olfaction, emotion & the amygdala: arousal-dependent modulation

Impulse: The Premier Journal for Undergraduate Publications in the NeurosciencesJune 2004, 1 (1): 1-58

Olfaction, Emotion & the Amygdala: arousal-dependentmodulation of long-term autobiographical memory and itsassociation with olfaction: beginning to unravel the Proustphenomenon?

Mark Hughes1

1University of Edinburgh, Centre for Neuroscience, No. 1 George Square, Edinburgh, UK

The sense of smell is set apart from othersensory modalities. Odours possess thecapacity to trigger immediately strongemotional memories. Moreover, odorousstimuli provide a higher degree of memoryretention than other sensory stimuli. Odourperception, even in its most elementalform - olfaction - already involves limbicstructures. This early involvement is notparalleled in other sensory modalities.Bearing in mind the considerableconnectivity with limbic structures, and thefact that an activation of the amygdala iscapable of instantaneously evokingemotions and facilitating the encoding ofmemories, it is unsurprising that the senseof smell has its characteristic nature.The aim of this review is to analyse currentunderstanding of higher olfactoryinformation processing as it relates to theability of odours to spontaneously cuehighly vivid, affectively toned, and often

very old autobiographical memories(episodes known anecdotally as Proustphenomena). Particular emphasis is placedon the diversity of functions attributed tothe amygdala. It’s role in modulating theencoding and retrieval of long-termmemory is investigated with reference tolesion, electrophysiological, immediateearly gene, and functional imaging studiesin both rodents and humans. Additionally,the influence of hormonal modulation andthe adrenergic system on emotionalmemory storage is outlined. I finish byproposing a schematic of some of thecritical neural pathways that underlie theodour-associated encoding and retrieval ofemotionally toned autobiographicalmemories.

Key Words: Olfaction; Emotion; Proustphenomenon; Amygdala.

Introduction

“Smell is a potent wizard that transports usacross thousands of miles and all the years wehave lived… odours, instantaneous andfleeting, cause my heart to dilate joyously orcontract with remembered grief.” HelenKeller (blind and deaf educator, 1880-1968)This quotation articulates a neurophysiologicalinteraction between the systems of olfaction,emotion and memory. Hughlings Jackson(1880), arguably the founder of neurology as ascience, termed this kind of forcedreminiscence as “a doubling ofconsciousness .” This occurrence is notuncommon during attacks of migraine or

epilepsy, in states of hypnosis or psychosis,and, less radically, in everybody in response tothe mnemonic stimulus of certain words,sounds, scenes, and especially smells. Odoursare rarely perceived in a purely objective andneutral way, often being loaded with appealingor repugnant feelings. Olfactory pathwaysinvolve anatomical structures that are alsoheavily implicated in emotion and memory.The amygdala is one such structure. Thechallenge is to propose a model of the neuralmechanisms that underlie encoding andretrieval of odour-cued emotionally tonedautobiographical memories (see figure one).

Pages 18-37Olfaction, Emotion and the Amygdala: Hughes

Proust Phenomena

The ability of odours to spontaneously cuehighly vivid, affectively toned, and often veryold autobiographical memories is known asthe Proust phenomenon (Chu & Downes,2000) and is derived from the experiences ofProust himself (1922):“And soon, mechanically, weary after a dullday with the prospect of a depressing morrow,I raised to my lips a spoonful of tea in which Ihad soaked a morsel of cake. No sooner hadthe warm liquid, and the crumbs with it,touched my palate than a shudder ran throughmy whole body, and I stopped, intent upon theextraordinary changes that were takingplace…I was conscious that it was connectedwith the taste of tea and cake, but that itinfinitely transcended those savours, couldnot, indeed, be of the same nature as theirs.”Such experiences are not limited to the realmof artistic license. A survey by the NationalGeographic (Gilbert & Wysocki, 1987) gavereaders a set of six odours on scratch-and-sniffcards. From a sample of 26 200 (takenrandomly from >1.5 million responses), ~55%of respondents in their 20s reported at leastone emotional memory cued by one of the sixodours.

Studying odour-cued memories

Many of the approaches used could bedescribed as experiments in odour-cuedcontext-dependent memory rather thanautobiographical memory. Context-dependentmemory is the effect whereby retrievinginformation in the same environment in which

it was encoded results in a better memoryperformance than encoding and retrieving indifferent contexts (Eich et al., 1985). Schab(1990) used ambient odour as a context,demonstrating that reinstating the same odourat test as that present at study appears toimprove memory performance in humans overconditions where the ambient odour is notreinstated at test. By manipulating odours thatwere appropriate or inappropriate to asituation, Herz (1997) has also shown inhumans that distinctive odours result in greatercontext effects.These studies indicate that ambient odourspresent at the time of an event can be encodedin parallel with event details and consequentlybe used as cues in the retrieval of those eventdetails. The types of memory commonlyreported in connection with Proust phenomenaare typically especially vivid, emotional andold. Therefore, several critical aspects of thesestudies do not address the core of Proustianphenomena. Firstly, the age of experiences islimited to time between study and test(typically only a few minutes). Secondly,investigators predetermine the target event,instead of allowing a subject to freely select anautobiographical episode that comes to mindin response to an odour. Lastly, the subject isnot always made aware of the presence of theodour during the retrieval phase. Mostexamples of Proustian recall involve anawareness of the odour.For empirical substantiation, other paradigmsthat more closely match the Proustian situationmust be applied. In the standard paired-associate learning paradigm, odours and otherstimuli are associated with target information.Later, the efficacy of each cue type iscompared in a cued recall test. Herz and

Figure 1 Simplified schematic illustrating neuroanatomical and systems-level interactions during the generation of odour-evoked emotional memories.


Cupchik (1995) asked subjects to associateodours or odour labels with emotionalpaintings. Later, they were presented with theodours or words and asked to describe thepainting with which each had been associated.A higher degree of emotional tone was foundin the recollections of paintings cued byodours than those cued by words. Here,subjects are actively using odour as an aid toretrieval of a past experience. However,autobiographical memories are usually formedthrough a passive encoding process: separateexperiential aspects of an incident becomeassociated without any purposeful effort onthe part of the subject. Therefore, there arelikely to be key differences in the underlyingprocesses recruited during encoding (and thenature of the resulting mnemonicrepresentation) in the above study as comparedwith naturally acquired autobiographicalmemories.

Examination of ‘true’autobiographical memories

Owing to problems of quantitativemeasurement and application of closeexperimental controls, only a limited numberof studies exist which have directly examinedProustian phenomena.Laird (1935) wrote about a survey of odours‘as revivers of memories and provokers ofthoughts in 254 living men and women ofeminence.’ More than four out of five of thosesurveyed reported olfactory-cued memoryexperiences that were described as vivid,emotional and old. 76% of women and 46% ofmen reported that odour-cued memories wereamongst their most vivid with only 7% ofwomen and 14% of men describing thesememories as hedonically neutral.Contemporary studies employ more scientificand empirical approaches, for example thesingle-cue comparison method. Data iscollected on memories that are retrieved inresponse to odours and are directly comparedwith those retrieved in response to stimulifrom other sensory modalities. Rubin et al.,(1984) found that odour-cued memories werethought of and spoken of with less frequencythan those cued by other stimuli. Contrary toProustian expectations, no differences werefound between different cue types, and

analysis of the pleasantness of memoriesyielded mixed and inconclusive results.Herz and Cupchik (1992) aimed to formulate acharacterization of odour-evoked memoriesrather than compare the efficacy of differentcue types. Their results characterize odour-evoked memories as highly emotional, vivid,specific, rare, and relatively old but leaveunanswered the question as to whether odour-evoked memories are, for example, moreemotional. Herz and Schooler (2002) havenow answered this question. The emotionaland evocative qualities of autobiographicalmemories cued by odours and visual cues werecompared using a new repeated measuresparadigm. Results demonstrated that memoriesevoked in the context of odours weresignificantly more emotional than thoserecalled in the context of the same cuepresented visually and by the verbal label forthe cue. This is the first unambiguousdemonstration that naturalistic memoriesevoked by odours are more emotional thanthose evoked by other cues.Aggleton and Waskett (1999) provided anexcellent demonstration of the potency ofolfactory cues. Their experiment focused on adisplay at a Viking museum (York) in whichunusual odours were used to enhance theimpact of the display. They asked whether re-presenting these Viking odours to individualswho had visited the museum a number of yearsago would allow them to remember moreabout the display than those presented withnon-Viking odours or a no-odour controlgroup. Mean memory performance for thosereceiving the Viking odour was numericallyhigher than for those receiving the non-Vikingodour and the no-odour control group.However, due to the expected large variationin performance, the difference failed to reachstatistical significance.This type of study design is getting closer to afully ecologically valid approach to Proustianphenomena. Naturally occurring experiencesare used and the association betweenexperience and ambient odour is passivelyformed. Several clear-cut results stronglysupport the proposition that odours a r eeffective reminders of autobiographicalexperience.The apparent emotional, vivid and deep-rootednature of odour-evoked memories can beexplained by two hypotheses. (1) The‘differential cue affordance value’ hypothesis


suggests that different sensory modalities varyin terms of cue affordance value (or theefficiency with which they accessautobiographical event details). Olfactory cuesare assumed to have a higher cue affordancethan other cue types. (2) The ‘differentialencoding bias’ hypothesis proposes that cuesdo not differ in terms of their efficacy inretrieving event details, but in terms of thetypes of event with which they becomeassociated and consolidated in memory.Olfactory details may become associated only,or preferentially, with more complex(emotional, personal, and unusual)autobiographical episodes. Consequently, thepresentation of an olfactory cue will biasretrieval towards more complex episodes.The encoding specificity principle (Tulving &Thomson, 1973) proposes that consolidatedmemory traces include not only targetinformation, but also salient contextualfeatures in the immediate environment inwhich the episode was experienced. Laterpresentation of such contextual features(which include odours) can therefore act asretrieval cues for related details that comprisethe original episode. If this alone were thecase, then all cue types would function in thesame manner and with equal effectiveness.However, assuming the superior ability ofolfactory cues to evoke emotionalautobiographical memories, encodingspecificity alone does not provide asatisfactory explanation. Here lies thechallenge.Olfaction is mediated by a number ofanatomical structures that are also heavilyimplicated in emotion and memory. Theolfactory bulb, for example, projects tostructures including the amygdala,hippocampus, and thalamus. These structureshave been shown to be involved in memoryfunction and the modulation of emotions.Proustian effects may also be linked to theability of odours to elicit affective reactions.Hinton and Healey (1993) compared reactionsto stimuli presented in visual, lexical, andolfactory modalities. Odours elicited by far themost affective reactions.Some (Herz, 1997; Aggleton & Waskett,1999) have ascribed the efficacy of odours inmemory retrieval, at least partially, to theacknowledged connection between emotionalarousal and the information associated withsuch affective reactions. This link is expected

to be mediated by the action of the amygdala(Cahill et al., 1995, 1996).Now equipped with a behavioural backgroundto odour-cued emotional memory, we move onto analyse the subject from a neurobiologicalperspective.

The amygdala and olfactoryprocessing

The amygdala refers to a highly differentiatedregion near the temporal pole of themammalian cerebral hemisphere. Burdachdiscovered and named the amygdala nucleus inthe early nineteenth century (Burdach, 1819-1822). Subsequent microscopic examinationbegan to reveal a progressively complexstructural differentiation. The extent of theamygdala’s outer border, and the number andclassification of its subdivisions, remainscontroversial today. Swanson and Petrovich(1998) suggest that the amygdala is neither astructural nor a functional unit. One part is aspecialized ventromedial expanse of thestriatum, a second part is caudal olfactorycortex (nucleus of the lateral olfactory tract,cortical nucleus, and postpiriform andpiriform-amygdalar areas) and a third part is aventromedial extension of the claustrum.Cell groups within the amygdala are derivedfrom different regions: their connectivity andthe distribution of neurotransmitters withineach grouping form the basis ofdifferentiation. Consequently, different groupscan be distinguished functionally into fourobvious systems – accessory olfactory, mainolfactory, autonomic and frontotemporalcortical.The amygdala receives highly processedsensory data and is in a position to integratethis data both within and across modalities.The only sensory modality with direct accessto the amygdala is smell (see figure two). Allother modalities connect with the amygdalavia a series of sensory association cortices.The amygdala can also influence autonomic,hormonal and motor function via itssubcortical connections.


Amygdala involvement in non-pheromonal forms of olfactoryrelated behaviour

Lesion studies involving rodents: Combinedelectrolytic lesions of the posterior lateralolfactory tract and anterior amygdala in ratsshowed no effect on the retention or learningof new olfactory discrimination tasks (Slotnik,1985). Bilateral amygdala lesions showed noeffect on learning of multiple odourdiscriminations (Eichenbaum et al., 1986).Failure of amygdala lesions to result inimpaired performance seems surprising given

the extensive connections of the olfactorysystem illustrated previously and the findingthat approximately 40% of rat amygdalaneurons respond to olfactory stimuli (Cain &Bindra, 1972). The lack of obvious deficitsappears to be caused, in part, by the existenceof parallel olfactory inputs to other regionsthat are able to counterbalance the destructionof amygdala olfactory routes. The olfactory

tests used to date must lack the requisitespecificity to pick up amygdala-dependentaspects of olfactory function.Electrophysiological studies: In one study 229neurons were recorded in the lateral andbasolateral nuclei of rats during olfactorydiscrimination learning (Schoenbaum et al.,1999). Of these, 60 were found to showselective responses depending on whether aparticular olfactory cue was paired with apleasant or an unpleasant outcome. Selectivityoften occurred early in learning and manyunits could be reversed, indicating that theamygdala neurons encoded the motivationalsignificance of the cues.

Immediate Early Gene (IEG) studies: c-fosimaging data supplements evidence forinvolvement of the basolateral amygdala(BLA) as olfactory learning progresses (Hesset al., 1997). Different stages of olfactorydiscrimination learning by rats were compared.Initial exploration (no specific odours) led toraised medial (medial, anterior, and posteriorcortical nuclei) and, to a lesser extent, raised

Main olfactory bulb

COApl TR PAA COAa

PA BLAp BMAp BMAa

Other main olfactory system Accessory olfactory system

(r) (r,i) (r) (d,r) SI

FS

ACB

CEAMEA

PB

LHAcl THpg

MDm

Medial prefrontal, agranular insular perirhinal, hippocampal cortex

Figure 2 Major neural inputs and outputs of the main olfactory system as they relate to the rat amygdala (adapted fromSwanson and Petrovich, 1998). Abbreviations: ACB, nucleus accumbens; BLAp, basolateral nucleus amygdala, posteriorpart; BMAa,p, basomedial nucleus amygdala, anterior, posterior parts; CEA, central nucleus amygdala; COAa,pl, corticalnucleus amygdala, anterior part, posterior part, lateral zone; d, medial hypothalamic defensive behavior system; FS, fundusof the striatum; i, medial hypothalamic ingestive behaviour system; LHAcl, lateral hypothalamic area, caudolateral part;MDm, mediodorsal nucleus thalamus, medial part; MEA, medial nucleus amygdala; PA, posterior nucleus amygdala; PAA,piriform-amygdala area; r, medial hypothalamic reproductive behavior system; SI, substantia innominata; THpg, thalamus,perigeniculate region (includes medial geniculate complex, posterior limiting nucleus, and parvicellular subparafascicularnucleus); TR, postpiriform transition area.


basolateral activation. In contrast, centralnucleus levels stayed low. Following initiationof cued olfactory discrimination training therewas high basolateral but relatively low medialactivation. In over-trained rats, the differencesin these amygdala regions were reduced.Balance of amygdala activation shifts in adynamic manner that depends on the demandsof the task and upon the stages of learning.Lesion studies involving primates: Thecontribution of the primate amygdala, and inparticular the human amygdala, to olfactoryprocessing remains poorly understood. Groupstudies have shown that comprehensive,unilateral temporal lobectomies can confuseolfactory identification (Rausch &Serafetinides, 1975) and the recall of odours(Rausch et al., 1977). However, theseexperiments provide limited direct evidencefor the specific involvement of the amygdala.As a rule, studies recording the effects ofsurgery to remove just the amygdala or partsof the amygdala have described no disruptionto the sense of smell (Scoville et al., 1953;Narabayashi, 1977), though these studies lackproper test details. A more systematic studyhas investigated odour discrimination in threegroups of unilateral temporal lobe resectioncases where the site of surgery was classed aseither (1) primarily neocortical, (2)amygdalohippocampectomy, or (3) anteriortemporal lobe with impingement on theamygdala or hippocampus (Jones-Gotman etal., 1997). All three groups were impaired,although no clear correlations could be drawnbetween the presence of an impairment andamygdala damage.Human electrophysiological studies: T h e s ehave confirmed the responsiveness ofamygdala neurons to odours (Andy & Jurko,1975) and shown that amygdala activity canlead to olfactory auras (Andy et al., 1975). Asmall body of single-case clinical studies alsoimplicates the amygdala in aspects of olfactoryprocessing. One case involves a woman whohad received an extensive, bilateralamygdalotomy that had involved theentorhinal cortex in one hemisphere (Andy etal., 1975). Although her detection of odourswas largely unchanged, her identification ofodours was impaired. Similarities are seen inpatient H.M., in whom the amygdala,entorhinal cortex, and hippocampus wereremoved bilaterally (Corkin et al., 1997).Odour discrimination, identification, and

matching are severely impaired, althoughH.M. can detect weak odours and appreciateodour intensity (Eichenbaum et al., 1983). Inboth instances, it is possible that thecombination of amygdala and entorhinalpathology is critical for identification. Theseand related case studies suggest that theinvolvement of the human amygdala inolfaction is often not made apparent by lesionstudies. Presumably, this reflects the presenceof multiple olfactory projection routes (i.e.parallel routes to orbitofrontal, thalamic, andother temporal regions), just as is seen inrodents. To further investigate involvement ofthe human amygdala in olfaction, the use offunctional imaging studies and thedevelopment of more selective tests isnecessary.Functional imaging studies: Zald and Pardo(1997) reported that exposure to a highlyaversive odourant produces a strong increasein cerebral blood flow as measured by positronemission tomography (PET) in bothamygdalae as well as in the left orbitofrontalcortex (see figure three).Interestingly, activity within the left amygdalawas associated significantly with subjectiveratings of aversiveness, suggesting thatamygdala activity is related to the perceivedhedonic value of the stimulus. Tests conductedwith pleasant smells (fruits, spices, and florals)only produced a non-significant increase inblood flow in the right anterioramygdala/periamygdaloid area. This studysuggests that amygdala activation occurs whenolfaction engages strong negative emotionalreactions.Aversive odours can affect amygdalaactivation via associative processes, as hasbeen shown by Schneider et al., (1999) in afunctional magnetic resonance imaging (fMRI)study. The presentation of a neutral face thathad previously been paired with an aversiveodour led to a decrease in amygdala activity ina group of 12 normal subjects, whereasincreased amygdala activation was found in agroup of 12 subjects suffering from socialphobia.Activation of the amygdala by aversive smellsand their associations may suggest other waysin which smells can influence memory.Aggleton’s real world study (Aggleton &Saunders, 2000) examined the long-termretention of specific smells by people who hadpreviously visited a Viking museum. People


who had visited the museum many yearspreviously almost always recalled that therehad been smells. However, only those thatwere independently rated as being most aversewere remembered consistently. It was alsofound that the same cues acted as contextualcues to help subjects to recall details of themuseum, despite having not visited it for manyyears (Aggleton & Waskett, 1999).

Intriguingly, this suggests that the amygdalaenhances learning when a subject isemotionally aroused, and that arousing smellsmight also be effective contextual cuesthrough a similar process.This pathway is a strong candidate for havinga role in the neural underpinnings of Proustianphenomena. An in depth look at arousal-dependent memory modulation by theamygdala follows.

The amygdala and modulation ofmemory storage: animal studies

There is now general agreement that theamygdala is involved in the learning ofemotionally arousing information (Cahill et

al., 1995). A substantial body of evidencesugges t s , however , tha t l a s t ingexplicit/declarative memory for emotionallyarousing information is stored in other brainregions and that the activation of the amygdalamodulates the storage of memory in thosebrain regions (e.g. Cahill & McGaugh 1998;Weinberger, 1995).

Hormonal modulation of memorystorage

Gerard (1961) proposed “…epinephrine…isreleased in vivid emotional experiences, suchthat an intense adventure should be highlymemorable.” He was the first to advance thishypothesis. Gold and van Buskirk (1975)found that systematic post-training injectionsof adrenaline administered to rats enhanceslong-term retention of inhibitory avoidance.These findings have been confirmed byexperiments using many different types oflearning tasks, all in support of the hypothesisthat endogenously released adrenalinemodulates memory storage (e.g. Liang, 2000;Williams, 2000). As with adrenaline, post-training administration of moderate doses of

Figure 3 (above and right) Cerebral activation duringaversive olfaction. Changes in regional cerebral blood flow(rCBF) are rendered in colour with white indicating thegreatest magnitude (Z score > 5) of activation. The relativepositions of coronal sections (A, B, and C) through thefrontal (a and b) and temporal (c) lobes are shownschematically to the right of this text. Maximal areas ofrCBF change are displayed superimposed on a standard T-1-weighted magnetic resonance image. The rCBF maximamap to the amygdala bilaterally and the left posteriorlateral orbitofrontal cortex. The right side of this figureshows the left side of the brain. Abbreviations: VCA,vertical line through anterior commissure; CACP,intercommissural line (reproduced with kind permissionfrom Zald & Pardo, 1997, © 1997 by PNAS).


glucocorticoids also induce dose- and time-dependent modulation of memory storage(Roozendaal & McGaugh, 1996a). Thesefindings strongly suggest that glucocorticoidsreleased by strongly emotionally arousingexperiences enhance the consolidation of thememory for those experiences (Roozendaal,

2000).

Amygda la med ia t ion o fneuromodulatory influences onmemory storage

Noradrenaline (NA) release in the brain mayregulate memory consolidation (Kety, 1972).Gold and van Buskirk’s finding (1978) thatadrenaline and footshock induce the release ofNA in the forebrain of rats has providedevidence consistent with Gerard and Kety’ssuggestions. Contemporary studies nowindicate that NA release within the amygdalais critically important in enablingneuromodulatory influences on memorystorage. Post-training intra-amygdala infusionsof β-adrenoceptor antagonist propranololblock adrenaline effects on memory storage(Liang et al., 1986) and post-training infusionsof NA or the β -adrenoceptor agonist

clenbuterol into the amygdala produces dose-dependent enhancement of memory storage(Hatfield & McGaugh, 1999). This stronglysupports the hypothesis that the amygdalaplays a critical role in memory consolidation.In the periphery, β-adrenoceptors are locatedon vagal afferents that project to the nucleus of

the solitary tract (NTS) (Schreurs et al., 1986).Projections from the nucleus of the solitarytract are known to release NA within theamygdala (Ricardo & Koh, 1978). Adrenalinedoes not readily pass through the blood brainbarrier. It seems likely that adrenaline effectsmemory storage by activation of peripheral β-adrenoceptors on vagal afferents. In support ofthis theory, inactivation of the NTS withlidocaine blocks adrenaline effects on memorystorage (Williams & McGaugh, 1993).Recent findings also indicate that theamygdala is involved in mediatingglucocorticoid influences on memory storage(Roozendaal, 2000). Lesions of the striaterminalis or of the amygdala block thememory enhancing effects of post-trainingsystematic inject ions of syntheticglucocorticoid dexamethasone (Roozendaal &McGaugh, 1996a,b). These glucocorticoid

Figure 4 Schematic summarizing the interactions of neuromodulatory influences in the basolateral amygdala onmemory storage. Abbreviations: α1 = α1-adrenoceptor; β1 = β1-adrenoceptor; cAMP = cyclic 3’5’ adenosinemonophosphate; GR = glucocorticoid receptor; NA = noradrenaline; NTS = nucleus of the solitary tract; OP =opioids (adapted from McGaugh et al., 2000).

Adrenaline

GABAergic agonists andantagonists

BASOLATERAL AMYGDALA

α- and β-adrenoceptor agonists andantagonists

Opioid agonists andantagonists

Peripherally acting β-adrenergicantagonists Corticosterone

vagusnerve

NTS

GR NAOP

STRIATERMINALIS?

Projections toother brainstructures

GABA cAMP

GR

β

α1

BLOOD-BRAIN BARRIER


effects involve noradrenergic activation of theamygdala.Noradrenergic activation in the amygdala isalso critical for other neuromodulatoryinfluences on memory storage. Post-trainingsystemic injections of opioid peptides andopiates generally impair memory storage andopiate antagonists enhance memory storage(Izquierdo & Diaz, 1983; McGaugh, 1989;McGaugh et al. , 1993). GABAergicantagonists and agonists administered post-training enhance and impair retention,respectively (Brioni & McGaugh, 1988;Brioni et al., 1989). Lesions of the amygdalaor stria terminalis block opioid peptidergic andGABAergic influences on memory storage(McGaugh et al., 1986). Also, and perhapsmost importantly, intra-amygdala infusions ofthe β-adrenoceptor antagonist propranololblock opioid peptidergic and GABAergiccontrol of memory storage (McGaugh et al.,1988). These neuromodulatory interactionswithin the amygdala are summarizedschematically in figure four.Studies that have not involved inhibitoryavoidance training have yielded similar resultsto those summarized above. Hence, there isevidence that amygdala influences on memorystorage are not restricted to the learning ofaversive information.

Modulation of long-term memory(LTM) in humans: adrenergicactivation and the amygdala

I begin here with the assumption that the brainmust possess a means to ‘weight’ or moderatethe actual information storage mechanisms inapproximate proportion to the importance ofthe information being stored (Gold &McGaugh, 1975). Without modulatorymechanisms, the brain would be unable todistinguish the important from the trivial inLTM. There appear to be, at minimum, twointeracting neurobiological componentsmediating the modulatory actions of emotionalarousal on memory: the adrenergic system andthe amygdala (especially BLA). An interactionbetween endogenous stress hormones activatedby emotional experiences and the amygdalaappears critical. According to animal studies,we can hypothesize that sufficientlyemotionally arousing (i.e. sympathetic nervous

system-activating) learning situations shouldengage amygdala participation in memoryformation, independently of whether theparticular emotions involved are positive ornegative. Hence, an emotionally tonedolfactory stimulus is ‘predisposed’ topreferential consolidation by the amygdala andtherefore improved storage.Although there is strong evidence in humansfor amygdala participation in LTM storage foremotionally arousing events, evidence for itsparticipation in memory retrieval, or for itsproduction of emotion, per se, is currently lessclear.

Emotionally influenced LTM: therole of the adrenergic system

Propranolol (a synthetic β-adrenergic receptorblocking agent) attenuates the enhanced LTMassociated with emotionally arousinginformation without affecting memory formore emotionally neutral information (Cahillet al., 1994). Hence, activation of theadrenergic system in humans is required forthe enhancing effect of arousal (eitheremotionally or physically induced) onmemory. The action of propranolol ismediated centrally (van Stegeren et al., 1998).Adrenergic stimulation can act in a retrogrademanner to enhance memory in humans(Soetens et al., 1995).

Emotionally influenced LTM: therole of the amygdala

Scoville and Millner examined memory in tenpatients, including some with relativelyselective amygdala damage and concluded“removal of the amygdala bilaterally does notappear to cause memory impairment”(Scoville & Millner, 1957). However, theirtests involved short-term memory, and/ormemory for non-emotionally arousingmaterial. These findings fit well with the‘memory modulation’ view of amygdalafunction, which predicts that amygdala activityshould not be critical for STM, or for LTM ofnon-emotionally arousing events or material.A rare patient with damage confined almostexclusively to her amygdala reported deficits


in memory for emotional material (Babinsky etal., 1993). Enhanced LTM associated with anemotionally arousing story is impaired inpatients with selective (Cahill et al., 1995) ornearly selective (Adolphs et al., 1997)amygdala damage. Amnesic patients withundamaged amygdalae show reasonably intactenrichment of memory for emotional datadespite their overall impaired memoryperformance (Hamann et al., 1997). Theemotional reactions of patients with amygdaladamage to emotionally provocative stimuli arenot significantly different from those ofcontrols even though LTM for emotionalmaterial is impaired (Cahill et al., 1996).Electrodermal responses to emotionallystressful events are intact in humans withamygdala damage (Adolphs et al., 1997). Onthe basis of these and PET studies described,Cahill et al., (1996) propose that amygdalaactivity in humans “may be more importantfor the translation of an emotional reactioninto heightened recall than it is for thegeneration of an emotional reaction per se.”Brain imaging studies: In one study, subjectsreceived two PET scans: one while viewing aseries of relatively emotionally arousing films,another while viewing a series of relativelyemotionally neutral films (Cahill et al., 1996).Memory for the films was tested in a surprisefree recall test three weeks later. Activity inthe right amygdala correlated very highly(r=0.93) with the number of emotional filmsrecalled three weeks later, but not with recallof the neutral films (see figure five).

Hamann and colleagues (1999) used PET tostudy cerebral blood flow while subjects

observed either pleasant or unpleasantarousing images. Memory for the images onemonth later correlated very highly withamygdala activity while viewing either thepleasant or unpleasant images. Amygdalaactivity whilst viewing emotionally neutralimages (including images of novel, but notemotionally arousing objects) did not correlatewith recall. The arousing dimension of thepictures was essential to amygdalaengagement in memory, as indicated inanother recent fMRI study (LaBar et al.,1998).Although activity of both the left and rightamygdala correlated with recall of arousingmaterial, the degree appeared greater on theright for negative slides (Hamann et al., 1999).Cahill et al. (1996), who used emotionallynegative material and found a unilateral effect,also suggest that differences in left versus rightamygdala participation in memory foremotionally arousing events may relate to theirpleasant/unpleasant nature. Left amygdalaactivation appears to be produced consistentlyby any of a variety of aversive stimuli used inhuman brain imaging studies (Cahill &McGaugh, 1998; Zald & Pardo, 1997, 2000 forolfactory examples). No compellingexplanation exists at present for suchasymmetries in amygdala function.Canli and colleagues (1999) reported thatamygdala activity in response to emotionallynegative pictures correlated with long-term (2-14 months) retention of pictures. In anindividual fMRI study conducted in

collaboration with the above, it was found thatincreasing the degree of emotional arousal

Figure 5 (A) Amygdala activity while watching a series of emotionally arousing films correlated very highly with long-term (three-week) recall of the films. (B) Amygdala activity in the same subjects while viewing a series of relativelyemotionally neutral films did not correlate significantly with recall (reproduced with kind permission from Cahill et al.,1996, © 1996 by PNAS).


induced by emotional pictures increases thedegree to which amygdala activity iscorrelated with long-term recall of pictures(Canli et al., 2000).In conclusion, evidence from several humanbrain imaging studies, including both PET andf M R I , c o n v e r g e s w i t h b o t hneuropsychological data and findings fromanimal research to support the view that theamygdala is involved preferentially with long-term storage of explicit memory foremotionally arousing events.

Does the amygdala participate inacquisition, retrieval or both?

The imaging studies discussed above clearlyi m p l i c a t e t h e a m y g d a l a i nencoding/consolidating processes, but do notaddress its role in the retrieval process. Somestudies report evidence consistent witha m y g d a l a p a r t i c i p a t i o n i n

encoding/consolidating but not retrievalprocesses. Two fMRI studies of amygdalainvolvement in aversive classical conditioningreported that the amygdala was active inresponse to a conditioned stimulus duringlearning of a conditioned association, but notonce the association was learned (Büchel eta l ., 1998; LaBar et al., 1998). A PETinvestigation reported amygdala activationwhile subjects initially viewed emotionalpictures, but not when retrieval of the pictureswas tested later (Taylor et al., 1998). Similarly

viewing, and presumably forming memoriesof, emotionally arousing films has beenreported to activate the amygdala (assessedwith PET), although retrieval of previouslyexperienced events did not (Reiman et al.,1997).Other reports do indicate activation ofamygdala during retrieval of emotional events.For example, Fink et al., (1996) reportedunilateral amygdala (and temporal lobe)

Figure 6 Functional anatomy of temporal activations during affect-laden autobiographical memory, as indicated by PET-measured changes in rCBF. This figure details the functional anatomy of temporal activations associated withautobiographical memory and their relationship to underlying anatomy. N.B. Activations are predominantly on the right(left image corresponds to subject’s left) and include temporomedial, temporolateral, and insular cortices. Red-cross hairindicates local maximum within area of activation (reproduced with kind permission from Fink et al., 1996, 1996 by theSociety for Neuroscience).


activation in subjects retrieving emotionallyladen autobiographical memories (see figuresix). Others report amygdala activation inpeople with post traumatic stress disorder(PTSD) when they recall the distressful eventthat caused their disorder (Rauch et al., 1996).There are many confounding variables in thistype of imaging study. In the study by Finkand coworkers (1996), the relative rCBFdifferences were monitored in subjects readingeither sentences containing affect-ladenautobiographical information or sentencescontaining semantically similar episodic butnon-autobiographical information. Materialpresented during the autobiographical phasewas (1) familiar to the subjects for many years,(2) affectively more significant, and (3) morepersonal. Any or all of these factors could be(partly) responsible the activations described.When considering evidence of amygdalaparticipation in retrieval, it is essential todistinguish between retrieval with or withoutconcomitant emotional arousal. According the‘memory modulation’ hypothesis, theamygdala should influence memory storageprocesses whenever a situation is sufficientlyemotionally arousing - independent of whetherthe emotion is produced by an external or aninternal event (such as a memory that inducesan emotional response, Cahill & McGaugh,1998). Indeed, Fink et al., (1996) reported thattheir subjects probably experienced substantialemotional arousal during retrieval. Similarly,retrieval of traumatic events in PTSD patientsis generally very emotionally stressful. Thusamygdala activation during retrieval occurringwith a concomitant emotional response may infact reflect amygdala participation in memorystorage, rather than retrieval, processes. Toclarify this issue, future studies need to bedesigned to determine whether concomitantemotional arousal is critical to amygdalaactivation at the time of retrieval.Recent imaging studies provide clues as to

potential brain regions with which theamygdala interacts to effect memory.Activation of the cholinergic basal forebrain(the ‘substantia innominata’) correlatedsignificantly with recall of emotional but notneutral films (Cahill et al., 1996). Animalresearch incriminates the hippocampal regionas a likely site of modulation by the amygdala(Packard & Teather, 1998). Data suggestingconcurrent amygdala and hippocampal regionactivation during emotionally arousing events(Cahill et al., 1996; Hamann et al., 1999) areconsistent with, while definitely do not prove,the idea that the amygdala affects memory viainfluences on the hippocampal region.Currently available imaging data, combinedwith neuropsychological investigations insingle cases with selective retrograde amnesia,suggest right hemispheric temporal regions asbeing key for autobiographical memoryecphory (the process by which retrieval cuesinteract with stored information so that animage or a representation of the information inquestion appears). Surrounding righthemispheric ‘satellite’ regions such as theamygdala, hippocampus-parahippocampus,posterior cingulate, and insular and prefrontalcortex support these regions. Such satelliteregions might be responsible for contributingadditional information, such as the emotional-affective components of a memory.

Modulatory circuits in olfactoryperception

Savic (2002) suggests that perception of odourcompounds is mediated by a set of coreregions (see figure seven). Depending on thetask associated with odour perception, coreregions are associated together with othercircuits in a parallel and hierarchical manner.Interestingly, some of the structures involvedhere correspond with those involved in

Figure 7 A hierarchical and parallel system for odourantprocessing based on data from tests of different cognitiveload. Regions shared by several olfactory tasks and theregions specific to a particular task are depicted, witheach pyramid indicating a separate task. The higher upthe pyramid, the more complex the task. The grey zoneindicates the olfactory core regions, activated by passivesmelling of odours. Abbreviations: OM, odour-recognition memory task; OD-I, discrimination of odourintensity task; OD-q, discrimination of odour qualitytask; Bil, bilateral; L, left; R, right (reproduced with kindpermission from Savic 2002).


autobiographical memory ecphory.

Discussion

In this section I propose a unification betweenvarious neural mechanisms, outlining some ofthe processes that may be involved inencoding and retrieval of emotionally tonedodour-evoked memories. Figures eight andnine aim to bring together a portion of thelarge and diverse body of evidence related toolfaction, emotion, and the amygdala.So why do odours act so effectively as cues forcertain emotionally toned, vivid, and oldautobiographical memories? At encoding, thesubject is emotionally aroused due to thecontext, resulting in systemic release ofadrenaline and corticosterone. Noradrenergicactivation of the BLA results in thesituation/memory being labelled as a priorityfor LTM storage. Hence, the episode ispreferentially encoded in brain areas via thestria terminalis. However, this arrangementalone does not suffice as an explanatoryconcept as to why odours are better than othersensory modalities at evoking those memoriesthat are particularly emotionally toned.Here, I think that it must be related to thedirect input of olfactory stimuli into theamygdala. There is a connection from the mainolfactory bulb to the post piriform area to theBLA. Do olfactory stimuli (present at the timeof arousal) therefore have the opportunity tobecome better associated with these emotionalmemories than stimuli from other sensorymodalities (which must first travel through aseries of cortical association areas beforeeventually reaching the amygdala)? The endresult of this encoding process is the formationof a memory trace with very strong olfactoryassociations. This is in agreement with the‘differential encoding bias’ hypothesiswhereby cues do not differ in terms of theirefficacy in retrieving event details, but interms of the types of event with which theybecome associated and consolidated inmemory.At retrieval, the presence of the same odour asthat present at encoding cues the recall of theoriginal episode. As this memory is soemotionally toned, sympathetic activationoccurs and there is concomitant retrieval ofmemory and emotion. Recall is centred on

right hemispheric temporal cortical areas, withsatellite regions contributing additionaldimensions to the retrieval experience. Theright amygdala is a common component inboth olfactory perception pathways andemotionally toned autobiographical memoryretrieval.

Future research direction

To test the idea described above, there is aneed to design experimental paradigms thatinvestigate whether odourant information doesindeed become more strongly associated withemotionally toned memories that aremodulated by the BLA.It would be interesting to see whether a personwith bilateral amygdala damage were able toappreciate the emotional aspect of odourschosen to evoke emotional responses. Thiswould indicate whether the amygdala isnecessary for the generation of emotion forolfactory stimuli. In addition, the subject oflaterality of amygdala function in relation toemotion and memory is evidently an importantarea for future research.Pragmatically, techniques currently used toinvestigate amygdala function are not capableof sufficient anatomical resolution orspecificity to distinguish between subregions.Improvements in spatial and temporalresolution are required.Experiments examining the covariance ofamygdala activity with additional brainregions in the context of emotionally arousingversus emotionally neutral situations shouldaid identification of a network of structuresthrough which the amygdala modulatesmemory.

Clinical applications

Several clinical disorders involve olfactoryimpairments and/or pathology in structuresassociated with higher level olfactoryprocessing. For instance, the olfactory systemshares a common neural substrate with manyof the cognitive and emotion processes that areabnormal in schizophrenia. Schizophrenicpatients exhibit abnormalities in certain


ENCODING

OE

Peripheral neuralmechanisms ofolfaction

Example of context of episodeduring encoding: As a child, thesubject is having an argument, isangry and ∴ emotionally aroused,resulting in release of adrenalineand corticosterone. The room hasbeen cleaned with a particularbrand of cleaning product that hasan unusual odour.N.B. In the case of Proustphenomena, emotional arousal isprobably primarily due not to anyhedonic dimension attributed tothe odour(s) present at encoding– but to the rest of the context.However, a particularly arousingodour could also have acontributory effect.

Peripheral releaseof adrenaline andcorticosterone viaHPA axis

Corticosterone

Adrenaline

GRGR NA

_

_1

cAMP

NTSStriaterminalis

Projectionsto otherbrainstructures

BBB

COApl TR PAA COAa

PA BLAp BMAp BMAa

LHAcl

PB

THpg

MDm

ACB

CEAMEA

VAGUSNERVE

BLA

FS

Other main olfactorysystem,Accessory olfactorysystem

Enlargement

Olfactory datainputs from theMOB via the TR

Cranial Nerve I

(r) (r,i) (r) (d,r)

MOB

Figure 8 Novel schematic illustrating some key components of the encoding process employed during the generation ofProustian memories. See discussion for details. Abbreviations: α1, α1-adrenoceptor; ACB, nucleus accumbens; β, β-adrenoceptor; BBB, blood brain barrier; BLAp, basolateral nucleus of amygdala- posterior part; BMAa,p, basomedialnucleus of amygdala– anterior, posterior parts; CEA, central nucleus of amygdala; COAa,pl, cortical nucleus of amygdala–anterior, posterior lateral parts; d, medial hypothalamic defensive system; FS, fundus of striatum; GR, glucosorticoidreceptor; LHAcl, lateral hypothalamic area – caudolateral part; MDm, mediodorsal nucleus thalamus; MOB, main olfactorybulb; MEA, medial nucleus of amygdala; NA, noradrenaline; NTS, nucleus of the solitary tract; OE, olfactory epithelium;PA, posterior nucleus amygdala; PAA, piriform-amygala area; PB, parabrachial nucleus; r, medial hypothalamicreproductive behaviour system; THpg, thalamus perigeniculate region; TR, postpiriform transition area.

To main olfactorysystem component ofamygdala

Olfactory datainputs becomemore stronglyassociated withthe episodeduring encoding

Odorant


Context of episode duringretrieval: Many years later,the individual comes across abroken bottle of the samecleaning product. The odourevokes the memory of thepast event (the argument)including i ts emotionaldimension.

Retrieval of autobiographicalmemory is dependent on arestricted key network ofp r e d o m i n a n t l y r i g h themispher ic tempora lcortical areas. Surroundingsatellite regions contributese lec t i ve ly to o therdimensions of the memoryretrieval.

RETRIEVAL

OE

Peripheral neuralmechanisms ofolfaction

MOB

Temporalcorticalareas

Amygdala

Posteriorcingulate

Prefrontalcortex

Hippocampus-parahippocampus

Insula

All structuresare right

hemispheric

Figure 9 Schematic illustrating some key components of the retrieval process employed during the recall of Proustianmemories. See discussion for further details.

Recall of emotionallycharged autobiographical

memory

Concomitantre-experience

of emotion

Brain regions involved inolfactory perception

Brain regionsinvolved inaffectively tonedautobiographicalmemory ecphory

Overlapping regions

Odorant


components of the olfactory evoked potential(Turetsky et al., 2003). Additionally, abnormalolfactory identification has been found inindividuals with obsessive compulsivedisorder (Goldberg et al., 1991). Continuedresearch and improved understanding of theneural underpinnings of higher level olfactoryprocessing will improve understanding ofthese disorders.One potential clinical effect of the proposedamygdala/adrenergic system memory-modulating mechanism is worth considering.The formation of excessively intrusivememories characteristic of PTSD may occur asa result of over-activation of this typicallyadaptive mechanism. If so, appropriatetreatment with β-blocking drugs immediatelyafter a person has experienced an emotionaltrauma should reduce/prevent the developmentof PTSD. Results from one small-scale clinicaltrial (Pitman et al., 2002) support the clinicalpossibility that a course of propranolol begunshortly after an acute traumatic experience iseffective in reducing PTSD symptoms 1 monthlater. Definitive conclusions, however, mustawait adequately powered studies with largersamples.

Conclusion

Odours are capable of evoking very emotionalmemories because of the way that suchmemories are encoded originally. Due to aspatial and temporal coincidence in the BLA,olfactory information has the capacity tobecome more strongly associated withemotional episodes. Consequently, only odourstimuli have the capacity to retrieve them.Returning to Proust, some important pointsneed to be clarified. When an odour evokes aProustian (emotionally charged) memory thisinvolves recollection with concomitant re-experience of the emotional side of thememory. It is not the case that the odour itselfis evoking an emotion; instead it evokes amemory with a strong emotional component.Similarly, at encoding, enhanced LTM storagemost probably occurs principally as a result ofemotional arousal unrelated to any hedonicdimensions attributable to the odour present atthe time. However, the presence of an odourthat is itself emotionally arousing wouldevidently contribute.

Chemical sense was originally used to moveaway from ‘bad’ (predators) and towards‘good’ (food). In evolutionary terms, limbicstructures grew out of the olfactory system.Consequently, the emotional dichotomybetween bad (death, danger, failure) and good(survival, love, reproduction) echoes thechemosensory one. Perhaps emotion beganmerely as an abstracted façade of primitiveolfactory instructions. This would explain whyodour has such a potent emotional cascade.

References

Adolphs R, Cahill L, Schul R, Babinsky R(1997) Impaired declarative memory foremotional material following bilateralamygdala damage in humans. Learning andMemory 4, 291-300.

Aggleton JP & Waskett L (1999) The abilityof odours to serve as state-dependent cuesfor real-world memories: can Viking smellsaid the recall of Viking experiences? BritishJournal of Psychology 90, 1-7.

Aggleton JP & Saunders RC (2000) TheAmygdala – what’s happened in the lastdecade? In: The Amygdala: A FunctionalAnalysis, ed. Aggleton JP, pp. 1-30, Wiley-Liss, New York.

Andy OJ & Jurko MF (1975) The humanamygdala: excitability state and aggression.In: Neurosurgical Treatment in Psychiatry,Pain, and Epilepsy, ed. Sweet WH, ObradorS, Martin-Rodriguez JG, pp. 417-427,University Park Press, Baltimore.

Andy OJ, Jurko MF, Hughes JR (1975) Theamygdala in relation to olfaction. ConfiniaNeurologica 37, 215-222.

Babinsky R, Calabrese P, Durwen HF,Markowitsch HJ, Brechtelsbauer D, HeuserL, Gehlen W (1993) The possiblecontribution of the amygdala to memory.Behavioural Neurology 6, 167-170.

Brioni JD & McGaugh JL (1988) Posttrainingadministration of GABAergic antagonistsenhances retention of aversively motivatedtasks. Psychopharmacology 96, 505-510.

Brioni JD, Nagahara AH, McGaugh JL (1989)Involvement of the amygdala GABAergicsystem in the modulation of memorystorage. Brain Research 487, 105-112.

Büchel C, Morris J, Dolan RJ, Friston KJ(1998) Brain systems mediating aversive


conditioning: an event-related fMRI study.Neuron 20, 947-957.

Burdach KF (1819-1822) Vom Baue undLeben des Gehirns, Leipzig.

Cahill L, Prins B, Weber M, McGaugh JL(1994) β-Adrenergic activation and memoryfor emotional events. Nature 371, 702-704.

Cahill L, Babinsky R, Markowitsch HJ,McGaugh JL (1995) The amygdala andemotional memory. Nature 377, 295-296.

Cahill L, Haier RJ, Fallon J, Alkire MT, TangC, Keator D, Wu J, McGaugh JL (1996)Amygdala activity at encoding correlatedwith long-term, free recall of emotionalinformation. Proceedings of the NationalAcademy of Science USA 93, 8016-8021.

Cahill L (1997) The neurobiology ofemot ional ly inf luenced memory:implications for the treatment of traumaticmemory. Annals of the New York Academyof Sciences 821, 238-246.

Cahill L & McGaugh JL (1998) Mechanismsof emotional arousal and lasting declarativememory. Trends in Neuroscience 21, 294-299.

Cain DP & Bindra D (1972) Responses ofamygdala single units to odors in the rat.Experimental Neurology 35, 98-110.

Canli T, Zhao Z, Desmond JE, Glover G,Gabrieli JDE (1999) fMRI identifies anetwork of structures correlated withretention of positive and negative emotionalmemory. Psychobiology 27, 441-452.

Canli T, Zhao Z, Brewer J, Gabrieli JDE,Cahill L (2000) Event-related activation inthe human amygdala associates with latermemory for individual emotionalexperience. Journal of Neuroscience 20,RC99 1-5.

Chu S & Downes JJ (2000) Odour-evokedautobiographical memories: psychologicalinvestigations of Proustian phenomena.Chemical Senses 25, 111-116.

Corkin S, Amaral DG, Gonzalez RG, JohnsonKA, Hyman BT (1997) H.M.’s medialtemporal lobe lesion: findings from magneticresonance imaging. Journal of Neuroscience17, 3964-3979.

Eich JE (1985) Context, memory andintegrated item/context imagery. Journal ofExperimental Psychology: Learning,Memory, & Cognition 11, 764-770.

Eichenbaum H, Morton TH, Potter H, CorkinS (1983) Selective olfactory deficits seen incase H.M. Brain 106, 459-472.

Eichenbaum H, Fagan A, Cohen NJ (1986)Normal olfactory discrimination learning setand facilitation of reversal learning aftermedial-temporal damage in rats:implications for an account of preservedlearning abilities in amnesia. Journal ofNeuroscience 6, 1876-1884.

Fink GR, Markowitsch HJ, Reinkemeier M,Bruckbauer T, Kessler J, & Heiss W-D(1996) Cerebral representation of one’s ownpast: neural networks involved inautobiographical memory. Journal ofNeuroscience 16, 4275-4282.

Gerard RW (1961) The fixation of experience.In: Brain Mechanisms and Learning,consulting ed. Fessard A, Gerard RW,Konorski J, pp. 21-35, Charles C. Thomas,Springfield, Illinois.

Gilbert AN & Wysocki CJ (1987) The smellsurvey results. National Geographic 172,514-525.

Gold PE & van Buskirk R (1975) Facilitationof time-dependent memory processes withpost trial epeniphrine injections. BehaviouralBiology 13, 145-153.

Gold PE & van Buskirk R (1978) Posttrainingbra in norepeniphrine concentrations:correlation with retention performance ofavoidance training with peripheralepeniphrine modulation of memoryprocessing. Behavioural Biology 23, 509-520.

Gold PE & McGaugh JL (1975) A single-trace, two process view of memory storageprocesses. In: Short-term Memory, ed.Deutsch J, Deutsch D, pp. 355-378,Academic Press, New York.

Goldberg ED, Goldberg RJ, Vannopen B(1991) Sense of smell and obsessionalbehaviour. American Journal of Psychiatry148, 1757.

Hamann S, Cahill L, McGaugh JL, Squire LR(1997) Intact enhancement of declarativememory by emotional arousal in amnesia.Learning and Memory 4, 301-309.

Hamann S, Elt T, Grafton S, Kilts C (1999)Amygdala activity related to enhancedmemory for pleasant and unpleasant aversivestimuli. Nature Neuroscience 2, 289-293.

Hatfield T & McGaugh JL (1999)Norepinephrine infused into the basolateralamygdala posttraining enhances retention ina spatial watermaze task. Neurobiology ofLearning and Memory 71, 232-239.


Herz RS (1997) Emotion experienced duringencoding enhances odor retrieval cueeffectiveness. American Journal ofPsychology 110, 489-505.

Herz RS & Cupchik GC (1992) Anexperimental characterization of odor-evoked memories in humans. ChemicalSenses 17, 519-528.

Herz RS & Cupchik GC (1995) The emotionaldistinctiveness of odor-evoked memories inhumans. Chemical Senses 20, 517-528.

Herz RS & Schooler JW (2002) A naturalisticstudy of autobiographical memories evokedby olfactory and visual cues: testing theProustian hypothesis. American Journal ofPsychology 115, 21-32.

Hinton PB & Healey TB (1993) Cognitive andaffective components of stimuli presented inthree modes. Bulletin of the PsychonomicSociety 31, 595-598.

Izquierdo I & Diaz RD (1983) Effect ofACTH, epinephrine, β-endorphin, naloxone,and of the combination of naloxone or β-endorphin with ACTH or epinephrine onm e m o r y c o n s o l i d a t i o n .Psychoneuroendocrinology 8, 81-87.

Jackson JH (1880) On right- or left-sidedspasm at the onset of epileptic paroxysms,and on crude sensation warnings, andelaborate mental states. Brain 3, 192-206.

Jones-Gotman M, Zatorre RJ, Cendes F,Olivier A, Andermann F, McMacking D,Staunton H, Siegel AM, Wieser, H-G (1997)Contribution of medial versus lateraltemporal lobe structures to human odouridentification. Brain 120, 1845-1856.

Kety S (1972) Brain catecholamines, affectivestates and memories. In: The Chemistry ofMood, Motivation, and Memory, edMcGaugh JL, pp. 65-80, Plenum Press, NewYork.

LaBar KS, Gatenby JC, Gore JC, LeDoux JE,Phelps EA (1998) Human amygdalaactivation during conditioned fearacquisition and extinction: a mixed-trialfMRI study. Neuron 20, 937-945.

Laird DA (1935) What can you do with yournose? Scientific Monthly 41, 126-130.

Liang KC, Juler R, McGaugh JL (1986)Modulating effects of posttrainingepinephrine on memory: involvement of theamygdala noradrenergic system. BrainResearch 368, 125-133.

Liang KC (2000) Epinephrine modulation ofmemory: amygdala activation and regulation

of long-term storage. In: Four Decades ofMemory: A Festschrift Honouring James L.McGaugh, ed. Gold PE & Greenough WT,American Psychological Association,Washington, DC.

McGaugh JL (1989) Involvement of hormonaland neuromodulatory systems in theregulation of memory storage. AnnualReview of Neuroscience 12, 255-287.

McGaugh JL, Introini-Collison IB, Juler RG,Izquierdo I (1986) Stria terminalis lesionsattenuate the effects of posttraining naloxoneand β-endorphin on retention. BehavioralNeuroscience 100, 839-844.

McGaugh JL, Introini-Collison IB, NagaharaAH (1988) Memory-enhancing effects ofposttraining naloxone: involvement of β-noradrenergic influences in the amygdaloidcomplex. Brain Research 446, 37-49.

McGaugh JL, Introini-Collison I, Castellano C(1993) Involvement of opioid peptides inlearning and memory. In: Handbook ofExperimental Pharmacology, Opioids, Parts1 and 2, ed. Herz A, Akil H, Simon EJ, pp.419-477, Springer-Verlag, Heidelberg.

McGaugh JL, Ferry B, Vazdarjanova A,Roozendaal B (2000) Amygdala: role inmodulation of memory storage. In: TheAmygdala: A Functional Analysis, ed. J.P.Aggleton, pp. 1-30, Wiley-Liss, New York.

Narabayashi H (1977) Stereotaxicamygdalectomy for epileptic hyperactivity-long range results in children. In: Topics inChild Neurology, ed. Blaw ME, Rapin I,Kinsbourne M, pp. 319-331, SpectrumPublications, New York.

Packard M & Teather L (1998) Amygdalamodulation of multiple memory systems:hippocampus and caudate-putamen.Neurobiology of Learning and Memory 69,163-203.

Pitman RK, Sanders KM, Zusman RM, HealyA R, Cheema F, Lasko NB, Cahill L, Orr SP(2002) Pilot study of secondary preventionof posttraumatic stress disorder withpropranolol. Biological Psychiatry 51, 189-192.

Proust M (Scott Monkrieff trans)(1922/1960)In: Swann’s Way, Part 2. pp. 58, Chatto &Windus, London.

Rauch SL, van der Kolk BA, Fisler RE, AlpertNM, Orr SP, Savage CR, Fischman AJ,Jenike MA, Pitman RK (1996) A symptomprovocation study of posttraumatic stressdisorder using positron emission


tomography and script-driven imagery.Archives of General Psychiatry 53, 380-387.

Rausch R & Serafetinindes EA (1975) Specificalterations of olfactory function in humanswith temporal lobe lesions. Nature 255, 557-558.

Rausch R, Serafetinindes EA, Crandall PH(1977) Olfactory memory in patients withanterior temporal lobectomy. Cortex 13,445-452.

Reiman EM, Lane RD, Ahern GL, SchwartzGE, Davidson RJ, Friston KJ, Yun LS, ChenK (1997) Neuroanatomical correlates ofexternally and internally generated humanemotion. American Journal of Psychiatry154, 918-925.

Ricardo J & Koh E (1978) Anatomicalevidence of direct projections from thenucleus of the solitary tract to thehypothalamus, amygdala, and otherforebrain structures in the rat. BrainResearch 153, 1-26.

Roozendaal B & McGaugh JL (1996a)Amygdaloid nuclei lesions differentiallyaffect glucocorticoid-induced memoryenhancement in an inhibitory avoidancetask. Neurobiology of Learning and Memory65, 1-8.

Roozendaal B & McGaugh JL (1996b) Them e m o r y - m o d u l a t o r y e f f e c t s o fglucocorticoids depend on an intact striaterminalis. Brain Research 709, 243-250.

Roozendaal B (2000) Glucocorticoids and theregulation of memory consolidation.Psychoneuroendocrinology 25, 213-238.

Rubin DC, Groth E, Goldsmith DJ (1984)Olfactory cuing of autobiographicalmemory. American Journal of Psychology97, 493-507.

Savic I (2002) Imaging of brain activation byodorants in humans. Current Opinion inNeurobiology 12, 455-461.

Schab FR (1990) Odors and the remembranceof things past. Journal of ExperimentalPsychology: Learning, Memory, &Cognition 16, 648-655.

Schneider F, Weiss U, Kessler C, Muller-Gartner HW, Posse S, Salloum JB, GroddW, Himmelman F, Gaebel W, Birbaumer N(1999) Subcortical correlates of differentialclassical conditioning of aversive emotionalreactions in social phobia. BiologicalPsychiatry 45, 863-871.

Schoenbaum G, Chiba AA, Gallagher M(1999) Neuronal encoding in orbitofrontal

cortex and basolateral amygdala duringolfactory discrimination learning. Journal ofNeuroscience 19, 1876-1884.

Schreurs J, Seelig T, Schulman H (1986) β2-Adrenergic receptors on peripheral nerves.Journal of Neurochemistry 46, 294-296.

Scoville WB, Dunsmore RH, Liberson WT,Henry CE, Pepe E (1953) Observations onmedial temporal lobotomy and uncotomy inthe treatment of psychotic states.Association Research into Nervous andMental Disorders 31, 347-373.

Scoville WB & Millner B (1957) Loss ofrecent memory after bilateral hippocampallesions. Journal of Neurology, Neurosurgery,and Psychiatry 20, 11-21.

Slotnik BM (1985) Olfactory discrimination inrats with anterior amygdala lesions.Behavioural Neuroscience 99, 956-963.

Soetens E, Casaer S, D’Hooge R, Hueting JE(1995) Effect of amphetamine on long-termre t en t i on o f ve rba l ma t e r i a l .Psychopharmacology 119, 155-162.

Swanson LW & Petrovich GD (1998) What isthe amygdala? Trends in Neuroscience 21,323-331.

Taylor S, Liberzon M, Decker L, MinoshimaS, Koeppe R (1998) The effect of emotionalcontent on visual recognition memory: aPET activation study. Neuroimage 8 , 188-197.

Tulving E & Thomson D (1973) Encodingspecificity and retrieval process in episodicmemory. Psychological Review 80, 352-373.

Turetsky BI, Moberg PJ, Owzar K, JohnsonSC, Doty RL, Gur RE (2003) Physiologicimpairment of olfactory stimulus processingin schizophrenia. Biological Psychiatry 53,403-411.

van Stegeren A, Everaerd W, Cahill L,McGaugh JL, Goeren L (1998) Memory foremotional events: differential effects ofcentrally versus peripherally acting beta-blocking agents. Psychopharmacology 138,305-310.

Weinberger NM (1995) Retuning the brain byfear conditioning. In: The CognitiveNeurosciences, ed. Gazzaniga MS, pp. 1071-1089 , MIT Press , Cambr idge ,Massachusetts.

Williams CL & McGaugh JL (1993)Reversible lesions of the nucleus of thesolitary tract attenuate the memory-modulating effects of posttraining


epinephrine. Behavioural Neuroscience 107,955-962.

Williams CL (2000) Brainstem contributionsto memory formation. In: Four Decades ofMemory: A Festschrift Honouring James L.McGaugh, ed. Gold PE & Greenough WT,American Psychological Association,Washington, DC.

Zald DH & Pardo JV (1997) Emotion,Olfaction, and the Human Amygdala:Amygdala activation during aversiveolfactory stimulation. Proceedings of theNational Academy of Science 94, 4119-4124.

Zald DH & Pardo JV (2000) Functionalneuroimaging of the olfactory system inhumans. International Journal ofPsychophysiology 36, 165-181.

Acknowledgements

I thank G Arbuthnott, K Campbell, NMcLeod, and E Wood for comments andChlo_.

Corresponding Author

Mark HughesUniversity of Edinburgh

[email protected] for Neuroscience ResearchAppleton Tower Level 7George SquareEdinburghEH8 9LEUK

olfaction, emotion & the amygdala: arousal-dependent modulation

Documents