Keywords
lucid dreaming - REM sleep - electroencephalography - consciousness
Introduction
In 1913, Dutch writer and physician Frederik Willem van Eeden suggested using the
term lucid dreaming (LD) to refer to dreams in which people know they are dreaming.[1] In the 1970s and 1980s, the phenomenon was recognized by the scientific community
with the help of laboratory studies in England and the USA.[2]
[3]
[4] Though LD's nature remains unclear,[5] it provides many application opportunities, such as interacting with computers while
asleep,[6] honing motor skills,[7]
[8] preventing nightmares,[9]
[10] solving problems,[11]
[12] improving waking mood,[13]
[14]
[15] reducing depression,[16] and eliminating chronic pain.[17] No less than 55% of people have experienced LD at least once in their lifespan,
and many of them experience it regularly.[18] This frequency may be 88% or more if other phenomena related to LD are considered.[19] For example, consciousness in rapid eye movement (REM) sleep is also linked to sleep
paralysis,[20]
[21] false awakenings,[22] and out-of-body experiences.[23]
[24]
[25] In 2004, we suggested that these phenomena could be united under the term phase state, as doing so may provide a more accurate picture of their nature.[26]
[27]
The number of LD studies has increased by 5.6% each year.[28] Lucid dreaming appears to be most often detected during REM sleep, with just a few
confirmed exceptions.[29]
[30]
[31] Later, it was found that LD was associated with increased activity in the occipito-temporal
cortex cuneus, prefrontal cortex, parietal lobules, and bilateral precuneus.[32] The connectivity between the temporoparietal junction and anterior prefrontal cortex
was also associated with LD, as was the volume of gray matter in the anterior prefrontal
cortex.[33]
[34]
As LD studies become more advanced, they require more optimized and effective methods
and technologies, especially regarding LD verification methods. In 1978, Hearne was
the first to confirm LD under laboratory conditions, which he did by observing preagreed
eyelid movements (PAEMs).[4] Later, LaBerge used the same method, and it became the gold standard for LD studies.[2] Though the method is not ideal,[35] almost all LD laboratory studies since those mentioned above have used polysomnographic
(PSG) observations, which include at least electroencephalography (EEG), electromyography
(EMG), and electrooculography (EOG). As a result, LD verification has become expensive
for independent researchers and requires PSG skills, which restricts the efficiency
of all such studies.
In 2021, in an attempt to simplify and make LD studies cheaper, it was suggested and
confirmed that REM sleep and consciousness can be detected at the same time using
only one EMG sensor. This was possible by preagreed chin movements (PACMs), even though
muscle atonia in the submentalis area exhibits less activity than in the distal muscles.[36] In this case, an EMG sensor detects REM sleep by its main feature in the form of
muscle atonia and then detects consciousness based on the residual electric activity
of three chin movements.[37] This PACMs method is effective when studies require many cords and sensors besides
PSG.[38] However, the main problem with this method is that it may be hard to differentiate
genuine muscle atonia from deep relaxation or the N3 stage of non-REM sleep.
It is unclear whether there is any way to verify LD that is as simple as the PACM
method. As any such method should disqualify false results, EOG cannot be used because
eye movements are easy to control willingly. At the same time, EEG could be used for
brain-computer interface, even during LD,[6] meaning that EEG may represent conscious actions. Moreover, even during PACMs testing,
there were distinctive EEG artifacts remaining from jaw movements during LD.[37] Furthermore, facial muscles present EMG activity during dreaming and LD, mostly
because speech and emotions in dreams, which can be controlled by will,[38]
[39] are only partially paralyzed.[40]
[41]
[42] This fact is usually regarded as a problem for EEG sensors—especially those located
close to the face—because electrical spikes from this area may create artifacts in
EEG data.
This situation begs the question of why preagreed EEG artifacts are not created intentionally.
In theory, this would be possible in locations where EEG sensors are close to facial
muscles, such as Fp1, Fpz, Fp2, AF7, AF3, Afz, AF4, and AF8 (10–20 system), and could detect raising the eyebrows. Therefore, it is possible
to transfer preagreed signals via EEG. Then, the same sensor simply needs to detect
REM sleep.
Hypotheses
The main hypothesis of this study is that PAFMs can be used to verify LD when only
EEG is available. This idea was evaluated by having a few LD practitioners test PAEMs
and PAFMs alongside one another under laboratory conditions. Confirmation of the hypothesis
would indicate that there is a simple way to verify LD in studies with few sensors
or for which PSG is unavailable. It would also make LD studies cheaper and more convenient
for volunteers. These implications are especially important in regard to sleep paralysis,
which has the same physiological attributes as LD, making studies on this phenomenon
more effective in some cases than in others because people could use PAFMs to report
cases of sleep paralysis.
Materials and Methods
Resources and Participants
The current experiment was performed by experienced LD practitioners under laboratory
conditions. The study approach was approved by the Phase Research Center ethical committee
review board (PRC-2022-03-30-01). Written informed consent about the study and its
methods was received from all volunteers. They also confirmed the absence of any psychological
or physiological health issues that could have been affected by the study tasks. All
participants confirmed that they were at least 18 years of age. The volunteers received
a financial reward for their participation, and their travel and accommodation expenses
were covered. No medical supplements were used to enhance LD attempts or the study
results.
Experimental Task
Participants were asked to follow these steps: A) to induce LD by any technique or
method; B) in LD, make three consecutive PAEMs to the left/right/left sides; C) after
making the PAEMs, make three consecutive PAFMs by raising the eyebrows; D) report
these LD actions upon awakening. Participants had up to 4 nights in the laboratory
to achieve the experimental task at least once. They were also allowed to use maintaining
and stabilizing techniques to make LD more stable and vivid. No questionnaires were
applied, as PSG sensors represent highly objective information.
Apparatus
Lucid dreaming was detected and verified using Encephalan-EEGR-19/26, with the following
settings: one EEG channel (Fpz and A2 positions from the 10–20 system; 50 Hz notch filter; 0.7–70 Hz band-pass filter),
two EOG channels (50 Hz notch filter; 0.7–70 Hz band-pass filter), and one chin EMG
channel (50 Hz notch filter; 16–70 Hz band-pass filter).
Results
Five volunteers participated in the present study (25–38 years old, all males). They
reported eight LDs. In one dream, PAEMs were not apparent, but PAFMs were distinctive;
in another dream, the opposite pattern emerged. In the other six cases, both LD verification
methods were distinctive. Preagreed frontalis movements left two different patterns
of artifacts on EEG, both of which could be observed in the LDs of one participant.
In three LDs, the most distinctive pattern represented a bold EEG artifact of high-frequency
waves with inconsistent low or average amplitudes that varied from LD to LD, but not
in one PAFMs set. Another type of PAFMs consisting of slow-sinusoidal, high-amplitude
EEG artifacts were detected in two LDs. In another two LDs, both patterns were mixed
in a single set of signals.
Volunteer #1 reported a dream that coincided with the most distinctive PAEM and PAFM
cycles. Preagreed frontalis movements were observed as high frequency, average-amplitude
EEG artifacts ([Figure 1]). Volunteer #2 reported a dream that coincided with a distinctive PAEMs and mixed
PAFMs cycles. This participant's PAFM were observed as the simultaneous presence of
high-frequency, average-amplitude EEG artifacts and slow-sinusoidal, high-amplitude
EEG artifacts, which changed each other ([Figure 2]). Volunteer #3 reported two LDs that coincided with distinctive and clear PAEMs
and PAFMs cycles. Preagreed frontalis movements were observed as high-frequency, low-
or average-amplitude EEG artifacts ([Figures 3] and [4]). Volunteer #4 reported two LDs that contained only distinctive PAEM or PAFM cycles.
In the first dream, only PAEMs were visible; in the second dream, only PAFMs were
visible, which were observed as slow-sinusoidal, high-amplitude EEG artifacts ([Figures 5] and [6]). Volunteer #5 reported two LDs that coincided with distinctive and clear PAEMs
and PAFMs cycles. In the first dream, PAFMs were visible as simultaneous high-frequency,
low-amplitude EEG artifacts and slow-sinusoidal, high-amplitude EEG artifacts, which
changed each other. In the second dream, PAFMs were visible only as slow-sinusoidal,
high-amplitude EEG artifacts ([Figures 7] and [8]).
Fig. 1 PAEM/PAFM Cycle of Volunteer #1. Note: The EOG sensors were not opposite.
Fig. 2 PAEM/PAFM Cycle of Volunteer #2.
Fig. 3 PAEM/PAFM Cycle #1 of Volunteer #3. Note: The EOG sensors were not opposite.
Fig. 4 PAEM/PAFM Cycle #2 of Volunteer #3.
Fig. 5 PAEM/PAFM Cycle #1 of Volunteer #4.
Fig. 6 PAEM/PAFM Cycle #2 of Volunteer #4.
Fig. 7 PAEM/PAFM Cycle #1 of Volunteer #5.
Fig. 8 PAEM/PAFM Cycle #2 of Volunteer #5.
Discussion
Other than eye movements and sleep atonia, two of the main features of REM sleep are
low alpha waves and dominant theta EEG waves.[43]
[44] This wave pattern is a defining feature of REM sleep but has some deviations in
LD.[35] In general, EEG could represent REM sleep as well as consciousness, and it cannot
be scammed. In a search for a simple and reliable method for verifying LD under laboratory
conditions, it was hypothesized that straining frontalis during LD may create distinctive
EEG artifacts during REM sleep, representing lucidity. This new method would make
LD studies much cheaper, as it does not require PSG and is more convenient than current
methods in some situations. This idea was tested by comparing the PAFMs method with
the traditional PAEMs method under laboratory conditions.
Hypothesis Confirmation
Neither the PAEMs nor the PAFM method was 100% efficient, but both were distinctive
in the same number of LDs. These results confirm our hypothesis by showing that PAFMs
can be used as an LD verification method, as it was visible in the form of EEG artifacts
in most of the reported LDs and coincided with verbal reports. As a result, if the
Fpz EEG sensor represents the dominance of theta waves with sawtooth waves adherent to
REM sleep,[43]
[44] and if there are consecutive artifacts, it could mean that LD occurred. This outcome
was apparent from previous studies, which show that facial muscles have residual electrical
activity during sleep.[36]
[37]
[38]
[39]
[40]
[41]
[42]
[45] The main contribution of the study is that it creates a clear procedure for using
this knowledge in a new practical way. As the PAFMs method requires only one EEG sensor,
it is more convenient than the PAEMs method for LD practitioners and researchers to
use. Furthermore, the PAFMs method is substantially cheaper than the traditional PAEMs
method using PSG.
The most important aspect of this study is the development of a new and simple method
for LD verification, which could promote LD studies and reduce their cost. As the
PAFM method requires only one EMG sensor, it takes little time to assemble and could
be applied by LD practitioners themselves, especially if dry-contact EEG sensors are
used. As sleep paralysis, out-of-body experiences, and false awakenings share similar
features with LD,[19] the PAFMs method could facilitate studies of all these phenomena.
PAFM Procedure
During the pilot tests and the study, itself, the implications of PAFMs appeared to
require more attention than those of PAEMs. All that is needed to detect PAEMs on
EOG is for the eyes to move in a left-right-left pattern; however, merely straining
the frontalis may not lead to the detection of PAFMs, or they may become less distinctive
on EEG. The results show that the following factors are important for detecting PAFMs:
A) there should be at least three PAEMs; B) frontalis movements must be done with
an emphasis on straining the muscle, not just on raising the eyebrows; C) frontalis
straining must be prolonged, approximately from 0.3 to 1 second; D) there must be
breaks (approximately 0.5–2 seconds) between frontalis movements; E) frontalis movements
should not be performed during unstable LD stages because this leads to immediate
awakening; and F) PAFM artifacts on EEG should coincide with verbal reports given
after LD.
Regarding EEG sensor positions, Fpz was used in the present study. In theory, Fp1, Fp2, AF7, AF3, AFz, AF4, and AF8 may be apt for PAFMs, but the artifacts could be less apparent.
The most unexpected result was that two types of EEG artifacts were observed when
the PAFMs method was used. Artifacts emerged as high-frequency waves with inconsistent
low- or average-amplitude or slow-sinusoidal, high-amplitude EEG artifacts. Both types
were recorded in similar LD cases and could miх with each other. The most obvious
explanation is that the physical frontalis movements could lead to slow-sinusoidal,
high-amplitude EEG artifacts. However, no such artifacts were observed during PAFMs
training, which was performed by actually raising the eyebrows. Currently, we cannot
explain this issue, and it is important to expect and accept both types of artifacts
while using PAFMs.
PAEMs, PACMs, or PAFMs
It is possible to compare three methods for verifying LD. The traditional PAEMs method
is the most complicated, expensive, and inconvenient, but it gives the most reliable
and detailed data, it can be performed in cases of unstable LD, and it is the easiest
to use during LD, as it requires only simple eye movements.[35] To date, the PAEMs method is the only LD verification method recognized by the scientific
community, making studies based on it more legitimate, whereas the PACMs method has
been accepted in only one study.[38] The PACMs and PAFMs methods are cheaper and more convenient to use while asleep.
In addition, they are less complicated, but both need concentration during LD. Both
can lead to awakening if performed in an unstable LD stage. Before choosing an LD
verification method, the pros and cons of each should be considered based on the specific
study conditions and goals.
All three methods can support each other to verify LD more reliably. This became apparent
when we analyzed the attempts of volunteer #4 ([Figures 5] and [6]). In one LD, only PAEMs were clearly distinctive, while only PAFMs were visible
in the second attempt. Without considering PAFMs, the second LD would not have been
verified. So, PACMs and PAFMs could be considered to support PAEM, thus providing
more reliable data than when considering PAEMs alone. As the PAEMs method can be used
during any LD stability stage, PACMs and PAFMs could not only highlight LD but also
verify its stable forms (they lead to awakening otherwise). The PACMs and PAFMs methods
are more convenient than the PAEMs method when the goal is to confirm rigid LDs.
PAEMs and PAFMs Failures
It is important to note that one volunteer (#4) experienced PAEMs and PAFMs failures
(in different LDs). It could be a coincidence, but this volunteer has participated
in some of our other studies, in which similar issues were observed. These failures
were related to conscious EMG or EOG signals during LD, sent by facial or body muscles.
While this is a different topic to study, two hypotheses could explain this situation.
First, unlike all other volunteers, this one lacked muscle development and led a lifestyle
without engaging in physical activity or sports. As a result, the weak development
of this participant's neuromuscular circuit could manifest while sending signals during
LD. Secondly, this volunteer's sleep atonia could be more profound than others'. However,
problems with the PAEMs method contradict this idea. Both hypotheses could explain
why the PAEMs and PAFMs methods may not work for some individuals.
Limitations
Alpha waves can be much higher during LDs than during REM sleep.[35] At least visually, these EEG patterns could be very close or similar to patterns
observed during relaxed wakefulness, meaning that the PAFM artifacts from LD could
be confused with wakefulness. Conversely, fake LDs could be reported when no chin
EMG is recorded to detect sleep atonia, which is the hallmark of REM sleep, even in
LD. Therefore, only PAFM artifacts in EEG patterns with low alpha waves can verify
LD if no other sensors are used.
Conclusions and Future Studies
Conclusions and Future Studies
The PAFMs method was tested in a search for a simple LD verification method. Two types
of distinctive EEG artifacts during REM sleep represent lucidity in dreams. This new
method could make LD studies cheaper and more convenient in some situations, resulting
in more studies on the topic.
Future studies could focus on explaining why PAFMs have two types of EEG artifacts
or how to clearly separate LD EEG patterns from the patterns observed during wakefulness.
Research could also look for simple ways to verify LD other than the PAEMs, PACMs,
and PAFMs methods. Such studies may generate new research and application opportunities
in many areas, enhancing the understanding of dreams and consciousness in general.
