Scentair.cl

European
Psychologist
Anoma lies in Cog nition: Olfactory Memory Anomalies in Cognition:
Olfactory Memory
Gesualdo M. Zucco
Department of General Psychology, University of Padova, Italy, and Department of Cognitive Sciences, San Raffaele
University, Milan, Italy
The two experiments presented in this paper examine the effects of strat- stimuli (environmental sounds). In the four sessions of the experiment, egies and interference tasks on odor recognition. In the first experiment they had to recognize the stimuli whether in a no-interference condition, (an extension of Lyman and McDaniel's study from 1986), participants or in an intramodality, or in two intermodalities interfering conditions. were asked to smell 30 odors and to perform different elaborative tasks for Consistently with the literature, interference affects recognition for vis- each of them such as: (1) providing a name or a short definition; (2) ual and acoustic material but has no effect on odor recognition. creating an image; (3) describing a specific life episode; (4) simply smell- The results of both experiments and some other anomalies in olfactory ing the odors. Results showed no effect of encoding tasks on the correct memory are discussed and tentatively integrated into a single model. The main assumption is that memory for odors represents a unique and sep- In the second experiment, participants were exposed to either 15 olfactory stimuli, 15 visual stimuli (photographs of human faces), or 15 acoustic Keywords: oder memory, consciousness, strategies and interferences Introduction
. Olfactory memory presents a relatively low initial ac-
quisition level compared to visual and verbal material.
Whether olfactory memory represents a separate and This led Engen (1982, 1991) to assume that odors are unique memory system, with different functional char- represented in memory as unitary and distinctive acteristics compared to the other sensory modalities, is events with little attribute redundancy. As evidence for an open issue (see e. g., Engen, 1991; Herz & Engen, 1996; this, Lawless (1978) found that common, complex Richardson & Zucco, 1989; Schab, 1991). Evidence to odors were encoded and remembered as well as simple support this hypothesis is the following: chemicals and abstract meaningless geometrical shapes. According to Lawless, odors are relatively fea- Odor recognition memory is only slightly influenced tureless stimuli, which means that immediate recogni- by the length of retention intervals. This was observed tion memory for odors is poor compared to visual and for short intervals, for example, a few seconds and verbal stimuli, the latter being richer in accessible fea- minutes (Bromley & Doty, 1996; Engen & Ross, 1973; tures. In Underwood's (1969) memory model, odors are Jehl, Royet & Holley, 1994; Jones, Moskowitz & But-ters, 1975) as well as for longer retention periods suchas a week (Lawless & Cain, 1975; Rabin & Cain, 1984; Gesualdo M. Zucco is Associate Professor at the Faculty of Medicine, Wood & Harkins, 1987; Zucco, 1983), a month (Lawless Department of General Psychology at the University of Padova, Italy.
& Cain, 1975), four months (Lawless, 1978), and over Research interests include perception and memory of odors, neuro- a year (Engen & Ross, 1973; Goldman & Seamon, psychology of olfaction, memory, mental retardation, and learningdisabilities.
1992). Usually the forgetting curve resulting fromthese experiments is relatively flat, unlike for verbal Correspondence concerning this article should be addressed to Ge-sualdo M. Zucco, Dipartimento di Psicologia generale, Via Venezia and visual material (see Ebbinghaus's seminal exper- 8, I-35100 Padova, Italy (tel. +39 49 8276507, fax +39 49 8276600, European Psychologist, Vol. 8, No. 2, June 2003, pp. 77-86 2002 Hogrefe & Huber Publishers stored by means of relatively inefficient forms of mne- short-term odor memory. Four groups of subjects were monic coding (see Richardson & Zucco, 1989). Also, asked to perform one of the following tasks, during the Schab, (1991) noted that "such stimuli, once encoded retention interval between odor acquisition and recogni- are characterized by a slower rate of forgetting, because tion: (1) to make free associations with the target odor, fewer features per stimulus means less interference (2) to make free associations with the name of a new from stimuli with the same or similar features" (p. 248).
distractor odor, (3) to smell a new distractor odor and to . Odor memory is very resistant to retroactive interfer- make verbal associations to it, and (4) control condition.
ence, i. e., to forgetting produced by subsequent learn- Authors found that recognition performance was best ing experiences (see Lawless & Engen, 1977), while when participants made free associations with the target strong effects of proactive interference are observed odor during the retention interval. Walk and Johns as- (Engen, 1987; Lawless & Engen, 1977). Generally sume that memory for odors and memory for stimuli in speaking, the absence of a retroactive interference ef- other modalities may share similarities.
fect could explain the endurance of olfactory traces In an another study, Rabin and Cain (1984) asked over time (see also Herz & Engen, 1996). Also, accord- their subjects to give a label to each substance of the ing to Schab (1991) odors acquired in experimental stimulus set. In the following recognition test (one week conditions and environmental odors are processed dif- later), authors found that naming the stimulus correctly ferently, and the second cannot interfere with the first.
(during both acquisition and retrieval) was positively This hypothesis, however, needs to be experimentally correlated with odor memory; other authors, however, failed in finding any memory facilitation by the use of . Odor memory seems to be unaffected by the familiar- verbal labels or rehearsal (Engen & Ross, 1973; Gabassi
ity of the substances used, as well as by the pleasant- & Zanuttini, 1983; Lawless & Engen, 1975).
ness of the substance (Engen & Ross, 1973; Lawless & A more direct manipulation of elaborative process- Cain, 1975). Furthermore, giving odors a meaningful ing was carried out by Lyman and McDaniel (1986).
label has no effect on the subsequent recognition of the Their study is examined in more detail here, since the odors (Lawless & Cain, 1975). Indeed, the relationship first experiment presented below is a replication thereof.
between odors and words seems to be very weak: Sub- Authors asked four groups of subjects to smell 30 jects show difficulties in the correct identification of odors and to perform one of the following tasks for each the odors (Cain, 1979, 1982; Engen, 1991; Engen & of them: (1) to provide a name and a short definition; (2) Pfaffmann, 1960; Sumner, 1962; Zucco & Aiello, 1996); to create an image; (3) to describe a specific life episode; verbal rehearsal and verbal interference tasks do not (4) simply to smell the odors for a subsequent recogni- affect their recognition (Gabassi & Zanuttini, 1983; En- gen, Kuisma & Eimas, 1973); the right hemisphere One week later subjects were asked to recognize the seems to be involved in the elaboration of odors (Ab- 30 odors out of a set of 60. Participants who had associ- hram & Mathai, 1983; Zatorre, Jones-Gotman, Evans, ated either names or personal episodes to odors gave the & Meyer, 1992; Zucco & Tressoldi, 1989).
lowest false alarm (FA) scores; control and visual-imag- . No differences emerge in recognition tasks for odor ery groups gave the highest FA scores, which were sig-
stimuli learned intentionally or incidentally (see En- nificantly different from those of the two former groups.
gen & Ross, 1973, for a seminal paper).
Episode vs naming and control vs visual-imagery com-parisons did not differ significantly.
The effect of strategies and interference tasks on odor Hit scores, too, did not differ significantly among recognition have been examined less thoroughly in the conditions. Nevertheless, the authors maintained that literature. The two experiments presented here aim to elaborative processing, such as naming odors or provid- ing life episodes for them, lead to the best recognitionperformance (p. 753, 760). However, such a statement isa little puzzling; in fact, their results would suggest thatstrategies were effective only in rejecting items that did Strategies and Odor Memory
Further, the authors observed that the "naming" Few studies have been carried out on the effects of strat- condition produced the best performance, while no dif- egies on odor recognition. Among these, Walk and Johns ferences emerged between "control" and "imagery" (1984) examined the effects of different instructions on groups. These results are at odds with the literature: It is European Psychologist, Vol. 8, No. 2, June 2003, pp. 77-86 Anomalies in Cognition: Olfactory Memory well known that visual codes give rise to a better mem- fects of elaborative encoding activities on odor memory ory performance than do verbal codes (Paivio, 1986)-and also that to give an odor a name is a verydifficult cognitive task (Engen, 1991; Richardson & Zuc-co, 1989) since the odor-name relationship is very weak Experiment 1
(Herz & Engen, 1996; Lawless & Cain, 1975; Zucco &Tressoldi, 1989) In a second paper, Lyman and McDaniel (1990) ob- served that the use of strategies such as processing a picture or a name of the source of an odor was effectivein odor memory (1st experiment). One group of subjects Forty-eight university students ranging in age from 20 were presented the photographs of the referent source to 30 years participated in the experiment (M = 24.7 for 30 odors (in addition to the odors themselves); a sec- years). They were randomly assigned to one of the fol- ond group was presented with the names of the 30 odors lowing conditions: visual-imagery, label-plus-defini- (in addition to the odors themselves); while a third tion, life-episode, and control. Both sexes were repre- group was presented both the names and the photo- sented (7 males and 5 females per condition). None of graphs (in addition to the odors themselves). A final two them had any impairment of the olfactory system.
groups (name only and odor only control groups) werepresented only the names of the odors or the 30 odors.
Analyses indicated that both visual and verbal elabora- Sixty olfactory stimuli contained in small glasses and tions improved recognition. However, results differ fitted with rubber plugs were used. The plugs were con- from those of the 1986 paper, where the use of imagery nected to a cotton swab wrapped at the end of a stick.
played no role in the recognition of odors. Further, the Test tubes were covered with white paper to prevent significant effect on hit scores was very small in its mag- participants from having visual cues. Substances were nitude. In the second experiment of the same paper Ly- almost all pleasant and natural, and were replaced every man and McDaniel instructed a first group of subject to 48 hours, so that concentration of the odors was kept imagine the scent of 20 stimuli presented verbally (e. g., under control. Thirty stimuli were used in the odor-ac- The odor of the word "banana"), while a second group quisition phase and the other 30 as distractors for the of subjects was instructed to conjure up 20 images of the recognition test. Odors were almost all the same as those same names. At recognition, subjects were exposed to 40 used by Lyman and McDaniel (1986), for example, al- odors and to 40 pictures and were required to distin- mond, Brut after-shave, clay, lemon, pipe tobacco, soap, guish which of them corresponded to the words they had seen previously. Subjects who had imaged olfactoryinformation of the words and recognized odors per- formed better than subjects who had imaged olfactoryinformation and recognized pictures. A similar pattern Participants were individually administered the tasks.
was observed for pictures. However, while significant They were informed that they had to smell 30 different differences among conditions were found for d' and FA odors, one after the other, and then, a week later, they scores, the results on hits showed no facilitation by ol- would have had to recognize such stimuli among a set factory imagery on recognition of odors and by visual of 60. Participants were asked to smell each odor for imagery on recognition of pictures. Herz and Engen about 15 seconds followed by 20 seconds during which (1996), reviewing this paper, pointed out that "The im- they had to perform the encoding tasks, that is: agery results were therefore dependent on differences in a) Visual imagery condition: Participants had to create an false alarm rates as a function of different encoding strat- image of the source of the odor. For instance, if they egies, and, as such, it is questionable how meaningful thought they were smelling beer, they may try to cre- they are" (p. 305). Thus, the role of strategies on odor ate an image of a glass (or a can) of beer in their mind.
memory is far from being understood.
b) Label-plus-definition condition: Participants were asked The first experiment presented here is a replication to think of a name and a short definition for the odor and extension of Lyman and McDaniel's first study (sim- that they were smelling. If they were not able to give pler in its experimental structure than the second). Any a name to the odor, they were asked to try and name methodological differences are described below. No ef- European Psychologist, Vol. 8, No. 2, June 2003, pp. 77-86 2002 Hogrefe & Huber Publishers c) Life-episode-condition: Participants were asked to try to remember a specific episode of their life related to the Mean scores of d', Hits and False Alarm as a function of d) Participants in the control group had only to smell each encoding conditions. (Lyman and McDaniel's study scores, odor. All associations made by participants in the three experimental groups were tape-recorded; sub- jects were strongly recommended to try to follow their One week later, subjects were administered the recogni- tion test. They were presented, in random order, 60 odors (30 target and 30 distractors), and they were asked to try to recognize those presented in the acquisition testby circling the words "new" or "old" on a booklet. Bothin the acquisition phase and here subjects were instruct-ed to close their eyes while smelling each odor.
plus-definition =.58) vs the proportions of correctly rec- Furthermore, subjects in the three experimental ognized odors plus the correct associations (see above), groups were asked to tell, what association they made were analyzed, by t-test, for each experimental group.
the week before, for each odor recognized.
None of these comparisons was significant.
The experiment took place in a well-ventilated room; the order of stimuli was randomized for each sub- ject. Responses were scored for accuracy.
The main result of the first experiment is that it fails to show any crucial effect on recognition of odors by pre-vious elaborative activities. This is particularly evident On the data, hit and false alarm rates and d' scores were when considering the results on hit rates: Subjects, in considered. Table 1 shows the results as a function of fact, did not differ in their ability to recognize odors, experimental conditions. For each of these measures a despite the use of different encoding strategies. The analyses performed on the correctly recognized odors (ANOVA) was carried out. The single factor "encoding plus the correct associations provide further support.
task" (control vs visual-imagery vs label-plus-definition The first indicates that no strategy was more effec- vs life-episode) reached a significant level for d': tive than another (proportion of hits were in fact almost F(3, 44) = 5.14, p <.004. Pair-wise post-hoc comparison the same for all the three experimental groups), while using Tukey test shows the following significant effects: the second suggests that the level of accuracy in recog- life-episode vs control (p <.01) and life-episode vs label- nizing odors is independent of the adoption of any strat- plus-definition (p <.01); the other comparisons were not egy. No differences indeed come out from intragroup significant. The ANOVA on hits was not significant (F = comparisons using t-test. Plausibly, if one is able to rec- 2.06). The ANOVA on false alarm reached significance: ognize a stimulus, but fails to recall the strategy used to F(3, 44) = 3.10, p <.04; but a Tukey pairwise post-hoc promote its recognition, this means that such a stimulus comparison did not show any significant effect.
was recognized per se (i. e., because of its own character- A further analysis was carried out on the correct istics) rather than on the basis of elaborative processes.
recognition of the odors for which participants were able However, as suggested by one of the reviewers, it would to recall also the associations made for them at acquisi- be interesting to see if the positive effect of encoding tion (i. e., the words, or the images, or the episodes).
strategies on recognition memory is related in the verbal On hit scores, a one-way between-subjects ANOVA and visual domains to good memory for the elabora- was performed. The analysis was not significant (F = .528). Mean proportion of hits, according to the experi- Both analyses were, then, crucial in testing Lyman mental conditions, were as follow: life-episode =.50; vis- and McDaniel's assumptions and suggest that strategies ual-imagery =.45; label-plus-definition =.42 are not effective cues for recognition of odors.
Finally, the proportions of correctly recognized Finally, a general d' significant effect was observed; odors (i. e., life-episode =.50; visual imagery =.55; label- but unlike Lyman and McDaniel's experiment, post-hoc European Psychologist, Vol. 8, No. 2, June 2003, pp. 77-86 Anomalies in Cognition: Olfactory Memory comparisons here yield better performances for the life- mental groups did not differ significantly from each oth- episode and imagery groups than for the label-plus-def- er. In both studies, however, the level of recognition and inition group. Thus, these data seem to be consistent recall scores was very low. Moreover, some aspects of the with those reported in the literature and discussed two studies are a little puzzling, e. g., in contrast to the above, namely, the weak odors-words link and the effi- literature the immediate and delayed recognition scores cacy of visual codes on memory. Moreover, analyses on differed significantly (first experiment); and it is not FA have shown a general significant effect, but-at odds clear why the authors used a recall procedure, this being with the two authors-post-hoc comparisons failed to find significant differences among conditions.
Thus, clear effects of interference on odor recogni- In my opinion, only if correct recognitions or hit tion were far from being demonstrated.
scores were to lead to clear differences among groups The effects of interference in short-term odor mem- would then unquestionable effects of encoding strate- ory were found in the study by Walk and Johns (1984) gies on odor recognition be actually demonstrated. Dif- quoted above. These authors observed that making as- ferences in FA and d' are not sufficient and meaningful sociations to an additional distractor odor during reten- conditions (FA is just a measure of the ability to reject tion interval led to the lowest recognition performance.
items that do not belong to the acquisition list, and d' is Lawless and Engen (1977), on the other hand, found a general measure of the accuracy and sensitivity in the clear effects of proactive, but not retroactive interference.
discrimination among old and new items).
Two groups of participants were required to learn oneset or two sets of 12 pictures associated with a single setof 12 odors. In the first session they had to associate the 12 odors with the first set of pictures; in the second ses-sion the experimental group was asked to associate the Studies on the effects of interference tasks on odor mem- 12 odors with the second set of pictures, while the con- ory are sparse. Furthermore, the results reported are trol group was asked to associate the 12 odors again with quite unclear, so that the issue of interference in odor the first set of pictures. Two weeks later no differences were observed between experimental and control group Perkins and McLaughlin Cook (1990) and more re- in remembering the first set of pictures. On the contrary, cently Annett, McLaughlin Cook, and Leslie (1995) in- the second set of pictures was remembered significantly vestigated whether memory for odors could be affected worse by the experimental group. Thus, the first associ- by concomitant interference tasks (see Baddeley, 1986).
ation to an odor seems to be relatively impervious to The rationale of the two studies was the following: If forgetting and difficult to modify through subsequent memory for odors represents a unique and separate memory system, then interference tasks should not af- Finally, Crowder and Schab (1995) found neither fa- fect the recognition of odors, whereas if olfactory mem- cilitation nor interference as a function of odor imagery ory is a part of a more general memory system, they on odor recognition (see also, Herz and Engen, 1996).
Perkins and McLaughlin Cook (1990) asked their subjects to smell 15 odors while performing one of the Experiment 2
following tasks: (1) rehearsing some digits (verbal sup-pression condition); (2) performing a visual game (visu-al suppression condition); (3) both (verbal + visual con- Is memory for odors affected by same-modality and/or ditions), and (4) simply smelling the odors for the sub- intermodality interference tasks? The second experi- sequent recognition and recall tasks (control condition).
ment aims at verifying the effects of interference on odor At immediate recognition (10 minutes later) the au- thors found that only tasks involving verbal suppressioninterfered with odor recognition; at delayed recognition, interference was observed also for the visual suppres- sion task. Annet et al. (1995), using the same methodwith a few changes (e. g., a more complex visual task), Thirty-six university students ranging in age from 20 to obtained different results than Perkins and McLaughlin.
24 years (M = 22.3 years), divided into three groups, took Particularly, at immediate recognition the three experi- part in the experiment. They were randomly assigned to European Psychologist, Vol. 8, No. 2, June 2003, pp. 77-86 2002 Hogrefe & Huber Publishers one of three experimental conditions. Both sexes were sisted of the old odor and two new odors (e. g., cinna- equally represented. Subjects had no impairment to the mon-tobacco-soap). For each of the three alternatives, olfactory, acoustic or visual systems.
the experimenter asked the question "Is this the odorthat you sniffed previously?" Participants were given a short rest every three subsequent presentations. Theposition of the target stimulus in the triplet was ran- A total of 270 olfactory, visual, and acoustic stimuli were used (90 per condition). Odors were almost all pleasant . Second session: The subject was exposed to a new set of
and natural (e. g., chocolate, cinnamon, leather, oregano, 15 odors and was shown, at the end of the presenta- pine, rose, rubber, and so on). They were contained in tion, some photographs representing human faces for small glasses covered with white paper to prevent par- 2½ minutes (if the interference task was visual). For ticipants from having visual cues for identification; sub- each photograph shown the subject had to say how stances were replaced every 48 hours, so their concentra- much he/she liked it, on a scale from 1 (not pleasant tions were controlled. The visual stimuli were black and at all) to 5 (very pleasant). At the end of such an inter- white photographs (9 × 13 cm) representing human fac- modality interference task, the subject was engaged in es of males and females seen in front. Their age ranged the olfactory recognition (about which he had previ- from about 18 years to 60 years. Finally, the acoustic stimuli were tape-recorded environmental sounds (e. g., . Third session: At the end of the acquisition of the 15
a telephone or a bell ringing, a child crying, a train hiss- olfactory stimuli, the subject was given a pair of ear- ing, a cat mewing, a hammer knocking, and so on). All phones (if the interference task was acoustic) and was stimuli selected were ecological and natural. They were asked to judge the pleasantness of some environmen- quite distinguishable, according to the judgment of tal sounds for 2½ minutes (intermodality interference).
Immediately after this, the subject performed the ol-factory recognition task.
. Fourthsession:Afterthepresentationofthe15olfactory
The experiment consisted of four sessions. In the first stimuli, the subject was administered the olfactory, in- (the no-interference condition), participants were asked to sit on a comfortable chair and to either sniff, listen (byearphones), or look at 15 stimuli; they were then request- In short, there were two intermodalities interference ses- ed to recognize the former stimuli from a set of 45 (se- sions, one intramodality interference session, and one lecting each target stimulus among three). In the other sessions, held one week apart to avoid learning effects, All the sessions (except the first) were counterbal- they were shown a new set of 15 stimuli followed by an interference task. During this task subjects either had to The experiment took place in a large and well-ven- listen, sniff, or look at other olfactory, visual, or acoustic tilated room. Responses were scored for accuracy. For stimuli (18 on average) for 2½ minutes. They were also each session the 30 distractors were randomly selected asked to judge each of them on a 5-point pleasantness from the set of 90 stimuli, with the restriction to not use scale. Finally, subjects were required to recognize the the same stimuli for subsequent sessions.
original set of 15 stimuli from 45. The interference taskswere intermodality (two) and intramodality (one).
For a better understanding of the experimental de- mands, let us consider a participant in the olfactory con-dition: Table 2 shows the mean number of correct recognition . First session (no-interference): The subject sat on a chair (out of 15) and standard deviations for each group. Cor-
and was asked to smell 15 odors, one after the other; rect responses underwent three one-way analyses of each stimulus was exposed for about three seconds, variance (ANOVA), with olfactory, visual, and acoustic with an interstimulus interval of about 6 seconds. Ev- conditions as factors, and no-interference vs olfactory ery five presentations the subject was given 15 seconds interference vs visual interference vs acoustic interfer- to rest to avoid any carry-over adaptation effects. At the end of the acquisition phase the subject was asked Only the acoustic and visual conditions reached a to recognize the target odors. The recognition set con- European Psychologist, Vol. 8, No. 2, June 2003, pp. 77-86 Anomalies in Cognition: Olfactory Memory Here, it was observed that despite adopting the strong-est interference task (as intramodality), no disruptive ef- fects on odor recognition appeared. Olfactory memory Mean scores and Standard deviations (in brackets) for thus does not seem to be affected by the use of strategies Acoustic, Visual and Olfactory modalities, as a function ofinterference.
(as observed in Experiment 1) or by interferences. Plau-sible reasons for this-as well as the possibility to think of olfactory memory as a distinctive memory sys- General Discussion
In this paper two experiments were carried out to inves-tigate both the effects of strategies and interference taskson the recognition of odors. Although strategies and in- terferences represent effective methods to improve or acoustic condition: F(3, 33) = 30.3, p < 001; disrupt verbal and visual memory (as observed also in visual condition: F(3, 33) = 26.3, p <.001; Experiment 2), they failed to show any effect on odor The question is: Why do odors behave differently A post-hoc analysis (Tukey test) was carried out to ex- from visual and verbal stimuli? My answer is a theoret- amine the effect of intramodality and intermodalities in- ical proposal that integrates the issues of the present ex- terference on visual and acoustic conditions (for α.01).
periments and some other peculiarities of olfactory In each analysis the within-modality condition shows the most interference; the between-modalities The main hypothesis lies in the assumption that condition show significantly less interference than the odors (1) do not give rise to a conscious representation within-modality condition, but more interference than of them and (2) could be stored in memory at a level below consciousness. The sequence from encoding tostorage of olfactory information could then occurs through the following steps:1) subject (S) is presented an odor; The general results of the second experiment indicatethat recognition performances for visual and acoustic 2) during encoding S is conscious that he/she is smell- stimuli are affected by interference tasks (see, e. g., Bad- deley, 1986; Brooks, 1967). In fact, in the no-interference 3) when the stimulus is taken away, S lacks a conscious condition recognition scores for acoustic and visual stimuli were significantly better than those in the inter- 4) the odor is, however, automatically stored in memory; modality interference condition, while the intramodali- ty interference condition produced significantly worse 5) at an implicit (unconscious) level of knowledge.
However, unlike visual and acoustic memory, olfac- The assumption that odors do not give rise to a conscious tory memory does not seem to be sensitive to interfer- representation thereof relies mainly on introspective re- ence effects: Neither same-modality nor different-mo- ports: All of us have experienced that recalling an odor is dalities interference scores differed significantly from an impossible or particularly hard task. Such a difficulty, however, could depend on the fact that people have no Engen et al. (1973) also observed that interference conscious representation of the stimuli experienced tasks (like counting backwards) also failed to show any (then, we could maintain that a conscious recall implies effect on recognition of odors. However, such a result a conscious representation). Recognition, however, is could be attributed either to the weak connection be- possible, and it is the only route for odor retrieval, since tween olfaction and language or to the fact that the task it is a match between a stimulus already stored in mem- demand was too easy, rather than to interference per se. ory with a newly encountered one. Obviously, people can European Psychologist, Vol. 8, No. 2, June 2003, pp. 77-86 2002 Hogrefe & Huber Publishers remember that an odor was encountered on one occasion proposal presented above could integrate them in a sin- and can also name it; but this does not mean that a con- gle interpretation. Such peculiarities (nearly all ob- scious representation or the recall of odors is possible. Of course, I am aware that speculations based on introspec- 1) Odor memory is slightly or at all affected by retention tive reports are not the best scientific demonstrations. But what is the connection between the results of both exper- 2) Odor memory is resistant to retroactive interference.
iments presented here and the analyses given above? 3) Generation of a meaningful label has no effect on sub- This link is indirect, because we have no methods to ob- serve directly whether a conscious representation of ol- 4) Verbal rehearsal and counting backwards do not af- factory stimuli actually exists. Hence, I would like to start by examining the effects of strategies and interfering 5) Incidental or intentional learning of odors gives rise tasks on memory for verbal and visual stimuli. It is well known that while the former improve their retrieval, the 6) As seen here, neither strategies nor interferences affect latter have a disruptive effect on them (see any book on memory, e. g., Baddeley, 1990, 1998).
I assume that an improvement or disruption of ver- The assumption that people lack a conscious represen- bal and visual material is possible only because subjects tation for odors could successfully explain any of these have access to their internal representation of such stim- effects, and it also satisfies the scientific criterion of par- uli. This means that, to be effective, interferences and simony. People can be conscious of olfactory stimuli on- strategies have to act on these conscious representations.
ly at the encoding and recognition stages, that is, when In short, if together with-or soon after-the presenta- they are concretely present. Between these periods con- tion of the word "dog" people are not able to create an scious access to the olfactory trace is not possible. There- image or to rehearse such a stimulus, then neither strate- fore, subsequent learning experiences cannot affect pre- gies nor interferences can affect recognition-and the viously learned odors for which we have no conscious stimulus is recognized on the basis of its own character- representation. Time passing has no effect on odor mem- istics. This is exactly what happened in the first experi- ory, for the same reason. Incidental or intentional tasks ment: No differences were found among conditions (pro- do not affect the recognition of stimuli that cannot be portion of hits were about the same for the three experi- rehearsed or elaborated, and so forth for all the other mental groups) and the level of recognition accuracy was independent from the use of any strategy (see discussion What kind of relationship is there between this pro- of Experiment 1). The main issue here, however, is still posal and other interpretations on the nature of odor that none of the strategies used actually affected recogni- tion of olfactory stimuli. In my opinion, this happened Engen (1982) suggests that odors are represented in because strategies have to act on the conscious represen- memory as unitary and distinctive events with little at- tations of the stimuli in order to be effective, and this in tribute redundancy. Odors, then, are learned holistically turn is not possible for the olfactory domain (probably, as (in an all-or-none fashion), which explains their relative- suggested from one of the reviewers (and I agree), olfac- ly low initial acquisition level and their resistance in tory representations are mainly perceptual, whereas vis- time (the interference caused from other olfactory stim- ual and verbal representations are also conceptual, the uli is minimal). Schab (1991), however, suggests that latter giving rise to conscious representations, while the odors acquired in an experimental set are processed dif- former do not). The same is true when adopting interfer- ferently from those encountered in other contexts. This ing tasks: Subjects of an experimental group do not rec- should explain the little or absent degree of interference ognize odors worse than a control group because they between the two conditions (experimental and nonex- lack a conscious representation of those odors. Indeed, to perimental) and the endurance of olfactory traces. The create a conscious representation of an odor (e. g., coffee) sturdiness of odor memory could be related to different is a very difficult task, compared to how easy it is to imag- ways of processing odors rather than to the separate na- ine a cup of coffee or to rehearse the word "coffee." Stor- ture of olfactory stimuli. As the author states, this hy- age and access to olfactory stimuli in memory, then, pothesis, however, needs to be tested experimentally.
should not imply an effort but be automatic.
Fundamentally, I agree with Engen's point of view In the introduction some peculiarities (or anoma- that odors could be stored as distinctive and unitary lies) of olfactory memory were discussed. I think the events, and I think that this interpretation does not con- European Psychologist, Vol. 8, No. 2, June 2003, pp. 77-86 Anomalies in Cognition: Olfactory Memory trast with, but could be integrate within, the model here An interesting point of view on the role of con- presented. The idea that there is no conscious represen- sciousness in cognition was advanced by Velmans tation for odors can account for their distinctiveness. In (1991). The author assumes that no human information my opinion, however, the latter is not sufficient to ex- processing is conscious, in the sense that consciousness plain all peculiarities of odor memory and has to be bet- neither takes part nor causally influences cognitive pro- cesses. Although the present paper claims that con- The claim that memory for odors is different from sciousness has no role only in processing of odors, Vel- visual and verbal memory was examined by Herz and man's contribution can be useful. The author in fact out- Engen (1996). They applied the criteria of multiple mem- lines some situations in which information may indeed ory systems (MMS) theories (see Roediger, Rajaram, & enter memory and be recalled without consciousness.
Srivinas, 1990; Sherry & Schacter, 1987; Schacter & Tul- Such conditions are hypnosis, blindsight, and "masked ving, 1994), according to which different memory sys- priming" studies. I think that the hypothesized unique- tems are governed by different and distinct mecha- ness of odor memory compared to other modalities can nisms. The authors maintain that despite the fact that be effectively tested under these circumstances (see, for not all the criteria of the MMS theories (i. e., functional instance, Olsson, Jonsson, & Faxbrink, in press, for a re- dissociation, stochastic independence, and independent neural systems) were satisfied, it remains difficult to Finally, the proposal presented here is difficult to in- consider memory for odors as memory in other modal- tegrate with any of the theories on consciousness (see Umiltà, 2000). Indeed, being either of the interactionist or Furthermore, within the discussion of whether odor unitary type they do not give attention to the possibility memory can be conceptually (or semantically) driven or of having unconscious representations of newly experi- perceptually driven (see, e. g., Herz & Engen, 1996; Ly- enced stimuli. However, a discussion on these issues is man & McDaniel, 1990; Schab, 1991) the above analy- beyond from the main purposes of this contribution.
ses-and the results of the present experiments-seemto suggest that odor memory is perceptual in nature.
Indeed, experimental manipulations do not affect recog- The author wishes to thank Dr. C. Caroppo and Dr. S.
Fior for their help in collecting the data, and Dr. S. De A last point concerns the possible relationship be- Cristofaro for comments on an earlier draft.
tween the theoretical proposal discussed in this paper This paper is dedicated to Prof. Trygg Engen and other cognitive theories, as well as the relationship (Brown University, Rhode Island, USA).
between this theory and other theories on consciousness(see Umiltà, 2000, for a recent review on consciousness).
These issues would have to be discussed in more detailin a separate paper; however, a few considerations can References
be made. The idea that there is no conscious representa-tion of odors can be connected somewhat to theories on Abraham, A., & Mathai, K. (1983). The effect of right temporal lobe implicit and explicit memory (see Graf & Schacter, 1985; lesions on matching of smells. Neuropsychologia, 21, 277-281.
Kihlstrom, Schacter, Cork, Hurt, & Behr, 1990; Schacter, Annett, J., McLaughlin Cook N., & Leslie, J. (1995). Interference 1987; Tulving & Schacter, 1990). Indeed, in implicit tasks, with olfactory memory by visual and verbal tasks. Perceptual such as fragment or stem completion, amnesic patients and Motor Skills, 80, 1307-1317.
show quite normal memory, despite the fact they fail to Baddeley, A. (1986). Working memory. Oxford: Clarendon.
recall that they were shown certain stimuli. Their access Baddeley, A. (1990). Human memory. Hove: Erlbaum.
Baddeley, A. (1998). Essential of Human memory. Hove: Psychology to memory traces is automatic (the same is true for odors). However, whether the acquisition of items (e. g., Bromley, S., & Doty, R. (1996). Odor recognition memory is better verbal and visual) in amnesic patients is automatic or under bilateral than unilateral test conditions. Cortex, 4, 25-40.
effortful, and whether these two conditions can give rise Brooks, L.R. (1967). The suppression of visualization by reading.
to different issues (as was not observed for odors) re- Quarterly Journal of Experimental Psychology, 19, 289-299.
mains to be demonstrated. Experiments on implicit Cain, W. (1982). Odor identification by males and females: Predic- memory in amnesic patients suggest simply that a stim- tion vs performance. Chemical Senses 7, 129-142.
ulus for which there is no recall can be stored, and that Ebbinghaus, H. (1885). Über das Gedächtnis [On memory]. Leipzig: access to it is possible only implicitly.
European Psychologist, Vol. 8, No. 2, June 2003, pp. 77-86 2002 Hogrefe & Huber Publishers Engen, T. (1982). The perception of odors. New York: Academic Press.
Rabin, M., & Cain, W. (1984). Odor recognition: Familiarity, iden- Engen, T. (1987). Remembering odors and their names. American tifiability and encoding consistency. Journal of Experimental Psychology: Learning Memory and Cognition, 10, 316-325.
Engen, T. (1991). Odor sensation and memory. New York: Praeger.
Richardson, J., & Zucco, G. (1989). Cognition and Olfaction: A re- Engen, T., & Pfaffmann, C. (1960). Absolute judgments of odor view. Psychological Bulletin, 105, 352-360.
quality. Journal of Experimental Psychology, 59, 214-219.
Roediger, H., Rajaram, S., & Srinivas, K. (1990). Specifying criteria Engen, T., & Ross, B. (1973). Long-term memory for odors with for postulating memory systems. In A. Diamond (Ed.), The de- and without verbal descriptions. Journal of Experimental Psy- velopment and neural bases of higher cognitive functions. New Engen, T., Kuisma, J., & Eimas, P. (1973). Short-term memory for Schab, F. (1991). Odor memory: Taking stock. Psychological Bulletin, odors. Journal of Experimental Psychology, 99, 222-225.
Gabassi, P., & Zanuttini, L. (1983). Riconoscimento di stimoli ol- Schacter, D. (1987). Implicit memory: History and current status.
fattivi nella memoria a breve termine [Recognition of olfactory Journal of Experimental Psychology, LMC, 13, 501-518 stimuli in short-term memory]. Giornale Italiano di Psicologia, Schacter, D., Tulving, E. (1994). What are the memory systems of 1994? In D. Schacter & E. Tulving (Eds.), Memory systems 1994. Goldman, W., & Seamon, J. (1992). Very long-term memory for odors: Retention of odor-name associations. American Journalof Psychology, 105, 549-563.
Shepard, R. (1967). Recognition memory for words, sentences and pictures. Journal of Verbal Learning and Verbal Behavior, 6, Graf, P., & Schacter, D. (1985). Implicit and explicit memory for new associations in normal and amnesic subjects. Journal ofExperimental Psychology, LMC, 11, 501-518 Sherry, D., & Schacter, D. (1987). The evolution of multiple mem- Herz, R., & Cupchik, G. (1995). The emotional distinctiveness of ory systems. Psychological Review, 94, 439-454 odor-evoked memories. Chemical Senses, 20, 517-528.
Sumner, D. (1962). On testing the sense of smell. Lancet, 2(n. 7262), Herz, R., & Engen, T. (1996). Odor memory: Review and analysis.
Psychonomic Bulletin and Review, 3, 300-313.
Tulving, E., & Schacter, D., (1990). Priming and human memory Jehl, C., Royet, J., & Holley, A. (1994). Very short-term memory for odors. Perception and Psychophysics, 56, 658-668.
Umiltà, C.A. (2000). Consciousness and conscious experience. In Jones, B., Moskowitz, H., & Butters, N. (1975). Olfactory discrim- K. Pawlik & M. Rosenzweig (Eds.), International handbook of ination in alcoholic Korsakoff patients. Neuropsychologia, 13, psychology. San Francisco: Sage.
Underwood, B. (1969). Attributes of memory. Psychological Review, Kihlstrom, J., Schacter, D., Cork, R., Hurt, C., & Behr, S. (1990).
Implicit and explicit memory following surgical anesthesia.
Velmans, M. (1991). Is human information processing conscious? Psychological Sciences, 1, 303-306 Behavioral and Brain Sciences, 14, 651-726 Lawless, H. (1978). Recognition of common odors, pictures and Walk, H., & Johns, E. (1984). Interference and facilitation in short- simple shapes. Perception and Psychophysics, 24, 493-495.
term memory for odors. Perception and Psychophysics, 36, Lawless, H., & Cain, W. (1975). Recognition memory for odors.
Chemical Senses, 1, 331-337.
Wood, J., & Harkins, S. (1987). Effects of age, stimulus selection Lawless, H., & Engen, T. (1977). Associations to odors: Interfer- and retrieval environment on odor identification. Journal of ence, mnemonics, and verbal labeling. Journal of Experimental Psychology, Human Learning and Memory, 3, 52-59.
Lyman, B., & McDaniel, M. (1986). Effect of encoding strategies on Zatorre, R., Jones-Gotman, M., Evans, A., & Meyer, E. (1992). Func- long-term memory for odors. Quarterly Journal of Experimental tional localization and lateralization of human olfactory cor- Lyman, B., & McDaniel, M. (1990). Memory for odors and odor Zucco, G. (1983, September). Esiste vicarianza sensoriale nel sordo? names: Modalities of elaboration and imagery. Journal of Exper- Una ricerca sul ricordo olfattivo in soggetti adulti sordi e udenti [Do imental Psychology, Learning, Memory and Cognition, 16, 656-664.
there exist vicarious sensations in the deaf? An investigation Olsson, M.J., Jonsson, F.U., & Faxbrink, M. (in press). Repetition of olfactory memory in deaf and hearing adult subjects]. Paper priming in odor memory. In C. Rouby, B. Schaal, D. Dubois, R.
presented at the Third Conference of the Italian Society of Sci- Gervais, & A. Holey (Eds.), Olfaction, taste and cognition. Lon- Zucco, G., & Aiello, L. (1996, August). Gender differences in the recall Paivio, A. (1986). Mental representations: A dual coding approach. of autobiographical memories cued by olfactory stimulation. Paper presented at the XXVI International Conference of Psychology, Perkins, J., & McLaughlin Cook, N. (1990). Recognition and recall of odors: The effects of suppressing visual and verbal encoding Zucco, G., & Tressoldi, P. (1989). Hemispheric differences in odor processes. British Journal of Psychology, 81, 221-226.
recognition. Cortex, 25, 607-615.
European Psychologist, Vol. 8, No. 2, June 2003, pp. 77-86

Source: http://www.scentair.cl/uploads/media/researchzuccocognitionolfactorymemory.pdf