Following bilateral hippocampal lesions, patient have more frequent déja vus and prescience episodes

Déjà vu and prescience in a case of severe episodic amnesia following bilateral hippocampal lesions. Jonathan Curot, Jérémie Pariente, Jean Michel Hupé, Jean-Albert Lotterie, Hélène Mirabel & Emmanuel J. Barbeau. Memory, Oct 6 2019. https://doi.org/10.1080/09658211.2019.1673426

ABSTRACT: Several studies pertaining to déjà vu have consistently made a connection with the perirhinal region, a region located below the hippocampus. This idea is strengthened by the fact that déjà vu is an erroneous sense of familiarity and that familiarity appears to largely depend on the perirhinal region in healthy subjects. In this context, the role of the hippocampus is particularly unclear as it is unknown whether or not it plays a role in the genesis of déjà vu. We report on the case of OHVR, an epileptic patient who suffers from severe episodic amnesia related to massive isolated bilateral damage to the hippocampus. In contrast, the perirhinal region is intact structurally and functionally. This patient reports frequent déjà vu but also another experiential phenomenon with a prominent feeling of prescience, which shows some of the characteristics of déjà vécu. She clearly distinguishes both. She also developed a form of synaesthesia by attributing affective valence to numbers. This study shows that déjà vu can occur in cases of amnesia with massively damaged hippocampi and confirms that the perirhinal region is a core region for déjà vu, using a different approach from previous reports. It also provides clues about a potential influence of hippocampal alterations in déjà vécu.

KEYWORDS: Déjà vécu, recollection, hippocampus, familiarity, recognition memory, perirhinal cortex, synaesthesia

Discussion

To our knowledge, this case study is the first to report on
the experience of déjà vu in a patient with severe episodic
amnesia and massive isolated hippocampal atrophy. In
addition, our patient also reported unusually frequent
experiential phenomena with a prominent feeling of prescience
as well as synaesthesia.

Two types of subjective experiences
In one of the most recent and exhaustive reviews on déjàexperiences
in epilepsy, Illman et al. (2012) suggested that
inconsistencies about déjà-experiences stem from a
problem of definition. These authors distinguished in particular
déjà vu, an inappropriate sense of familiarity, from
déjà vécu, an erroneous sensation of recollecting contextual
information. These qualitatively different phenomenological
experiences may rely on functionally independent
neural substrates: déjà vu – a “pure” familiarity experience –
that relies on a network of brain structures that process
familiarity, including most notably the perirhinal cortex,
and déjà vécu – a “recollective” experience – that relies
on a network centred on the hippocampus (Illman et al.,
2012). OHVR could clearly differentiate between two types
of subjective phenomena: one without prescience that
appears to match the definition of déjà vu and another
associated with prescience. One question is whether or
not the prescience experience is related to déjà vécu. In
the following sections, we discuss these two types of experience
in OHVR along with their possible neural correlates.

Déjà vu in OHVR
OHVR clearly reported the frequent occurrence of “pure”
déjà vu without any prescience. She spontaneously insisted
on the idea that in this case she was a “spectator” which
matches the ideas of Illman et al. (2012) about déjà vu.
She also firmly differentiated it from the other type of
her experiential phenomena. Whether OVHR experiences
déjà vu of a similar type to that experienced by healthy
subjects is a matter to be discussed. One study compared
déjà vu in epileptic patients and healthy subjects (Warren-
Gash & Zeman, 2014) and arrived at the conclusion that
déjà vu was qualitatively similar but that in epileptic
patients it was also associated with distinct features,
including derealisation (a sensation in which the external
environment appears unfamiliar, with other people
appearing like actors and the world appearing to be twodimensional
or like a stage set, as defined by the DMS-IV
and ICD-10. For a review see Hunter, Sierra, & David,
2004). OHVR’s déjà vu seems to fit well with this report.
She was aware that her sense of familiarity was inappropriate
(Brázdil et al., 2012; Illman et al., 2012). The knowledge
that the sensation is wrong is a major aspect of the déjà vu
sensation (O’Connor & Moulin, 2010). Our patient’s description
was consistent with those by O’Connor and Moulin
(2008) and Martin et al. (2012) as she explicitly reported
that her inappropriate sense of familiarity could initially
start with specific objects (e.g., “glasses” or “a pen”) and
secondarily and quickly expanded to the entire situation
and environment. Her déjà vu lasted a few seconds as in
control subjects. However, the feeling of derealisation
associated with her déjà vu was more pronounced than
for control subjects.
What makes OHVR unusual is that she suffers from
severe episodic amnesia due to massive and isolated bilateral
lesions of the hippocampus. Despite the lack of specific
experimental tasks to evaluate recollection and familiarity
specifically, OHVR’s case is highly suggestive of an episodic
amnesia: performance on context-rich and relational
memory tests was severely impaired, both clinically and
on neuropsychological tests, and with both verbal and
visual material (standard score on both the Logical
Memory and the Family Picture subtests delayed recall:
1). Such results match the finding that OHVR’s hippocampi
were severely damaged. In contrast, the performance on
context-free memory tests such as semantic memory (standard
score on the Information subtest: 9) or recognition
memory tests, where performance can rely on familiarity
(Face recognition subtest or Doors recognition), showed
that these were considerably preserved. Context-free
memory largely depends on anterior subhippocampal
structures (Barbeau, Pariente, Felician, & Puel, 2011; Jonin
et al., 2018; Vargha-khadem et al., 1997), including the perirhinal
cortex. The perirhinal cortex appeared to be preserved
in OHVR both structurally (MRI) and functionally
(PET). Therefore, this case report suggests that déjà vu
can occur in the absence of recollection and a functional
hippocampus and appears to correspond with the hypothesis
by Spatt (2002) and Illman et al. (2012) that déjà vu
depends on the activation of a neocortical familiarity
system. A debate between the respective roles of the
entorhinal cortex and the perirhinal cortex still remains to
be solved however. In the study by Bartolomei et al.
(2004), déjà vu was induced more often by stimulation of
the entorhinal than the perirhinal cortex. However, the
exact effect (i.e., local excitation or inhibition) of these electrical
stimulations is largely unclear and it is difficult to infer
from this the exact role of each structure. In OHVR, hypometabolism
of the right entorhinal cortex was observed.
Overall, these observations seem to confirm that the
rhinal region is as a core region in déjà vu but the role of
the entorhinal and perirhinal cortices in this experiential
phenomenon must still be clarified.
At first sight, OHVR’s case seems to suggest that recollection
or the hippocampus plays no role in déjà vu.
According to this hypothesis, it is a dysfunction within the
perirhinal region / familiarity system that induces déjà vu,
possibly in relation to epileptic activities in this region.
However, an alternative hypothesis is that it is a dysfunction
between the perirhinal region and the hippocampus
that is at the origin of déjà vu. Although OHVR’s hippocampi
were massively damaged in comparison to the
usual standards, the possibility must be considered that
some neuronal activity remains in this structure, enough
to trigger erroneous signals that are misinterpreted in the
perirhinal region. A hypothesis stemming from this
finding is that the more severe the atrophy of the hippocampi
(in isolation, i.e., with preserved subhippocampal
structures), the more déjà vu there might be. However,
this specific hypothesis should be tested in future studies
in epileptic patients, which would help to clarify whether
or not the hippocampus plays a role in déjà vu.

A subjective experience with a prominent feeling of
prescience
Our patient also frequently experienced a second type of
subjective phenomenon with a prominent feeling that
she describes as feeling like she could predict the future.
Some features of this subjective feeling match the
definition of prescience and corresponds with the conceptual
definition of déjà vécu by Illman et al. Experimental
evidence following the re-creation of a déjà vu sensation
in a virtual environment suggests that déjà vu may be
related to an illusion of prediction (Cleary & Claxton,
2018). Interestingly, despite making a clear distinction
between both phenomena, OHVR also acknowledged
that the phenomenon with prescience could occur sometimes
with pure déjà vu, but it was always after déjà vu,
which could be suggestive of evidence of certain dynamics
between these two experiential phenomena. To some
extent, OHVR’s case also supports the idea of some proximity
between déjà vu and prescience as both coincided
chronologically with the development of anterograde
amnesia and hippocampal dysfunction. This seems to fit
data from healthy subjects since approximately 16% of
them report feelings of prescience at least once a year
while having a feeling of déjà vu. Another 17% also know
about this feeling in association with déjà vu (Mumoli et al.,
2017). From this, it can be logically deduced that if familiarity
is experienced first and if this feeling persists over a
few seconds, it could lead to the feeling that the future
could be (or should be) predicted. Therefore, déjà vu
with prescience would be more intense and last longer.
This hypothesis can easily be tested in healthy subjects.
Nonetheless, other theoretical concepts include prescience
as a feature that helps to separate two déjà-experiences,
déjà vu and déjà vécu. Supposedly, déjà vu is devoid
of any feeling of prescience whereas déjà vécu integrates it
in association with emotions, source or context and
content (Illman et al., 2012). OHVR spontaneously insisted
on a distinction between her “déjà vu” and this other
feeling she could not name precisely (she sometimes
named it “prediction”, sometimes “support”). To the question:
“Are they two different phenomena for you?” she
answered: “Yes”, although she never used the terms déjà
vu or déjà vécu to describe this phenomenon. To the question:
“Are they concomitant or not necessarily linked?” she
also responded: “Sometimes, but not necessarily”. In fact,
OHVR’s feelings of premonition featured some aspects of
déjà vécu. She spontaneously emphasised that she was
an “actor” during prescience (while being a “spectator”
during déjà vu). This agency component is a feature of
autonoetic consciousness during recollection (Metcalfe &
Son, 2012; Piolino et al., 2003; Tulving, 1985) and this
state of consciousness may correspond to a dysfunctional
recollection during déjà vécu. She also clearly experienced
an erroneous sensation of time, which was absent from her
déjà vu experiences. However, and this is inconsistently
with Illman’s definition of déjà vécu, she never related
this experience to a “prior experience”. In brief, OHVR
clearly makes a distinction between two different types
of experiential phenomena, déjà vu, and a second one,
which resembles déjà vécu. This second phenomenon is
close to but does not exactly fit the definition of déjà
vécu (Illman et al., 2012). Further investigations of déjà
vécu are therefore needed.
This is an important discussion as the conceptual
entity of déjà vécu lacks a clear neuroanatomical substrate.
Therefore, OHVR’s second phenomenon could be
a way to approach the neural substrate of déja-vécu
and its components. To date, only 4 cases of prescience
have been described and studied in detail, but they were
not related to déjà vécu (Sadler & Rahey, 2004). These
authors found prescience phenomena in only 3 of 927
epileptic patients and they also quoted a fourth case
reported by Gloor et al. (1982). Prescience in experiential
phenomena seems to be infrequent but may have been
underestimated because of a lack of detailed reports.
Interestingly, Sadler and Rahey (2004) also asserted
that prescience must be distinguished from déjà vu,
since the 4 patients they described could clearly make
a distinction between the two phenomena. One of
their patients had an unknown aetiology and two had
bitemporal interictal spikes (two had bilateral temporal
seizures and only right discharge-induced prescience,
for two no seizures were recorded). Our patient has
similar epileptic characteristics (bitemporal epilepsy of
unknown origin).
During the transient phenomena associated with prescience,
she experienced an erroneous relation to time,
giving her the brief impression of being able to foresee
the future. A damaged hippocampus, severely impaired
in her case, could support such a symptomatology and
an erroneous temporal arrangement of the sequence of
events. Projection into the future is an inseparable mechanism
of episodic memory (Addis & Schacter, 2012; Klein,
2013). The recovery of an episodic autobiographical
memory involves a mental travel back to the past, but individuals
also become aware of their own identity in the
future (Tulving, 2002, 2005; Tulving, Voi, Routh, & Loftus,
1983). Tulving (1985) and Klein, Loftus, & Kihlstrom (2002)
have shown that amnesic patients may have difficulty imagining
the future. Hassabis, Kumaran, and Maguire (2007),
in a functional MRI study, demonstrated that the construction
of new scenes involves a network comprising the hippocampus,
the parahippocampal gyrus, the retrosplenial
cortex and the posterior parietal cortex. Therefore, the
same brain regions are activated when patients think of
past or future events (Addis, Moscovitch, & McAndrews,
2007; Okuda et al., 2003). Consequently, the episodic
system may contribute significantly to imagining the
future (Addis et al., 2007). Addis and Schacter (2012)
suggested that 3 processes might be particularly dependent
on the hippocampus: (1) allowing access to details
stored in memory to develop new scenarios, (2) combining
these different details in a spatio-temporal context (3)
encoding simulation of a future project in memory so
that it can influence future behaviours (Addis & Schacter,
2012). In addition, “time cells” that encode different successive
moments of an experiment and recognize the time
intervals between each episode have been discovered in
the hippocampus (Eichenbaum, 2013; MacDonald,
Lepage, Eden, & Eichenbaum, 2011). Therefore, the hippocampus
might be an essential structure for learning
sequences of events, allowing the brain to distinguish
memories for conceptually similar but temporally distinct
episodes, but also to associate temporally contiguous representations
linked to independent experiences (Ranganath
& Hsieh, 2016). Considering the roles played by the
hippocampus in the projection into the future and the temporal
organization of memory representations, we hypothesize
that prescience might be directly related to
hippocampal dysfunction in parallel with the preservation
of subhippocampal structures, at least in epileptic patients.
In fact, OHVR’s cognitive profile is the opposite of the
impaired familiarity/intact recollection of epileptic patients
who experienced déjà vu as reported in the study by
Martin et al. (2012). One possibility could be that in OHVR
a feeling of familiarity might be an indication of impending
retrieval, i.e., interpreted as prescience, which might never
happen due to hippocampal damage.

Acquired synaesthesia in OHVR?
In our patient, her disease (of unknown aetiology) was also
associated with the awareness of synaesthesia together
with a strong interest for numbers approximately three
years after onset. For example, she reported that she developed
the habit of counting all the time until she reached a
number she liked and therefore experienced a pleasurable
emotion. Synesthetic associations are thought to be
acquired most often during childhood and they may constitute
a variant of childhood memories (Witthoft &
Winaver, 2013). This so-called developmental synaesthesia
is a particular feature of subjective experience, considered
as non-pathological and shared by only a fraction of the
population. The estimated prevalence varies widely
according to the definition and criteria, from a few to
approximately ten percent (Chun & Hupé, 2016; Rouw &
Scholte, 2016; Simner, 2012; Simner & Carmichael, 2015;
Watson et al., 2017). Many synaesthetes only become
aware of their particularity late in life, when they learn
about the phenomenon or start to pay special attention
to their inner life. This could have been the case for our
patient, her attribution of affective valence for numbers
having been kept at the sub-conscious level until then.
However, most synaesthetes report that they experienced
their synesthetic associations “as far as they can remember”,
even while acknowledging that they had not been
conscious of them. Our patient clearly reported not
having any synaesthetic experience before her epileptic
episodes, suggesting a causal link. However, her memory
difficulties may also have prevented her from remembering
her childhood clearly enough.
Acquired synaesthesia in adults is rare and has been
reported in very different contexts: following psychotropic
and drug ingestion (Luke & Terhune, 2013; Sinke et al.,
2012), migraine (Alstadhaug & Benjaminsen, 2010; Podoll
& Robinson, 2002), after neuropathology involving the
optic nerve and/or chiasm (Afra, Funke, & Matsuo, 2009)
or blindness (Armel & Ramachandran, 1999; Niccolai,
Jennes, Stoerig, & Van Leeuwen, 2012), after a head
injury (but without visible lesions on MRI) (Brogaard,
Vanni, & Silvanto, 2013) or a thalamic stroke (Ro et al.,
2007; Schweizer et al., 2013). In most cases, no focal brain
lesion could be identified and the diversity of the cases
makes a common neurological substrate unlikely. In
addition, acquired variants of synaesthesia seem to be
qualitatively different from developmental synaesthesia,
involving mostly low-level sensory triggers rather than
learned symbols as observed for our patient and in developmental
synaesthesia (Ward, 2013).
To date, no epidemiological study has tested whether
there is a higher prevalence of synaesthetes in epileptic
patients. Similarly, no case of acquired synaesthesia has
been reported after hippocampal lesions, or formally
associated with déjà vu. The neural correlates of developmental
synaesthesia remain to be identified (Hupé &
Dojat, 2015), and therefore, there is no indication as to
whether or not the hippocampus and subhippocampal
regions are involved. Whether or not there is any causal
link between hippocampal atrophy and the appearance
of synaesthesia in our patient therefore remains an open
question. This case clearly calls for further investigation
of synaesthesia in (temporal lobe) epilepsy in future
studies, in relation, or not in relation to déjà vu.
Temporal and extra-temporal remapping within memory
networks have been demonstrated in TLE patients (verbal
and visual encoding memory tasks during functional MRI)
(Alessio et al., 2013; Sidhu et al., 2016). Compensatory
brain activation can be observed in healthy areas, like the
contralateral hippocampus in reaction to hippocampal sclerosis
or extratemporal regions such as the cuneus or the
anterior cingulate cortex (Sidhu et al., 2016). The growing
attention of OHVR to the affective valence and personifications
of numbers could be part of a compensatory cognitive
strategy. It would be interesting to examine whether
synaesthesia occurs more often in epileptic patients as it
could be a way to explore its neural substrates.
Nevertheless, an alternative and speculative explanation
might be possible considering the strikingly preserved
amygdala volume in OHVR, despite the bilateral
loss of hippocampi. These major nodes for emotional processing
are bilaterally intact in OHVR. Amygdala and hippocampal
complex are the core nodes of two independent
memory systems, respectively emotional (such as fear) conditioning
and declarative memory. These systems tightly
interact when emotional stimuli are encountered and
when complex emotional memories are created or
retrieved (Phelps, 2004). The parietal cortex, essential for
the number processing (Dehaene, Piazza, Pinel, & Cohen,
2003), is also apparently preserved in OHVR (an apparently
normal volume, normal PET metabolism). Therefore, we
can hypothesise that there is some reorganisation and a
sort of overinvestment or dysregulation of such preserved
networks after bilateral hippocampal lesions to cope with
severe memory deficits. The implication of amygdala in
emotions and valence could suggest that a functional
network has been released from normal inhibition, reactivating
long forgotten childhood associations of numbers
with affective valence.


Conclusion

Our report suggests that the perirhinal region plays a critical
role in déjà vu. It also suggests that déjà vu can occur in
patients with severe memory impairment and massive hippocampal
damage. However, it does not entirely resolve
the issue of whether or not some relation between the
perirhinal region and the hippocampus is necessary for
déjà vu to occur as some reports suggest (Bartolomei
et al., 2012) or whether the hippocampus is involved in
déjà vu at all. OVHR experienced what she thought to be
two distinct phenomena, one related to déjà vu, the
second related to prescience and resembling déjà vécu,
which supports the idea that the two should be distinguished.
Whether these depend on the dysfunction of
different brain areas or are related to a continuum,
remains to be investigated. However, OHVR’s profile
highly suggests that hippocampal dysfunction is needed
for the emergence of feelings of prescience, and more
broadly for déjà vécu.