aversive gustatory stimulation

Brain (1998), 121, 11431154

Aversive gustatory stimulation activates limbiccircuits in humansDavid H. Zald,1,2 Joel T. Lee,1 Kevin W. Fluegel1 and Jose V. Pardo1,2

1Cognitive Neuroimaging Unit, Psychiatry Service, Correspondence to: Jose V. Pardo, Cognitive NeuroimagingMinneapolis Veterans Affairs Medical Center, 2Division of Unit (11P), VAMC, One Veterans Drive, Minneapolis,Neuroscience Research, Department of Psychiatry, MN 55417, USAUniversity of Minnesota, Minneapolis, Minnesota, USA

SummaryAnimal studies implicate the amygdala and its connectionsin the recognition of aversive stimuli. A recent PETstudy demonstrated that the human amygdala and leftorbitofrontal cortex show substantial increases in regionalcerebral blood flow (rCBF) during exposure to aversiveodourants. To examine if aversive gustatory stimulisimilarly activate these regions, nine healthy women tastedan aversive saline solution, pure water and chocolatewhile rCBF was measured with PET. The aversive salinecondition, when contrasted with the water condition,

Keywords: amygdala; medial thalamus; orbitofrontal; PET; taste

Abbreviations: BA 5 Brodmann area; rCBF 5 regional cerebral blood flow

IntroductionElectrophysiological and lesion studies in a number ofmammalian species indicate that the amygdala plays a criticalrole in evaluating the affective significance of stimuli inmany sensory modalities (LeDoux, 1987; Davis, 1992).Recently, we demonstrated that humans show large bilateralincreases in amygdala activity during exposure to highlyaversive odourants (Zald and Pardo, 1997). This responseexceeded that seen with pleasant or neutral odourants.Although the amygdala receives particularly direct projectionsfrom the olfactory system, it seems unlikely from the animalliterature that the response of the human amygdala to aversivestimuli is confined to the olfactory domain.

Like the olfactory system, the gustatory system possessesrelatively direct projections to the amygdala. Substantialgustatory information reaches the lateral, basal and centralnuclei of the amygdala directly from the insular primarygustatory region (Turner, 1980). Additional gustatoryafferents may also reach the amygdala from the caudolateralorbitofrontal cortex and gustatory responsive nuclei in thebrainstem and thalamus (Norgren, 1976; Beckstead et al.,1980; Turner and Herkenham, 1981; Yasui et al., 1984;Amaral et al., 1992). Consistent with these multipleprojections, several discrete nuclei within the amygdala Oxford University Press 1998

increased activity in the right amygdala, left anteriororbitofrontal cortex, medial thalamus, pregenual anddorsal anterior cingulate, and the right hippocampus. Theright amygdala, left orbitofrontal cortex and pregenualcingulate remained significantly activated when salinewas compared with chocolate. The present results indicatethat the amygdala and orbitofrontal cortex respond toaversive stimuli in both the olfactory and gustatorymodalities, and highlight the role of the pregenualcingulate in negative emotional processing.

possess cells with gustatory responses (Azuma et al., 1984;Scott et al., 1993; Yasoshima et al., 1995). Electro-physiological and lesion data indicate that the amygdala isprobably not necessary for basic sensory perception anddiscrimination of taste (Aggleton, 1992; Scott et al., 1993)However, lesions of the amygdala in animals critically impairtheir ability to directly associate gustatory stimuli with stimulifrom other sensory modalities (Kennie and Nagel, 1973;Arthur, 1975; Mikulka et al., 1977; Gaffan and Murray,1990). Moreover, bilateral amygdalectomies in non-humanprimates produce marked alterations in feeding behaviourincluding the ingestion of substances that normal animalsfind unpalatable (Kluver and Bucy, 1937). Based upon thisbehavioural evidence and our previous findings with olfactorystimuli, we hypothesized that exposure to aversive gustatorystimuli would significantly increase activity within the humanamygdala.

We similarly hypothesized that aversive gustatory stimuliwould increase activity within the orbitofrontal cortex. Theorbitofrontal cortex receives significant projections from boththe primary gustatory cortex and the amygdala (Amaral et al.,1992; Baylis et al., 1995; Carmichael and Price, 1995; Zaldand Kim, 1996a), placing it in a position to code or process

1144 D. H. Zald et al.

information about the motivational value of gustatory stimuli.Lesions of the orbitofrontal cortex in non-human primatesproduce impairments that resemble those observed followingamygdala lesions, including deficits in associating gustatoryreinforcers with neutral stimuli and KluverBucy-likeabnormalities in eating behaviour (Butter et al., 1969; Ursinet al., 1969; Gaffan and Murray, 1990; Baylis and Gaffan,1991). Furthermore, orbitofrontal cortex activity (particularlyin the left hemisphere) has been observed in a number ofPET studies involving negative emotional inductions (Pardoet al., 1993; Rauch et al, 1994, 1995; Fischer et al.,1996), including exposure to aversive odourants (Zald andPardo, 1997).

To examine whether an unconditioned aversive gustatorystimulus would activate the amygdala and orbitofrontal cortexin humans, we measured regional cerebral blood flow (rCBF)with PET while subjects tasted an aversive saline solution.Attempting to taste water served as the primary controlcondition, and as an additional gustatory control condition,subjects were scanned while tasting chocolate. This allowedan examination of whether regions activated by saline aresimilarly activated by other gustatory stimuli, and whethersaline, perceived as highly unpleasant, differentially activatelimbic regions relative to a highly pleasant gustatory stimulus.

MethodsSubjectsTen right-handed female subjects (mean age 37 years, range2562 years) were studied with PET while tasting eithersaline solution, pure water or chocolate. Only women werestudied in order to maximize the likelihood of intenseemotional experiences (Shields, 1991), and for consistencywith our previous study of aversive olfaction which utilizedonly women (Zald and Pardo, 1997). All subjects gaveinformed consent approved by the VAMC Human SubjectsCommittee and Radioactive Drug Research Committee. Onesubject was excluded due to motion during the salinecondition (see below) leaving a total of nine subjects withvalid data for all conditions.

Experimental designThe stimulus in the aversive saline condition consisted of a5% solution of iodized NaCl dissolved in deionized distilledwater at room temperature. The stimulus for the pure watercondition consisted of deionized distilled water at roomtemperature. The chocolate stimulus consisted of 3 g ofHersheys Symphony Chocolate (Hersheys, Hershey, Pa.,USA). In the saline and pure water conditions, 3 ml of thefluid was injected ~5 s before the start of scan acquisitioninto the mouth through a small plastic cannula held betweenthe teeth. An additional 2 ml of fluid was slowly injectedinto the subjects mouth during the course of the next 45 s.Prior to fluid injection, subjects received the following

instructions: You are about to receive a liquid in your mouth.Close your eyes, and see if you can taste anything. Whenyou feel the liquid in your mouth, swish it around a coupleof times and then allow your tongue to rest. If there is toomuch fluid in your mouth, go ahead and swallow. In thechocolate condition, the chocolate was placed on the tip ofthe subjects tongue 5 s before the start of scan acquisition,and subjects were instructed, Close your eyes, and see ifyou can taste anything. When you feel something on yourtongue, close your mouth, swish it around a couple of times,and then allow your tongue to rest. After each condition,subjects rated the stimulus for pleasantnessunpleasantnesson an 11-point visual analogue scale with anchors at 0(extremely unpleasant), 5 (neutral) and 10 (extremelypleasant), and intensity, also on an 11-point visual analoguescale with anchors at 0 (undetectable) and 10 (extremelyintense). Subjects were informed that they would receive anunpleasant taste during one scan condition, but were blindto the scan number for that condition, the identity of thestimulus and to its degree of unpleasantness. Subjects alsodid not know what other gustatory stimuli they would receivefor the other scans. However, when subjects completed thesaline condition before the water and/or chocolate condition,they were told that they would not receive another unpleasanttaste. Five of the subjects with valid scans received thechocolate condition before the saline condition, while fourof the subjects received the saline first. Because of the oddnumber of subjects providing valid scans, it was not possibleto fully counterbalance the conditions. However, post hocanalysis provided no evidence of an order effect.

PET imaging and analysisThe rCBF was estimated from normalized (1000 counts)tissue radioactivity using a Siemens ECAT 953B camera(Knoxville, Tenn., USA) with septae retracted; a slow-bolusinjection of H215O (an initial dose of 814 MBq or 22 mCiinfused at a constant rate over 30 s; Silbersweig et al., 1993)was followed by a 90 s scan acquisition beginning uponradiotracer arrival into the brain (10 min inter-scan intervals).Images were initially reconstructed with a 3D reconstructionalgorithm using a 0.5 cycles-per-pixel Hanning filter (Kinahanand Rogers, 1989) and attenuation correction using a 2Dtransmission scan. Measured coincidences were corrected forrandom events and electronic dead time, but not for decayor scatter. Following reconstruction, images were visuallyinspected for motion artefacts using ANALYZE (BRU, MayoFoundation, Rochester, Minn., USA). Movement betweenscans was empirically quantified through examination of shiftand rotation parameters from automated co-registration files(Minoshima et al., 1992). One subject was excluded becauseshe moved her head substantially during the course of thesaline condition. This problem did not occur with the othersubjects, but initial examination of motion artefacts andregistration data indicated that some of the subjects movedbetween scans. Such movement may cause artefacts in ventral

Limbic responses to saline 1145

regions of the brain as a result of misalignment of emissionscans with transmission scans (Huang et al., 1979). Toeliminate this potential source of error, we implemented arealignment procedure similar to that described by Anderssonet al. (1995); before attenuation correction, each emissionscan was coregistered to the subjects first emission scan,using Automated Image Registration (Woods et al., 1992).Normalization for global activity, co-registration within eachstudy session, placement of the intercommissural line fromimage fiducials, and linear warping of each subjects scansto a reference stereotactic atlas (Talairach and Tournoux,1988) were subsequently accomplished with softwareprovided by Minoshima and co-workers (Minsohima et al.,1992, 1993, 1994). Images were blurred with a 3-pixel 3DGaussian filter producing a final image resolution of ~10 mmfull-width at half-maximum and a mapping resolution of ,2mm (Fox et al., 1986).

Pixel-wise subtractions were performed to determineactivations occurring in the saline condition relative to thewater condition. Statistical analysis employed the globalvariance of all intracerebral pixels (Worsley et al., 1992). Athreshold of P , 0.0005 (equivalent to a Z-score of 3.3) wasselected for the analysis of the contrast between salineand pure water conditions. This threshold is slightly moreconservative than the P , 0.001 cutoff frequently used inpixel-wise analyses of PET images. The more stringent P ,0.0005 threshold was based on a bootstrapping study of 60subjects scanned twice while resting with their eyes closed.This bootstrapping analysis indicated that, on average, theimaging techniques used in this study produce approximatelyone false positive focus (emerging due to chance) with asample size of nine subjects. Because the goal of the salineversus chocolate comparison was to elucidate the nature ofactivity in these areas rather than to perform anotherexploratory analysis, a threshold of P , 0.001 was adoptedfor areas already identified in the other comparison. However,regions which failed to reach significance in the initial salineversus water comparison are only reported if they meetthe P , 0.0005 criterion in the saline versus chocolatecomparison.

ResultsAll nine subjects with valid scans rated the saline solutionas highly aversive (mean 5 0.9; range 02.5 on the 11-pointscale) and highly intense (mean 5 8.9; range 610). Incontrast, only one subject reported detecting any taste duringthe pure water condition and this was described as barelydetectable. The chocolate was rated as highly pleasant(mean 5 9.0; range 810) and moderately intense (mean 56.8; range 58.5). These ratings indicate that the chocolateand saline conditions were well matched for hedonic strength(i.e. absolute deviation from a neutral rating of 5.0), butthe saline was slightly more perceptually intense than thechocolate. Subjects frequently described experiencing muscletension during the saline condition. None of the subjects

Table 1 Locations of increased rCBF when tasting 5%saline versus pure water

Area x y z Z-score

Right medial thalamus 1 22 9 4.2Right amygdala 26 1 16 3.8Left pregenual cingulate (BA 32) 1 39 0 3.7Right cingulate cortex (BA 24) 3 10 40 3.7Left cingulate cortex (BA 24) 6 22 36 3.5Left pregenual cingulate (BA 24) 3 30 0 3.5Left orbitofrontal cortex (BA 11) 24 41 7 3.4Right hippocampus 26 15 20 3.4

Stereotactic coordinates (mm) identify the location of the rCBFmaxima according to the atlas of Talairach and Tourneau (1988).x 5 mediallateral position relative to the midline (righthemisphere positive); y 5 anteriorposterior position relative tothe anterior commissure (anterior positive); z 5 inferiorsuperiorposition relative to the intercommissural plane (superior positive).

Table 2 Locations of rCBF maxima when tasting salineversus chocolate

Area x y z Z-score

Left pregenual cingulate (BA 24) 1 32 0 4.2Right motor cortex (BA 4) 57 15 38 3.9Right insula 33 14 7 3.7Right motor cortex (BA 4) 46 15 34 3.5Left orbitofrontal cortex (BA 11) 21 39 7 3.5Right amygdala 26 8 18 3.1

reported feeling disgust or fear, but they all indicated adislike for the saline. In comparison, subjects frequentlystated that the chocolate relaxed them, and several subjectsasked for additional chocolate upon completion of the scansession.

Table 1 shows brain regions activated in the contrastbetween tasting saline and tasting pure water. The largestrCBF increase in the saline condition was localized to themedial thalamus in the region of the dorsomedial and midline(intermediodorsal) nuclei. The peak of this diencephalic focusoccurred in the right hemisphere, but the focus extended intothe left hemisphere and may represent a bilateral response(see Fig. 1). Tasting saline also induced significant rCBFincreases in the pregenual and dorsal anterior cingulate,left orbitofrontal cortex, right amygdala and right anteriorhippocampus. Figure 2 shows the location of rCBF increasesin the right amygdala and hippocampus. The focus in theleft orbitofrontal cortex involved an anterior medial regionlocated along the superior apex of the H-shaped orbitalsulcus (Mai et al., 1997). Also a modest non-significantactivation was localized to the right orbitofrontal cortex inBrodmann area (BA) 11 (at coordinates x 5 17, y 5 37,z 5 14; Z-score 5 3.0), indicating that the orbitofrontalcortex response to aversive saline was not completelylateralized.

Many of these regions exhibited statistically significantactivity in the contrast between the saline and chocolate


Fig. 1 Medial thalamic and cingulate activation arising in the contrast between tasting aversive salineand tasting pure water. The image displays activity above a significance threshold of P , 0.001 (hottercolours denote greater activation) superimposed on a surface rendering of the medial wall of the brain.The image only displays the maximum activity occurring within 4 pixels (9 mm) of the medial wall.The left hemisphere of the brain is displayed at the top of the figure.

conditions (see Table 2). Foci emerging in this comparisonincluded the pregenual cingulate, left orbitofrontal cortex andright amygdala. Figure 3 shows the location of the leftorbitofrontal cortex focus in the saline versus chocolatecomparison. Subjects also demonstrated hippocampal

(x 5 28, y 5 19, z 5 18; Z-score 5 2.8) and thalamicactivity (x 5 1, y 5 22, z 5 7; Z-score 5 2.5), but neitherof these reached statistical significance at the P , 0.001cutoff. Unexpectedly, the saline condition producedsignificantly greater activity than the chocolate condition in


Fig. 2 Right amygdala and hippocampal activation arising in thecontrast between tasting aversive saline and tasting pure water.The image displays rCBF activation above a threshold ofP , 0.001 (hotter colours denote greater activation) superimposedon the corresponding transverse slice (z 5 16) of a standard T1-weighted MRI. The hippocampal maxima localized ~4 mm belowthis slice. The right side of the MRI corresponds to the left sideof the brain.

the right anterior insula in the vicinity of presumed primarygustatory cortex, even though both conditions involvedgustatory stimulation. The primary motor cortex (BA 4) alsoshowed increased activity in this contrast, suggesting thatthere was more tongue movement during the saline condition.

The contrast between the chocolate and pure waterconditions revealed no significant increases in activity in theamygdala, orbitofrontal cortex or pregenual cingulate. It maybe noted that some subjects showed rCBF increases in theright amygdala in the chocolate condition, but these increaseswere inconsistent, and several of the subjects showed rCBFdecreases. Interestingly, the two subjects showing the greatestamygdala responses (11.5% and 13.9% rCBF increases) inthe chocolate versus water contrast also showed the greatestrCBF increases (22.1% and 22.6%) in the contrast betweenthe saline and water conditions. The most robust increase inthe chocolate versus water contrast was localized to theposterior cingulate [BA 31; two peaks, at x 5 12, y 5 40, z 5 36 (Z-score 5 4.6) and x 5 3, y 5 42, z 5 36(Z-score 5 4.5)]. A dorsal anterior cingulate focus alsoemerged in BA 24 (x 5 6, y 5 24, z 5 38; Z-score 54.3) at similar coordinates to those appearing in the contrast

Fig. 3 Surface rendering of the ventral aspect of the braindemonstrating the location of the left orbitofrontal cortex focus inthe saline versus chocolate comparison. Only rCBF increases withsignificance greater than P , 0.001 are shown with hotter coloursdenoting greater activation. The right side of the imagecorresponds to the left side of the brain.

between the saline and water conditions. Therefore, the dorsalanterior cingulate responds similarly to both pleasant andaversive gustatory stimuli, which hence cancel each other inthe contrast between the saline and chocolate conditions.Finally, the chocolate condition produced activation in theright dorsomedial thalamic region which failed to reachstatistical significance (x 5 6, y 5 15, z 5 9; Z-score 5 2.8).

The restricted range of hedonic (pleasantnessunpleasantness) ratings of saline limits the utility ofperforming parametric or correlational analyses betweenrCBF and psychoperceptual ratings. Five subjects rated thestimulus as extremely unpleasant (i.e. hedonic ratings of 0.0)and four subjects rated it as moderately unpleasant (ratingsof 2.02.5). To determine whether the perceived degree ofaversiveness influenced rCBF, we split subjects into thosewith moderately unpleasant and extremely unpleasantexperiences, based on their subjective ratings to saline. Thepercentage rCBF increase for the two groups was calculatedby averaging all pixels within spherical (4.5 mm radius)regions of interest. These regions of interest were centred onthe peak coordinates defined by the saline minus watercomparison, except in the case of the right amygdala, wherethe region of interest was centred on the peak coordinatesfrom the saline minus chocolate comparison to ensure thatthe region of interest remained limited to the amygdala


Table 3 Percentage rCBF increase within regions ofinterest for the saline versus pure water in subjectsexperiencing the saline as extremely aversive versusmoderately aversive

Region Group with Group withextremely moderatelyunpleasant unpleasantexperience (n 5 5) experience (n 5 4)

Right hippocampus 14.5 4.4Pregenual cingulate 9.8 1.5Right amygdala 11.1 10.4Left orbitofrontal cortex 11.4 10.7Medial thalamus 13.4 14.1

proper. Table 3 lists the rCBF increases for activity withinpregenual cingulate, medial thalamic, orbitofrontal, amygdalaand hippocampal regions of interest for the two subjectgroups. The hippocampus showed the most dramaticdifference between the two groups, increasing 10% more inthe extremely unpleasant group than in the moderatelyunpleasant group. The pregenual cingulate only showedsubstantial rCBF increases in the subjects who perceived thestimulus as extremely aversive. In contrast, the rightamygdala, medial thalamus and left orbitofrontal cortex allshowed robust responses regardless of whether subjects ratedsaline as moderately or extremely unpleasant.

To determine whether additional regions were specific tothe sub-sample experiencing the saline as extremely aversive,we performed a post hoc pixel-wise comparison between thesaline and water conditions for the five subjects rating thesaline as extremely aversive. This increased the detectionsensitivity which might have been decreased in the full groupcomparison. Several right hemisphere foci emerged in thiscomparison which had not reached statistical significance inthe full group analysis. The largest of these additionalresponses arose in the right frontal pole (x 5 21, y 5 50,z 5 2; Z-score 5 3.9). A right pregenual cingulate focus(BA 32; x 5 10, y 5 28, z 5 9; Z-score 5 3.8) appearedin addition to the previously reported left cingulate focus.This confirms that the right hemisphere pregenual cingulateactivity observed in Fig. 2 most probably reflects bilateralactivity rather than blurring from a left hemisphere focus.The analysis also revealed bilateral medial thalamic activation(x 5 6, y 5 19, z 5 9; Z-score 5 3.7 and x 5 1, y 522, z 5 9; Z-score 5 3.8). Finally, a right anterior insularregion was activated in this condition (x 5 35, y 5 10,z 5 2; Z-score 5 3.6), with four of the five subjects showinga .12% increase in right insular rCBF. These analysesindicate that some brain regions only respond robustly whensaline is experienced as highly aversive, and these rCBFresponses may be obscured when subjects with more moderatehedonic responses are pooled with subjects with moredramatic subjective responses to saline.

To examine the functional interactions between limbicregions activated by saline, we examined the correlation

between rCBF within regions of interest selected from thesaline and water contrast. Non-subtracted rCBF for regionsof interest within the saline and water conditions wereseparately submitted to correlational analysis. Interpretationof such an analysis must be considered exploratory due tothe small sample size and the multiple structures activatedin the contrast between saline and water conditions.Nevertheless, such analyses provide valuable information ininstances where brain regions are functionally coupled (Zaldet al., 1998). Regions of interest were defined as describedabove; however, instead of examining the change in rCBFbetween conditions, rCBF was calculated separately withinthe saline and water conditions. These analyses revealed atight coupling of activities in the left pregenual cingulate andthe left anterior orbitofrontal cortex during the saline condition(r 5 0.81, P , 0.01). In contrast, the activity in these regionswere not significantly correlated in the water condition(r 5 0.27, P . 0.10). No other regions showed significantfunctional coupling in the saline or water condition, althoughactivity in the right amygdala showed a tendency to be linkedwith that in the right hippocampus (r 5 0.55, P , 0.10) andthe left orbitofrontal cortex (r 5 0.51, P , 0.10) in thesaline condition.

DiscussionThe current study demonstrates that exposure to an aversivegustatory stimulus activates a network of limbic structuresinvolving the amygdala, pregenual cingulate and orbitofrontalcortex. Subjects showed increased activity in these regionswhen tasting aversive saline, relative to attempting to tastepure water and to tasting chocolate. Because the saline andpure water conditions were matched for temperature andvolume of fluid, the observed activations cannot be attributedto non-gustatory sensory aspects of intraoral stimulation (e.g.somatosensory or thermal coding). These activations probablydo not reflect basic gustatory processing, since these areasremained significantly activated when saline was contrastedwith another gustatory stimulus. Furthermore, the ability ofsaline, relative to chocolate, to activate these regions doesnot appear to reflect differences in the hedonic strength ofthe two stimuli, since subjects rated both stimuli as strong.Rather, the rCBF increases in the amygdala, pregenualcingulate and anterior orbitofrontal cortex appear to reflectspecifically the recognition, experience and/or response tothe aversive quality of the saline.

The ability of an aversive gustatory stimulus to activatethe amygdala robustly converges with our previous findingthat exposure to aversive odourants activates the amygdala(Zald and Pardo, 1997). It is also consistent with a recent PETstudy of unpleasant visual stimulation, which demonstrated anincrease in amygdala rCBF (Lane et al., 1997b). Takentogether, these findings support the hypothesis that thehuman amygdala plays a multimodal role in the recognition,evaluation and/or response to aversive stimuli. In both thepresent gustatory study and our previous olfactory study (Zald


and Pardo, 1997), subjects rated the stimuli as moderately toextremely aversive and highly arousing. We have not observedsubstantial increases in amygdala activity when subjects rateodours or tastes as only mildly unpleasant. The strength ofthe subjective experience of aversiveness may underlie theability of these gustatory and olfactory paradigms to activatethe amygdala strongly. This interpretation converges withanimal studies which indicate that amygdala lesions criticallyimpair the acquisition of conditioned emotional responses toneutral stimuli that are paired with highly aversive or highlyarousing stimuli, while leaving conditioning to mildlyunpleasant, non-arousing stimuli unimpaired (Cahill andMcGaugh, 1990).

Unexpectedly, the amygdala activation occurred only inthe right hemisphere. This contrasts with our previous findingof bilateral amygdala activation during aversive olfaction(Zald and Pardo, 1997). The reason for this lateralizedresponse is unclear, although it may reflect a right dominancefor some aspects of gustatory processing. Patients withright anterior temporal lobectomies have been observed todemonstrate elevated taste recognition thresholds for citricacid relative to patients with left anterior temporallobectomies (Small et al., 1997a; but see Henkin et al.,1977). This pattern of right dominance for anterior temporallobe processing of citric acid has also been confirmed withPET (Small et al., 1997a). In the present study, a similarpattern of laterality was evident in the insula, which onlyshowed increased activity in the right hemisphere.

The peak focus in the amygdala region for the salineminus water comparison lies at the extreme anterior end ofthe amygdala as defined by the atlases of Talairach andTournoux (1988) and Mai et al. (1997). Given the spatialresolution of current PET techniques and inter-subjectanatomical variability, this focus may include cortex slightlyanterior to the amygdala. Indeed, we have observed activityin the vicinity of the pyriform cortex (anterior to theamygdala) in some taste experiments (D.H.Z. and J.V.P.,unpublished observations). The right anterior temporal lobefocus reported by Small et al. (1997a) in subjects tastingcitric acid also falls slightly anterior and medial to theTalairach coordinates for the amygdala. However, activity inthe anterior amygdala/pyriform area largely cancels whentasting saline is compared with tasting chocolate, leaving adiscrete focus centred more posteriorly within the amygdala.This more posterior focus falls solidly within the Talairachand Tournoux (1988) and Mai et al. (1997) coordinatesfor the amygdala, and remains clearly distinct from thehippocampal focus which lies 11 mm posterior.

Recordings of single units in non-human primates suggestthat gustatory reward may also activate the amygdala (Azumaet al., 1984; Ono and Nishijo, 1992). However, we did notobserve consistent evidence of amygdala activation in thecontrast between the chocolate and water conditions.Furthermore, the aversive saline condition activated theamygdala at similar levels when compared with either wateror chocolate. Although some subjects showed rCBF increases

in the amygdala region of interest when tasting chocolaterelative to tasting water, these increases were inconsistentacross subjects. The possibility that water to some extentobscured the ability to observe amygdala responses tochocolate requires consideration. A few amygdala cells showresponses to water in electrophysiological studies of non-human primates (Nishijo et al., 1988). Indeed, we haveobserved moderate rCBF increases in the right amygdalaregion in subjects attempting to taste water relative to a restingcondition (D.H.Z. and J.V.P., unpublished observation). Thismay reflect the role of water as a reinforcer in its own right.Nevertheless, saline activates this area substantially morethan water does, whereas chocolate does not. A smallpercentage of amygdala cells also respond to intra-oralthermal or tactile information (Azuma et al., 1984; Nishijoet al., 1988; Scott et al., 1993). Such thermal and tactilefactors were well matched in the saline versus pure watercomparison, but not adequately controlled in comparisonsinvolving chocolate and either water or saline. Despite thispotential problem, the greater amygdala activation duringaversive as compared with pleasant gustatory stimulationconverges with a growing literature demonstrating greaterinvolvement of the human amygdala in negative emotionalprocessing than positive emotional processing (Gloor, 1990;Adolphs et al., 1995; Breiter et al., 1996; Irwin et al., 1996;Ketter et al., 1996; Morris et al., 1996; Young et al., 1996;Lane et al., 1997b; Scott et al., 1997).

In addition to activating the right amygdala, saline activatedseveral closely interconnected limbic and paralimbic regions.The amygdala, pregenual cingulate, orbitofrontal cortex andmedial thalamus (dorsomedial and midline nuclei) all connectwith one another through direct and mostly reciprocalconnections (Pandya et al., 1981; Vogt and Pandya, 1987;Amaral et al., 1992; Neafsey et al., 1993; Ray and Price,1993; Van Hoesen et al., 1993; Groenewegen and Berendse,1994; Baylis et al., 1995; Carmichael and Price, 1995; 1996;Zald and Kim, 1996a; Bachevalier et al., 1997). The areasactivated by saline thus conform closely to a widespreaddistributed limbic/paralimbic network.

Anatomical and functional studies highlight the closefunctional interaction between the orbitofrontal cortex andamygdala (Zald and Kim, 1996a, b; Zald et al., 1998). Theability of aversive saline to activate the human orbitofrontalcortex converges with electrophysiological data from non-human primates demonstrating the presence of orbitofrontalcortex cells with specific responses to aversive saline (Thorpeet al., 1983; Rolls et al., 1990). Single cell recordings inmonkeys suggest that the human orbitofrontal cortex shouldalso possess cells responsive to rewarding gustatory stimulisuch as chocolate (Thorpe et al., 1983; Rolls et al., 1990).However, we failed to observe significant increases in theorbitofrontal cortex during exposure to chocolate comparedwith water. In fact, half of the subjects showed greater rCBFwithin the anterior orbitofrontal region of interest whentasting water than when tasting chocolate. Furthermore,


comparison of saline with chocolate instead of water did notreduce the magnitude of the left orbitofrontal cortex focus.

The orbitofrontal cortex (particularly the left orbitofrontalcortex) emerges as one of the most frequently activatedregions during aversive sensory and psychologicalexperiences (Pardo et al., 1993; Rauch et al., 1994; 1995;Fischer et al., 1996; Zald and Pardo, 1997). However, thespecific foci activated within the orbitofrontal cortex tend tovary depending on the sensory modality and experimentalparadigm, perhaps reflecting the large differences in sensoryand limbic afferents that distinguish orbitofrontal cortexsubregions (Carmichael and Price, 1995, 1996). It may benoted that the left orbitofrontal cortex focus in the currentstudy appears too anterior to represent the caudolateralorbitofrontal cortex area identified as secondary gustatorycortex in monkeys (Rolls, 1989; Rolls et al., 1990). Thisanterior-medial region is more likely to represent aheteromodal association region which receives gustatoryinformation secondary to the caudolateral orbitofrontal cortex(Rolls and Baylis, 1994; Carmichael and Price, 1995; 1996).

The medial thalamus emerged as the region with the largestmagnitude activation in the saline versus water comparison.Like lesions of the orbitofrontal cortex and amygdala,dorsomedial thalamic lesions disrupt the ability to form directassociations between visual stimuli and gustatory reinforcers(Gaffan and Murray, 1990). Thalamic lesions that includethe dorsomedial nucleus frequently disrupt the acquisition ofavoidance learning and other conditioned emotional responses(Buchanan and Powell, 1993; Gabriel, 1993). Human dataalso indicate that medial thalamic areas are activated duringexposure to visual stimuli perceived as disgusting (Laneet al., 1997a; Paradiso et al., 1997). However, the dorsomedialthalamic activity appears less clearly specific to the aversivecondition. Chocolate (when compared with water) producedmoderate, albeit non-significant, increases in right medialthalamic rCBF, and the comparison between saline andchocolate failed to produce a statistically significant activationin this region. Furthermore, this area showed robust rCBFincreases (.11%) regardless of whether subjects found thesaline moderately or extremely aversive. This indicates thatthe medial thalamus probably plays a relatively balanced rolein the processing of gustatory stimuli, which is not restrictedby the intensity or direction of the hedonic experience. Thisconclusion converges with evidence that visual stimuli thatinduce positive emotions often produce rCBF increases inthe thalamus similar to those with visual stimuli that inducenegative emotions (Lane et al., 1997a, b; Reiman et al.,1997). It also appears consistent with a recent case report inwhich a dorsomedial thalamic lesion in a 68-year-old womanproduced changes in the hedonic perception of tastes,particularly causing previously pleasant stimuli to beperceived as neutral or unpleasant, without altering basicolfactory and gustatory identification (Rousseaux et al., 1996).

Exposure to aversive saline induced activity in multipleportions of the cingulate cortex. However, only the activityin the inferior, pregenual portions of the cingulate (BA 24

and adjacent BA 32) remained significant when saline wascompared with chocolate, indicating a valence-specificresponse (unpleasant distinct from pleasant). Furthermore,this region appeared heavily influenced by the strength ofthe subjective experience with large rCBF increases (.6%)in both the left and right pregenual cingulate occurring onlyin subjects who found the stimulus extremely aversive. Thelarge dependence of pregenual cingulate activity upon thehedonic significance of the stimulus argues against a role inbasic taste processing. This inferior pregenual cingulateregion has previously been implicated in affective behaviour(Devinsky et al., 1995). Animal studies implicate this regionin aspects of avoidance learning and emotional conditioning(Buchanan and Powell, 1993; Gabriel, 1993). Recentneuroimaging studies of anxiety, fear, dysphoria anddepression also implicate the inferior-medial frontal lobe inunpleasant emotional states (George et al., 1995; Rauch et al.,1995, 1996; Drevets et al., 1997; Mayberg et al., 1997; Shinet al., 1997). The current findings thus converge with bothanimal and human data, in highlighting the importanceof the inferior-medial frontal cortex in negative emotionalexperiences. In contrast, the more dorsal anterior and mid-cingulate regions did not exhibit a similar level of specificityto aversive gustatory stimulation. For instance, the dorsalanterior cingulate responded equally well to both chocolateand saline in comparison to water.

The hippocampal activation was unexpected. Although thehippocampus is directly or indirectly connected with severalother areas activated by saline, including the amygdala(Amaral, 1986; Saunders et al., 1988), midline thalamus(Groenewegen and Berendse, 1994), pregenual cingulate andanterior medial orbitofrontal cortex (Vogt and Pandya, 1987;Van Hoesen et al., 1993; Carmichael and Price, 1995; Zaldand Kim, 1989), it has not been commonly observed inprevious neuroimaging studies of emotion. Interestingly, ina recent PET study, there was coactivation of the left amygdalaand left hippocampus in subjects viewing unpleasant pictures(Lane et al., 1997b). In both the study of Lane et al. (1997b)and the present study, the hippocampal activation occupiedan anterior portion of the hippocampus, and occurredipsilateral to the amygdala activation. The present data furthersuggest that hippocampal activation primarily occurredamong subjects who found the saline extremely aversive.This could reflect a gating process through which highlyarousing stimuli preferentially gain access to hippocampalmnemonic processes. Data from both humans and animalshighlight the importance of the amygdala for the emotionalenhancement of memory (Markowitsch et al., 1994; Cahillet al., 1995, 1996; McGaugh et al., 1996). The current dataraise the possibility that the amygdala accomplishes thisenhancement by modulating hippocampal processing whenstimuli are perceived as extremely aversive or arousing.

The right anterior insula showed activations in both thecontrast of saline versus chocolate, and the contrast of salineversus water, when the latter contrast was limited to subjectswho perceived the saline as extremely aversive. This insular


focus is consistent with the location of primary gustatorycortex as identified in non-human primates (Yaxley et al.,1990; Norgren, 1990). The insular focus in the saline versuswater contrast arose despite the ability of water to activatethis region relative to resting conditions (D.H.Z. and J.V.P.,unpublished observations). Whether this response reflectstaste processing itself, or a neural correlate specific to highlyaversive gustation, remains unresolved, since the insularactivation failed to reach significance when all subjects wereincluded in the analysis. That this insular focus again emergedin the saline minus chocolate comparison (even thoughboth conditions involved gustatory stimulation) supports thehypothesis that insular activity reflects the aversive nature ofsaline. However, a recent PET study demonstrated greaterinsular activity during exposure to basic tastants than duringcombined gustatory/olfactory stimulation (Small et al.,1997b). Since chocolate possesses both gustatory andolfactory properties, it remains possible that the observedinsular activity in the saline versus chocolate comparisonreflects basic taste processing, rather than being a specificcorrelate of aversive gustation.

A potential limitation of the current study involves the useof a single continuous stimulus presentation per 90 s scanperiod. Cells in the gustatory system frequently show rapidhabituation during extended exposure to stimuli. This mayexplain the inconsistencies in anterior insular activity andthe lack of significant rCBF increases in the frontal operculum.While the continuous presentation of gustatory stimuli mayreduce activation in primary gustatory cortex, the longduration of stimulation may increase the intensity of theaversive experience. Subjects frequently commented that thelength of the scan made the saline condition more aversivebecause they were unable to stop or ignore the unpleasantsensation. This protracted experience may thus aid the abilityto activate areas involved in the hedonic processing of tastantsrelative to regions processing their primary sensory properties.

In conclusion, tasting aversive saline activates a networkof interconnected limbic and paralimbic structures. Thepresent data indicate that several of these areas participatein the evaluative or motivational aspects of the aversiveexperience rather than reflecting basic sensory coding ofgustatory stimuli. Disruption of such evaluative ormotivational processes by lesions in these regions could leadto an inability to recognize and learn which foods areunpalatable (as occurs in the KluverBucy syndrome).Moreover, the capacity of these regions to respond duringexposure to aversive stimuli may underlie their role inavoidance learning and emotional conditioning.

AcknowledgementsWe wish to thank our volunteers and the staff of the PETImaging Service. This work was supported, in part, by theDepartment of Veterans Affairs, NARSAD and the MinnesotaObesity Center (P30 DK5045602). D.H.Z. was supportedby a NRSA grant (1 F32 MH1164101A1).

ReferencesAdolphs R, Tranel D, Damasio H, Damasio AR. Fear and thehuman amygdala. J Neurosci 1995; 15: 587991.

Aggleton JP. The functional effects of amygdala lesions in humans:a comparison with findings from monkeys. In: Aggleton JP, editor.The amygdala: neurobiological aspects of emotion, memory, andmental dysfunction. New York: Wiley-Liss; 1992. p. 485503.

Amaral DG. Amygdalohippocampal and amygdalocorticalprojections in the primate brain. In: Schwarcz R, Ben-Ari Y, editors.Excitatory amino acids and epilepsy. Adv Exp Med Biol, Vol 203.New York: Plenum Press; 1986. p. 317.

Amaral DG, Price JL, Pitkanen A, Carmichael ST. Anatomicalorganization of the primate amygdaloid complex. In: Aggleton JP,editor. The amygdala: neurobiological aspects of emotion, memory,and mental dysfunction. New York: Wiley-Liss; 1992. p. 166.

Andersson JL, Vagnhammar BE, Schneider H. Accurate attenuationcorrection despite movement during PET imaging. J Nucl Med1995; 36: 6708.

Arthur JB. Taste aversion learning is impaired by interpolatedamygdaloid stimulation but not by post-training amygdaloidstimulation. Behav Biol 1975; 13: 36976.

Azuma S, Yamamoto T, Kawamura Y. Studies on gustatory responsesof amygdaloid neurons in rats. Exp Brain Res 1984; 56: 1222.

Bachevalier J, Meunier M, Lu MX, Ungerleider LG. Thalamic andtemporal cortex input to medial prefrontal cortex in rhesus monkeys.Exp Brain Res 1997; 115: 43044.

Baylis LL, Gaffan D. Amygdalectomy and ventromedial prefrontalablation produce similar deficits in food choice and in simple objectdiscrimination learning for an unseen reward. Exp Brain Res 1991;86: 61722.

Baylis LL, Rolls ET, Baylis GC. Afferent connections of thecaudolateral orbitofrontal cortex taste area of the primate.Neuroscience 1995; 64: 80112.

Beckstead RM, Morse JR, Norgren R. The nucleus of the solitarytract in the monkey. Projections to the thalamus and brain stemnuclei. J Comp Neurol 1980; 190: 25982.

Breiter HC, Etcoff NL, Whalen PJ, Kennedy WA, Rauch SL,Buckner RL, et al. Response and habituation of the human amygdaladuring visual processing of facial expression. Neuron 1996; 17:87587.

Buchanan SL, Powell DA. Cingulothalamic and prefrontal controlof autonomic function. In: Vogt BA, Gabriel M, editors.Neurobiology of cingulate cortex and limbic thalamus: acomprehensive handbook. Boston (MA): Birkhauser; 1993.p. 381414.

Butter CM, McDonald JA, Snyder DR. Orality, preference behavior,and reinforcement value of nonfood object in monkeys with orbitalfrontal lesions. Science 1969; 164: 13067.

Cahill L, McGaugh JL. Amygdaloid complex lesions differentiallyaffect retention of tasks using appetitive and aversive reinforcement.Behav Neurosci 1990; 104: 53243.

Cahill L, Babinsky R, Markowitsch HJ, McGaugh JL. The amygdalaand emotional memory [letter]. Nature 1995; 377: 2956.


Cahill L, Haier RJ, Fallon J, Alkire MT, Tang C, Keator D, et al.Amygdala activity at encoding correlated with long-term, free recallof emotional information. Proc Natl Acad Sci USA 1996; 93:801621.

Carmichael ST, Price JL. Limbic connections of the orbital andmedial prefrontal cortex in macaque monkeys. J Comp Neurol1995; 363: 61541.

Carmichael ST, Price JL. Connectional networks within the orbitaland medial prefrontal cortex of macaque monkeys. J Comp Neurol1996; 371: 179207.

Davis M. The role of the amygdala in fear and anxiety. [Review].Annu Rev Neurosci 1992; 15: 35375.

Devinsky O, Morrell MJ, Vogt BA. Contributions of anteriorcingulate cortex to behaviour. [Review]. Brain 1995; 118: 279306.Drevets WC, Price JL, Simpson JR Jr, Todd RD, Reich T, Vannier M,et al. Subgenual prefrontal cortex abnormalities in mood disorders[see comments]. Nature 1997; 386: 8247. Comment in: Nature1997; 386: 76970.

Fischer H, Wik G, Fredrikson M. Functional neuroanatomy ofrobbery re-experience: affective memories studied with PET.Neuroreport 1996; 7: 20816.

Fox PT, Mintun MA, Raichle ME, Miezin FM, Allman JM, VanEssen DC. Mapping human visual cortex with positron emissiontomography. Nature 1986; 323: 8069.

Gabriel M. Discriminative avoidance learning. In: Vogt BA, GabrielM, editors. Neurobiology of cingulate cortex and limbic thalamus:a comprehensive handbook. Boston (MA): Birkhauser; 1993.p. 478523.

Gaffan D, Murray EA. Amygdalar interaction with the mediodorsalnucleus of the thalamus and the ventromedial prefrontal cortex instimulus-reward associative learning in the monkey. J Neurosci1990; 10: 347993.

George MS, Ketter TA, Parekh PI, Horwitz B, Herscovitch P, PostRM. Brain activity during transient sadness and happiness in healthywomen. Am J Psychiatry 1995; 152: 34151.

Gloor P. Experiential phenomena of temporal lobe epilepsy: factsand hypotheses. [Review]. Brain 1990; 113: 167394.Groenewegen HJ, Berendse HW. The specificity of the nonspecificmidline and intralaminar thalamic nuclei. [Review]. Trends Neurosci1994; 17: 528.

Henkin R, Comiter H, Deio P, ODoherty D. Defects in taste andsmell recognition following temporal lobectomy. Trans Am NeurolAssoc 1977; 102: 14650.

Huang SC, Hoffman EJ, Phelps ME, Kuhl DE. Quantitation inpositron emission computed tomography: 2. Effects of inaccurateattenuation correction. J Comput Assist Tomogr 1979; 3: 80414.

Irwin W, Davidson RJ, Lowe MJ, Mock BJ, Sorenson JA, Turski PA.Human amygdala activation detected with echo-planar functionalmagnetic resonance imaging. Neuroreport 1996; 7: 17659.

Kennie ED, Nagel JA. Failure to form a learned taste aversion inrats with amygdaloid lesions. Bull Psychonomic Soc 1973; 2: 1556.

Ketter TA, Andreason PJ, George MS, Lee C, Gil DS, Parekh PI,

et al. Anterior paralimbic mediation of procaine-induced emotionaland psychosensory experiences [see comments]. Arch GenPsychiatry 1996; 53: 5969. Comment in: Arch Gen Psychiatry1997; 54: 7635.

Kinahan PE, Rogers JG. Analytic 3D image reconstruction usingall detected events. IEEE Trans Nucl Sci 1989; 36: 96486.

Kluver H, Bucy PC. Psychic blindness and other symptomsfollowing bilateral temporal lobectomy in rhesus monkeys. Am JPhysiol 1937; 119: 3523.

Lane RD, Reiman EM, Ahern GL, Schwartz GE, Davidson RJ.Neuroanatomical correlates of happiness, sadness, and disgust. AmJ Psychiatry 1997a; 154: 92633.

Lane RD, Reiman EM, Bradley MM, Lang PJ, Ahern GL, DavidsonRJ, et al. Neuroanatomical correlates of pleasant and unpleasantemotion. Neuropsychologia 1997b; 35: 143744.

LeDoux JE. Emotion. In: Mountcastle VB, Plum F, editors.Handbook of physiology, Sect. 1, Vol. 5, Pt 1. Bethesda (MD):American Physiological Society; 1987. p. 41959.

Mai JK, Assheuer J, Paxinos G. Atlas of the human brain. SanDiego (CA): Academic Press; 1997.Markowitsch HJ, Calabrese P, Wurker M, Durwen HF, Kessler J,Babinsky R, et al. The amygdalas contribution to memorya studyon two patients with UrbachWiethe disease. Neuroreport 1994; 5:134952.

Mayberg HS, Brannan SK, Mahurin RK, Jerabek PA, Brickman JS,Tekell JL, et al. Cingulate function in depression: a potentialpredictor of treatment response. Neuroreport 1997; 8: 105761.

McGaugh JL, Cahill L, Roozendaal B. Involvement of the amygdalain memory storage: interaction with other brain systems. [Review].Proc Natl Acad Sci USA 1996; 93: 1350814.

Mikulka PJ, Freeman FG, Lidstrom P. The effect of trainingtechnique and amygdala lesions on the acquisition and retention ofa taste aversion. Behav Biol 1977; 19: 50917.

Minoshima S, Berger KL, Lee KS, Mintun MA. An automatedmethod for rotational correction and centering of three-dimensionalfunctional brain images. J Nucl Med 1992; 33: 157985.

Minoshima S, Koeppe RA, Mintun MA, Berger KL, Taylor SF,Frey KA, et al. Automated detection of the intercommissural linefor stereotactic localization of functional brain images. J Nucl Med1993; 34: 3229.

Minoshima S, Koeppe RA, Frey KA, Kuhl DE. Anatomicstandardization: linear scaling and nonlinear warping of functionalbrain images. J Nucl Med 1994; 35: 152837.

Morris JS, Frith CD, Perrett DI, Rowland D, Young AW, CalderAJ, et al. A differential neural response in the human amygdala tofearful and happy facial expressions. Nature 1996; 383: 8125.

Neafsey EJ, Terreberry RR, Hurley KM, Ruit KG, Frysztak RJ.Anterior cingulate cortex in rodents: connections, visceral controlfunctions, and implications for emotion. In: Vogt BA, Gabriel M,editors. Neurobiology of cingulate cortex and limbic thalamus: acomprehensive handbook. Boston (MA): Birkhauser; 1993.p. 20623.


Nishijo H, Ono T, Nishino H. Topographic distribution of modality-specific amygdalar neurons in alert monkey. J Neurosci 1988; 8:355669.

Norgren R. Taste pathways to hypothalamus and amygdala. J CompNeurol 1976; 166: 1730.

Norgren R. Gustatory system. In: Paxinos G, editor. The humannervous system. San Diego: Academic Press; 1990. p.84561.

Ono T, Nishijo H. Neurophysiological basis of the KluverBucysyndrome: responses of monkey amygdaloid neurons to biologicallysignificant objects. In: Aggleton JP, editor. The amygdala:neurobiological aspects of emotion, memory, and mentaldysfunction. New York: Wiley-Liss; 1992. p. 16790.

Pandya DN, Van Hoesen GW, Mesulam MM. Efferent connectionsof the cingulate gyrus in the rhesus monkey. Exp Brain Res 1981;42: 31930.

Paradiso S, Robinson RG, Andreasen NC, Downhill JE, DavidsonRJ, Kirchner PT, et al. Emotional activation of limbic circuitry inelderly normal subjects in a PET study. Am J Psychiatry 1997; 154:3849.

Pardo JV, Pardo PJ, Raichle ME. Neural correlates of self-induceddysphoria [see comments]. Am J Psychiatr 1993; 150: 7139.Comment in: Am J Psychiatry 1994; 151: 7845.

Rauch SL, Jenike MA, Alpert NM, Baer L, Breiter HC, SavageCR, et al. Regional cerebral blood flow measured during symptomprovocation in obsessive-compulsive disorder using oxygen 15-labeled carbon dioxide and positron emission tomography [seecomments]. Arch Gen Psychiatry 1994; 51: 6270. Comment in:Arch Gen Psychiatry 1995; 52: 15960.

Rauch SL, Savage CR, Alpert NM, Miguel EC, Baer L, BreiterHC, et al. A positron emission tomographic study of simple phobicsymptom provocation. Arch Gen Psychiatry 1995; 52: 208.

Rauch SL, van der Kolk BA, Fisler RE, Alpert NM, Orr SP, SavageCR, et al. A symptom provocation study of posttraumatic stressdisorder using positron emission tomography and script-drivenimagery. Arch Gen Psychiatry 1996; 53: 3807.

Ray JP, Price JL. The organization of projections from themediodorsal nucleus of the thalamus to orbital and medial prefrontalcortex in macaque monkeys. J Comp Neurol 1993; 337: 131.

Reiman EM, Lane RD, Ahern GL, Schwartz GE, Davidson RJ,Friston KJ, et al. Neuroanatomical correlates of internally generatedhuman emotion. Am J Psychiatry 1997; 154: 91825.

Rolls ET. Information processing in the taste system of primates.[Review]. J Exp Biol 1989; 146: 14164.Rolls ET, Baylis LL. Gustatory, olfactory, and visual convergencewithin the primate orbitofrontal cortex. J Neurosci 1994; 14:543752.

Rolls ET, Yaxley S, Sienkiewicz ZJ. Gustatory responses of singleneurons in the caudolateral orbitofrontal cortex of the macaquemonkey. J Neurophysiol 1990; 64: 105566.

Rousseaux M, Muller P, Gahide I, Mottin Y, Romon M. Disordersof smell, taste, and food intake in a patient with a dorsomedialthalamic infarct. Stroke 1996; 27: 232830.

Saunders RC, Rosene DL, Van Hoesen GW. Comparison of the

efferents of the amygdala and the hippocampal formation in therhesus monkey: II. Reciprocal and non-reciprocal connections. JComp Neurol 1988; 271: 185207.

Scott SK, Young AW, Calder AJ, Hellawell DJ, Aggleton JP,Johnson M. Impaired auditory recognition of fear and angerfollowing bilateral amygdala lesions. Nature 1997; 385: 2547.

Scott TR, Karadi Z, Oomura Y, Nishino H, Plata-Salaman CR,Lenard L, et al. Gustatory neural coding in the amygdala of thealert macaque monkey. J Neurophysiol 1993; 69: 181020.

Shields SA. Gender in the psychology of emotion: a selectiveresearch review. In: Strongman KT, editor. International review ofstudies on emotion. New York: Wiley; 1991. p. 22745.

Shin LM, Kosslyn SM, McNally RJ, Alpert NM, Thompson WL,Rauch SL, et al. Visual imagery and perception in posttraumaticstress disorder: a positron emission tomographic investigation. ArchGen Psychiatry 1997; 54: 23341.

Silbersweig DA, Stern E, Frith CD, Cahill C, Schnorr L,Grootoonk S, et al. Detection of thirty-second cognitive activationsin single subjects with positron emission tomography: A new low-dose H215O regional cerebral blood flow three-dimensional imagingtechnique. J Cereb Blood Flow Metab 1993; 13: 61729.

Small DM, Jones-Gotman M, Zatorre RJ, Petrides M, Evans AC.A role for the right anterior temporal lobe in taste quality recognition.J Neurosci 1997a; 17: 513642.

Small DM, Jones-Gotman M, Zatorre RJ, Petrides M, Evans AC.Flavor processing: more than the sum of its parts. Neuroreport1997b; 8: 39137.

Talairach J, Tournoux P. Co-planar stereotaxic atlas of the humanbrain. Stuttgart: Thieme; 1988.

Thorpe SJ, Rolls ET, Maddison S. The orbitofrontal cortex: neuronalactivity in the behaving monkey. Exp Brain Res 1983; 49: 93115.

Turner B, Herkenham M. An autoradiographic study of thalamo-amygdaloid connections in the rat [abstract]. Anat Rec 1981;199: 260A.

Turner BH, Mishkin M, Knapp M. Organization of theamygdalopetal projections from modality-specific corticalassociation areas in the monkey. J Comp Neurol 1980; 191: 51543.

Ursin H, Rosvold HE, Vest B. Food preference in brain lesionedmonkeys. Physiol Behav 1969; 4: 60912.

Van Hoesen GW, Morecraft RJ, Vogt BA. Connections of the monkeycingulate cortex. In: Vogt BA, Gabriel M, editors. Neurobiology ofcingulate cortex and limbic thalamus: a comprehensive handbook.Boston (MA): Birkhauser; 1993. p. 24984.Vogt BA, Pandya DN. Cingulate cortex of the rhesus monkey: II.Cortical afferents. J Comp Neurol 1987; 262: 27189.

Woods RP, Cherry SR, Mazziotta JC. Rapid automated algorithmfor aligning and reslicing PET images. J Comput Assist Tomogr1992; 16: 62033.

Worsley KJ, Evans AC, Marrett S, Neelin P. A three-dimensionalstatistical analysis for CBF activation studies in human brain [seecomments]. J Cereb Blood Flow Metab 1992; 12: 90018. Commentin: J Cereb Blood Flow Metab 1993; 13: 10402.


Yasoshima Y, ShimuraT, Yamamoto T. Single unit responses ofthe amygdala after conditioned taste aversion in conscious rats.Neuroreport 1995; 6: 24248.

Yasui Y, ltoh M, Mizuno N. Projections from the parvocellular partof the posteromedial ventral nucleus of the thalamus to the lateralamygdaloid nucleus in the cat. Brain Res 1984; 292: 1515.

Yaxley S, Rolls ET, Sienkiewicz ZJ. Gustatory responses of singleneurons in the insula of the macaque monkey. J Neurophysiol 1990;63: 689700.

Young AW, Hellawell DJ, Van De Wal C, Johnson M. Facialexpression processing after amygdalotomy. Neuropsychologia 1996;34: 319.

Zald DH, Kim, SW. The anatomy and function of the orbitalfrontal cortex, I. Anatomy, neurocircuitry, and obsessive-compulsivedisorder. J Neuropsychiatry Clin Neurosci 1996a; 8: 12538.

Zald DH, Kim, SW. Anatomy and function of the orbital frontalcortex, II. Function and relevance to obsessive-compulsive disorder.J Neuropsychiatry Clin Neurosci 1996b; 8: 24961.

Zald DH, Kim SW. The orbitofrontal cortex. In: Salloway S,Malloy P, editors. The frontal lobes and neuropsychiatric illness.Washington (DC): American Psychiatric Press 1998. In press.Zald DH, Pardo JV. Emotion, olfaction, and the human amygdala:amygdala activation during aversive olfactory stimulation. Proc NatlAcad Sci USA 1997; 94: 411924.Zald DH, Donndelinger MJ, Pardo JV. Elucidating dynamic braininteractions with across-subjects correlational analyses of PET data:the functional connectivity of the amygdala and orbitofrontal cortexduring olfactory tasks. J Cereb Blood Flow Metab 1998. In press.

Received September 11, 1997. Revised December 26, 1997.Accepted January 21, 1998

aversive gustatory stimulation

Documents