For over a century, the phenomenon of déjà vu has attracted much interest, and
in recent times, it has been studied by researchers in various scientiﬁ c ﬁ elds
(e.g., Brown, 2003, 2004; Sno & Linszen, 1990). Empirical studies of déjà vu
phenomena have used interviews and questionnaires with normal people as well
as psychiatric patients (e.g., Neppe, 1983; Sno & Linszen, 1990; Sno, Schalken,
de Jonghe, & Koeter, 1994). In this study, we used experiments and surveys to
examine human metacognition and the déjà vu phenomenon. This chapter is
divided into four parts. First, we deﬁ ne déjà vu using notions from the ﬁ elds
of psychology and psychiatry. Second, we propose that déjà vu phenomena
involve a component of memory monitoring and that a metacognitive approach
(e.g., Chambres, Izaute, & Marescau, 2002) appears to be most suitable for its
study. Third, we present highlights of the data obtained from our questionnaire
(Kusumi, 1994, 1996) and experimental studies (Matsuda & Kusumi, 2001).
Fourth, we propose a déjà vu model based on an analogical reminding
mechanism. Finally, we discuss some of the implications of a déjà vu model
that involves an adaptive metacognitive mechanism.
Deﬁ nition of Déjà Vu
Déjà vu experiences have been described in many works of ﬁ ction including
those by Dickens, Tolstoy, Proust and Hardy (Sno, Linszen, & de Jonghe, 1992).
However, psychological studies of déjà vu in mainstream memory research
are rare (e.g., Brown, 2003, 2004). Déjà vu experiences have been primarily
studied as memory disorders (e.g., illusions, hallucinations, schizophrenia,
temporal lobe epilepsy (TLE) in the ﬁ elds of psychiatry and psychoanalysis
(e.g., Neppe, 1983). Researchers have noted that most people experience déjà
vu only when they are extremely fatigued; for the average individual, déjà vu is
In K.Fujita & S. Itakura (Eds.) 2006 Diversity of Cognition: Evolution, Development, Domestication, and Pathology. Kyoto University Press
a rare and abnormal memory experience. However, this study explains the déjà
vu experience as a normal metacognitive mechanism. This approach proposes
that déjà vu occurs during an analogical reminding process (e.g., Wharton,
Holyoak, & Lange, 1996) in which a present experience automatically reminds
an individual of similar past experiences. Therefore, the déjà vu experience is
generated by similarities between a present experience and corresponding past
We will ﬁ rst deﬁ ne déjà vu based on the ﬁ ndings of cognitive research.
Déjà vu is a French term meaning already seen. It refers to ‘any subjectively
inappropriate impression of familiarity of a present experience with an undeﬁ ned
past experience’ (Neppe, 1983). The term inappropriate familiarity is deﬁ ned
as a form of false recognition in which one experiences a strong sense of
familiarity with new events or objects. In déjà vu, successful reality monitoring
(e.g., Johnson, Hashtroudi, & Lindsay, 1993) enables one to determine that an
event is actually new although it may feel old. Such judgements of familiarity
are based on metacognitive monitoring.
Recently, Brown (2004) classiﬁ ed scientiﬁ c explanations of déjà vu into
four categories: dual-processing explanations (two cognitive functions that are
momentarily out of synchrony), neurological explanations (brief dysfunction
in the brain), memory explanations and double-perception explanations (brief
break in one’s ongoing perceptual processing). This study focuses on memory
explanations. Brown wrote, ‘Memory interpretation assumed that some
dimension(s) of the present setting is actually objectively familiar, but the source
of familiarity is not explicitly recollected’ (p. 127). This study presents a new
memory explanation of déjà vu, integrating three metacognitive components in
order to explain the mechanisms involved in analogical reminding (Figure 1).
The three metacognitive components are as follows:
1. Preliminary feelings of strong familiarity for a present experiences,
involving a process of implicit memory
2. Similarity and dissimilarity judgements between the present and a
retrieved past experience made after a search of explicit memory
3. Reality monitoring for the retrieved experience (prototype event), which
is a decision on whether or not the present experience is identical with a
As illustrated in Figure 1, we postulate that the déjà vu phenomenon stems
from ordinary metamemory mechanisms. Roedier (1996) suggested that déjà
vu and jamais vu
are illusions of metacognition. Brown (2004) theorized that
the déjà vu experience is a pure metamemory experience that is unconnected
with the empirical world (i.e., an identiﬁ able eliciting stimulus or a veriﬁ able
behavioural response to corroborate the subjective state). However, this study
postulates that déjà vu is a common and adaptive metacognitive process based
on the similarity between present and past experiences.
Some researchers have differentiated between types of déjà vu: auditory and
visual déjà vu, event and place déjà vu, reactive (external, i.e. precipitated by
brain mechanism) and endogenous (internal, i.e. elicited by brain mechanism)
déjà vu and normal and pathological déjà vu (Brown, 2004). This study
differentiates between place and person déjà vu—two types of déjà vu that are
familiar to most subjects. Moreover, it is easy to compare the stimuli and the
underlying cognitive processes associated with these types of déjà vu.
This chapter addresses the following three research questions. First, what
types of déjà vu experiences are common in normal university students?
Second, what are the metacognitive and analogical reminding mechanisms used
when déjà vu is experienced in daily life and in the laboratory? Finally, is déjà
vu the result of an adaptive metacognitive mechanism? This paper highlights
the data obtained from two questionnaire surveys (Kusumi, 1994, 1996) and
one from an experiment (Matsuda & Kusumi, 2002).
Survey Data of Déjà Vu Experiences
Several questionnaire studies have been conducted on the phenomenon of déjà
vu. Neppe (1983) conducted interview surveys with normal subjects, subjects
with TLE and patients of schizophrenia. He developed a screening questionnaire
for 11 déjà vu experiences (place, situation, doing, happening, meeting, saying,
hearing, thinking, reading, dreams, etc.) and a qualitative questionnaire (57
Judgments of similarity and dissimilarity between
the present and a retrieved past experience
Reality monitoring for the present experience
Deja vu experience
items concerning frequency, duration, feelings, etc.). The results indicated that
in normal people, there are two kinds of déjà vu, namely, associative déjà vu
and subjective paranormal déjà vu. Neppe found that in the average person,
associative déjà vu tended to be vague and poorly remembered, was often
triggered by the environment, was initially characterized by partial familiarity,
lasted for a short duration and lacked outstanding qualitative features. This type
of déjà vu can be explained by the mechanisms of normal memory (Neppe, 1983,
p. 249). The second type of déjà vu, which occurred in subjective paranormal
experiences, was characterized by time-dissociations and outstanding qualitative
features (Neppe, 1983, p. 254). Neppe’s results also indicated that there were
some differences between subjects with TLE and schizophrenia; both déjà vu
and jamais vu occurred more frequently in the TLE group. Sno, Schalken,
de Jonghe and Koeter (1994) reﬁ ned Neppe’s questionnaire items; Adachi,
Adachi, Kimura, Akamatsu and Kato (2001) translated them into Japanese
and checked their reliability and validity using normal Japanese subjects and
those with schizophrenia. However, these surveys did not explore the cognitive
mechanisms of déjà vu such as similarity and time intervals between the source
and déjà vu experiences.
Kusumi’s research (1994, 1996) explored the déjà vu experience based on
a metacognitive and analogical reminding mechanism that traced a present
experience to similar past experiences. In one study (Kusumi, 1994), 202
Japanese university students completed an original questionnaire on déjà vu
experiences and analogical reminding. The participants were asked the following
questions: (a) Have you ever been in a new place and felt as if you had been
there before? Or have you ever gone somewhere for the ﬁ rst time and yet felt
it was familiar? (b) Have you ever met someone for the ﬁ rst time and felt as if
you had met that person before? (c) If yes, when and where did you have your
last experience of these particular feelings? (d) What were the surroundings and
the cues for the experience? (e) Did you identify a similar past experience? (f)
If yes, when and where did you have the similar past experience?
Figure 2 shows the results obtained from the 202 participants. The déjà vu
experience was observed to be a common phenomenon. Place déjà vu, which
occurred when people visited a new place and felt as if they had been there
before, was observed in 63% of the participants. Person déjà vu, which occurred
when people met a new person and felt as if they had met that person before,
was observed in 35% of the participants. Of the individuals who had experienced
person déjà vu, 89% could identify the exact situation in which it had occurred.
Of the participants who had experienced place déjà vu, 61% could identify the
exact situation in which it had occurred. Of the 61% who could identify an
exact place déjà vu, only 23% could identify the past source experience that
might have triggered the déjà vu experience. On the other hand, of the 89%
who could identify an exact person déjà vu, 36% could identify the past source
person. Identifying a person requires precise information and exact matching
(e.g., face and name). In contrast, place déjà vu occurs when a source memory
is vague. Brown (2003, p. 404) explained these judgement processes based on
a source-monitoring framework (Johnson, Hashtroudi, & Lindsay, 1993). For
example, you meet a new person and feel a strong sense of familiarity; yet, you
know that you have never met this person before. The déjà vu experience could
arise from the conﬂ ict between two types of source-monitoring processes: the
judgement based on your general knowledge and episodic memory (of never
having met before) conﬂ icts with the heuristic judgement based on the vivid
representation of personal appearance from past experiences, which implies
familiarity. You then search your memory for a similar person in order to resolve
the strange familiarity. Subsequently, you remember an old friend who is similar
to the person you have just met. In such a case, even after a considerable length
of time, you might remember the strange experience of the similarity between
the new and old person. On the other hand, consider the situation wherein you
visit a new place and feel a strong sense of familiarity although you know that
you have never before visited this place. You search your memory for similar
places to resolve this strange feeling of familiarity; however, you do not
remember any particular place that is similar to the current one. In this case,
after a considerable length of time, you might not have an exact memory of
this strange experience.
Place déjà vu experiences reported by the participants occurred 3 days to 10
years before the study was conducted. The reported source experiences occurred
2 to 17 years before the study was conducted. The time interval between place
déjà vu experiences and their source (original) experiences was 2 months to 12
years. Person déjà vu experiences reported by participants occurred 1 month to
of Deja vu
N 59 (.39)
Y 63 (.89)
N 8 (.11)
Y 18 (.23)
N 69 (.77)
N 27 (.43)
12 years before the study and their source experiences occurred 1 to 16 years
prior to it before the study. The time interval between person déjà vu experiences
and their original experiences was 1 to 14 years.
Participants recalled original experiences that occurred more than 1 to 17
years before the study was conducted. Brown (2004) reported that individuals’
estimates of the length of time since the original experience were distributed
evenly across days, weeks, months and years. He suggested that future
questionnaires should include a more detailed query on this topic. Kusumi
(1994) used detailed questionnaires to ask participants about the content and
time of original experiences and then calculated the retention time between the
source and déjà vu experiences. The results obtained by Kusumi indicated that
déjà vu experiences are based on very long-term or autobiographical memory.
Figure 3 shows similarity ratings between the source experience and the
déjà vu experience on a 7-point scale, with –3 corresponding to very dissimilar
and 3 corresponding to very similar. The perceived similarity ratings between
source and target experiences of place déjà vu ranged from 1 to 3. When rating a
person déjà vu, participants rated appearance similarity higher than personality
similarity; appearance seems to be a stronger cue for reminding people of source
experiences. Most source persons were acquaintances whom the subjects had
not seen for a long time, for example an old classmate or a distant relative.
In Kusumi (1996), 104 Japanese undergraduates completed a questionnaire
on place déjà vu experiences. They rated the frequency of déjà vu experiences
for 13 places and three situations on a 5-point scale (never, once, twice, three,
for these experiences (e.g., perceptual cues, atmosphere, weather and mood).
Figure 4 indicates that déjà vu experiences occurred frequently during
conversations and dreams, when walking down a street, visiting old-style
a) Similarity Ratings of Place Deja vu
b) Similarity Ratings of Person Deja vu
-3 -2 -1 01 2
M = 1.9
-3 -2 -1 01 2
villages, walking in a park, visiting a school and visiting a temple or shrine.
In examples of place déjà vu, there exists a high degree of similarity between
elements in types of scenes. For example, many parks in Japan resemble each
other. Thus, people frequently view similar park scenes. Consequently, a typical
scene of a park is constructed in their memory. When visiting a new place, the
new scene tends to match the typical scene in a person’s memory, leading to a
feeling of familiarity.
Sixty-two percent of the participants reported experiencing déjà vu during
a dream. The mechanism of dream déjà vu is similar to place déjà vu; dreams
are based on compressed past real experiences that are similar to typical scenes.
Therefore, a dream fragment of a typical scene tends to match earlier dream
experiences and evoke a feeling of familiarity. Brown (2002) also claimed that
dream memory fragments might trigger a déjà vu experience when similar
situations are encountered while awake. Moreover, Brown’s literature review
(2004) reported that the frequency of déjà vu experience has a weak positive
correlation (rs = .22–.30) with dream memory (e.g., recall, vividness and dream
Déjà vu in conversation occurs when individuals feel that they have heard the
words in a conversation before. Based on responses from university students,
Brown (2004) found that approximately 50% of déjà vu experiences occur in
Number of Participants (N=104)
Over 6 Times
the company of friends. The sources of conversation déjà vu are similar past
experiences, such as similar participants and surroundings. These similarity
factors affect the conversation content, thus increasing the likelihood of a match
between a current conversation and stored conversations and the impression of
Firstly, the results of our survey indicate that déjà vu experiences are
common in normal people; 72% of participants experienced déjà vu. This ratio
supports the ﬁ gures obtained 32 surveys of non-clinical subjects (M = 68%,
Mdn = 70%) and is slightly higher than the ﬁ gures reported in nine surveys of
neuropsychiatric patients (M = 55%, Mdn = 65%) (Brown, 2003). Secondly,
place déjà vu experiences are based on typical scenes in stored memory (Figure
4). These scenes are easy to match with a new experience, leading people to
ﬁ nd the new experience familiar. Thirdly, the number of matching cues between
source and new experiences in place déjà vu increases the sense of similarity
and familiarity, and may lead to a feeling of déjà vu, although, logically, people
still realize that they are experiencing a new situation.
Experimental research on mere exposure effects has provided new experimental
paradigms through which the psychological processes of déjà vu can be
Matsuda and Kusumi (2001) have investigated how prototypical stimuli and
exposure frequency affect déjà vu experiences for scenes by using old-new
and nostalgia judgements in a paradigm of mere exposure (Bornstein, 1989;
Zajonc, 1980). Forty-three university students participated in this experiment.
In the study phase, 54 photographs of obscure temples were displayed for 1 s
each at four levels of exposure frequency (0, 1, 3, 6 times). Other participants
had previously judged these temples as having low, moderate or high typicality.
In the test phase, participants judged typicality, familiarity, liking, beauty and
nostalgia for each photograph using a 9-point scale and also participated in a
recognition test of new and old items.
The typicality of scenes affected the false recognition of new scenes. The
false alarm rate was 46% for highly typical scenes, 29% for moderately typical
ones and 16% for atypical scenes. Exposure frequency increased participants’
responses of ‘old’ in the recognition test. Mean judgements pertaining to
familiarity and nostalgia for highly and moderately typical stimuli were higher
than those for atypical stimuli. Exposure frequency had an effect on both
judgements; a higher exposure frequency led to higher ratings of familiarity
and nostalgia. There was a high correlation between judgements of familiarity,
nostalgia, liking and beauty (rs(516) = .30–.70, p < .01), suggesting that
the effect of mere exposure on liking extends to judgements of beauty and
Figure 5 presents the results of structural equation modeling (SEM). The
analysis suggested that the typicality of stimuli and frequency of exposure
had a positive inﬂ uence on the formation of prototypes, which in turn directly
promotes feelings of knowing (familiarity and nostalgia), and then affects
positive judgement (liking and beauty). Similar SEM results were produced in
the 1-week delay and artiﬁ cial-picture condition (Matsuda & Kusumi, 2003),
as well as in the incidental-learning and artiﬁ cial-picture condition (Kusumi
& Matsuda, 2004). Seamon, Brody and Kauff (1983) also tested whether
familiarity affected preference judgement using a subliminal mere exposure
paradigm. They presented 10 geometric shapes, each repeated ﬁ ve times with
an exposure duration of 5 ms. They found that subliminal exposure enhances
positive affective evaluation without conscious recognition.
Brown (2004) examined another déjà vu process based on affective responses
by means of perceptual ﬂ uency. Reber, Winkielman and Schwarz (1998) found
that the manipulation of ﬂ uency (ﬁ gure-ground contrast, etc.) led to an enhanced
positive effect (liking, prettiness, etc.) for a particular stimulus. If this positive
affect is misidentiﬁ ed as familiarity, then the stimulus could lead to a déjà vu
experience through the following four steps: (a) perceptual ﬂ uency, (b) positive
effect, (c) familiarity and (d) déjà vu (Brown, 2004, p. 165).
Figure 6 shows the typicality and analogical reminding model of a déjà vu
experience. Our results suggest that the déjà vu experience is based on similarity
GFI=0.949, AGFI=0.890, RMSEA=0.116
between typical source experiences and new experiences; similar source
experiences are compressed in memory into a single typical representation.
This typical representation shares similarities with each original experience.
During the construction of typical representations, unimportant cues or indices
are omitted and important cues are preserved. One study revealed that the four
most common retrieval cues were perceptual attributes (64%, as a percentage of
the participants), atmosphere (41%), weather (26%) and mood (26%) (Kusumi,
1996). A matching process between cues would determine familiarity and could
result in a déjà vu experience. When people experience a new situation, matched
retrieval cues between a typical representation and the new representation
increase, thus increasing the feeling of familiarity with the new scene and leading
to a feeling of déjà vu. This model is consistent with the gestalt familiarity
explanation of déjà vu (Brown, 2004); the gestalt of a present experience (the
general visual organization of the elements in a scene) conﬁ gures similarity to
a previous experience, triggering a déjà vu experience.
Our study proposed a typicality and analogical reminding model of déjà vu
based on data from two surveys and one experiment. The model provides a
satisfactory account of the déjà vu phenomenon, in which a person experiences
an inappropriate feeling of familiarity. In the ﬁ eld of psychiatry, the déjà vu
phenomenon has been treated as a memory disorder. However, we conclude
Perceptual Cue (0.64)
that this phenomenon is based on normal metacognitive mechanisms because
(a) 70% of normal adults experience the phenomenon; (b) déjà vu experiences
of locations, highly nostalgic experiences and false alarms of recognition often
involve prototypical scenes stored in memory; and (c) feelings of familiarity
increase as exposure frequency, typicality and the number of cues that match
between past and new experiences increase.
The phenomenon of déjà vu appears to be a component of adaptive
human behaviour and sheds light on human metacognitive mechanisms. Déjà
vu phenomena are based on the metacognitive components of feelings of
familiarity and recognition memory. Both ontogenetically and evolutionarily,
recognition memory develops earlier than recall memory (Todd & Gigerenzer,
2000). Gigerenzer (2000) theorized that familiarity is the principal heuristic
during the initial, rapid stage of retrieval. How are déjà vu phenomena related
to recognition heuristics? When individuals feel a sense of familiarity with a
present experience or problem, they retrieve past experiences or problems by
matching cues. They evaluate the similarity and dissimilarity between the two
experiences and then transfer useful information from past experiences to the
present one. This process performs the same function as analogical problem
solving (Holyoak & Thagard, 1995), in which a similar old problem provides
a solution to a new one. This metacognitive mechanism appears to be adaptive
in humans and is an aspect for future research.
This paper has presented a model that integrates three metacognitive components
to explain the mechanisms involved in analogical reminding, and the reporting
of relevant empirical data. These three metacognitive components are (a)
feelings of knowing and monitoring reality during new events, (b) judgements of
similarity and dissimilarity between new and retrieved events and (c) monitoring
of reality in prototype events. The model provides a satisfactory account of the
déjà vu phenomenon, in which individuals experience an inappropriate feeling
of familiarity with a current situation because they erroneously believe that a
similar situation has occurred in the past. In the ﬁ eld of psychiatry, the déjà vu
phenomenon has been treated as a memory disorder. However, we conclude
that this phenomenon is based on normal memory mechanisms in view of
the following results: (a) 70% of normal adults experience the phenomenon,
(b) prototypical scenes stored in memory are frequently involved in déjà vu
experiences of locations (i.e. ‘I have been here before’) and (c) the feeling of
familiarity increases as the number of cues that match between past and new
experiences increase. Thus, this phenomenon appears to be a part of adaptive
human behaviour and sheds light on human metamemory and knowledge
Adachi, N., Adachi, T., Kimura, M., Akanuma, N., & Kato, M. (2001).
Development of the Japanese version of the inventory for déjà vu
experiences assessment (IDEA). Clinical Psychiatry, 43, 1223–1231.
Bornstein, R. F. (1989). Exposure and affect: Overview and meta-analysis of
research, 1968–1987. Psychological Bulletin, 106, 265–289.
Brown, A. S. (2003). A review of the déjà vu experience. Psychological
Brown, A. S. (2004). The déjà vu experience. East Sussex, England: Psychology
Chambres, P., Izaute, M., & Marescaux, P. (2002). Metacognition: Process,
Gigerenzer, G. (2000). Adaptive thinking: Rationality in the real world. New
York: Oxford University Press.
Holyoak, K. J., & Thagard, P. (1995). Mental leaps: Analogy in creative thought
Cambridge: MIT Press.
Johnson, M. K., Hashtroudi, S., & Lindsay, D. S. (1993). Source monitoring.
Kusumi, T. (1994, July). Déjà vu phenomena by analogical reminding. Paper
presented at the 11th annual meeting of the Japanese Cognitive Science
Society, Tokyo, Japan.
Kusumi, T. (1996, March). Situational factors of déjà vu experiences:
Representational similarities in autobiographical memory and dream.
Paper presented at the 7th annual meeting of the Japan Society of
Developmental Psychology, Tokyo, Japan.
Kusumi, T. (1998, March). Déjà vu experiences: An explanation based on
1st Tsukuba International Conference on Memory, Tsukuba, Japan.
Matsuda, K., & Kusumi, T. (2001, November). Scene typicality inﬂ uences the
mere exposure effect in affective judgments. Paper presented at the 42nd
annual meeting of the Psychonomic society, Orlando, FL.
Matsuda, K., & Kusumi, T. (2003, January). A mere exposure effect for the
concept formation II: The effect of duration on the typicality and the
affective judgments, Paper presented at the 4th Tsukuba International
Conference on Memory, Tsukuba, Japan.
Matsuda, K., & Kusumi, T. (2004, August). The mere exposure effect on
incidentalconcept formation. In K. Forbus, D. Gentner, & T. Regier (Eds.),
Proceedings of the 26th Annual Conference of the Cognitive Science
Society (p. 1598). NJ: Lawrence Erlbaum Associates.
Neppe, V. M. (1983). The psychology of déjà vu: Have I been here before?
Johannesburg, South Africa: Witwatersrand University Press.
Reber, R., Winkielman, P., & Schwarz, N. (1998). Effects of perceptual ﬂ uency
on affective judgments. Psychological Science, 9, 45–48.
Roediger, H. L., III. (1996). Memory illusions. Journal of Memory and
Seamon, J. G., Brody, N., & Kauff, D. M. (1983). Affective discrimination of
stimuli that are not recognized: II. Effect of delay between study and test.
Bulletin of the Psychonomic Society, 21, 187–189.
Sno, H. N., & Linszen, D. H. (1990). The déjà vu experience: Remembrance of
things past? American Journal of Psychiatry, 147, 1587–1595.
Sno, H. N., Linszen, D. H., & de Jonghe, F. (1992). Art imitates life: Déjà
vu experiences in prose and poetry. British Journal of Psychiatry, 160,
déjà vu experiences: A review. Behavioural Neurology, 5, 155–160.
Sno, H. N., Schalken, H. F. A., de Jonghe, F., & Koeter, M. W. J. (1994). The
inventory for déjà vu experiences assessment: Development, utility,
reliability, and validity. Journal of Nervous and Mental Disease, 182,
smart. Behavioral & Brain Sciences, 23, 727–780.
Wharton, C. M., Holyoak, K. J., & Lange, T. E. (1996). Remote analogical
reminding. Memory & Cognition, 24, 629–643.
Zajonc, R. B. (1980). Feeling and thinking: Preferences need no inferences.
American Psychologist, 35, 151–175.
1. The jamais vu experience involves an objectively familiar situation that