Mental Time Travel [MTT]

1. INTRODUCTION

The term Mental time travel (MTT) has first been introduced by Suddendorf
and Corballis in 1997 and has since become an accepted and important topic of
research in the scientific community. MTT refers to the ability of humans to mentally
project themselves into the past or the future - to re-live or pre-live events[1].
It is therefore a remarkably flexible tool for humans to predict, plan and shape any
future situation. While, without a doubt, it provides a tremendous advantage for
natural selection, the purpose of many aspects of this ability still remain unclear,
as well as the fact, that it does not seem to have evolved in any other species (see
section 5). Whereas the systems and neural correlates for future travels have only
recently become subject of research, MTT to the past has been extensively studied
under a different name: episodic memory (EM).

2. A CLOSER DEFINITION OF MENTAL TIME TRAVEL
Before discussing MTT in detail, it should be clearly set apart from related, but
different, phenomena.
Let’s focus on past MTT for a moment, since for the memory systems, a widely
used taxonomy has already been formulated. Non-declarative memory (e.g. conditioning)
may also serve as a predictor of the near future, but it is bound to the
perception of stimuli. This bottom-up system through perception clearly does not
belong into the domain of MTT. Declarative memory on the other hand, is a topdown
system, not bound to perception and provides a greater flexibility. It is further
divided into semantic and episodic memory: Whereas the episodic memory contains
the personal experience of an event, the semantic memory merely holds the
knowledge extracted from this event. Recollecting where and when one has learned
about an information, is different from solely knowing that fact. This property of
the episodic memory has been named autonoetic consciousness by Tulving (1985)
and is essentially what MTT is about. In addition to retrospection however, Suddendorf
proposed that there is also a prospective counterpart to episodic memory -
therefore forming the notion of MTT as the capacity to both re- and pre-experience
personal events[1]. Whether or not it is valid (in a neuroanatomical view) to merge
those two abilities into one system is discussed in section 4.
For further clarification, it is helpful to look at some psychological notions that
might be confused with MTT. For example, daydreaming (mind wandering) may
involve MTT, but does not necessarily involve the self as a character. It can rather
contain purely fictitious characters and improbable events. Similarly, prospective
memory, defined as “memory for actions to be performed in the future such as
remembering to give a message to a friend or remembering to take medication”[2],
does not require the person pre-living the task.

3. METHODS OF INVESTIGATING MENTAL TIME TRAVEL
A classical approach in memory research is the use of questionnaires, which is
still the only viable method for numerous research aims. For instance, Berntsen
(2008) could show that involuntary MTT occurs frequently in daily life, simply by
equipping her participants with a notebook and instructions to take notes of time
travel occurrences[2]. In contrast, even sophisticated imaging experiments focused
only on well controlled, voluntary MTT. Questionnaires certainly contain a great
deal of difficulties, such as the requirement for cooperation and reliability of the
subjects or the objective judgment of the highly heterogeneous data. However they
can provide unique information about the content of mental representations, the
environmental stimuli triggering MTT or the impact on mood.
In brain imaging experiments, the subject is usually presented with stimuli (e.g.
key words) for which he or she has to imagine a past or future event. The recorded
information, depending on the task, can contain a combination of behavioural data
(reaction time, error rate) and different imaging methods, usually EEG, PET or
fMRI. Due to their coarse temporal resolution, PET and fMRI cannot provide precise
information about the temporal activity pattern but can provide the brain regions
involved with high spatial resolution. For identifying the dynamics of activity,
EEG is the method of choice (usually measuring slow cortical potentials (SCP)
or evoked potentials)[3].

4. NEURAL CORRELATES OF MENTAL TIME TRAVEL
In the past, functional MRI studies have thoroughly identified the brain regions
involved in autobiographical memory. Recent research comparing future with past
MTT has found the same patterns of activity as in those previous studies. Therefore
supporting the view of a common neuronal system for both past and future
MTT, which might be counter intuitive, due to the fundamental asymmetry of those
two components (past is fact whereas future is fiction). The fMRI results involve
a widespread left-predominant cerebral activity and three regions with particular
high activation: The medial prefrontal cortex (mPFC), the medial temporal lobes
(particularly the hippocampi) and posterior regions [4]. Using sophisticated analysis
methods, EEG measurements can also be applied to create a voltage topography
map (evoked potential map), which in a recent study by Arzy et al. (2008) has
shown to be consistent with the regions described in fMRI studies. In particular,
three regions could be localized that are activated 300-600 ms after stimulus onset
(i.e. a key word as mentioned earlier): The left anteromedial temporal cortex, the
right temporoparietal junction and the occipitotemporal cortex [5]. Surprisingly,
he did not mention prefrontal regions anywhere in this study. Most likely, they are
important at a later stage in the MTT process.
The left hippocampus is related to the retrieval of context-dependent autobiographical
memory, i.e. it acts as an index-like code to retrieve storage of personal
specific events from different cortical sites. Unlike the right hippocampus which is
only associated with place context, the left hippocampus is involved in both place
and person conditions. It is therefore interpretable as a necessary first step to create
the representation of a past or future event [4]. The occipitotemporal cortex is
known to be involved in recall through visual imagery and clinical studies have described
amnesia with damage to this structure. The temporoparietal junction plays a crucial role in self-other distinction, self-location in space and - with this new data - also self-location in time [5].
Though minor differences in activity patterns between past and future MTT
could be observed using EP mapping, I will focus on the measurement of slow
cortical potentials (negative potential shifts at the scalp, reflecting activation of underlying
cortical regions) to illustrate this difference. By measuring SCPs during
the generation of experienced or imagined mental representations, it is possible to
clearly distinguish those two components of MTT. Experienced memories were associated
with significantly larger negative dc shifts over occipito-temporal regions,
reflecting the access of autobiographical imagery. In contrast, a more intense activation
of the left PFC for imagined events was observed, which may reflect the
generation of generic imagery in those networks [3].
In light of recent research, it is therefore hypothesized that MTT does not only
involve memory mechanisms, but mechanisms of mental imagery and self-location
as well [5].
Many of the brain regions discussed here, have undergone profound developments
that are unique for humans, therefore neurophysiological results may support
the view that animals are not capable of MTT. A matter that will be discussed
next.

5. MENTAL TIME TRAVEL IN ANIMALS?
Evolutionary considerations.

An evolutionary explanation for MTT, in particularfor episodic memory, cannot easily be provided. EM is highly selective and retains only a tiny fraction of experiences (you may encounter dozens of people each day,
but only remember a few). At the same time, a vast amount of details such as
the precise location in space and time, the weather conditions, the persons present,
the words exchanged, and so forth are remembered. In a comment to Suddendorfs
work, Dessalles described it as a “waste of storage” in a computational perspective,
since it does not allow generalization of the learned content (as in semantic memory)
and it is highly unlikely that a person will ever encounter a situation matching
all those parameters again [1]. An other astonishing fact is, that much of our cortical
mass is devoted to EM, making it a seemingly important system. At the same
time, its constructive nature makes it prone to errors and even in intact systems
autobiographical memories can be entirely false (confabulation), let alone in cases
of amnesia where EM is the most vulnerable, fragile system [3].
The closely related brain areas and the similarities between past and future MTT
might explain EM as a side-effect or prerequisite (providing the raw-material) for
creating mental presentations of the future (which in an evolutionary point of view
are obviously more valuable). As with other human attributes, such as theory of
mind or language, MTT is considered a product of the hominins being forced into
a “cognitive niche” by danger from predators as a result of climate changes 2.5
million years ago [1].
Animal research

Any research about cognitive evolution has the intrinsic problem,
that we can only observe animal behaviour, but we can never know what their
mental state is, i.e. whether they are mentally time traveling anywhere.
Experimenting with episodic memory is especially challenging, since the result
is never conclusive. Previously, animals could be shown to perform extraordinarily
on certain memory tasks, but simply knowing something does not imply that
the animal actually remembers when and where it has acquired this knowledge.
Thus, future MTT is suggested to be the more promising subject of research, since
it would essentially provide the selective advantage and hence be visible to evolution
(or clever researchers) [1]. The Bischof-Köhler hypothesis (1985) states, that
animals can only act on present needs or drive states, as opposed to future needs.
At first sight, this seems incorrect as exemplified by the caching of food, the manufacturing
of tools by some animals or other spectacular accomplishments under
experimental conditions. However, I will refrain from quoting any of this previous
work, since a closer examination of the results always allows alternative explanations
based on instinct or associative learning. Thus far, in more personal, flexible
situations involving non-instinctive behaviours, the Bischof-Köhler hypothesis has
not yet been falsified. A nonverbal test for MTT has to meet many criteria, such as
controlling the current need state (e.g. that the animal is not thirsty at the present)
and providing the opportunity to to secure a future need (e.g. obtaining a drink for
a thirst inducing future situation) involving species-untypical behaviour [1].
Sophisticated experiments are presently carried out and fascinating new insights
are to be expected in the near future.