insular cortex of the rat using electrical and chemical microstimulation and measured changes in

autonomic variables such as arterial blood pressure, heart rate, and gastric motility. The efferent pathways mediating the autonomic responses originating from the insular cortex are also of great interest. Previous neuroanatomical experiments using anterogradely and retrogradely transported tracer substances demonstrated direct connections with several subcortical sites involved in autonomic control (Saper, '82; Mesulam and Mufson, '82; Shipley and Sanders, '82; Shipley, '82; Ruggiero et al., '87). However, the differential role of these pathways in producing various autonomic responses has not been explored. In order to demonstrate the precise

Abbreviations IV

VII

x

XI1 ACC ac ACo AH AHi AI Amb APT BL BLa BLP BLv BM BNST

cc

Ce CG

CI CL CM CPU CP cu Cuf DI DLG DLL DT En ec ECu emPB f fr FStr G g GI GP Gr HDB IC ic icp IL Ins ILL I&M I0 IP La LC LD LDT LH lot LP IPB LRt

trochlear nucleus facial nucleus dorsal motor nucleus of vagus hypoglossal nucleus accumbens nucleus anterior commissure anterior cortical amygdaloid nucleus anterior hypothalamic area amygdalohippocampalarea agranular insular cortex ambiguous nucleus anterior pretectal nucleus basolateral amygdaloid nucleus anterior basolateral amygdaloid nucleus posterior basolateral amygdaloid nucleus ventral basolateral amygdaloid nucleus basomedial amygdaloid nucleus bed nucleus of stria terminalis corpus callosum central amygdaloid nucleus central grey claustrum centrolateral thalamic nucleus central medial thalamic nucleus caudate-putamen cerebral peduncle cuneate nucleus cuneiform nucleus dysgranular insular cortex dorsal lateral geniculate nucleus dorsal nucleus of lateral lemniscus dorsal tegmental nucleus endopiriform nucleus external capsule external cuneate nucleus external medial parabrachial subnucleus fornix fasciculus retroflexus fundus striati gelatinosus thalamic nucleus genu of facial nerve granular insular cortex globus pallidus gracile nucleus nucleus of horizontal limb of diagonal band inferior colliculus internal capsule inferior cerebellar peduncle infralimbic cortex insular cortex intermediate nucleus of lateral lemniscus intralaminar and midline nuclear group of the thalamus inferior olivary nucleus interpeduncular nucleus lateral amygdaloid nucleus locus coeruleus laterodorsal thalamic nucleus laterodorsal tegmental nucleus lateral hypothalamic area lateral olfactory tract lateral posterior thalamic nucleus lateral parabrachial nucleus lateral reticular nucleus

LS LV LVe MB

MD Me MG mcp met ml mot MoV MS mt mPB MVe NST ot

ox

PB PBG

PF PH Pir Pn Po PLCO PMco POmc PRh PV PVH PY Re RAmb RFpc rf RtT

sc

SCP

scpx Sm

so

SPF st stt ST t tt vc Vi vo VDB VL VLG VLL VM

VMH VPL VPM VPPC VTG 21

lateral septa1nucleus lateral ventricle lateral vestibular nucleus mamillary body mediodorsal thalamic nucleus medial amygdaloid nucleus medial geniculate nucleus middle cerebellar peduncle mesencephalic trigeminal tract medial lemniscus motor root of trigeminal nerve motor trigeminal nucleus medial septum mamillothalamic tract medial parabrachial nucleus medial vestibular nucleus nucleus of solitary tract optic tract optic cbiasm parabrachial nucleus parabigeminal nucleus parafascicular thalamic nucleus prepositus hypoglossal nucleus piriform cortex pontine nucleus posterior thalamic nuclear group posterolateral cortical amygdaloid posteromedial cortical amygdaloid nucleus magnocellular preoptic nucleus perirhinal cortex paraventricular thalamic nucleus paraventricular hypothalamic nucleus pyramis reuniens thalamic nucleus retroambiguous nucleus parvocellular reticular formation rhinal fissure reticulotegmental nucleus of the pons superior colliculus superior cerebellar peduncle superior cerebellar peduncle deucussation somatosensory cortex superior olivary nucleus subparafascicular thalamic nucleus solitary tract stria terminalis subthalamic nucleus spinal trigeminal tract tenia tecta caudal spinal trigeminal nucleus interpolar spinal trigeminal nucleus oral spinal trigeminal nucleus vertical limb of diagonal band ventrolateral thalamic nucleus ventral lateral geniculate nucleus ventral nucleus of lateral lemniscus ventromedial thalamic nucleus ventromedial hypothalamic nucleus ventral posterolateral thalamic nucleus ventral posteromedial thalamic nucleus parvocellular division of ventroposterior thalamic nucleus ventral tegmental nucleus zona incerta

AUTONOMIC SITES AND PATHWAYS IN THE INSULAR CORTEX

357

efferent projections of sites shown to produce cardiovascu- began 0.5 mm above the insular cortex and was given every lar responses in the microstimulation study, a second series 0.2 mm until the electrode was 0.5 mm below the ventral of experiments was done in which injections of the antero- aspect of the insular cortex. The stimulus consisted of 20 to gradely transported tracer Phaseoleus vulgaris leucoagglu- 30 second trains of positive 1 ms pulses at 50 Hz and 200 tinin (PHA-L) were made at physiologically characterized 4.The indifferent electrode was an alligator clip attached sites. Some preliminary results of this investigation have to the exposed scalp muscles. At the site of the peak been reported in an abstract form only (Yasui et al., '90). response in each track, current threshold and the most effective stimulating frequency were determined. In nine animals, at the most responsive site in the track, the METHODS micropipette electrode was removed from the brain, the saline was ejected under pressure and the pipette filled with Microstimulationofthe insularcortex a 1M solution of DL-H acid and lowered again to the same These experiments were done in 44 male Sprague-Dawley depth. Injections of DL-H (50 to 100 nl) were made to or Wistar rats anesthetized with either a-chloralose (60 determine the autonomic response to activation of cell mgkg, i.pJ or urethane (1.5 g k g i.p.1. The chloralose was bodies in the region. The bottom of each electrode track was supplemented every 1to 2 hours with 20 mgkg injections marked either by the iontophoretic injection of 2% Poni.v. The femoral artery was cannulated for the measuretamine blue (10 &Anegative current for 10 minutes) or by ment of arterial blood pressure using a Statham P23D making a lesion (1mA for 20 seconds). pressure transducer and continuously monitored on a At the end of each experiment the animals were perfused Grass model 7 polygraph. Mean arterial pressure was with 100 ml of 0.9% saline followed by 250 ml of a 10% monitored by filtering the pulsatile arterial pressure with a formalin solution. The brains were removed, 50 pm sec0.5 Hz high-frequency filter. The heart rate was determined tions were cut on a freezing microtome, mounted on glass from the pulse pressure using a Grass 7P44 tachograph and slides, stained with a 0.125% thionin solution, and covercontinuously monitored on the polygraph. The femoral vein slipped. The sections were examined under the light microwas cannulated for the administration of drugs. A cannula scope for the location of the stimulation tracks in the was inserted through a tracheostomy for the monitoring of insular cortex and the stimulation sites were drawn with respiration in some animals or for artificial ventilation with the aid of a camera lucida drawing attachment. a Harvard small animal respirator in others. Respiratory rate and volume was monitored with a Fleisch 0000 pneumotachograph and a Grass PT5 volumetric pressure transCardiovascularresponsesand ducer. The integrated tidal volume was measured by a PHA-L injections Grass 7PlOF integrator. In 24 rats gastric motility was also In 20 male Sprague-Dawleyrats anesthetized with a-chlomonitored by the endogastric insertion of a Foley catheter (14 Fr, 5 ml) inflated with 2 ml saline and connected to a ralose microstimulation was used to locate a cardiovascular Statham P23D pressure transducer. A rectal thermoprobe responsive site in the insular cortex. The anterograde was inserted and a Yellow Springs temperature controller axonal transport substance PHA-L was then injected at was used to maintain the body temperature of the animal at these sites to trace their efferent connections. In these animals a femoral catheter was inserted for monitoring 37 & 0.1"C. arterial blood pressure and deriving heart rate as described The animals were placed in a David Kopf stereotaxic frame and the cortex dorsal to the insular area was exposed for the first set of experiments. An endotracheal tube was through small burr holes. Single glass capillary micropi- inserted to monitor respiration using the Fleisch pneumotpettes containing either a 3 M NaCl solution or 1 M achograph as described above. A Yellow Springs temperaD,L-homocysteic acid (DL-H) solution were inserted into ture controller was used to maintain rectal temperature at the region of the insular cortex. There were three groups of 37 0.1"C. The animal was placed into a stereotaxic frame and the animals. The first group of rats (n = 26) was anesthetized with a-chloralose and spontaneously breathing while moni- cortex overlying the insula exposed using a dental drill to toring respiration. This group was used to determine the make a burr hole of approximately 2 mm diameter. The location of cardiovascular, gastric, and respiratory sites in microelectrode was inserted into the insular cortex using the insular cortex. The second group (n = 7) was anesthe- the coordinates from the previous series of experiments tized with a-chloralose and paralysed with pancuronium that had resulted in the peak changes in arterial blood bromide (2 mgkg i.v.) and ventilated. The duration of the pressure and heart rate. The brain was stimulated, as paralysis was approximately 20 minutes and the injections before, using 10 to 20 second trains of 1ms pulses at 50 Hz of pancuronium bromide had to be continuously repeated. and 200 pA. The electrode was lowered from 0.5 mm above When the animal recovered from paralysis it was routinely the insular cortex and stimulation was applied every 0.2 tested to ensure adequate levels of anesthesia before each mm. When an arterial pressure response was obtained, the repeat injection of the pancuronium bromide. This second electrode was lowered in 0.1 mm increments until the peak group of animals was used to determine if the cardiovascu- of the response determined. At this point the electrode was lar responses were indirectly the result of changes in withdrawn vertically from the brain, the 3 M NaCl solution respiration. The third group of animals (n = 11)was anes- was ejected under pressure and the micropipette filled with thetized with urethane and was not paralysed. This group a 2.5% PHA-L solution in 0.1 M phosphate buffer at pH 7.4. of animals was used to determine whether the responses to The micropipette was lowered to the site where the peak of stimulation in the insular cortex were specifically related to the response had been obtained and the accuracy of the placement was confirmed by quickly repeating the electrical the use of a-chloralose. The insular cortex was systematically stimulated along stimulation to reproduce the cardiovascular response. one to five electrode tracks per animal, positioned 0.3 mm PHA-L was iontophoresed with a 6 FA positive current apart in the rostral-caudal direction, extending from 3.0 applied with a 7 seconds on, 7 seconds off cycle for 15 to 30 mm anterior to 1.0 mm posterior to bregma. Stimulation minutes. Following the injection of the PHA-L the wounds

*

358

were sutured closed, the endotracheal tube removed, and the rat was allowed to recover from anesthesia. Following 10 to 14 days survival the animal was reanesthetized with chloral hydrate (350 m@g, i.p.1 and perfused with 100 ml 0.9% saline followed by 4% paraformaldehyde and 0.2%glutaraldehyde in 0.1 M phosphate buffer, pH 7.4, followed by 10% sucrose in 0.1 M phosphate buffer. The brains were removed and 50 pm sections were cut on a freezing microtome and a 1-in-2 series was processed for visualization of the transported PHA-L. Two methods of PHA-L staining were used in different rats. In the first method the sections were incubated for 60 minutes at room temperature in phosphate-buffered saline (PBS), pH 7.4, containing 0.25% triton and 3% normal swine serum (PBS-swine).The sections were then incubated with PBSswine containing goat anti-PHA-L (Vector; 1:lOOO) for 24 hours at 4°C. The sections were then rinsed in PBS, incubated for 60 minutes in PBS-swine containing swine anti-goat IgG-peroxidase (Tago; 1:50),rinsed in PBS, then incubated in phosphate buffer containing diaminobenzidine (0.5 mg/ml) and H,O, (0.01%), rinsed in phosphate buffer, mounted on glass slides, and coverslipped with Histoclad. In the second method, using the ABC kit (Vector Labs), the sections were rinsed in PBS, incubated for 60 minutes at room temperature in PBS containing 0.25%triton and 3% normal rabbit serum (PBS-rabbit), rinsed in PBS, incubated in PBS-rabbit containing goat anti-PHA-L (1: 1000) for 24 hours at 4"C, rinsed in PBS, incubated for 3 hours in PBS-rabbit containing biotinylated rabbit antigoat IgG, rinsed in PBS, incubated for 60 minutes in PBS containing avidin-biotin-horseradish peroxidase, rinsed in PBS, incubated in phosphate buffer containing diaminobenzidine (0.5 mg/ml) and H,O, (0.01%),rinsed in phosphate buffer, mounted on glass slides, and coverslipped with Histoclad. The alternate series of adjacent sectionswere mounted on glass slides and stained with thionin (0.125%),coverslipped, and used to add cytoarchitectural details to the camera lucida drawings of the PHA-L-labeledprojections.

RESULTS Autonomicresp~nses from the insularcortex Cardiovascular responses. There were two discretely localized, reproducible, patterns of cardiovascular response elicited from the insular cortex. The first was a pressor response accompanied by an increase in heart rate (Fig. 1A). The pressor response was invariably sustained throughout the duration of the stimulus train. The example in Figure 1A demonstrates this response in a track in the posterior insular cortex at a level immediately rostral to the crossing of the anterior commissure. The peak of the response was located in cortical layer V in the dysgranular insular cortex. The second type of response was a decrease in arterial blood pressure and heart rate (Fig. 1B). The peak of this response was evoked in the posterior dysgranular insular cortex at a rostral-caudal level that coincided with the crossing of the anterior commissure (Fig. 1B). A minimum stimulation current of 25 FA was necessary to elicit a change in arterial pressure ( 2 10 mmHg) for both the pressor and depressor responses. In addition, both pressor and depressor responses were elicited by excitation of cell bodies only with injections of 50 to 100 nl of 1 M D,L-homocysteicacid at the same sites from which similar responses were elicited with electrical stimulation (Fig. 2). Smaller doses of D,L-homocysteicacid were unable to elicit

Y. YASUI ET AL. responses. This is likely due to the reasonably large anteriorposterior distribution of the pressor and depressor neurons in the insular cortex. Our experience has indicated that much smaller injections (5-20 nl) are acceptable elsewhere in the central nervous system where the cell populations are more compact. These chemicalresponses develop slowly, beginning approximately 20-40 seconds after the injection and continuing from 1to several minutes thereafter. As illustrated in Figure 3, the majority of the peak pressor responses were localized to layer V of the posterior dysgranular and agranular insular cortex with a few sites also located in the ventral part of the granular insular cortex. The pressor sites had a relatively restricted rostralcaudal distribution located at the level of the rostral edge of the joining of the anterior commissure. The depressor responses were mainly located at sites in the posterior insular cortex caudal to those giving pressor responses. The majority of depressor sites were located in layer V of the dysgranular and ventral granular insular cortex with a few sites extending into the agranular region. There were occasional pressor responses located adjacent to the external capsule at the rostral-caudal level of the depressor responses. As will be discussed later, these likely represent stimulation of the efferent pathway from the pressor region. Fifty-six tracks lying medial, lateral, anterior, or posterior to these sites were unable to elicit cardiovascular responses. Figure 3 also indicates that both the pressor/tachycardic and depressorlbradycardic responses were evoked in all three groups (spontaneously breathing chloralose anesthetized rats, paralysed and artificially ventilated chloralose anesthetized rats, and rats anesthetized with urethane). Thus, the responses were not indirectly the result of respiratory changes or specific to a single anesthetic. Gastric responses. In 24 animals gastric motility was also monitored during stimulation of the insular cortex t o determine whether other autonomic responses were topographically related to the cardiovascular sites. Regular rhythmic contractions of the stomach were obtained with the endogastric balloon catheter in spontaneously breathing and paralysed, artificially ventilated rats. Both excitation and inhibition of gastric tone were obtained with stimulation current 5 200 FA. Examples of the typical changes in gastric motility are indicated in Figure 4. Stimulation of gastric inhibitory areas in the insular cortex resulted in a cessation of spontaneous contractions that at the most responsive sites usually outlasted the duration of the stimulus train (Fig. 4A). Sites in the insular cortex that were excitatory for gastric motility usually resulted in both an increase in the strength and duration of the contractions. This combination of an increase in both strength and duration often resulted in a sustained contraction for up to 30 seconds following a 20 second stimulus train as illustrated in Figure 4B. The distribution of sites affecting gastric motility is shown in Figure 4D. It can be seen that the gastric inhibitory areas were located primarily in the rostral granular insular cortex while the excitatory region was immediately caudal and ventral in the rostral dysgranular and ventral granular insular cortex. Although the gastromotor sites were more anterior than the cardiovascular areas, there was some overlap of these responses in the posterior insular cortex. In fact, it was possible to obtain both changes in arterial blood pressure and gastric motility at a few sites. Figure 4C illustrates a stimulation site from

AUTONOMIC SITES AND PATHWAYS IN THE INSULAR CORTEX

359

Fig. 1. Examples of cardiovascular responses elicited by microstimulation ofthe insular cortex. A Illustrates a site from which an increase in arterial pressure accompanied by tachycardia was obtained from layer V of the dysgranular insular cortex immediately rostral to the joining of the anterior commissure. B: Illustrates a depressor response and bradycardia obtained from layer VI of the ventral granular

posterior insular cortex at a level coinciding with the joining of the anterior commissure. In each case the first column shows the pulsatile the second the mean arterial pressure (MAP) arterial pressure (AP), and the third the heart rate (HR). Stimulation consisted of 1ms pulses delivered at 50 Hz for the time indicated by the bar under each column. i-vi, cortical layers.

which both arterial pressor and gastric contraction responses were obtained.

Efferentsfrom pressor sites. The response and pattern of PHA-L labeling from an injection into a pressor site is shown in Figure 5 (R724). This injection is one of eight into pressor sites in the insular cortex. The anterograde labeling resulting from this injection is representative of that obtained from other injections into pressor sites. Wherever the projections from this injection differ from other injections is indicated in the following description. This injection is located in the center of the pressor region at the rostral edge of the joining of the anterior commissure in the agranular insular cortex and the ventral part of the dysgranular insular cortex. Within the insular cortex, labeled fibers ran rostrally to innervate the dysgranular and

Efferent pathwaysfrom cardiovascularsites in the insularcortex Iontophoretic injections of PHA-L into sites in the insular cortex from which changes in arterial blood pressure and heart rate were obtained were made in 18 rats. PHA-L injections were made in eight rats at the site of a pressor response and in ten animals at a depressor site. The pressor and depressor sites gave rise to distinct patterns of efferent projections. These are illustrated by reference to two representative experiments.

C

200

-

DLH

L

Fig. 2. Illustrates examples of cardiovascular responses elicited by microstimulation of the insular cortex (B,E) or microinjections of D,L-homocysteicacid (DL-H) to activate cell bodies (C,F). A illustrates the stimulation site for a pressor response to electrical (B)and to chemical (C) stimulation in the rostral part of the posterior insular cortex. D shows the stimulation site for a depressor response to

2001

A

1

DLH

-

I

1-

-I

J

-1 -

electrical (E) and to chemical (F)microstimulation in the caudal part of the posterior insular cortex. Electrical stimulation in each case consisted of 1 ms pulses delivered at 50 Hz for 20 seconds. Chemical stimulation consisted of 100 nl of 1.0 M DL-H. The response to chemical stimulation has a more gradual onset and longer duration than that to electrical stimulation.

F

D

20s

0

82

W

AUTONOMIC SITES AND PATHWAYS IN THE INSULAR CORTEX

361

+ 0.5

0.0

-0.5

Fig. 3. Line drawings of coronal sections through the rat brain indicating the distribution of the peak pressor and depressor sites from each track through the insular cortex. Only those tracks from which a change in arterial pressure t 10 mmHg was obtained with stimulation intensities 5200 (LA are included on the brains sections. Fifty-six nonresponsive tracks surrounding these active sites are not shown.

Circles indicate responses in animals anesthetized with chloralose, triangles indicate responses in animals anesthetized with chloralose and paralysed and ventilated, and squares indicate responses in rats that were anesthetized with urethane. Filled symbols indicate the peak pressor response for a single tract, while open symbols indicate peak depressor responses.

some agranular regions of the anterior insular cortex, mainly terminating in layers I and V, with less dense innervation of layers I1 and 111. A few fibers were seen in layer I of the somatosensory cortex. Other fibers entered the corpus callosum to innervate the homologous area of the contralateral insular cortex. Two distinct efferent pathways were seen exiting the insular cortex (Fig. 5D). One of these ran in a ventral direction, entering the piriform cortex and the olfactory tubercle, giving rise to numerous terminal processes. This

pathway continued ventrally and then rostromedially through the nucleus accumbens (Fig. 5B) to terminate densely in the deep layers of the infralimbic cortex (Figs. 5B and 6A). A few fibers were also seen in deep layer I running parallel to the medial surface of the infralimbic cortex. Other cortical projections included fibers running caudally from the insular cortex into the ventral perirhinal cortex and the piriform cortex (Fig. 5E). The second pathway leaving the insular cortex exited medially providing a dense termination in the fundus of the

Y. YASUI ET AL.

362

A

20s B

1

2

3

4

5

6

I

2 3

4 5 6

C

Fig. 4. Gastric motility responses and the sites from which they originate in the insular cortex. A: A n example of a decrease in gastric motility evoked by electrical microstimulation. The peak of the response as indicated by the triangle in the brain section is located in layer 5 in the rostral anterior part of the dysgranular insular cortex. It can be seen that the inhibition of gastric motility begins 3 to 5 seconds after the start of the stimulation and outlasts the stimulation by approximately 20 s at the peak of the response shown in panel 4.B: An example of an increase in gastric motility elicited by electrical stimulation. The peak of the response (triangle) is located in layer V of the caudal anterior part of the dysgranular insular cortex. The stimulation evokes an increase in both the strength and duration of the gastric contrac-

tions resulting in a sustained contraction. C : Illustrates an example of a site in the caudal anterior insular cortex from which both an increase in gastric motility and a pressor response was obtained. D: A series of line drawings of coronal sections through the forebrain of the rat indicating the peak of the responsive sites in each tract from which changes in gastric motility were obtained with stimulation current 5 200 WA. Inhibition of gastric motility was obtained primarily from the granular area of the most rostral region of the anterior insular cortex. Increases in gastric motility were obtained primarily from the granular area of the caudal half of the anterior insular cortex. In A, B, and C the traces from top to bottom represent pulsatile arterial pressure (AP), mean arterial pressure (MAP), heart rate (HR), and gastric motility (GM).

striatum (Fig. 5C,D). Some fibers continued medially through the ventral part of the globus pallidus to provide innervation of the dorsal and ventral lateral divisions of the bed nucleus of the stria terminalis (Fig. 5C,D and 6B). Not all pressor sites provided as dense an innervation of the dorsal divisions of the bed nucleus of the stria terminalis although moderate to dense innervation was a consistent finding. Other fibers continued in the midline crossing of

the anterior commissure to terminate contralaterally in the bed nucleus of the stria terminalis and the fundus of the striatum. Within the amygdala there was dense terminal innervation throughout the central nucleus (Figs. 5E and 6C,D). Substantial terminal fields were also found in the ventral and posterior basolateral nuclei and basomedial and anterior cortical nuclei of the amygdala (Fig. 5E,F).

AUTONOMIC SITES AND PATHWAYS IN THE INSULAR CORTEX

363

Labeled fibers also were traced medially through the ansa depressor sites are noted. The injection site was just caudal peduncularis into the internal capsule and into the dienceph- to the level of the decussation of the anterior commissure, alon. Some fibers entered the thalamus, innervating the in layer V of the dysgranular insular cortex (Figs. 7 and 8D). paraventricular nucleus, medial parts of the mediodorsal, The cardiovascular response and efferent projections from a and the parafascicular nuclei, as well as the central grey depressor site are shown in Figure 8 (R840).Fibers leaving matter (Fig. 5E,F,G). A few fibers were found bilaterally the injection site by a dorsal rostral pathway resulted in with an ipsilateral predominance in the visceral relay terminal labeling in the adjacent parts of the insular and nucleus, the parvocellular ventroposterior nucleus of the somatosensory cortices (Fig. 8B,C,D). Within the insular thalamus (Fig. 5F),and in the subparafascicular and lateral cortex there was also a heavy projection mainly to layers I, subparafascicular nuclei (Fig. 5F). Injections into other 11-111, and V of the anterior dysgranular and granular areas pressor sites exhibited a more dense innervation of the (Fig. 8B,C).Contralaterally, there was a homologous,heavily midline intralaminar nuclei than that illustrated in Fig- labeled terminal plexus in layers I, 11, and 111. In the ure 5 . somatosensory cortex dense terminal innervation was seen Other labeled fibers entered the ipsilateral lateral hypo- rostrally in the superficial part of layer I and in layers V and thalamic area (Figs. 5F,G and 6E), giving rise to two VI (Fig. 8B,C). A similar but less intense labeling pattern terminal fields. One terminal plexus was located in the most was seen in the contralateral somatosensory cortex. The lateral part of the lateral hypothalamic area immediately dense terminal labeling in layer I of the somatosensory dorsal to the medial edge of the internal capsule and medial cortex was not consistently observed following injections to the subthalamic nucleus (Fig. 5F,G). A second terminal into depressor sites. In one case there was little or no field was found in the ventrolateral quadrant of the tuberal anterograde labeling in layer I. However, as most injections lateral hypothalamic area (Fig. 5F). A few fibers also ran did result in labeling in this region it was considered part of medially into the lateral parvocellular paraventricular nu- the projection pattern. cleus of the hypothalamus (Fig. 5E). A few labeled fibers took a ventral pathway from the Descending fibers entered the brain stem through the depressor site, innervating the deep layers of the olfactory caudal continuation of three pathways: from the paraven- tubercule and piriform cortex, as well as the ventral part of tricular thalamus into the central grey matter, from the the nucleus accumbens (Fig. 8B). Some turned rostromediinternal capsule into the pyramidal tract, and from the ally, entering the infralimbic cortex where they gave rise to medial forebrain bundle into the central tegmental fields a moderate terminal plexus in the deep part of layer I as (Fig. 5H). Three main sites of termination were identified well as sparse arborization into the deeper layers (Figs. 8B in the brain stem: the periaqueductal grey matter, the and 9A). parabrachial nucleus, and the nucleus of the solitary tract. A caudally projecting pathway provided innervation to The fibers in the central grey matter were observed ipsilat- the perirhinal and piriform cortices and amygdala (Fig. erally in the ventral part of the periaqueductal grey with 8E,F). In the perirhinal cortex dense terminal labeling was only occasional contralateral labeled fibers (Fig. 5H). concentrated along the ventral bank of the rhinal fissure in Within the parabrachial nucleus there was substantial layer I with moderate terminal labeling in layer I1 and a few terminal labeling in the ipsilateral ventral lateral, central fibers in layers 111 and V (Figs. 8E,F, and 9C). Labeling in lateral, medial, and external medial subnuclei (Figs. 51,J the contralateral perirhinal cortex was also observed. Modand 6F; see also Moga et al., '90). Particularly dense erate numbers of fibers were seen in the posterior piriform innervation was seen in a restricted region of the most cortex (Fig. 8F). Within the amygdala a moderate terminal projection was medial part of the medial parabrachial subnucleus (Fig. 55). In the contralateral parabrachial nucleus much lighter observed to the dorsal lateral region of the central nucleus labeling was found in the same subnuclei. A few fibers were (Fig. 8E). Many terminals were located laterally in the also observed in the rostral pole of the ipsilateral locus posterior basolateral nucleus and a few fibers were observed in the basomedial nucleus (Fig. 8E,F). The projections to coeruleus (Fig. 55). In the posterior part of the nucleus of the solitary tract the amygdala were bilateral with an ipsilateral predomithe terminal labeling was located bilaterally in the commis- nance. The medial pathway from the depressor site resulted in s u r d and ventrolateral subdivisions (Fig. 5M). More rostrally there was anterograde labeling bilaterally in the heavy labeling in the most ventral part of the caudatemedial and rostral subdivisions of the nucleus of the putamen (Figs. 8C,D, and 9B) bilaterally, but with an solitary tract (Fig. 5L). A few fibers were also seen in the ipsilateral predominance. Moderate terminal labeling was observed in the fundus of the striatum (Fig. 8C,D). The parvocellular reticular formation (Fig. 5K,L,M) Efferents from depressor sites. A typical depressor dorsal and ventral lateral parts of the bed nucleus of the site response and projection pattern is shown in Figures 7 stria terminalis received only a very light efferent projecand 8. This injection is one of ten into depressor sites in the tion. There was intense innervation of the ventral basal thalaposterior insular cortex. The efferent projections from this site represent those observed from other injections into mus (Fig. 8F,G). This projection was predominantly ipsilatdepressor sites. In the following description any projections era1and resulted in dense collectionsof fibers and terminals that are not a consistent feature of the efferents from located in the perimeter of the visceral parvocellular ventro-

Fig. 5 (appears on pages 364-365). A series of drawings illustrating labelled fibers after a PHA-L injection into a site in the insular cortex at which electrical microstimulation elicited an increase in arterial pressure. A: The cardiovascular response to stimulation is shown at several depths of penetration through the insular cortex. The first column

represents the pulsatile arterial pressure trace (AP)and the second represents the heart rate (HR).The star indicates the site of iontophoretic injection of PHA-L at the peak of the response. B-M: Line drawings of coronal sections of the rat brain showing the PHA-L injection site (D) and the anterograde transport.

AP

r

* trrrrrmr L

200 r

Figure 5

HA

E FE-

AUTONOMIC SITES AND PATHWAYS IN THE INSULAR CORTEX

G

Figure 5 continued

365

366

Fig. 6. Photomicrographs of PHA-L-labeled fibers and terminals resulting from an injection (R724)into a pressor site. A Ventral half of the infralimbic cortex showing terminals primarily in the deep layers. Scale = 200 Fm. B: Terminals in the lateral part of the bed nucleus of the stria terminalis. Scale = 100 pm. C: Dense labeling of terminals in

Y. YASUI ET AL.

the central nucleus of the amygdala. Scale = 200 pm. D: Higher magnification of the field shown in C . Scale = 100 p.m. E: Fibers and terminals in the lateral hypothalamic area medial to the subthalamic nucleus. Scale = 100 pm. F: Moderate numbers of fibers and terminals in theventral lateral and central parabrachial subnuclei. Scale = 100 +m.

AUTONOMIC SITES AND PATHWAYS IN THE INSULAR,CORTEX

367

Fig. 7. Photomicrograph of the PHA-L injection (R840)into a depressor site in the insular cortex. Neurons transporting the PHA-L have their cell bodies located primarily in layers 111-V in the dysgranular insular cortex. Scale = 1mm.

posterior thalamic nucleus and extending dorsally and caudally into the ventral part of the posterior thalamic nucleus (Figs. 8F,G, and 9D).Some fibers were seen in the most medial part of the ventroposterior medial nucleus, the ventral part of the parafascicular nucleus and the central grey matter. Labeled fibers entered the lateral hypothalamic area from the ansa peduncularis but terminal arborization followinginjections in depressor sites was restricted to only a few fibers located along the medial edge of the internal capsule and medial to the subthalamic nucleus (Fig. 8F,G). Small numbers of fibers continued caudally into the brain stem through the ventrolateral part of the periaqueductal grey matter, the central tegmental fields and the pyramidal tract (Fig. 8H). The anterograde labeling following other injections into the depressor sites did not result in any apparent label in the periaqueductal grey matter. Terminal labeling in the ipsilateral parabrachial nucleus was diffusely distributed throughout the medial and ventral lateral subnuclei (Figs. 8I,J, and 9E). A few fibers were also seen in the central lateral and occasionally in the external medial subnuclei as well as contralaterally in the medial and the ventral lateral subnuclei. Occasional fibers with terminal boutons were observed in the rostral part of the locus coeruleus (Figs. &J and 9E). Terminal labeling in the medulla was found mainly in three sites. The projection to the nucleus of the solitary tract terminated primarily in the contralateral ventral lateral division (Fig. 8L,M). An efferent projection was also seen to terminate in the dorsal part of the caudal spinal trigeminal nucleus with some fiber labeling extending laterally into the marginal zone (Figs. 8M and 9F). Finally, some fibers were consistently observed in the parvocellular reticular formation medial to the spinal trigeminal nucleus. These projections were bilateral with contralateral predominance.

DISCUSSION Our results demonstrate that specific patterns of cardiovascular response are highly localized to certain regions of the posterior insular cortex. Pressorltachycardic responses are elicited from a zone at the same antero-posterior level as the midline joining of the anterior commissure. Immediately caudal to the pressor site is a region from which depressoribradycardic responses can be obtained.

tio on of visceral motor responses A number of previous investigators have demonstrated changes in arterial pressure following stimulation of the insular cortex. In many of those studies large stimulating current were used and there was an absence of accurate histological localization (for review see Cechetto and Saper, '90). Only one recent investigation combined electrical and chemical microstimulation with rigorous histological localization (Ruggiero et al., '87). These investigators demonstrated pressor responses in the posterior insular cortex at a location similar to that seen in our experiments. Ruggiero and colleagues did not describe a depressor region in the insular cortex. However, in the course of their experiments they were able to elicit decreases in arterial pressure (personal communication). Comparison of the location of the cardiovascular responses in this study with the cardiorespiratory sensory area we previously identified (Cechetto and Saper, '87a) reveals a striking correspondence. Arterial baroreceptor and chemoreceptor activation result in changes in neuronal activity in the posterior part of the granular insular cortex primarily in layers I11 and IV (Cechetto and Saper, '87a). The cardiovascular responses to electrical or chemical stimulation are obtained at the same anteroposterior level of the insular cortex, but slightly more ventrally in the dysgranular and agranular regions. It is unlikely that this difference is due to small inaccuracies in localization, as we used similar methods in both studies, and injected tracers in both types of experiments directly

AP

HR

E

t 150r

Fig. 8. A series of drawings to show labeled efferents after an injection of PHA-L into a depressor site in the caudal part of the posterior insular cortex. A.Illustrates the response evoked by microstimulation. The peak of the response in the descending track is indicated by

a star. B-M: Line drawings of coronal sections through the rat brain indicating the PHA-L injection site (D) and the labeled fibers and terminals resulting from the injection.

AUTONOMIC SITES AND PATHWAYS IN THE INSULAR CORTEX

Figure 8 F-M

369

370

Fig. 9. Photomicrographs of the PHA-L fiber labeling and terminals resulting from an injection into a depressor site in the rat brain (R840). A: Superficial layer of the ventral half of the infralimbic cortex showing terminal plexus in the deep part of layer I. Scale = 100 pm. B Terminal labeling in the fundus of the striatum dorsal lateral to the anterior commissure. Scale = 100 pm. C: Dense terminal labeling in the rhinal fissure in primarily layer I in the perirhinal cortex. Scale = 200 pm. D Collections of dense terminal labeling surrounding the visceral parvo-

Y. YASUI ET AL.

cellular ventroposterior nucleus and extending dorsally into the posterior thalamic nucleus. Scale = 200 km. E: Fibers and terminals in the medial parabrachial subnucleus ventral to the superior cerebellum peduncle and immediately lateral to the mesencephalic trigeminal tract. Scale = 100 pm. F: Fibers and terminals in the caudal spinal trigeminal nucleus immediately adjacent to the spinal trigeminal tract. Scale = 100 Km.

AUTONOMIC SITES AND PATHWAYS IN THE INSULAR CORTEX

371

Fig. 10. Summary diagram of the efferent projections from pressor and depressor sites in the insular cortex. The efferents from pressor sites originate from the solid circle while the projections from the de-

pressor sites begin at the open circle. Thick solid lines point to areas that are most heavily innervated while the thin solid lines and broken lines indicate brain areas receiving progressively less innervation.

into responsive sites. Rather, the granular and dysgranular visceral sensory areas may have local cortico-corticalconnections with the agranular field, which has been proposed as a site of limbic-autonomicconvergence (Saper, '82). Similarly,gastromotor responses were located more anteriorly in the insular cortex at a site corresponding to the antero-posterior level of the gastric sensory area. However, neurons whose activity was altered in response to activation of gastric mechanoreceptors and gustatory receptors (Cechetto and Saper, '87a) were located just dorsal to our gastromotor sites. In the present experiments we made small injections of the anterograde tracer PHA-L at cardiovascular active sites to determine the efferent connections from these areas. The efferents from the pressor and depressor sites (summarized in Fig. 10) had many terminal fields in common. More importantly, certain differences emerged that may explain the opposite cardiovascular responses elicited from these two areas. Five major sites of differential innervation were identified.

Both pressor and depressor, as well as gastromotor responses, have been reported during electrical stimulation of the infralimbic cortex (Kaada, '51; Burns and Wyss, '86; Hurley-Guis et al., '86). It has been suggested that the infralimbic area may serve as a visceral motor cortex (Hurley-Guis et al., '86) corresponding to the insular visceral sensory area. Although the interconnectivity of the insular and infralimbic areas would support such a view, our combined functional-anatomicaldata suggest a more complex relationship. The heaviest projection to the deep layers of the infralimbic cortex arises from the insular pressor area, while most responses to electrical stimulation of the infralimbic area are depressor (Burns and Wyss, '86; Hurley and Saper, unpublished observations). Hence, we would predict that if the insular projection to the infralimbic area plays a role in blood pressure control, it would most likely contact inhibitory interneurons in the infralimbic area. Conversely, the depressor sites in the insular cortex provide very little innervation of the infralimbic area, and thus most likely use other pathways to elicit cardiovascular changes.

hfralimbic cortex The insular pressor area provides a dense projection to the infralimbic cortex, particularly to its deeper layers, while the projection from the depressor sites was less intense and mainly involved layer I. The deep layers of the infralimbic cortex, in turn, project to numerous central autonomic structures, including the lateral hypothalamic area, the parabrachial nucleus, the nucleus of the solitary tract, the ventrolateral medulla, and even the interomediolateral cell column of the spinal cord (Tucker and Saper, '85; Room et al., '85; Hurley-Gius et al., '86; Hurley et al. '91).

Bed nucleus of the stria t e The insular pressor area selectively projects to the lateral part of the bed nucleus of the stria terminalis. This cell group provides descending projections to several central autonomic cell sites, including the lateral hypothalamic area, the parabrachial nucleus, and the nucleus of the solitary tract (Berk and Finkelstein, '81; Holstege et al., '85; Luiten and Room, '80; Ross et al., '81; Moga et al., '89, '90). There have been relatively few studies of the role of the bed nucleus of the stria terminalis in cardiovascular control (for review see Moga et al., '89), but the stimulation

Y. YASUI ET AL.

372 studies that have included this area have reported mainly depressor responses (Kabat et al., '35; Hilton and Spyer, '71; Faiers et al., '75). Consequently, it seems likely that if the insular area projection to the bed nucleus of the stria terminalis plays any role in the pressor response, it may be largely inhibitory.

Lateralhypothalamus The pressor and depressor sites in the insular cortex project to distinct terminal fields in the lateral hypothalamic area. Fibers from the pressor area terminate in two zones at tuberal and premammillary levels, one located dorsally along the medial edge of the internal capsule and subthalamic nucleus, and the other situated ventrolaterally, adjacent to the tuberomammillary nucleus. The insular depressor area mainly innervates a portion of the lateral hypothalamic area located dorsomedial to the internal capsule and subthalamic nucleus, at premammillary levels. The projection to the lateral hypothalamic area from the insular cortex has not previously been characterized in detail (Saper, '82). While the lateral hypothalamic area gives rise to descending projections to the parabrachial nucleus, nucleus of the solitary tract, ventrolateral medulla, and interomediolateral cell column in the spinal cord (Saper et al., '76; Hoysoya, '80; Luiten et al., '861, only the dorsal field innervated by the insular pressor area contributes to these pathways. On the other hand, the remaining two insular terminal fields may contribute to autonomic control by means of local connections within the hypothalamus. Despite numerous electrical stimulation studies showing both pressor and depressor responses from the lateral hypothalamus (e.g., Kabat et al., '35; Enoch and Kerr, '67), surprisingly little is known about the role of neuronal cell bodies in this area in control of arterial blood pressure. A recent chemical stimulation study has shown that most of the responses to electrical stimulation were probably due to activating the many fiber pathways that traverse the lateral hypothalamus (Spencer and Loewy, '88). In fact, these investigators were only able to identify depressor responses when stimulating the lateral hypothalamic area with glutamate. The most responsive site was similar in location to the terminal field from the insular depressor area, suggesting that this projection may play an important role in producing the depressor response. On the other hand, the doses of glutamate used in these studies were quite large (approximately 7,500 pmoles), which might have caused depolarization block rather than excitatory discharge of highly sensitive neurons. Hence, the possibility exists that pressor responses might be obtained from the lateral hypothalamic area using other excitatory amino acids or different dosages. Cechetto and Chen ('90a) examined the responses elicited by insular stimulation during synaptic blockade, induced by local cobalt injection into the lateral hypothalamus. They found that the cobalt injection blocked sympathetic nerve discharge during insular cortex stimulation. Thus, it seems likely that the sympatheticresponses to insular stimulation requires synaptic relay in the lateral hypothalamic area.

Parabrachialnucleus The projection from the insular cortex to the parabrachial nucleus primarily innervates the medial, external medial, ventral lateral and central lateral subnuclei, and the "waist" of the superior cerebellar peduncle (Moga et al.,

'90); this study). Moga et al. ('90) found that rostral insular sites project more heavily to the ventral lateral subnucleus, while more caudal insular projections (at the level at which cardiovascular responses were obtained) project more heavily to the medial and external medial subnuclei and the "waist" area. None of these sites, however, have descending projections (Saper and Loewy, '80; Fulwiler and Saper, '84; Herbert et al., 'go), and electrical microstimulation does not produce blood pressure responses at these sites in the rat (Holmes et al., '87). The caudal insular injections reported by Moga and colleagues ('90) also projected sparsely to the Kolliker-Fuse nucleus, which provides descending projections to the nucleus of the solitary tract, the rostral ventrolateral medulla, and the sympathetic preganglionic cell column in the spinal cord. However, electrical stimulation in the Kolliker-Fuse has not been associated with cardiovascular responses (Holmes et al., '87). On the other hand, electrical or chemical microstimulation in the region of the external lateral subnucleus causes robust pressor responses, and the area just lateral to the external lateral nucleus projects to the pressor area in the rostral ventrolateral medulla (Herbert et al., '90). These observations suggest that the parabrachial nucleus is not a likely relay for the insular influence over sites in the medulla and spinal cord that are involved in cardiovascular control. As might be expected, therefore, cobalt injections into the parabrachial nucleus have not blocked sympathetic responses evoked by stimulation of the insular cortex (Cechetto and Chen, '90).

Nucleus of the solitarytract The insular projection to the nucleus of the solitary tract is organized in a topographic fashion (Saper, '82; Ruggiero et al., '87; this study). The rostral insular cortex, including the granular taste area, projects only to the rostral (gustatory) part of the nucleus of the solitary tract, while the rostral part of the posterior insular cortex, including the pressor area, projects to the medial part of the nucleus of the solitary tract, including its cardiovascular component (see Herbert et al., '90 for review). In contrast, the slightly more caudally located insular depressor area projected mainly to the ventrolateral (respiratory) subdivision of the nucleus of the solitary tract. The portion of the nucleus of the solitary tract receiving the projection from the insular pressor area also is the site of termination for arterial baro- and chemo-receptor afferents (see Spyer, '81; Calaresu et al., '84 for reviews). However, electrical or chemical stimulation in this area produces mainly depressor responses (Talman and Reis, '80). Also, synaptic blockade with cobalt injection a t this site is not able to block insular sympathetic responses (Cechetto and Chen, '90a). Hence, it seems unlikely that the projection to the nucleus of the solitary tract is responsible for the insular pressor response, although it may contribute to the blockade of the reflex bradycardia that increases in blood pressure would ordinarily produce. A sixth possible site for mediating the insular pressor response is the rostral ventrolateral medulla. Although this area received only a small number of PHA-L-labeled terminals in our experiments, it is possible that the injections may have missed cells in the insular cortex with more extensive projections to the ventrolateral medulla. There is a large body of evidence that neurons in the rostral ventrolateral medulla provide direct excitatory inputs to

AUTONOMIC SITES AND PATHWAYS IN THE INSULAR CORTEX sympathetic vasomotor neurons, mediating both many tonic and phasic mechanisms for blood pressure control (Ross et al., '84; Sun and Guyenet, '87). Injections of cobalt into the rostral ventrolateral medulla can block the insular sympathoexcitatory response (Cechetto and Chen, '90b). Some of the projections from insular cardiovascular sites, such as to the parabrachial nucleus and the nucleus of the solitary tract, may be involved in modulation of visceral sensation. Both the nucleus of the solitary tract and the parabrachial nucleus are obligate relays in the visceral sensory pathway to the insular cortex (see Cechetto and Saper, '87a,b). The insular depressor area also projects into layer I of the neighboring primary somatosensory cortex, as well as innervating the peripheral zone of the ventroposterior medial and lateral thalamic nuclei and marginal zone of the caudal spinal trigeminal nucleus. The latter two sites are believed to be particularly important for nociception (Knifki and Vahle-hinz, '87; Vahle-Hinz et al., '87; Yokota et al., '85; Sessle et al., '86; Dallel et al., '88). Hence, the insular cortex may be capable of modulating nociceptive and other somatosensory afferents, as well as visceral sensation. Although, these pathways may be relatively unimportant for autonomic control in the anesthetized animal, they may assume greater prominence in the waking state during natural behaviors.

ACKNOWLEDGMENTS This research was supported by USPHS grants HL36872 (DFC)and NS22835 (CBS)and by a grant from the Ontario Heart and Stroke Foundation. D.F.C. is the recipient of a Canadian Heart Foundation Scholarship.

Berk, M.L., and J.A. Finkelstein (1981) M e r e n t projections to the preoptic area and hypothalamic regions in the rat brain. Neuroscience 6t16011624. Burns, S.M., and J.M. Wyss (1986) The involvement of anterior cingulate cortex in blood pressure control. Brain Res. 34Ot71-77. Calaresu, F.R., J. Ciriello, M.M. Caverson, D.F. Cechetto, and T.L. Krukoff (1984) Functional neuroanatomy of central pathways controlling the circulation. In T.A. Kotchen, C.P. Guthrie (eds): In Hypertension and The Brain. Mount Kisco, Ny: Futura Publications, pp. 3-21. Cechetto, D.F., and C.B. Saper (1987a) Evidence for a viscerotopic sensory representation in the cortex and thalamus in the rat. J. Comp. Neurol. 2622745. Cechetto, D.F., and C.B. Saper (198%) Organization of visceral sensory thalamus in the rat. Neurosci. Abstr. 13t728. Cechetto, D.F., and C.B. Saper (1990) Role of the cerebral cortex in autonomic function. In A.D. Loewy, K.M. Spyer (eds): In The Autonomic Nervous System: Central Regulation of Autonomic Function. Oxford: Oxford University Press, pp. 208-223. Cechetto, D.F., and S.J. Chen (199Oa) Subcortical sites mediating sympathetic responses from the insular cortex in the rat. Am. J. Physiol. (Regulatory Integrative Comp. Physiol.) 258: R245-255. Cechetto, D.F., and S.J. Chen (1990b) The ventral lateral medulla mediates insular sympathetic responses. Neurosci. Abstr. 15:595. Dallel, R., P. Raboisson, P. Auroy, and A. Woda (1988)The rostral part of the trigeminal sensory complex is involved in orofacial nociception. Brain Res. 448t7. Delgado, J.M.R. (1960) Circulatory effects of cortical stimulation. Physiol. Rev. 40:146-171. Enoch, D.M., and F.W.L. Kerr (1967) Hypothalamic vasopressor and vesicopressor pathways. I. Functional studies. Arch. Neurol. 16.290306. Faiers, A.A., F.R. Calaresu, and G.J. Mogenson (1975) Pathway mediating hypotension elicited by stimulation of the amygdala in the rat. Am. J. Physiol. 228: 1358-1366.

373

Fulwiler, C.E., and C.B. Saper (1984) Subnuclear organization of the efferent connections of the parabrachial nucleus in the rat. Brain Res. Rev. 7:229-259. Hall, R.E., R.B. Livingston, and C.M. Bloor (1977) Orbital cortical influences on cardiovascular dynamics and myocardial structure in conscious monkeys. J. Neurosurg. 46:638-647. Herbert, H., M.M. Moga, and C.B. Saper (1990) Connections of the parabrachid nucleus with the nucleus of the solitary tract and medullary reticular formation in the rat. J. Comp. Neurol. 293~540-580. Hilton, S.M., and K.M. Spyer (1971) Participation of the anterior hypothalamus in the baroreceptor reflex. J. Physiol. 218t271-293. Hoff, E.C., J.F. Kell, Jr., and M.N. Carroll, Jr. (1963) Effects of cortical stimulation and lesions on cardiovascular function. Physiol. Rev. 43r68114. Hoffman, B.L., and T. Rasmussen (1953) Stimulation studies of insular cortex of Macaca mulatta. J. Neurophysiol. 16,343-351. Holmes, H.R., Herbert, H. and C.B. Saper (1987) Parabrachial subnuclei in respiratory control. Neurosci Abstr. 13,1586. Holstege, G., L. Meiners, and K. Tan (1985)Projections of the bed nucleus of the stria terminalis to the mesencephalon, pons, and medulla oblongata in the cat. Exp. Brain Res. 58t379-391. Hoysoya, Y. (1980) The distribution of spinal projection neurons in the hypothalamus of the rat, studied with the HRP method. Exp. Brain Res. 40:79-87. Hurley-Gius, K.M., D.F. Cechetto, and C.B. Saper (1986) Spinal connections of the infralimbic autonomic cortex. Neurosci. Abstr. 12t538. Hurley, K.M., H. Herbert, M.M. Moga, and C.B. Saper (1990) Connections of the infralimbic cortex of the rat: I. Efferent projections. (in preparation). Kaada, B.R. (1951) A study of responses from the limbic, subcallosal, orbito-insular piriform and temporal cortex, hippocampus-fornix and amygdala. In Somato-motor, Autonomic and Electrocorticographic Responses to Electrical Stimulation of "Rhinencephalic" and Other Structures in Primates, Cat and Dog. A.W. Broggers Boktrykkeri, Oslo. Acta Physiol. Scand. 24:Suppl. 83, pp. 1-285. Kabat, H., H.W. Magoun, and S.W. Ranson (1935) Electrical stimulation in the forebrain and midbrain. Arch. Neurol. Psychiat. 34t931-955. K n i E i , K.D., and C. Vahle-Hinz (1987)The periphery of the cat's ventroposteromedial nucleus (WM,): Nociceptive neurones. In Besson, J.M., G. Guilbaud, and M. Peschanski (eds): Thalamus and Pain. New York: Elsevier Science Publ. pp. 245-257. Luiten, P.G.M., and P. Room (1980)Interrelations between lateral, dorsomedial and ventromedial hypothalamic nuclei in the rat: An HRP study. Brain Res. 190:321-332. Luiten, P.G.M., G.J. ter Horst, and A.B. Steffens (1986) The hypothalamus, intrinsic corrections and outflow pathways to the endocrine system in relation to the control of feeding and metabolism. Progress Neurobiol. 28t1-54. Mesulam, M.-M., and E.J. Mufson (1982) Insula of the old world monkey. 111: Efferent cortical output and comments on function. J. Comp. 212t38-52. Moga, M.M., C.B. Saper, and T.S. Grey (1989) The bed nucleus of the stria terminalis: Cytoarchitecture, immunohistochemistry and projections to the parabrachial nucleus in the rat. J. Comp. Neurol. 283t315-322. Moga, M.M., H. Herbert, K.M. Hurley, Y. Yasui, T.S. Grey, and C.B. Saper (1990) Organization of cortical, basal forebrain and hypothalamic aff'erents to the parabrachial nucleus in the rat. J. Comp. Neurol. 295:624661. Penfield, W., and M.E. Faulk, J r . (1955) The insula: Further observation on its function. Brain 78:445470. Room, P., F.T. Russchen, H.J. Groenewegen, and A.H.M. Lohman (1985) Efferent connections of the prelimbic (area 32) and the infralimbic (area 25) cortices: An anterograde tracing study in the cat. J. Comp. Neurol. 242:40-55. Ross, C.A., D.A. Ruggiero, and D.J. Reis (1981) Afferent projections to cardiovascular portions of the nucleus of the tractus solitarius in the rat. Brain Res. 223t402-408. Ross, C.A., D.A. Ruggiero, T.H. Joh, and D.H. Park (1984) Rostral ventrolatera1 medulla: Selective projections to the thoracic autonomic cell column from the region containing C1 adrenaline neurons. J. Comp. Neurol. 228:16&185. Ruggiero, D.A., S. Mraovitch, A.R. Granata, and M. Anwar (1987) A role of insular cortex in cardiovascular function. J. Comp. Neurol. 257t189. Saper, C.B., A.D. Loewy, L.W. Swanson, and W.M. Cowan (1976) Direct hypothalamo-autonomicconnections. Brain Res. I I7:305-312.

374 Saper, C.B., and A.D. Loewy (1980) Efferent connections of the parabrachial nucleus in the rat. Brain Res. 197.991-317. Saper, C.B. (1982) Convergence of autonomic and limbic connections in the insular cortex of the rat. J. Comp. Neurol. 21Ot163-173. Sessle, B.J., J.W. Hu, N. Amano, and G. Zhong (1986) Convergence of cutaneous, tooth pulp, visceral, neck and muscle afferents onto nociceptive and non-nociceptive neurones in trigeminal subnucleus caudalis medullary dorsal horn and its implications for referred pain. Pain 27:219. Shipley, M.T. (1982) Insular cortex projection to the nucleus of the solitary tract and brain stem visceromotor regions in the mouse. Brain Res. Bull. 8:139-148. Shipley, M.T., and M.S. Sanders (1982) Special senses are really special: Evidence for a reciprocal, bilateral pathway between insular cortex and nucleus parabrachialis. Brain Res. Bull. 8t493-501. Spencer, S.E.,W.B. Sawyer, and A.D. Loewy (1988) L-glutamate stimulation of the zona incerta in the rat decreases heart rate and blood pressure. Brain Res. 458:72. Spyer, K.M. (1981) Neural organization and control of the baroreceptor reflex. Rev. Physiol. Biochem. Pharmacol. 8823-124. Sun, M.K., and P.G. Guyenet (1987) Arterial baroreceptor and vagal inputs to sympathoexcitatory neurons in rat medulla. Amer. J. Physiol. 252; R699.

Y. YASUI ET AL. Talman, W.T., M.H. Perrone, and D.J. Reis (1980) Evidence for L-glutamate as the neurotransmitter of baroreceptor afferent nerve fibers. Science 209:813-815. Tucker, D.C. and C.B. Saper (1985) Specificity of spinal projections from hypothalamic and brainstem areas which innervate sympathetic preganglionic neurons. Brain Res. 360:159-164. Vahle-Hinz, C., I. Freund, and K.D. Kniffki (1987) Nociceptive neurons in the ventral periphery of the cat thalamic ventroposteromedial nucleus. In Schmidt, F.S., H.G. Schaible, and C. Vahle-Hinz (eds): Fine Merent Nerve Fibers and Pain. pp. 4 4 1 4 5 . Von Euler, U S., and B. Folkow (1958) The effect of stimulation of autonomic areas in the cerebral cortex upon the adrenaline and noradrenaline secretion from the adrenal gland in the cat. Acta Physiol Scand. 42313-320. Wall, P.D., and G.D. Davis (1951) Three cerebral cortical systems affecting autonomic function. J. Neurophysiol. 14:507-517. Yasui, Y., C.B. Saper, and D.F. Cechetto (1990) Cardiovascular area of the insular cortex and its efferent projections in the rat. Neurosci. Res. (suppl.) (in press). Yokota, T., N. Koyama, and N. Matsumoto (1985) Somatotopic distribution of trigeminal nociceptive neurons in ventrobasal complex of cat thalamus. J. Neurophysiol. 53r1387.