You are on page 1of 12

CONSCIOUSNESS AND COGNITlON 4, 410-421 (1995)

A Novel Demonstration of Enhanced Memory Associated


with Emotional Arousal

1
LARRY CAHILL AND JAMES L. MCGAUGH

Center for the Neurobiology of Leaming and Memory and Department of Psychobiology,
University of California, Irvine, California 92717-3800

The relationship between emotional arousal and long-term memory is addressed in two
experiments in which subjects viewed either a relatively emotionally neutral short story
(presented as a brief slide show) or a closely matched but more emotionally arousing story
and were tested for retention of the story 2 weeks later. Experiment I provides an essential
replication of the results of Heuer and Reisberg ( 1990) and illustrates the common interpre-
tive problem posed by the use of different stimuli (slides) in the neutral versus emotional
stories. In Experiment 2, identical slides (and sequence) were used in both the neutral and
arousal stories. Two different stories were created by varying the narration that accompanied
each slide. In both experiments, subjects who viewed the arousal story both experienced
a greater emotional reaction to the story than did the subjects who viewed the neutral
story, and subsequently exhibited enhanced memory for the story. Subjects in Experiment
2 who viewed the arousal story also recalled more slides than did the subjects who viewed
the neutral story. This effect was greatest for story phase 2, the phase in which the emotional
slide narration occurred. Because this enhanced retention of the story slides cannot be
explained by any differences in the slides themselves, the results provide new evidence
to support the contention that emotional arousal influences long-term memory in normal
human subjects. :D 1995 Academic Press. Inc.

INTRODUCTION
Considerable evidence suggests that emotional arousal influences memory. Con-
trolled laboratory investigations as well as ''field studies'' of memory in human sub-
jects have demonstrated that emotional arousal can be associated with altered memory
for the emotional events (e.g., Christianson & Loftus, 1987; Heuer & Reisberg, 1990;
Melzter, 1930; Revelle & Loftus, 1992; Yuille & Cutshall, 1986). Although both ap-
proaches have merits, laboratory investigation offers obvious advantages over field
studies in the control of experimental variables and assessment of memory accuracy.
In a typical laboratory study, subjects view either a set of stimuli designed to evoke
an emotional response or control stimuli designed to be as similar as possible to the
emotional stimuli, but lacking their emotional impact. Thus, for example, subjects
may view an emotionally arousing story or a story that is closely matched in con-
tent, complexity, style, and comprehensibility, but lacking the emotional impact
(Christianson, 1984; Christianson & Loftus, 1987; Heuer & Reisberg, 1990; Burke,
Heuer, & Reisberg, 1992). Memory for the stories is then tested. By using stories that
are matched as completely as possible except for the emotional reactions produced,

1
To whom reprint requests should be addressed. E-mail: lcahill@parker.bio.uci.edu; Fax: (714) 824
2952.
410
1053-8100/95 $12.00
Copyright I995 by Academic Press, Inc
All rights of reproduction in any form reserved
EMOTION AND MEMORY 411

investigators may more confidently conclude that observed differences in memory for
the two stories are due to the emotional reactions elicited when they were presented.
There exists, however, a problem (Christianson & Loftus, 1987) inherent in each
published experiment to date of which we are aware. Subjects in the emotional condi-
tions viewed stimuli (typically pictures) that were not viewed by subjects in the neu-
tral conditions. Of course, stories which produce different emotional reactions must
by necessity differ in many respects. However, the use of different stimuli for subjects
in the neutral and arousing conditions presents an interpretive problem: It is possible
that some features of the to-be-remembered (TBR) items other than their differing
emotional consequence (for example, differential novelty of the elements shown)
account for the subsequent retention differences. In other words, memory differences
between the neutral and emotional groups may result from "story effects" (Burke
et al., 1992), such as the novelty or inherent memorability of the stimuli used, rather
than "emotion effects."
The primary experiment described in this report (Experiment 2) attempted to cir-
cumvent this problem (at least for the visual TBR material) by using the same slides to
generate two different stories-one relatively neutral and one much more emotionally
arousing. The stories were matched as closely as possible for complexity and compre-
hensibility, and differed primarily in the emotional reaction they produced. The use
of identical visual elements for both the neutral and arousing stories assures that any
differences in memory for the visual elements cannot be due to differences inherent
in the visual material. The stories were designed to be very similar to those already
used by earlier investigators (Heuer & Reisberg, 1990), who reported enhanced long-
term memory associated with emotional arousal in their paradigm. We first replicated,
in part, the study of Heuer and Reisberg ( 1990) using somewhat different procedures.
The results of this experiment (Experiment I) indicate that their findings are robust,
but also clearly indicate the need to control for potential "story effects" which may
confound interpretations of altered memory based on emotional arousal. The issue
of emotions' effects on central versus peripheral story information investigated by
Heuer and Reis berg ( 1990) is not germane to the primary issue of interest here and
is therefore not addressed.

EXPERIMENT 1
Methods
Subjects. Twenty undergraduate students ( 17 female, 3 male, average age 23.6
years) served as subjects and received course credit for their participation in this
experiment. Nine subjects (8 female, l male) were randomly assigned to the neutral
condition, and 11 (9 female, 2 male) were assigned to the arousal condition.
Materials and procedure. The stimulus materials were, in part, those used by Heuer
and Reis berg ( 1990), who generously provided copies of the slides used in their
study. The subjects viewed one of two versions of a simple story-a neutral or an
arousal version-each consisting of 12 slides, each accompanied by one sentence of
narration played on tape. Both stories depicted a mother taking her young son to visit
his father at work. In the neutral version the father is an auto mechanic repairing a
car. In the arousal version the father is a surgeon working on victims of an automobile
412 CAHILL AND MCGAUGH

accident. In both stories the son watches the father work. The two stories differed
primarily in the middle 5 slides. It was in these slides (which will be referred to
subsequently as phase 2) that the emotional elements were introduced into the
"arousal" story. These consisted mostly of 2 graphic slides, I of surgeons operating
on the exposed chest cavity of a victim, the other of a victim's badly scarred and
disfigured legs. The middle 5 slides (phase 2) in the neutral condition depicted scenes
of the father's garage, at which the father was repairing a car. For both groups, the
first 3 slides (phase 1) and the last 4 slides (phase 3) were identical, and the narration
for both was either identical or closely matched.
The subjects were tested individually. Upon arrival, each was first asked to wash
their hands in preparation for "physiological measurements" during the testing and
was then seated at a small table on which sat a slide viewing screen (I 9 by 14 cm)
surrounded by a black board (80 X 80 cm) used to help focus the subject's attention
on the viewing screen. They were then asked to read a prepared statement explaining
that we were interested in studying ''physiological responses that occur in response
to a variety of different visual and auditory stimuli and situations" and that they
would likely find some slides "pleasant, some unpleasant, and some basically neu-
tral.'' They were instructed only to attend to each slide for the duration of its presenta-
tion and to view the show "as you would a movie." They were told that the slide
presentation would last only 4 min and that their continuous attention to it was neces-
sary for the accurate measurement of physiological responses. To convince the sub-
jects that we were interested in physiological responses, patch electrodes leading to
a hidden recording device were attached to the middle and index fingers of the sub-
ject's right hand (using Velcro strips) after cleaning the fingers with 95% ethanol.
Subjects were instructed to keep finger movements to a minimum, as movements
would "disrupt the recording of the skin responses." During the slide presentation,
the experimenter sat behind the partition around the slide screen and pretended to
be recording skin responses.
Subjects first viewed a series of nine unrelated slides. This was done to acclimate
the subjects to the testing situation. As Heuer and Reisberg ( 1990) did not use prelimi-
nary slides, our procedure differed from theirs in this respect. After a delay of approx-
imately 5 min, the story slides were presented. Each slide was shown for 20 s, with
the taped narration that accompanied each slide beginning immediately after the slide
appeared. At the end of the story presentation, the subject was asked to rate on a
scale of O to 10 (with O as "not emotional" and 10 as "highly emotional") how
emotional they found the story to be by marking the scale at the appropriate point.
The subject was asked to return in 2 weeks, at which time they expected to view a
new set of slides. They were given no indication that their memory would be tested.
The subjects were also asked not to discuss the experiment with other students who
may be participating in the experiment, as this would jeopardize the results.
On their return 2 weeks later, the subject was informed that their memory of the
story viewed 2 weeks earlier would be tested. Prior to the start of memory testing,
the subjects were asked whether they had anticipated a memory test despite our efforts
to hide this fact (all reported that the memory test was completely unsuspected). The
testing was divided into two phases: a free recall test and a recognition test. For the
free recall test, subjects were asked to recall as much as possible of the story they
EMOTION AND MEMORY 413

viewed and their responses were recorded on tape for later analysis. Specific instruc-
tions were given to recall as much as possible of the general story line, as well as
any specific details they remembered having seen or heard. When they indicated that
they had recalled as much as possible, they were encouraged once again to recall
any other details (such as color of clothing, etc.) that they could and to describe in
as much detail as they could any slides that they remember having seen. This latter
procedure often elicited more information from the subjects, despite their earlier state-
ment that they had recalled as much as possible. After the free recall test, subjects
were given a recognition test consisting of four-choice multiple choice questions (five
to eight questions per slide). The questions and possible answers were read to the
subjects, who were asked to give a reply even if forced to guess. The questions
pertained both to story elements presented visually as well as in the narration. The
order of the questions was the same for each subject. With one exception, the ques-
tions and possible answers relating to the first 3 (phase I) and the last 4 (phase 3)
slides were identical. The questions relating to the middle 5 (phase 2) slides were
matched as closely as possible and were in many cases identical. Each subject was
told prior to the test that the order of the questions would follow the story sequence
as it had been presented. Furthermore, each was kept informed of the number of the
slide (of 12 total) to which the questions referred. After completion of the recognition
test, the subjects were informed of the intent of the study.
Although it was not possible to score the results of the recall test ''blind'' to the
experimental group (because of obvious differences in the stories). the results were
scored as uniformly as possible. Two judges listened to the taped responses of all
subjects and created a list of acceptable story elements. The responses of each subject
were then scored against this list. Most of the responses given in the free recall test
could clearly be identified as story elements (e.g., "the mother had short hair" or
"the father worked in a hospital"); consequently, there was a very large (>%90)
degree of agreement between the two judges on what constituted acceptable story
elements. A small percentage of the responses were either clear factual errors or were
considered too vague (e.g., "it was a typical suburban neighborhood," "the father
looked Polish'') to represent a specific memory of a story element. In such cases the
responses were excluded from the analysis. Decisions about the small percentage of
questionable responses were decided by the second judge (LC.)

Results of Experiment 1
Emotionality ratings of stories. The arousal and neutral groups differed signifi-
cantly in their mean ratings of the emotionality of the stories. The mean (SEM)
rating for the arousal group was 5.36 (0.47) versus 2.31 (0.67) for the neutral
group [t( 17) = 3.83, p < .05]. Thus, the stories effectively produced different levels
of emotional arousal in the two groups.
Free recall test. The arousal group recalled more story elements ( 18 2.01) than
did the neutral group (13 1.78), although the difference between the groups did
not reach statistical significance [t(18) = 1.82, 0.05 < p < .IO].
Recognition test. There was no overall difference between the two groups on the
recognition test. Neutral subjects correctly answered an average of 52% of the ques-
414 CAHILL AND MCGAUGH

70
-0-- neutral
Si'
l1J
en
* _,._ arousal
.!.
1it: 60

8
'E

i C
50

m
:::E
40
Phase 1 Phase 2 Phase 3

F10. I. Phase-by-phase analysis of the recognition test results in Experiment l. *p < .02 from the
phase 2 neutral group value.

tions, and the arousal subjects were correct on 57% [t(] 8) = l.66, n.s.]. However,
a phase-by-phase comparison of the results revealed a difference between the groups
for one phase of the story. As shown in Fig. 1, the arousal subjects answered signifi-
cantly more questions pertaining to phase 2 of the story correctly than did the neutral
subjects [t( 18) = 2.81, p < .02]. The same analysis showed no differences between
the groups for either story phase 1 or phase 3.
Finally, we examined the recognition test results on a slide-by-slide basis. These
results are shown in Fig. 2. This analysis clearly suggests that the improved phase
2 performance of the arousal subjects on the recognition test depended heavily, but
not exclusively, on the results from one slide (No. 7), the picture of the bloody legs
(arousal story) or engine part (neutral story).

Discussion of Experiment 1
Using the same stimulus materials used by Heuer and Reis berg ( 1990), but with
somewhat different procedures, we obtained results comparable to theirs in certain
key respects. First, the two groups clearly differed in their emotional reaction to the
stories, with those viewing the arousal story expressing a substantially greater emo-
tional reaction than those viewing the neutral story. Second, although there was no

100

i
en
90

.:J. 80 ---0-- Neutral


u 70
----- Arousal
~
8 60
E
~ 50
8.
.,
C 40
j 30

20
1 2 3 4 5 6 7 8 9 10 11 12

Phase 1 Phase 2 Phase 3


FIG. 2. Slide-by-slide analysis of the recognition test results in Experiment I.
EMOTION AND MEMORY 415

overall difference between the neutral and arousal story groups in recognition test
performance, a phase-by-phase analysis showed that the arousal subjects had better
retention of story phase 2, the phase in which the emotional events were introduced
in the arousal story, than did the neutral subjects.
Because the procedures used in this experiment differed in several respects of from
those of Heuer and Reis berg ( I 990; for example, the time of slide exposure), the
findings presented here suggest that the general findings are robust and reliable (at
least to the extent the analyses overlap). However, as indicated by Fig. 2, it also
appears that much of the enhanced phase 2 retention performance resulted primarily
from one slide, number 7. For the arousal story, this slide was a picture of bloody,
scarred legs of a young person lying on a table, whereas for the neutral story this
slide was a piece of an automobile engine lying on the ground. It is distinctly possible
that the former slide was remembered better than the latter for reasons unrelated to
the emotional reaction each produced. It seems likely that a picture of a person's
badly disfigured legs (seen by the arousal subjects) might be, to most people, more
memorable than a piece of an automobile engine (the comparable picture seen by
the neutral subjects) for reasons not related to the relative emotional impact of the
pictures. The former picture may be more memorable because of its relative novelty,
or perhaps because it deals with familiar human anatomy as opposed to automobile
engine parts, which may be very unfamiliar to many subjects. If such explanations
account for the differential retention of slide 7, the general conclusion that emotional
arousal produced the enhanced phase 2 performance in this paradigm would be seri-
ously questioned. This problem could be avoided (at least for the visual stimuli) if
the same pictures were used in stories which, although closely matched in form,
differed in the emotional reaction they produced. Experiment 2 addresses this issue.

EXPERIMENT 2
In Experiment 2, identical pictures were shown to subjects in both the neutral and
arousal conditions, but differential emotional arousal was induced by narrations that
accompanied the pictures. To achieve this, slides that could reasonably be interpreted
in at least two different ways (at least for phase 2) were used. In fact, the use of
identical visual elements (film clips) as the basis for two different stories has been
used successfully in two unpublished studies of which we are aware (see Burke et
al., I 992, for a description of these studies). The two stories used in Experiment 2
were similar to those of Experiment l in content and design. Both stories were sep-
arable into three phases, and both stories differed primarily in the second phase, in
which the emotionally arousing elements were introduced in the arousal version.
Because the visual elements used in both stories were identical, differences in reten-
tion of those elements cannot be attributed to intrinsic differences in the visual ele-
ments themselves, as is possible for the results Experiment 1.

Methods
Subjects. Eighteen undergraduate students (13 women, 5 men, average age 20.9
years) participated in this experiment (9 subjects in the neutral group, 9 subjects in
the arousal group). Each received course credit for participating in the experiment.
416 CAHILL AND MCGAUGH

Materials and procedure. The procedure used was identical to that of Experiment
l, with the following exceptions. First, no slides other than those used in the memory
testing were presented. Second, in addition to rating the emotionality of the story,
each of the subjects were asked to judge how well they felt they had understood the
story they viewed by making a mark on a scale of Oto 10, with O being "not at all"
and 10 being "completely." Finally, the slides were projected onto a larger screen
(22.5 X 16.5 cm) than that used in Experiment I.
A total of 12 slides (IO of which were taken from Experiment I) were used as the
basis for two stories. The stories were created by varying the taped, one-sentence
narration that accompanied each slide (see Appendix). In both stories, a mother takes
her young son to visit his father at the hospital where he works. In the neutral version,
the son watches as the staff conduct a practice disaster drill, after which the mother
leaves to pick up her other child at preschool. In the arousal version, the boy is badly
hurt in an automobile accident, and the surgeons struggle to save the boy's life. The
same pictures which in the neutral version were described as relatively routine aspects
of a disaster drill watched by the boy are described in the arousal version as attempts
to save the boy. Thus for example, a picture of badly scarred legs is described in the
neutral version as merely an actor made up to appear injured for the disaster drill,
but in the arousal version is described as the boy after his feet had been reattached
by surgeons. Care was taken to match the two stories as much as possible in complex-
ity and comprehensibility, as well as in the grammatical and syntactic structure of
the sentences. As in Experiment 1, the narration accompanying the first four slides
was identical in both stories (phase I) and highly similar for the last three slides
(phase 3, some slight differences were necessary to maintain story continuity). The
narration differed between the two stories primarily during the middle five slides
(phase 2).
As in Experiment I, subjects returned 2 weeks later expecting to see a new set of
slides, but instead received a memory test consisting first of free recall followed by
a recognition memory test. The recognition test consisted of 80 questions for both
groups, 5 to 8 for each slide. As in the first experiment, one correct and three plausible
alternatives were given for each question, and questions pertaining to both visual
and narrative aspects of the stories were asked. For the analysis of individual slides
recalled, subjects were credited with remembering a slide if they described some
feature that was only visible in that slide and not in any other slide or mentioned in
the narrative. Because of the high degree of agreement between the two judges and
the low overall incidence of ambiguous responses in Experiment 1, a single judge
(LC.) scored the recall tapes in this experiment.
Results
Subjects in the arousal condition rated their story as significantly more emotional
than did those in the neutral condition. The mean (SEM) rating for the arousal story
was 6.28 (0.71) and for the neutral story 2.94 (0.62) [t(l6) = 3.53, p < .Ol]. In
contrast to the results of emotional ratings, the two groups did not differ in the ratings
of the understandability of the stories. The average rating of understandability by the
neutral subjects was 8.28 (0.76) versus 8.94 (0.64) for the arousal subjects [t(l6)
= 0.67, n.s.].
EMOTION AND MEMORY 417
75
* --0- Neutral Story
_.,.._ Arousal Story

35 - ' - - ~ - - - ~ - - - - . . - -
2 3

StoryPhase
F1G. 3. Phase-by-phase analysis of the recognition test results from Experiment 2. *p < .02 from
phase 2 neutral group value.

Recognition test. As in Experiment I, the two groups did not differ in total recogni-
tion scores. The mean number of questions correctly answered (SEM) was 42.2
( 2.11) for the neutral subjects and 45.1 ( 1.25) for the arousal subjects [t(l 6) =
1.18, n.s.). However, a phase-by-phase analysis revealed a clear difference between
the two groups for phase 2 of the stories (Fig. 3). Subjects in the arousal group
answered significantly more questions pertaining to story phase 2 correctly than did
the neutral subjects [t(l 6) = 2.77, p < .02). The two groups did not differ in recogni-
tion test performance for either phase 1 or phase 3.
Comparison of the phase-by-phase performance within each group demonstrated
a clear difference between the neutral and arousal groups. A one-factor (phase)
ANOV A showed no effect of phase for the neutral group, but a highly significant
effect of phase for the arousal group [F(2, 24) = 10.42, p < .001]. Individual t test
comparisons indicated that the percent of questions correctly answered by subjects
in the arousal group was significantly higher for phase 2 compared to both phase I
[t(l6) = 4.93, p < .001) and phase 3 [t(l6) = 3.27, p < .05].
Free recall test. Subjects in the arousal group recalled significantly more total
information than did subjects in the neutral group. The mean (SEM) number of
elements recalled for each group were: 23 ( 1.94) for the arousal subjects and 15.6
(2.45) for the neutral subjects [t(16) = 2.38, p < .05].

4
* --er- Neutral Story
irn Arousal Story

,,,.::!.
.!!
~
.
IC
"'
:E
in
2

C
ca
i
0
Phase 1 Phase 2 Phase 3
F1G. 4. Phase-by-phase analysis of the free recall of slides in Experiment 2. *p < .001 from phase 2
neutral group value.
418 CAHILL AND MCGAUGH

"'
= 9
neutral
0 8
..."'
(1,1 7
arousal
Q.
....
0
6

...
11,1 4
~ 3
s
=
=
0
0 2 3 4 5 6 7 a 9 10 11 12

phase 1 phase 2 phase 3

F1G. 5. Total number of persons (of nine in each group) recalling each of the twelve slides used in
Experiment 2.

There was a highly statistically significant difference between the two groups in
the number of individual slides recalled. Overall, arousal subjects recalled a mean
of 2.56 more slides than the neutral subjects [3.22 for neutral, 5. 78 for arousal, t(l6)
= 3.89, p < .01]. Furthermore, as can be seen in Fig. 4, most of this difference
resulted from greater recall of phase 2 slides. Although arousal subjects recalled more
slides on average from each story phase than did the neutral subjects, the difference
between the groups was significant only for phase 2 slides [t(l6) = 4.4, p < .001].
Finally, Fig. 5 shows the number of subjects (of nine in each group) who recalled
each of the 12 slides. Most of the slides recalled better by the arousal subjects oc-
curred in phase 2. One apparent exception to this rule is slide 3, the picture of the
father at work in the hospital, which was recalled by six of nine arousal subjects,
but only one of nine neutral subjects. The slide recalled by the most subjects in both
groups was number 9, the picture of the disfigured legs. Six of nine neutral, and nine
of nine arousal, subjects, recalled this slide. In contrast, slide number 7, a pseudo-
colored image of a brain, was very poorly recalled even by the arousal subjects.

DISCUSSION
Overall, the results of Experiment 2 parallel those of Experiment 1. In both experi-
ments, subjects who viewed the arousing story experienced a substantially larger
emotional reaction than did the subjects who viewed the neutral story. In addition,
in both experiments the subjects who viewed the arousal story recalled more story
elements than did the subjects who viewed the neutral story (although this difference
was statistically significant only in Experiment 2). Finally, although in both experi-
ments no overall difference betweeen the neutral and arousal groups in recognition
test performance was found, a phase-by-phase analysis of the recognition test results
in both experiments clearly showed enhanced memory for phase 2 of the arousal
stories, in which the emotional elements were introduced.
The most significant new finding of Experiment 2 is the improved free recall of
individual slides by the arousal compared to neutral subjects. Recall of the same
information (visual elements contained in the slides) was improved by altering the
description of what the slides depicted from one more mundane to one much more
emotionally compelling. Furthermore, the enhanced memory for the slides was due
EMOTION AND MEMORY 419

largely to an increase in phase 2 recall, the phase in which the emotional descriptions
occurred. The enhanced memory cannot be due to any difference in the content of the
slides themselves and must therefore have been due to some aspect of their assigned
meaning. Inasmuch as the subjects expressed equal degrees of comprehension for
both stories, it seems likely that the better memory for the slides seen in subjects in
the arousal group was due to the greater emotional reaction to the arousal story.
Several additional findings support the view that the emotional arousal, and not
some other feature of the stories, was responsible for the enhanced memory for the
emotional story. First, in both experiments, the subjects reporting a greater emotional
reaction to their story (the arousal subjects) also showed better memory. Second, in
both experiments, improved memory was primarily associated with phase 2 for each
story, the phase containing the emotional material. This was true for recognition
testing in both experiments and in the analysis of slide recall in Experiment 2. No
differences between the groups were found for memory for either phase l or 3 in
either experiment. Thus it appears that under these conditions emotional arousal pro-
duced no, or at best limited, anterograde or retrograde effects on memory.
It may be argued that the arousal story in Experiment 2 contained a clearer central
focus (the boy being hurt) than did the neutral story and that it was this factor, not
the greater emotional story content, that produced better retention in the arousal sub-
jects. Thus for example, a picture of wrecked cars (slide 5) was described in the
neutral story as something that simply attracted the boy's interest, whereas in the
arousal story the cars are the cause of his injuries. In fact, this slide was recalled by
many more arousal than neutral subjects (see Fig. 5). However, in a separate analysis
(not reported) we have found that the proportion of story elements considered (by
independent judges) to be central versus peripheral to the main story line are highly
similar for both stories in Experiment 2. Thus, it does not appear to be the case that
the two stories differed greatly in the degree to which they contained a central focus.
Furthermore, any potential differences between the stories in this respect cannot ac-
count for the within-story differences in phase-by-phase performance that were evi-
dent in the arousal, but not neutral story groups. That is, the superior retention of
the phase 2 (relative to phases I and 3) arousal story elements cannot be explained
by any potential differences in the neutral and arousal stories.
The degree to which memories of emotional events are conscious or unconscious
is currently the subject of much debate. Indeed, considerable evidence indicates that
some form of memory for an emotionally arousing event can persist in the absence of
conscious recall of the event (for examples, see Tobias, Kihlstrom, & Schacter, 1992).
In the most often cited anecdote, Claparede ( 1911, cited in Tulving, 1983) painfully
pricked the hand of an amnesic patient while shaking her hand. Shortly thereafter, the
patient refused to shake his hand, although she had no conscious recollection of the
painful incident and had difficulty explaining why she was reluctant to shake his hand.
The patient apparently retained some memory of the event, but this memory was not
consciously accessible. While compelling, demonstrations such as this do not prove
the existence of a separate class of unconscious memory for emotional events (LeDoux,
1992); rather, they may simply represent specific examples of a broader class of uncon-
scious memory most commonly referred to as implicit memory (Tobias et al., 1992).
In contrast, substantial evidence indicates that emotional arousal influences conscious
420 CAHILL AND MCGAUGH

recall (e.g., Heuer & Reisberg, 1990; Yuille & Cutshall, 1986), and the present results
support this interpretation. In our view, the primary effect of emotional arousal on
memory is to alter the degree to which memories (especially long-term memories) of
the emotional events are accessible to conscious free recall.
In the final analysis, the findings presented here, and others like them, indicate
only that changes in memory (in this case, improved long-term memory) are associ-
ated with emotional arousal. No experiment(s) to date has demonstrated that an effect
on memory is unequivocally due to emotional arousal and not "other" factors (atten-
tion, novelty, etc.), nor is any single experiment likely to do so. In fact, emotional
arousal may influence memory, at least in part, because of its affects on these "other"
factors. However, experiments such as those described here provide additional sup-
port for the general conclusion that emotional arousal induced in controlled settings
does in fact, either directly or indirectly, affect memory in humans.

ACKNOWLEDGMENTS
The authors thank Laureen Tews for her assistance in conducting this study. Research supported by NIA Institu-
tional National Research Service Award AG00096 (LC.) and USPHS MHl 2526 from NIMH and NIDA (JLM).

APPENDIX
Sentences Accompanying the Slide Presentation in Experiment 2

Slide Neutral version Arousal version


1) A mother and her son are leaving home in A mother and her son are leaving home in
the morning. the morning.
2) She is taking him to visit his father's work- She is taking him to visit his father's work-
place. place.
3) The father is a laboratory technician at Vic- The father is a laboratory technician at Vic-
tory Memorial Hospital. tory Memorial Hospital.
4) They check before crossing a busy road. They check before crossing a busy road.
5) While walking along, the boy sees some While crossing the road, the boy is caught in
wrecked cars in a junk yard, which he finds a terrible accident, which critically injures
interesting. him.
6) At the hospital, the staff are preparing for a At the hospital, the staff prepare the emer-
practice disaster drill, which the boy will gency room, to which the boy is rushed.
watch.
7) An image from a brain scan machine used in An image from a brain scan machine used in
the drill attracts the boy's interest. a trauma situation shows severe bleeding
in the boy's brain.
8) All morning long, a surgical team practiced All morning long, a surgical team struggled
the disaster drill procedures. to save the boy's life.
9) Make-up artists were able to create realistic- Specialized surgeons were able to re-attach
looking injuries on actors for the drill. the boy's severed feet.
10) After the drill, while the father watched the After the surgery, while the father stayed
boy, the mother left to phone her other with the boy, the mother left to phone her
child's pre-school. other child's pre-school.
11) Running a little late, she phones the pre- Feeling distraught, she phones the pre-school
school to tell them she will soon pick up her to tell them she will soon pick up her
child. child.
12) Heading to pick up her child, she hails a taxi Heading to pick up her child, she hails a taxi
at the number nine bus stop. at the number nine bus stop.
EMOTION AND MEMORY 421

REFERENCES
Burke. A., Heuer, F., & Reisberg, D. ( 1992). Remembering emotional events. Memory and Cogni-
tion, 20, 277-290.
Christianson, S. (1984). The relationship between induced emotional arousal and amnesia. Scandanavian
Joumal of Psychology, 25, 147-160.
Christianson, S., & Loftus, E. ( 1987). Memory for traumatic events. Applied Cog11iti1e Psychology, 1,
225-239.
Heuer, F.. & Reisberg, D. (1990). Vivid memories of emotional events: The accuracy of remembered
minutiae. Memory and Cognition, 18(5), 496-506.
Ledoux, J. E. ( l 992). Emotion as memory: Anatomical systems underlying indelible neural traces. In
S-A. Christianson (Ed.), Handbook of emotion and memo/): Research and theory (pp. 269-288).
Hillsdale, NJ: Erlbaum.
McGaugh, J. L. ( 1992). Affect, neuromodulatory systems and memory storage. In S-A. Christianson
(Ed.), Handbook of emotion and memol): Research and theol) (pp. 245-268). Hillsdale, NJ: Erl-
baum.
Meltzer, H. ( l 930). Individual differences in forgetting of pleasant and unpleasant experiences. The
Joumal of Educational Psychology, 21(6), 399-409.
Revelle. W., & Loftus, D. (1992). The implications of arousal effects for the study of affect and memory.
In S-A. Christianson (Ed.), Handbook of emotion and memory: Research and theory (pp. 113-
150). Hillsdale, NJ: Erlbaum.
Tobias, B. A.. Kihlstrom, J. F., & Schacter, D. L. (1992). Emotion and implicit memory. In S-A. Chris-
tianson (Ed.). The Handbook of emotion and mem<>I)': Research and theol)' (pp. 67-92). Hillsdale,
NJ: Erlbaum.
Tulving. E. ( 1983 ). Elements of episodic memo I)". Oxford: Oxford Univ. Press.
Yuille, J.C.. & Cutshall, J. L. ( 1986). A case study of eyewitness memory of a crime. Joumal of Applied
Psychology, 71(2), 291-30 I.
Received May 31, 1995