Content uploaded by Rachel Herz
Author content
All content in this area was uploaded by Rachel Herz
Content may be subject to copyright.
Emotion experienced during encoding
enhances odor retrieval cue effectiveness
RACHEL S. HERZ
Monell Chemical Senses Center
The American Journal of Psychology
Winter 1997, Vol. 110, No. 4, pp. 489–505
Content in the AJP database is intended for personal, noncommercial use only.
You may not reproduce, publish, distribute, transmit, participate in the transfer
or sale of, modify, create derivative works from, display, or in any way exploit
the AJP content in whole or in part without the written permission of the
copyright holder.
To request permission to reprint material from The American Journal of
Psychology, please find us online at:
http://www.press.uillinois.edu/about/permission.html
or email us at:
UIP-RIGHTS@uillinois.edu
© 2002 by the Board of Trustees of the University of Illinois
Emotion experienced during encoding
enhances odor retrieval cue effectiveness
RACHEL S. HERZ
Monell Chemical Senses Center
Emotional potentiation may be a key variable in the formation of odor-associ-
ated memory. Two experiments were conducted in which a distinctive ambient
odor was present or absent during encoding and retrieval sessions and subjects
were in an anxious or neutral mood during encoding. Subjects’ mood at re-
trieval was not manipulated. The laboratory mood induction used in Experi-
ment 1 suggested that anxiety might increase the effectiveness of an odor re-
trieval cue. This trend was confirmed in Experiment 2 by capturing a naturally
stressful situation. Subjects who had an ambient odor cue available and were
in a preexam state during encoding recalled more words than subjects in any
other group. These data are evidence that heightened emotion experienced
during encoding with an ambient odor can enhance the effectiveness of an
odor as a cue to memory.
Odor-evoked memories are typically distinguished from other stimulus-
evoked memories by their emotional potency (Laird, 1935; Herz &
Cupchik, 1992). The possibility of such differences is consistent with the
unique neural interconnections between olfactory areas of the central
nervous system and the amygdala–hippocampal complex of the limbic
system. Only two synapses separate the olfactory nerve from the amygda-
la, critical for the expression and experience of emotion (Aggleton &
Mishkin, 1986) and human emotional memory (Cahill, Babinsky,
Markowitsch, & McGaugh, 1995); and only three synapses separate the
olfactory nerve from the hippocampus, involved in the selection and
transmission of information in working memory, short-term and long-
term memory transfer, and various declarative memory functions
(Eichenbaum, 1996; Schwerdtfeger, Buhl, & Gemroth, 1990; Staubli, Ivy,
& Lynch, 1984, 1986). The fact that no other sensory system makes this
kind of direct and intense contact with the neural substrates of emo-
tion and memory has long been advanced as indirect support for the
emotional distinctiveness of odor-evoked memory.
Recently, behavioral data have become available (Herz & Cupchik,
1995; Herz, 1996). We paired 16 emotionally evocative paintings, as to-
be-remembered (TBR) items, with eight odors and eight odor names,
AMERICAN JOURNAL OF PSYCHOLOGY
Winter 1997, Vol. 110, No. 4, pp. 489–505
© 1997 by the Board of Trustees of the University of Illinois
emotion and odor encoding 491
as associated memory cues, in an incidental learning procedure. Paint-
ing recall and the emotionality of memory were tested 48 hr later by
cued recall. The results showed that odor-evoked memories were more
emotional than memories evoked by words, but that the accuracy of
memory was the same with both cue types. More recently, memories
associated with cues represented in either olfactory, tactile, or visual
form were compared (such as the smell of an apple, the feel of an ap-
ple, and the sight of an apple), showing again that odor cues produced
the most emotional memories but that there were no differences in
recall accuracy as a function of sensory cue type (Herz, 1996). Togeth-
er, these findings illustrate that although odors do not differ from oth-
er cues in their capacity to retrieve events from memory, odor-evoked
memories are significantly more emotionally loaded.
Associative learning mechanisms play a major role in the development
of odor responses. Engen (1988) observed that children aged 4 did not
show any differentiation in their response to either butyric acid (ran-
cid butter odor) or amyl acetate (banana odor), but by age 8 had come
to model cultural norms. Similarly, preweanling rat pups who were ex-
posed to a peppermint odor while being stroked or receiving food
showed selective preference and increased 2–deoxy-glucose uptake re-
sponse to peppermint odor later in life (Sullivan & Leon, 1986, 1987).
Based on the importance of associative mechanisms in olfactory learn-
ing and the unique connection between olfaction and emotion, it was
proposed that if an odor is first experienced in an emotionally salient
context, it will be an especially effective memory cue (Herz, 1992). The
reasoning is as follows: Emotional experiences result in higher levels of
activation in the amygdaloid complex, the amygdala has been shown to
be critical for emotional memory (Cahill, Prins, Weber, & McGaugh,
1994; Cahill et al., 1995), and olfactory neurons synapse directly into
the amygdala. Thus, if odor encoding takes place in an emotionally
heightened state, increased limbic activation may cause the connection
between an odor and event to become more tightly fused (for exam-
ple, by long-term-potentiation mechanisms) than if encoding occurs in
a nonemotional state. As a result of the specific odor–event integration
that takes place during encoding, the odor becomes an especially effec-
tive retrieval cue for the associated event.
The present research tested the hypothesis that an ambient odor will
be a more effective retrieval cue if it is encoded in an emotionally charged
state than if encoded in an emotionally neutral state. Two experiments
were conducted in which a distinctive ambient odor was either present
during both the incidental learning session for a list of neutral nouns and
the free recall session or was never present, and subjects were either in
492 herz
an anxious mood during the encoding session or in a neutral mood. At
retrieval, subjects were always in a nonmanipulated mood.
EXPERIMENT 1
Experiment 1 evaluated whether a laboratory induction of anxiety ex-
perienced during the encoding of words in the presence of a distinc-
tive ambient odor would increase the efficacy of that odor as a retriev-
al cue. Anxiety was chosen as the mood to be manipulated for the
following reasons: Unpleasant affective states are more threatening and
hence generally more meaningful than pleasant affective states (Aver-
ill, 1976; Rozin & Fallon, 1987), items learned in arousing contexts (the
ascending arm of the Yerkes–Dodson curve; see Hebb, 1955) are bet-
ter remembered than items learned in nonarousing contexts (Craik &
Blankstein, 1975; Eysenck, 1976), and experimental inductions of anx-
iety are well documented and have proved to be effective in other set-
tings (see Heatherton, Herman, & Polivy, 1991).
Subjects participated in two experimental sessions (encoding and
retrieval) separated by 48 hr. Word list learning was incidental during
the encoding session, and memory for the words was tested by free re-
call 48 hr later at the retrieval session.
METHOD
Design and procedures
A 2 × 2 between-subject factorial design, with odor condition (violet leaf present
at both encoding and retrieval, no odor present) and mood at encoding (anx-
iety induction, neutral induction) as the independent variables, was adhered
to. Twelve subjects were randomly assigned to the four experimental groups,
with an equal number of males and females in each. A full factorial design,
manipulating odor presence/absence at encoding and retrieval, was not
deemed necessary because previous work has clearly shown that an odor is an
effective retrieval cue only if it is present at both encoding and retrieval (Herz,
1997; Schab, 1990).
Subjects
Forty-eight University of Toronto undergraduates (24 males and 24 females)
participated in Experiment 1 in exchange for course credit. All subjects were
individually tested by the same experimenter and were in good respiratory
health. Five of the 48 subjects replaced subjects who did not return for the
retrieval session. When this occurred, a new subject was contacted and run in
the appropriate encoding and retrieval sessions. Of the 5 supplanted subjects,
2 replaced subjects in the no-odor/anxiety group (1 male, 1 female), 1 in the
emotion and odor encoding 493
odor-present/anxiety group (male), 1 in the odor-present/neutral group (fe-
male), and 1 in the no-odor/neutral group (male).
Encoding session
The encoding session room was 7 × 9 ft, carpeted, and illuminated by bright
overhead fluorescent lights. It was furnished with a filing cabinet, a desk, two
chairs, and an old polygraph machine. There were no wall decorations. After
subjects were seated and informed consent obtained, the experimenter ex-
plained that the purpose of the study was to examine the effects of context, both
external environment and internal feeling states, on thought processes. Partic-
ipants in the odor-present conditions were then alerted to the presence of the
ambient odor with the following comment so that a correct attribution between
the odor and the environment would be made (Fernandez & Glenberg, 1985;
Herz, 1996):
You may have noticed that this room has a certain smell to it. This is just how some
of the rooms in this building happen to smell. The reason I draw your attention to
it is because this experiment has to do with context, and smell is one aspect of the
context you are in.
To further ensure that subjects attended to the ambient odor, all subjects
were given a room environment questionnaire (REQ) (Herz, 1997), which
asked for scalar ratings of the room’s lighting, temperature, odor, appearance,
and general comfort. Ratings obtained on the REQ were not statistically ana-
lyzed. After subjects completed the REQ, mood manipulation took place, fol-
lowing which Eich’s autobiographical event generation procedure (Eich,
Macaulay, & Ryan, 1994) was conducted as the incidental word learning task.
For this task, subjects were presented with 16 common, concrete, semantically
unrelated, and affectively neutral English nouns (such as pencil, airplane, key),
selected from the Brown and Ure (1969) word norms as to-be-remembered
(TBR) items. Subjects were read each target word and asked to describe, in a
few sentences, an event that had happened to them that the word reminded
them of. Subjects were told that the event had to be a specific incident (as
opposed to an everyday occurrence) and from at least 1 month before. These
restrictions were imposed to ensure that the words were not superficially pro-
cessed. The experimenter wrote down each event recounted. No time limit was
given for event recollections, so the interval between TBR words was subject
paced and variably determined. Most subjects completed the incidental work
learning task within 20 min. Subjects were then thanked and asked to return
in 2 days for further testing. No mention of future memory tests was ever made.
Ambient odor manipulations. The ambient odor used in the odor-present
conditions was violet leaf. Violet leaf is an unfamiliar and mildly unpleasant
odor (Herz & Cupchik, 1992). It was chosen to be hedonically congruent with
an unpleasant mood. A moderate level of odor intensity was achieved by plac-
ing 12 diethyl phthalate polypropylene pellets saturated with 10% violet leaf
solution in three bowls (4 pellets in each bowl) around the room (Herz, 1997).
One bowl was placed near the subject’s chair, and two other bowls were placed
at equidistant room locations. A floor fan set on low helped circulate the air.
494 herz
The experimenter assessed the room before each subject’s entry to ensure that
the smell was consistent, and refreshed the odor pellets as necessary. Rooms
were completely aerated when there was any change in ambient environment
caused by either subject artifact (such as perfume) or a change in odor exper-
imental conditions. In the odor-absent conditions, there was no manipulated
room scent.
Mood assessment. A mood matrix (affect grid; Russell, Weiss, & Mendelsohn,
1989) was used to measure mood. The mood matrix is composed of nine in-
tersecting columns (horizontal axis) and rows (vertical axis). The horizontal
axis of the matrix corresponds to varying degrees of pleasure, ranging from
extremely unpleasant feelings in the far left column to extremely pleasant feel-
ings in the far right column. The vertical axis corresponds to varying degrees
of arousal, ranging from extremely high arousal in the top row to extremely
low arousal in the bottom row. To indicate mood, the subject places an “X” at
the appropriate location on the matrix. One mark yields two scores that can
range from –4 to +4, one for pleasantness and the other for arousal. Negative
scores indicate unpleasant mood or below-average arousal; positive scores in-
dicate pleasant mood or above-average arousal. A mark in the center of the
matrix (0) corresponds to average (neutral) pleasantness and arousal. Mood
matrix ratings were obtained at both the encoding and retrieval sessions.
Anxious mood induction. Anxiety was induced in the laboratory by using a
“speech threat” manipulation (Heatherton et al., 1991). Subjects were told that
they would have to give 2–min speeches in front of classmates who would crit-
icize them for dysfluencies or ungrammatical style, and that their performance
was a serious measure of verbal fluency. After making this announcement, the
experimenter left the subject alone, saying that she was going to look for the
classmates. The experimenter then returned 5 min later and informed the
subject that the classmates were confused as to the time of arrival and would
be there in about 20 min (that is, after the incidental word learning task). The
subject then rated his or her mood on a mood matrix and the incidental word
learning task began.
Neutral mood induction. Subjects in the neutral mood condition were in-
formed that they would be left alone in the laboratory for 15 min, and that they
should just sit and relax during this time. Several general-interest magazines
were placed on the table adjacent to the subject, and subjects were told they
could read if they chose to. This 15–min waiting procedure was used on the
basis of pilot testing which showed that neutral mood could be reliably manip-
ulated in this way. Subjects were told that the waiting period was for familiar-
ization with the room environment and experimental control. When the ex-
perimenter returned after 15 min, the subject indicated his or her mood using
a mood matrix and the incidental word learning task began.
It should be noted that all subjects spent at least 25 min in the encoding
session room. The somewhat longer exposure to odorant of subjects in the
odor-present/neutral-mood group compared with subjects in the odor-present/
anxious-mood group was not expected to affect olfactory perception in any
detectable way (Dalton, 1996).
emotion and odor encoding 495
Retrieval session
Forty-eight hr later the subject returned for the retrieval session. The retrieval
session room was adjacent to the encoding room, and though similar in size
and general furnishings was clearly distinguishable in specific features. No
mention of ambient odor was made at the retrieval session, regardless of the
ambient odor condition, and no mood was induced. At the start of the retriev-
al session, subjects filled out a new mood matrix and REQ. Since mood was not
manipulated, all subjects were expected to be in an “average” (nonanxious)
mood. Subjects were then given a surprise free recall test for the words pre-
sented during the encoding session, for which they were asked to try to recall
as many words as they could from the first session, in any order, within 10 min.
Word-associated event reminiscences were encouraged to facilitate target word
recall. Subjects spoke the words and any associated memories aloud and the
experimenter wrote down what was recounted. Only recalled target words were
scored. Subjects were then fully debriefed.
RESULTS AND DISCUSSION
Pleasure and arousal ratings
Pleasure and arousal mood matrix ratings were examined in 3–factor
anovas separately for the encoding and retrieval session data. The fac-
tors were odor presence, mood at encoding, and subject sex. At encod-
ing, there was a significant main effect of mood for both pleasure, F =
45.19, p < 0.01, and arousal, F = 15.83, p < 0.01. Mean pleasantness and
arousal ratings for subjects in the anxious and neutral mood groups
were –2.00 and 1.96, and 1.00 and 0, respectively. These results confirm
that subjects who had undergone the anxiety manipulation felt signifi-
cantly more unpleasant and aroused than subjects in the neutral mood
condition. No effects caused by odor presence or subject sex were ob-
tained on the pleasure and arousal ratings and there were no signifi-
cant interactions. At retrieval, no significant differences in ratings of
pleasure or arousal were observed as a function of any of the indepen-
dent variables. Mean scores on pleasure and arousal were 0.87 and 0.83,
and 1.0 and 0.96, for subjects who were formally in anxious or neutral
moods, respectively.
Word recall
The number of words correctly recalled out of 16 for each subject was
evaluated in a 2–factor anova, with odor presence and mood at encod-
ing as the independent variables. A significant main effect for odor
presence was obtained, F(1,44) = 5.78, p < 0.05 (MS
E
= 5.2) (see Figure
1). Subjects who experienced an ambient odor at both encoding and
Figure 1. The effects of laboratory-induced anxious mood or neutral mood at
encoding, and ambient odor presence or absence at both encoding and retriev-
al on recall for words. Word recall is expressed in percentages (± SEM).
emotion and odor encoding 497
retrieval recalled more words (M = 54%) than subjects who did not have
an odor context cue available (M = 44%).
Although the mood by odor interaction was not statistically reliable
(p = 0.45), a trend in the expected direction was observed (see Figure
1). Subjects who underwent the anxiety induction at encoding and had
an ambient odor cue available appeared to recall more words than sub-
jects in the other groups. Effect size (ES) and power were determined
for the interaction term (Cohen, 1988), and an ES of 0.11 with power
of 0.10 was obtained. Thus, the probability that statistically significant
results could be achieved in this experiment was only 10%. Moreover,
it was considered that, despite the differences in mood ratings between
the anxiety and neutral mood groups, the laboratory mood inductions
used may not have been sufficient to produce differential effects on
odor encoding.
EXPERIMENT 2
Experiment 1 suggested that anxiety might potentiate ambient odors
as memory cues. It was suspected that this effect would have been stron-
ger if the mood manipulation had produced more intense mood group
differences. Experiment 2 was therefore designed to explore whether
a more realistic and personally meaningful mood could enhance odor
context cue effects. One common situation for students that induces
substantial natural anxiety is exams (Smith & Ellsworth, 1987). To cap-
ture natural mood fluctuations using this scenario, the encoding ses-
sions for Experiment 2 were run in either the hour just before a mid-
term exam (anxious mood group) or in the hour before a routine class
day (neutral mood group). Retrieval sessions were all conducted in the
hour before a routine class in the same course.
METHOD
Design and procedures
A 2 × 2 between-subject factorial design was used, with odor condition (violet
leaf present at both encoding and retrieval, no odor context cue) and mood at
encoding determined by class condition (preexam, routine class day) as the
independent variables. Subject sex was not included as a variable because of the
unequal numbers of males and females who were available for testing, and be-
cause Experiment 1 indicated that subject sex did not contribute to the data.
Subjects
Forty University of British Columbia undergraduates (17 males, 23 females)
participated in Experiment 2 in exchange for course credit. All subjects were
498 herz
tested by the same experimenter. Subjects were students in a 1–semester evening
course that was offered in the fall and spring terms. Subjects in the odor-present/
preexam and odor-present/routine-class groups were from the fall course, and
subjects in the no-odor/preexam and no-odor/routine-class groups were from
the spring course. The order of subject groups tested was odor-present/preex-
am, odor-present/routine-class, no-odor/preexam, no-odor/routine-class. Thus,
in addition to explicit appeals for confidentiality made during subject debriefing,
the order of groups tested ensured that subjects in the key experimental groups
would not be affected by discussion with classmates. To minimize variations in
other extraneous academic demands that may have influenced performance be-
tween groups, subjects in each course were tested in the middle of the term and
within 2 weeks of each other. Subjects were tested in groups (initially compris-
ing 15–20 students) and were in good respiratory health. Several subjects in each
condition did not return to complete the retrieval phase of the experiment, but
a minimum of 10 subjects remained in each group. To minimize violations to
anova assumptions, group sizes were equated to 10 (based on the minimum
group size) using a randomization procedure.
Encoding and retrieval session conditions
A classroom (26.5 × 13.5 ft), in a separate building from where the psychology
class was held, was used for the encoding and retrieval sessions. There were
approximately 40 individual folding desk chairs in the room, as well as one large
wooden desk. Two of the walls had blackboards, one had windows, and one was
bare. Because subjects were recruited from an evening course that met only
once a week, the interval between encoding and retrieval sessions was 7 days.
At the encoding session, subjects were met by the experimenter in front of
their psychology classroom and escorted to the room used for testing. Subjects
were given the same explanation for the purpose of the study as given to sub-
jects in Experiment 1. Once subjects were seated (subjects were free to sit in
any of the individual chairs), the mood matrix and REQ were administered.
Subjects were then given instructions for a written version of the incidental word
learning task. A written version was used to accommodate group testing. Sub-
jects received sheets with the same 16 words printed on them as were auditori-
ly presented to subjects in Experiment 1. Each printed word was followed by
several blank lines for subjects to fill in with their autobiographical event gen-
erations. When subjects had finished filling in the sheets, they were free to leave
the encoding session.
At the retrieval session, 1 week later, subjects were met by the experimenter
and escorted to the classroom used for the experiment. Once all subjects were
seated, the mood matrix and REQ were administered. Then subjects were giv-
en blank sheets and instructed to write down as many of the target words (and
associations) from the encoding session as they could remember. Only recalled
target words were scored. Ten min were allocated for this task, and subjects were
required to remain seated for the entire time. At the end of the session, sub-
jects were fully debriefed and dismissed.
Ambient odor. The ambient odor used was violet leaf (10%). Ambient room
scent was achieved as previously described.
emotion and odor encoding 499
Mood manipulations. Experiment 2 exploited the natural academic stresses
and nonstressful routines of a psychology class to influence the mood that sub-
jects experienced. Subjects designated as undergoing the anxiety manipulation
were students who participated in the encoding session in the hour just before
their only midterm exam in a psychology course. To further increase anxiety
in the preexam groups, the experimenter soberly reminded the students of the
seriousness and imminence of their upcoming exam. Subjects designated as
undergoing a neutral mood manipulation were students who participated in
the encoding session in the hour just before a routine class in the same psy-
chology course. The experimenter did not make any additional mood com-
ments to these subjects.
RESULTS AND DISCUSSION
Pleasure and arousal ratings
Pleasure and arousal mood matrix ratings from the encoding and re-
trieval sessions were examined in separate 2–factor anovas for the en-
coding and retrieval sessions, respectively. The independent variables
were odor presence and mood at encoding. At encoding, there was a
significant main effect of mood for both pleasure (F = 25.41, p < 0.01)
and arousal (F = 10.29, p < 0.01). Mean pleasantness and arousal rat-
ings for subjects in the preexam and routine class groups were –1.35 and
0.80, and 1.70 and –0.80, respectively. This confirms that subjects in the
preexam groups felt significantly more unpleasant and aroused than
subjects in the routine class groups. At retrieval, no significant differ-
ences on pleasure and arousal ratings were observed as a function of
previous mood condition or ambient odor presence. Mean scores on
pleasure and arousal were –0.25 and 0.70, and 0.70 and 0, for subjects
who had been in the preexam and routine class groups, respectively. No
effects or interactions due to odor presence were found for the plea-
sure or arousal ratings.
Word recall
The numbers of words correctly recalled out of 16 for each subject were
evaluated in a 2–factor anova, with odor presence, and mood at encod-
ing, as the independent variables. Figure 2 shows that the mood-by-odor
interaction was significant, F(1,36) = 6.09, p < 0.01; MS
E
= 5.03; see Fig-
ure 2. Newman-Keuls post hoc comparisons (p < 0.05) confirmed that
subjects who were anxious during encoding and had an ambient odor
cue available recalled more words (M = 59%) than subjects in any oth-
er group. Mean word recall for subjects in the no-odor/anxious-mood,
odor-present/neutral-mood, and no-odor/neutral-mood groups were
33%, 46%, and 41%, respectively. These means did not differ significant-
Figure 2. The effects of naturally occurring anxious mood or neutral mood at
encoding, and ambient odor presence or absence at both encoding and retriev-
al on recall for words. Word recall is expressed in percentages (± SEM).
emotion and odor encoding 501
ly. Effect size and power values were calculated for the interaction term,
yielding an ES of .42 and power of .68. The large ES (Cohen, 1988) and
statistically significant F ratio obtained for the odor-by-mood interaction
demonstrate that a salient emotional state experienced in the presence
of an ambient odor can substantially enhance the effectiveness of an
ambient odor as a memory cue. A main effect of odor presence was also
obtained, F(1,36) = 12.94, p < 0.01, showing that, in general, the pres-
ence of an ambient odor cue enhanced word recall (53% versus 37%).
GENERAL DISCUSSION
Experiments 1 and 2 showed that word recall was higher when an am-
bient odor cue was available during both encoding and retrieval than
when no ambient odor was present. This finding is consistent with evi-
dence that distinctive odors are effective context cues (Cann & Ross,
1989; Herz, 1997; Schab, 1990; Smith, Standing, & de Man, 1992). More
importantly, subjects in Experiment 2 who were anxious during encod-
ing and had an odor context cue available recalled more words than
subjects in any other group. Notably, subjects’ mood at retrieval did not
mediate this effect.
The present results demonstrate that heightened emotion experi-
enced during encoding with an ambient odor can enhance the effec-
tiveness of an odor as a retrieval cue. This finding supports the theory
that emotional potentiation is a key variable in the formation of odor-
associated memory. It was suggested at the outset of this article that such
a finding might be explained by olfactory–amygdala interactions. The
premise is that emotional activation increases amygdala activity, which
because of the direct connections between the olfactory bulb and the
amygdala–hippocampal complex intensifies the association between an
odor and an event in memory. The amygdala is necessary for the expe-
rience of emotion (Aggleton & Mishkin, 1986), and the hippocampus
is a critical mediator for learning and memory associated to context
(Eichenbaum, 1996; LeDoux, 1994; Maren & Fanselow, 1995). Thus
amygdala–hippocampal circuitry in olfactory cognition may be related
as follows. Hippocampal connections may be necessary for the associa-
tion of odor to context (that is, an ambient odor in a particular envi-
ronmental and experiential context), whereas amygdala connections are
necessary for the emotional component of the memory to be formed
and experienced (anxiety in the presence of an ambient odor). Direct
recording from relevant brain areas during odor + emotion encoding
and retrieval episodes or pharmacological blockade experiments such
as those described by Cahill and colleagues (1994) would greatly inform
this hypothesis.
502 herz
It is noteworthy that ambient odors appear to be robust contextual
reminders (Cann & Ross, 1989; Herz, 1996; Schab, 1990; Smith et al.,
1992). Context-dependent memory (CDM) research is known for incon-
sistent findings in studies where other physical cues, such as colors,
sounds, and physical environments, have been manipulated (Smith,
1988). Perhaps odors lend themselves to contextual associability better
than do other stimuli because of their unique interaction with the cor-
tical substrates involved in contextual learning.
Anxiety (arousal) experienced at encoding alone cannot account for
the data obtained in Experiment 2 because anxious subjects who did
not have an ambient odor cue available showed poor word recall. It is
reasonable to surmise that because anxiety was induced by an upcom-
ing threatening event, subjects were distracted from the word-learning
phase of the experiment and hence later showed poor word recall. What
is interesting is that in the presence of a useful contextual cue (such as
an ambient odor), the memory deficits produced by this distraction at
encoding were overridden. Stated otherwise, subjects may require a
contextual cue as a memory aid for peripheral information (that is, the
words to which subjects were exposed in the experiment) to a central
event, when that central event is emotionally meaningful (such as an
impending exam). This possibility is compatible with the hypothesized
role of the hippocampus in olfactory–contextual learning and with
Smith’s (1995) mental context hypothesis, which states that many inci-
dental events become represented in memory in association with a fo-
cal set of stimuli.
Several methodological aspects of this research merit further discus-
sion. First is the issue that different retention intervals were used in
Experiment 1 (2 days) than in Experiment 2 (7 days). Do longer inter-
vals between encoding and retrieval strengthen cuing effects? Eich et
al. (1994) directly tested this question and found that mood-dependent
word recall was substantially worse after 7 days than after 2 days for the
mood-congruent memory groups (but not the mood-incongruent
groups), showing that, if anything, longer retention intervals can dimin-
ish cuing effects rather than enhance them. Notably, in the present
research, the average level of word recall did not vary greatly between
the two experiments (means were 49% and 45% in Experiments 1 and
2, respectively). Thus, it is unlikely that the longer retention interval in
Experiment 2 facilitated the odor–memory effect.
A second issue is that the same room was used for the encoding and
retrieval sessions in Experiment 2, and different rooms were used for
encoding and retrieval in Experiment 1. It might therefore have been
that room artifacts other than the ambient odor manipulated were used
by subjects as retrieval cues in Experiment 2. However, if room cues
emotion and odor encoding 503
other than the ambient odor were involved in memory, then recall
should not have varied between the groups in Experiment 2 in any sys-
tematic way. Moreover, comparable odor-based CDM effects have been
found in studies when the same room was used at the encoding and
retrieval sessions (Schab, 1990) and when different rooms were used
(Herz, 1997). Thus, it seems reasonable to conclude that extraexperi-
mental room cues did not bias the results obtained in Experiment 2.
Third, it might be argued that the hedonic congruency between the
ambient odor and mood (both unpleasant) contributed to the findings.
Odors have been shown to have mood-altering effects (Ehrlichman &
Bastone, 1992; Schiffman, Sattely-Miller, Suggs, & Graham, 1995). Thus,
violet leaf (a mildly unpleasant odor) may have contributed to being in
an unpleasant mood. This possibility is refuted in the present experi-
ments by the finding that at both encoding and retrieval, subjects’ mood
was not affected by the presence of the ambient odor alone. Schab
(1990) also found that the hedonics of an ambient odor were not re-
lated to CDM effects, as both pleasant (chocolate, apple–cinnamon) and
unpleasant (mothball) odors produced positive outcomes. Nonetheless,
it is possible that an odor vastly different in hedonics from one’s mood
state might be jarring, and thus more attention might be paid to it (Jac-
coby & Craik, 1979). Such a situation would be comparable to the mem-
ory benefits observed when learning and retrieval take place in the
presence of a distinctive ambient odor, as compared to a nondistinctive
ambient odor (Herz, 1997). Direct tests of the relationship between
mood and odor hedonics would now be informative.
A final consideration is that the significant results reported here were
obtained with anxiety as the operative mood state. It is therefore possi-
ble that the findings were caused by the specific effects of experienc-
ing negative emotion or high arousal, as opposed to the general phe-
nomenon of heightened emotional activity (good, bad, calming, or
energizing). Experiments separating the relative contribution of arousal
from pleasantness should be conducted so that the parameters of pleas-
antness and arousal that mediate odor cue potentiation can be defined.
Furthermore, whether emotion can generally enhance the effectiveness
of potential context cues irrespective of the sensory modality through
which they are perceived, or whether olfaction is singularly affected by
such manipulations, must be explored.
Notes
The author expresses thanks to Shelley Hodder for collecting the data obtained
in Experiment 1; Sharon Zeitlin for providing the laboratory space used in
Experiment 1; Eric Eich for providing methods, materials, and critical insight;
504 herz
Russ Mason, Steve Smith, Paula Hertel, and Howard Eichenbaum for helpful
commentary; Larry Clark and Pamela Dalton for statistical consultation; and
International Flavors and Fragrances for donating the violet leaf odorant used
in this research. This research was conducted under an NSERC postdoctoral
fellowship.
Correspondence concerning this article should be addressed to Rachel Herz,
Monell Chemical Senses Center, 3500 Market Street, Philadelphia, PA 19104
(E-mail: herz@pobox.upenn.edu). Received for publication May 12, 1996; re-
vision received August 1, 1996.
References
Aggleton, J. P., & Mishkin, M. (1986). The amygdala: Sensory gateway to the
emotions. In R. Plutchik & H. Kellerman (Eds.), Emotion: Theory, research
and experience: Vol 3. Biological foundations of emotion. Orlando, FL: Academic
Press.
Averill, J. R. (1976). Emotion and anxiety: Sociocultural, biological, and psy-
chological determinants. In M. Zuckerman and C. D. Spielberger (Eds.),
Emotions and anxiety: New concepts, methods, and application. Hillsdale, NJ:
Erlbaum.
Brown, W. P., & Ure, D. M. J. (1969). Five rated characteristics of 650 word
association stimuli. British Journal of Psychology, 60, 233–249.
Cahill, L., Babinsky, R., Markowitsch, H. J., & McGaugh, J. L. (1995). The
amygdala and emotional memory. Nature, 377, 295–296.
Cahill, L., Prins, B., Weber, M., & McGaugh, J. L. (1994). Beta-adrenergic acti-
vation and memory for emotional events. Nature, 371, 702–704.
Cann, A., & Ross, D. A. (1989). Olfactory stimuli as context cues in human
memory. American Journal of Psychology, 102, 91–102.
Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2d ed.). Hills-
dale, NJ: Erlbaum.
Craik, F. I. M., & Blankstein, K. R. (1975). Psychophysiology and human mem-
ory. In P. H. Venables & M. J. Christie (Eds.), Research in psychophysiology
(pp. 388–417). New York: Wiley.
Dalton, P. (1996). Odor perception and beliefs about risk. Chemical Senses, 21,
447–458.
Ehrlichman, H., & Bastone, L. (1992). The use of odour in the study of emo-
tion. In S. Van Toller & G. H. Dodd (Eds.), Fragrance: The psychology and
biology of perfume. Amsterdam: Elsevier.
Eich, E., Macaulay, D., & Ryan, L. (1994). Mood dependent memory for events
of the personal past. Journal of Experimental Psychology: General, 123, 201–
215.
Eichenbaum, H. (1996). Olfactory perception and memory. In R. Llinas & P.
Churchland (Eds.), The mind-brain continuum (pp. 173–202). Cambridge,
MA: MIT Press.
Engen, T. (1988). The acquisition of odor hedonics. In S. Van Toller & G. H.
Dodd (Eds.), Perfumery: The psychology and biology of fragrance. New York:
Chapman & Hall.
emotion and odor encoding 505
Eysenck, M. W. (1976). Arousal, learning and memory. Psychological Bulletin, 83,
389–404.
Fernandez, A., & Glenberg, A. M. (1985). Changing environmental context
does not reliably affect memory. Memory & Cognition, 13, 333–345.
Heatherton, T. F., Herman, C. P., & Polivy, J. (1991). Effects of physical threat
and ego threat on eating behavior. Journal of Personality and Social Psycholo-
gy, 60, 138–143.
Hebb, D. O. (1955). Drives and the CNS. Psychological Review, 62, 243–254.
Herz, R. S. (1992). The relationship between odor and emotional memory. Unpub-
lished Ph.D dissertation, University of Toronto, Canada.
Herz, R. S. (1996). A comparison of olfactory, visual and tactile cues for emo-
tional and non-emotional associated memories. Chemical Senses, 21, 614–
615.
Herz, R. S. (1997). The effects of cue distinctiveness on odor-based context
dependent memory. Memory & Cognition, 25, 375–380.
Herz, R. S., & Cupchik, G. C. (1992). An experimental characterization of odor-
evoked memories in humans. Chemical Senses, 17, 519–528.
Herz, R. S., & Cupchik, G. C. (1995). The emotional distinctiveness of odor-
evoked memory. Chemical Senses, 20, 517–528.
Jaccoby, L. J., & Craik, F. I. M. (1979). Effects of elaboration of processing at
encoding and retrieval: Trace distinctiveness and recovery of initial con-
text. In L. S. Cermak & F. I. M. Craik (Eds.), Levels of processing in human
memory (pp. 1–22). Hillsdale, NJ: Erlbaum.
Laird, D. A. (1935). What can you do with your nose? Scientific Monthly, 41, 126–
130.
LeDoux, J. E. (1994). Emotion, memory and the brain. Scientific American, 270,
50–57.
Maren, S., & Fanselow, M. S. (1995). Synaptic plasticity in the basolateral
amygdala induced by hippocampal formation stimulation in vivo. The Jour-
nal of Neuroscience, 15, 7548–7564.
Rozin, P., & Fallon, A. E. (1987). A perspective on disgust. Psychological Bulle-
tin, 94, 397–412.
Russell, J. A., Weiss, A., & Mendelsohn, G. A. (1989). Affect grid: A single-item
scale of pleasure and arousal. Journal of Personality and Social Psychology, 57,
493–502.
Schab, F. R. (1990). Odors and the remembrance of things past. Journal of Ex-
perimental Psychology: Learning, Memory and Cognition, 16, 648–655.
Schiffman, S. M., Sattely-Miller, E. A., Suggs, M. S., & Graham, B. G. (1995).
The effect of pleasant odors and hormone status on mood of women at
midlife. Brain Research Bulletin, 36, 19–29.
Schwerdtfeger, W. L., Buhl, E. H., & Gemroth, P. (1990). Disynaptic olfactory
input to the hippocampus mediated by stellate cells in the entorhinal cor-
tex. Journal of Comparative Neurology, 194, 519–534.
Smith, C. A., & Ellsworth, P. C. (1987). Patterns of appraisal and emotion re-
lated to taking an exam. Journal of Personality and Social Psychology, 52, 475–
488.
506 herz
Smith, D. G., Standing, L., & de Man, A. (1992). Verbal memory elicited by
ambient odor. Perceptual and Motor Skills, 74, 339–343.
Smith, S. M. (1988). Environmental context-dependent memory. In G. M.
Davies & D. M. Thomson (Eds.), Memory in context: Context in memory. New
York: Wiley.
Smith, S. M. (1995). Mood is a component of mental context: Comment on
Eich (1995). Journal of Experimental Psychology: General, 124, 309–310.
Staubli, U., Ivy, G., & Lynch, G. (1984). Hippocampal denervation causes rap-
id forgetting of olfactory information in rats. Proceedings of the National
Academy of Sciences USA, 81, 5885–5887.
Staubli, U., Ivy, G., & Lynch, G. (1986). Studies on retrograde and anterograde
amnesia of olfactory memory after denervation of the hippocampus by
entorhinal lesions. Behavioral Neurology and Biology, 46, 432–444.
Sullivan, R. M., & Leon, M. (1986). Early olfactory learning induces an en-
hanced olfactory bulb response in young rats. Developmental Brain Research,
27, 278–282.
Sullivan, R. M., & Leon, M. (1987). One-trial olfactory learning enhances ol-
factory bulb responses to an appetitive conditioned odor in 7–day-old rats.
Developmental Brain Research, 35, 307–311.
x
