Posted comment on ´Sleep fragmentation and lucid dreaming` written by J.Gott, M. Rak, L. Bovya, E. Peters, C.F.M. van Hooijdonk, A. Mangiaruga, R. Varatheeswaran, M. Chaabouf, L. Gorman, S. Wilson, F. Weber, L. Talamini, A. Steiger and M. Dresler and published in Consciousness and Cognition 84 (2020) 102988
SUMMARY
Gott and colleagues describe in their article the results of their study investigating the association between sleep fragmentation and lucid dreaming. The authors wanted to look at whether or not there is a direct correlation between the two and whether or not lucid dreaming is an artefact of deleterious or disrupted sleep or vice versa.
The study consisted of 4 different investigations. Study 1 looked at the association between sleep fragmentation and lucid dreaming. This required 202 participants (aged 18-63 with recruitment via internet forums on lucid dreaming) performing an online survey. The participants were asked to rate on a scale of 0 to 8 (0 – never, 8 – almost every night) dream recall frequency and lucid dreaming. (Lucid dreaming was defined according to Schredl and Erlacher that the individual has to be aware they are dreaming and with this awareness can control the dream actions or to observe passively the course of the dream). The scale was then re-coded using a class mean system, eg. 0 to 0, 1 to 0.042 etc. Continuity of sleep was also rated by the participants on a scale from 1 to 5 with 1 being highly continuous and 5 highly fragmented. Finally in Study 1 the participants were asked to self-report on how many times they were awake and to assess subjectively the perceived link between their sleep fragmentation and their lucid dreaming (eg. Do you have more lucid dreams during 1: highly continuous……… sleep?).
Study 2 was carried out to assess the association between dream lucidity and sleep rhythms. In this case, 22 volunteers were interviewed about their lucid dream frequency during monophasic and polyphasic sleep rhythms. Ten participants made up the ´in-house` group and these were subjected to a normal monophasic sleep schedule for 4 weeks before going into a polyphasic sleep routine of 6 naps of 20 minutes each every 24 hours with no extended night sleep period. Sleep timing was recorded by sleep diaries, actigraphy and weekly questionnaires including Altman Self-Rating Mania Scale (ASRM), the Beck Depression Inventory (BDI-V2) and for complaints about cognitive disturbance, FLei. Some participants also underwent 24 hours of polysomnography (SOMNOwatch), blood sampling every 30 minutes and cognitive testing for declarative and procedural memory performance, fluid reasoning and psychomotor vigilance before changing to the polyphasic sleep routine. The polyphasic sleep routine was scheduled to be maintained for up to 8 weeks with a final set of testing in the last 24 hours. Early termination required the participants to fill in a questionnaire. For the remaining participants, interviews were carried out within 3 months of returning to a regular monophasic sleep rhythm. The other 12 participants taking part in the study formed the ´external` group. These were participants who stated that they had previously tried a similar polyphasic sleep routine although there was no systematic testing or other procedures undertaken. The participants of the ´external` group were interviewed by email. Participants of both the ´in-house` and ´external` group recorded their dream experiences for both the polyphasic sleep phase and the following monophasic sleep phase. Frequency of dreaming and lucid dreaming was rated on a 6 point scale for the polyphasic period with 0 as never and 5 after each nap/24 hour and on a 5 point scale for the monophasic period with 0 as never and 4 as several dreams per day. Lucid dreaming was rated as given for Study 1 and then re-coded with a slightly different scale to Study 1 (eg. 0 to 0, 1 to 0.03). Participants were asked to rate their sleep quality (eg. difficulties in falling asleep and difficulties to sleep through) in the polyphasic period up to the monophasic period. The ratings were between 1 and 5 with a rating of 1 being ´much better`, 2 ´better` going up to 5 as ´much worse`.
In Study 3 the association of self-rated sleep quality and longitudinally assessed lucid dreaming was investigated. Forty-two participants aged 18 to 34 that fulfilled the Gott and team`s initial inclusion criteria were first asked to fill out a survey during the first week of study which consisted of a demographics questionnaire and a questionnaire about dream frequency, general attitude to dreams and level of control. The participants if allowed to continue then completed once the Pittsburgh Sleep Quality Index (PSQI) and then daily for six weeks the questionnaire on dream lucidity (DLQ) as well as questionnaires on sleep and mood. The PSQI consisted of 19 questions relating to the sleep habits of the participants (eg. How often during the past month have you had trouble sleeping?) during the past month. The PSQI questions were grouped into 7 component scores each with a range of 0 to 3 with 0 as ´no difficulty` and 3 as ´severe difficulty`. The groups were: subjective sleep quality, sleep latency, sleep duration, habitual sleep efficiency, sleep disturbances, use of sleeping medication and daytime dysfunction. The scores were then added to form a global score with the higher scores relating to worse sleep quality. The DLQ questionnaire consisted of 12 areas assessing awareness of the lucid dreaming (awareness that dream characters/objects are not real, awareness of dreaming, awareness that the physical body is asleep), different types of control (changing dream scenes/characters/events, breaking physical laws, deliberately choosing an action) and memory recall (of intentions and of waking life). The participants rated each on a 5 point scale (0 – not at all, 1 – just a little, 4 – very much) and then the scores were added together and the mean calculated.
For Study 4, 30 volunteers selected according to study criteria (eg. had a dream recall frequency of greater than 3 dreams per week) were subjected to polysomnography and dream lucidity questionnaires in order to investigate the association between arousal features and lucid dreaming. The Lucidity and Consciousness in Dreams Scale (LuCiD) was used which consisted of 28 statements (eg. Whilst dreaming, I was aware of the fact that the things I was experiencing in the dream were not real.) which the participants had to rate on a 5 point scale as to agreement or disagreement (eg. 0 – strongly disagree, 5 – strongly agree). Polysomnography was also carried out for 1 night. The EEG data was appropriately processed and then sleep was scored in 30 second time periods according to standard scoring systems by at least 2 independent scorers. Arousals were analysed automatically and then manually quantified and transitions (awakenings) between sleep and wake were counted manually. The authors then analysed the number of wake-REM sequences (WREM) which was defined as the number of times a participant successfully entered or re-entered into REM from an arousal state within 5 minutes. The arousal state was defined as either a continuous wake period that spanned at least 2 epochs or as a single wake epoch or arousal that resulted in transition through sleep Stages 1 and 2 before returning to REM. (This analysis was carried out because of views that dream lucidity achieved by the wake-back-to-bed strategy could be explained by waking levels of prefrontal activity persisting into subsequent REM sleep – Dresler). The analysis was carried out manually and differences between ratings given by the scorers were settled by taking the average. Automatic analysis was carried out using an R script. The authors also looked at in this study the Number of Awakenings (NOA) and Total Wake Time (TWT) which were also assessed manually and automatically using an R script. The first was initially assessed only for the last two hours of the sleeping phase before final awakening when REM sleep incidence was believed to be higher, dream recall more robust and the occurrence of lucid dreams more, but later analyses used varied time windows (eg. a 4 hour window).
All studies were subjected to appropriate statistical analyses, eg. for Study 2, the Wilcoxon signed rank-test and paired t-test, Pearson correlations for Study 4 and in some cases, significance was not in agreement between the different methods used, eg. for Study 2.
Gott and colleagues continued their article with a report of their study results. Results from Study 1 about the association between sleep fragmentation and lucid dreaming showed that even though not all questions were answered by all participants, on average the participants remembered their dreams approx. 16 plus or minus 11 times per month. The participants experienced a lucid dream almost once a week (0.31 plus or minus 0.42) with the average number of lucid dreams per month at 2.7 plus or minus 5.7. Most participants rated their sleep as normally continuous (N=78) or quite fragmented (N=51). The majority of participants reported 2 or less awakenings per night (N=147) whereas 46 reported between 2 and 5 and only 5 reported having 6 or more. Therefore, analysis of the results using partial correlation showed no significant association between lucid dreaming frequency and the amount of sleep fragmentation (r=0.05), but a significant association between reported lucid dream frequency and the number of awakenings (r=0.25). The authors concluded that their findings were therefore at this stage inclusive. Looking at the results on an individual basis, the authors found that participants did report increased lucid dream frequency when their sleep was determined as fragmented. The authors gave three possible explanations for this discrepancy. The first possible explanation was the accuracy to which participants determined and reported their own sleep fragmentation, eg. dissatisfaction with sleep leading to the reporting of higher levels of fragmentation relative to number of actual awakenings. The second explanation given was that the types of fragmentation that are likely to produce the poorest sleep quality and subjective dissatisfaction (eg. multiple arousals from light sleep) could be different to those fragmentation types that produce lucidity. The final possible explanation given was that the participants were more likely to remember the next morning when reporting arousals from REM whilst multiple arousals from NREM sleep would be described as a single, continuous period. This would result in an under-reporting of arousals relative to sleep fragmentation, subjective and objective.
Results from Study 2 showed that the duration of the polyphasic sleep rhythm differed between the ´in-house` group (3-44 days, mean 16.7 plus or minus 12.9) and the ´external` group (3-220 days, mean 41.9 plus or minus 68.3). Together, the participants reported 2.64 plus or minus 1.90 dreams per day and 0.5 plus or minus 0.92 lucid dreams per day during the polyphasic sleep rhythm whilst afterwards in the monophasic phase there were reports of only 0.81 plus or minus 1.01 per day of dreams and 0.05 plus or minus 0.09 of lucid dreams. Different statistical analyses produced varying results. The paired t-tests showed significant differences for both absolute and relative lucid dreaming frequency between the polyphasic and monophasic sleep routines whereas a Pearson correlation test did not. Even normalising with the generalised dream frequency did not change the statistical results. The authors interpreted their results as a positive indication that participants with their polyphasic sleep schedules experienced an increased absolute number of lucid dreams as well as an increased proportion of lucid dreams in relation to general dream recall frequency. The authors explained this observation as possibly being due to the more scheduled routine of longer wakefulness preceding an increased number of sleep periods with the polyphasic sleep in comparison to the brief periods of wakefulness that characterise fragmented sleep during normal monophasic sleep patterns. It was also suggested that there is a higher likelihood for REM periods following the wakefulness in the polyphasic sleep routine. Therefore, both factors could increase the probability of a still activated prefrontal cortex (PFC) in the sleep REM episode and hence, the chance of a lucid dream would be increased.
The results from Study 3 showed that the mean global PSQI score measured before the 6 week test period was 4.48 plus or minus 1.72 which indicated that most participants of the study were not clinically defined as having sleep disturbances. The authors carried out a more in-depth analysis of the mean score for the PSQI question 5b (During the past month how often have you had trouble sleeping because you wake up in the middle of the night or early morning? Possible answers – 0 – not during the past month, 1 – less than once a week to 3 – three or more times a week). In this case the mean global PSQI score was found to be decreased to 1.38 plus or minus 0.9 which also indicated that there were few reports of sleep disturbance. The DLQ results gave a mean score of 5.46 plus or minus 6.63. Statistical analysis showed no significant correlation between the nocturnal awakenings as measured by PSQI question 5b and lucid dreaming measured by the DLQ score and no significant correlation between the global PSQI scores and the average DLQ score. This indicated to the authors that there are no correlations between nocturnal awakenings, overall sleep quality and dream lucidity if measured by the PSQI. The suggestion was then made that if the association does exist then it could not be measured by the PSQI tool or that the disturbance is reasonably unlikely to be deleterious according to clinical definitions.
From their Study 4 experiments Gott and colleagues found that their study participants on average experienced 1.17 plus or minus 1.07 wake-REM (WREM) sequences in the chosen time window of last two hours before awakening. There was a significant correlation between the number of WREM sequences and the lucidity factor Insight of the LuCiD. However, there was no correlation between Insight and Number of Awakenings or Total Awake Time. The WREM sequences also did not correlate to any other LuCiD scales such as Realism, Thought and Control although this last factor came close to the threshold. There were also no correlations between the LuCiD scales and either Number of Awakenings or Total Awake Time with the exception of Realism which had a negative correlation to the Number of Awakenings. When the time period was changed to the last 4 hours of sleep, the correlation between WREM and dream lucidity was stronger (3.57 plus or minus 2.29) and the threshold was maximised. Varying the length of the wake-REM sequence to between 1 and 20 minutes did not change the effects substantially although the effect was maximum between 5 and 10 minutes with the highest r value at 10 minutes and the lowest at 2 minutes. The authors concluded that their results indicated a potential correlation between dream lucidity, sleep patterns and physiological processes and that further investigations should involve more optimised experimental set-ups (eg. 4h time period, 10-minute WREM sequences).
Gott and colleagues concluded their article with a discussion of their results. Their results showed that there were slight differences when looking at the connection between sleep fragmentation and lucid dreaming because self-assessed numbers of awakenings, polyphasic sleep and physiologically validated wake-REM sleep transitions were associated with lucid dreaming, but neither self-assessed sleep quality nor physiologically validated number of awakenings were. Their investigations were carried out on the premise that sleep interruptions increase the likelihood of increased activity in the PFC which persists into the REM sleep stage and this is why lucid dreaming occurs. This view is supported by other evidence that sleep onset REM episodes occur in healthy subjects particularly after periods of sleep interruption during the night (Fukuda), during daytime naps (Bishop) and that sleep fragmentation seen in narcolepsy and intentional sleep disruptions that promote lucid dreaming also demonstrate REM periods rapidly following wake periods.
Reasons why the activity in the prefrontal cortex would be sustained during the return to REM sleep were then discussed. A physiological explanation was given as being due to a process of physiological inertia (Hobson and McCarley) with a higher number of wake-REM sequences comparatively maximising the likelihood of this happening (bottom-up model). This explanation is supported by narcolepsy with hypothalamic abnormalities in Orexin projecting neurons predicting dream lucidity (Rak). However, it was said that this explanation was not entirely satisfactory since narcolepsy is also associated with greater than average dream content indicating heightened cognitive function and therefore, further research would have to be carried out to establish a link between higher lucidity and wake-REM sequences in narcolepsy. Therefore, from a physiological perspective it could only be concluded that sleep disruptions may increase the chance of wake-REM sequences.
Gott and colleagues then went on to discuss a psychological top-down model which would explain the narcolepsy observations and other study results such as using alarm snooze buttons, late night gaming and voluntary disruption of sleep. This model suggested that spontaneous metacognitive activity produces lucid dreaming in reported cases and sleep is more likely to be disrupted through ´failed` lucidity attempts that activate and arouse the entire brain as an unintended consequence. In order to validate this view the authors suggested that there has to be a wider understanding of the actual physiology distinct from subjective reports.
The authors then discussed a third possible model which involves both the psychological and physiological models. This model came about because some waking studies showed that diminished cognitive performance from sleep fragmentation and derivation can lead to functional independence of brain areas that are part of the default mode network (DMN), eg. precuneus linked to lucid dreaming, virtual representation and mind wandering. It implies that high level cognitive content is allowed to occur when the physiological processes that constrain it become voluntarily or involuntarily disrupted. This view supports theoretical models of brain function that portray the mind in terms of primary and secondary consciousness processes (based on Edelman – Dresler, Voss, Hobson). It also supports the models describing psychosis (Limosani, Dresler, Scarone) which suggest that sleep-based stressors negatively impact global brain function and cause certain types of brain function to experience heightened activation. However, no definitive conclusion could be made with relation to the physiological, psychological or physiological/psychological models and the authors stated that further investigations of the neurophysiological correlates is required.
Therefore, Gott and colleagues concluded that deleterious sleep does not seem to have an important effect on dream lucidity. They reported differences in subjective and objective findings with dream lucidity being associated with self-reported number of awakenings per night and polyphasic sleep, but not subjective degrees of fragmentation or dissatisfactory sleep even though participants had the subjective impression that dream lucidity had occurred most often in sleep that was quite fragmented. The authors did hypothesise though that increasing the chance of experiencing a lucid dream occurs when there is rapid re-entry into REM sleep from a waking or aroused state (the ´wake-back-to-bed` method, WREM sequence) and that fragmented or polyphasic sleep might potentiate these occurrences. However, Gott and colleagues conclude that although the physiological architecture of sleep might be the cause of the lucid dreaming and may manifest as a consequence of poor sleep there is no requirement for it. This implied that previous thoughts on potential detrimental effects of lucid dreaming on sleep quality or sleep function are unfounded and supports the research of others which report that there is no correlation between poor sleep and lucid dreaming once nightmare content is controlled for. Gott and collegues suggested that their experimental set-up of wake-back-to-bed and unusual sleep scheduling could be used in any future research in order to elucidate the association between lucid dreaming and sleep fragmentation.
COMMENT
What makes this topic interesting is that it suggests that there can be flexibility in the sleep cycle without affecting memory performance. We are used to the idea that detrimental sleep patterns and tiredness can have adverse effects on memory, but Gott and colleagues` article about lucid dreaming indicates that not all changes in the sleep cycle are negative. Lucid dreaming differs from ´normal` dreaming in that the individual can be physiologically asleep while at the same time aware that they are dreaming and can intentionally perform diverse actions (Baird). Individuals can also remember their dreams and actions long enough after waking to report them. These characteristics are in conflict with the usual case of dreaming that occurs in the REM phase of the sleep cycle where individuals are just ´passive` observers and after waking normally forget the content of their dreams within a very short amount of time.
The lack of effect of lucid dreaming on memory performance implies that for some people and some circumstances, the explanations of why we have sleep in the physiological form that it is are not strictly conclusive. The general theories for why humans must sleep range from psychological reasons, eg. allows the ´mind` to rest (eg. Naiton, Berry, Webb) – not seen with lucid dreaming since the individual is aware during the lucid dreaming phase; evolutionary ecology reasons (Meldis) – not particularly relevant since only humans lucid dream; and for this blog in particular the physiological/neurobiological reasons, eg. Oswald restoration theory (Lawton, Alison, van der Waaf) – restoration of the brain`s physiological structure to allow the most efficient working and performance levels – not completely valid since there are differences in brain area firing and connectivity in lucid dreaming than normal REM dreaming. Memory function itself is described as particularly dependent on the sleep cycle in the physiological form that it is and there is plenty of evidence that tiredness or sleep deprivation have deleterious effects on memory and other cognitive capabilities performances. We know that there is a need for the sleep cycle in various physiological aspects of neural representation, memory storage and consolidation. For example: the whittling down of neuronal connections formed during waking hours to make room for new memories in the following waking period so that only the strongest firing and connections survive (Tononi); the stabilisation and consolidation of memories (Yoo) which may include the shift from the hippocampus firing to long term memory storage in cortical areas (Mehti, Ji) and the reactivation of firing patterns in REM to aid memory consolidation and stabilisation against interference (Deuker, Oudiette); as well as basic neuronal housekeeping functions such as the repair of damage from free radicals which occurs in the NREM stage (Lawton). Sleep also appears to influence memory content with for example, sleep preferentially favouring the consolidation of memories that are relevant for future behaviour (Wilhelm); and ´low-value` associations demonstrate lower recall accuracy following a 90 minute nap than ´high value` associations (Oudiette). In addition, sleep also appears to affect performance, eg. good sleepers perform prospective memory tasks better (Fabbrio) and sleep directly after training stabilises motor sequence consolidation (Nettersheim). However, lucid dreaming, which alters some of the physiological aspects of the sleep cycle does not appear to affect memory performance and therefore, those aspects of the sleep cycle that are changed in lucid dreaming appear to have some ´play` or flexibility in their physiological structural make-up and function without having a consequential effect on long-term memory.
If we look at the sleep cycle we can see where these areas of flexibility relating to lucid dreaming exist. A quick summary of the sleep cycle is that it consists of 5 (although some report 4 with one less NREM stage) stages which demonstrate definitive patterns of brain area activation, neuronal firing and brain waves. Stages 1-4 are the NREM stages followed by the REM dreaming stage (Stage 5). The total duration of the sleep cycle is around 90 minutes (BRAC rhythms – Lloyd) with each stage lasting varying lengths of time (2 mins /5-15 mins/ 5 mins/ 20-40 mins/ 5-15mins).
The NREM stages take up the most time in the sleep cycle and are important in general for neuronal ´housekeeping` such as replenishing the synapse, inserting the receptors responsible for reception and plasticity. Each stage is associated with specific brain waves and firing patterns that make them identifiable. Stage 1 is the period of transition from wake to sleep and can be easily disrupted. This stage is identified by the relatively unsynchronised beta and gamma waves (frequency 12-30 and 25-100) normal for the wake state shifting to the more synchronized, but slower alpha (8-13) and then to theta (4-7HZ). The theta waves are indicative of Stage 2 where conscious awareness fades. Stage 2 serves to protect sleep and suppresses responses to externally sourced stimuli. It also aids in the consolidation of memory and information processing. This stage demonstrates distinctive neuronal firing such as occasional sleep spindles (also known as sigma waves) and high amplitude K complex firing. The sleep spindles of 8-14Hz are generated by the thalamic pacemaker and last about half a second. They show subject specific activity and can be separated into slow and fast components (Cox) whereas the high amplitude K complex are theta waves (4-7Hz) with short negative high voltage peaks followed by a slower positive complex and then a final negative peak with each complex lasting 1-2 mins.
Stage 2 NREM is followed by Stage 3 (and Stage 4 by the old nomenclature) known as deep, delta or slow wave sleep (SWS) because this stage is typified by slow, large amplitude firing (less than 1HZ). Stage 3 occurs in longer periods during the first half of the night and makes up 15-20% of the total sleep time. These stages are characterized by delta brain waves (0.5 – 4HZ), which in Stage 4 is greater than 50% of the total and in these stages, the individual is unresponsive to the outside environment and is unaware of any sounds or stimuli.
One type of specific firing is also seen in these early NREM stages and these are the sharp wave ripples (150-250HZ). This type of firing is observed in hippocampus to cortex connectivity and hippocampus to amygdala connectivity and are particularly associated with memory.
Therefore, in terms of memory according to the ´dual process hypothesis` these NREM stages are responsible for non-declarative, hippocampus dependent memory including emotional memory and also because of sleep spindle activity in consolidation of the declarative memory type, motor memory. The NREM stages play roles in the reactivation of memory information (ie. replay) in the hippocampal-cortical circuits and specific connectivity between appropriate brain areas. In the case of context dependent emotional memory (eg. spatial memory combined with an aversive component) the specific connectivity between the hippocampus and basolateral amygdala involves SWR (Girardeau). Arousal and consciousness appear to involve the feedforward thalamocortical connectivity of the NREM stages (Herrera) and memory consolidation and re-organisation of prefrontal cortical networks is associated with hippocampus and cortical connectivity (Maingret). Episodic memory and consolidation also appear to be associated with SWR where it is suggested that neural representation replay and SWR are tightly interconnected (Jahnke). Other types of firing, eg. the sleep spindles and high amplitude K complex, both observed in Stage 2 are also associated with memory performance. For example, fast type sleep spindles in the task-related motor region leads to predicted overnight enhancement of learnt procedural skills (Tamminen) and motor skills consolidation (Walker) as well as improved word learning skills (with sleep spindles -Tamminen; naps with increased spindles, but no change in motor learning – Mednick). Aiding memory consolidation also appears to be the function of high amplitude K complex (theta) firing which serves to protect the sleep state and suppress responses to outside stimuli. Specific firing for Stage 3 is also important to the provision of conditions for effective memory storage and consolidation. Slow waves seen in Stage 3 appear to aid coordination of interregional cortical communication (Cox) and memory consolidation (SW and the thalamocortical network- Wei; SW modulation of strength and precision of memory seen with odour perception and post training sleep – Barnes; and SW and sleep spindles in word learning – Tamminen).
Therefore, brain memory functions are reliant on a number of physiological processes, eg. connectivity, brain waves, distinctive firing during the NREM stages and when this ´balance` is correct then performance is at its most effective. This is demonstrated in the following examples. The first where hippocampus connectivity to the cortex which is required for memory consolidation (Maingret). The timed electrical stimulation leads to reinforcement of connectivity between the two areas and relies on sharp wave ripples, cortical delta waves and sleep spindles. Such firing and connectivity are associated with beneficial re-organisation of the prefrontal cortical networks (Maingret) and increased PFC responsiveness to tasks and high recall performance the day following sleep. It is also seen where there is connectivity between the neocortex and hippocampus-entorhinal complex with sleep spindles and theta oscillations in the hippocampus (Sullivan). This involves phase locking sleep spindles of the CA1 to spiking of putative principal cells and interneurons in the CA1, CA3, DG and specific layers of the EC. It is also the case with other firing types with connectivity between the hippocampus and basolateral amygdala being associated with consolidation of context dependent emotional memory, but characterized by SWR (Giradeau).
However, having described memory dependence on NREM sleep functioning we have to say that there appears to be no association between correctly identified lucid dreaming and the NREM stages per se. That does not mean that dreams have not been reported for the NREM stages but definitive identification is difficult. Some individuals have reported dreams or ´sleep mentations` when waking up from NREM sleep in 70% of awakenings (Baird) although these NREM ´dreams` are less emotional, less visually vivid and more ´thought-like` than the reported REM lucid dreams. There are also reports of ´lucid dreams` in NREM stages of going from wake to sleep (Stage 1) and in consolidated light sleep (Stage 2), but these ´lucid dreams` are short and appear phenomenologically different to the REM based ones and more importantly are not normally associated with the identifying left-right-left-right (LRLR) eye signals. Also, some yoga practitioners report ´lucid dreams` in all sleep stages, but again these cannot be verified. Therefore, these reported ´dreams` cannot be given as definitive proof of lucid dream activity in the NREM stages and so, it can be assumed that the NREM stages in lucid dreamers exhibit the same functions, brain area connectivity and firing as non-lucid dreamers. This would imply that long-term memory requirements that involve NREM stages do not differ between non-lucid dreamers and lucid dreamers and that the physiological processes being undertaken are the same.
There is, however, one difference at the end of the NREM stages or the transition between NREM and the next stage, REM between the lucid dreamers and the non-lucid dreamers and that is that the non-lucid dreamer is in a status where he has no conscious awareness for internal thought nor conscious awareness of his environment. Therefore, this means a shift and a re-awakening of conscious awareness in the following REM stage of the lucid dreamer in contrast to the non-lucid dreamer. Before we look at the physiological basis of the REM lucid dreaming we will consider what this means. There are two possibilities regarding consciousness and lucid dreamers. The first is that lucid dreamers come to the REM stage in the non-conscious state (like non-lucid dreamers) and in the REM stage as they slip into their lucid dream shift to the conscious state by some physiological change. The second option is that all through the NREM stages the lucid dreamer maintains a form of low-level consciousness which shifts to a higher level in the REM stage and their lucid dream. Although there is some talk about primary and secondary levels of consciousness (Dresler, Voss), what actually goes on from a neurochemical basis and here with relevance with memory is much more complicated and we should consider that there is no global consciousness state, just areas of ´non-conscious` processing and an area of ´conscious` processing. Therefore, since there appears to be no differences in the NREM stages between the non-lucid dreamers and the lucid dreamers it is more likely that the second option is valid and the lucid dreamer shifts from non-conscious processing to conscious awareness representing the lucid dream whereas the non-lucid dreamer remains in the non-conscious processing state for the REM stage.
For both lucid dreamers and non-lucid dreamers the REM stage of the sleep cycle is said to be important for memory function, but differences in the physiology of the REM stages between the two must mean that there is flexibility within it to allow no deleterious effects to be observed. In non-lucid dreamers, the REM stage is the stage where individuals experience dreams and this stage is characterized by its fast beta brain waves. This type of brain wave is in non-sleep conditions associated with attention particularly alerting with feedforward firing to cortical subregions via the prefrontal cortex (Tingley) and therefore, it is of no surprise as being reported as being involved in motor movements and the motor cortex (Moisa) and reward expectancy and hippocampal-ventral striatal connectivity (Lansink). It is also therefore, of no surprise when insomniacs exhibit fast gamma and beta waves caused by the motor cortex activity – a situation that can be alleviated by tDCS or TMS (Finkbeiner). In sleep, however, beta waves are also seen unsynchronized in the NREM Stage 2 as conscious awareness fades and are part of the brain wave ´mix` observed with unsynchronized gamma, occasional sleep spindles and high amplitude K complex (theta). This brain wave ´mix` is part of the connectivity and integration of brain areas into cortical networks that aid the consolidation of brain memories (Cox).
In the REM stage, it can be assumed then that the beta waves observed are associated with the creation of the ´dreams` which are activated neuronal assemblies leading to neural representations made up of characteristics from real experienced events that are stored and recalled (dreams are described as having the same visual quality as waking perception rather than imagination – Nir) and reactivated random characteristics from a number of events that are presented like a mosaic from which the individual ´weaves` a ´story` from his first-person perspective. Bearing in mind then what is going on in the brain, in general, REM sleep in non-lucid dreamers has similar properties to the waking state, eg. beta brain waves are again associated with the connectivity of the firing networks, there are strong theta oscillations and similar acetylcholine levels. However, there are differences: there are lower levels of norepinephrine and 5HT; some connectivity is different such as the CA1 of the hippocampus receives inputs from the CA3 and entorhinal cortex, but the CA1 pyramidal cells responds more effectively to CA3 during waking and to the EC input during REM sleep; also there is reduced coherence in the oscillating networks (including theta) between the hippocampal-prefrontal and thalamo-cortical circuits during REM sleep than in both NREM sleep stages and waking. This suggests that the connectivity between the hippocampal and prefrontal regions decouples in REM sleep. This would explain the disconnect between bottom-up processing, spontaneous type firing, lack of conscious awareness from stimulus from the external environment and top-down memory reactivation, imagery type firing and controlled input. It also implies that the ´dreams` in non-lucid dreamers are the result of ´lower level` brain area activation removed from the ´control` of the upper level cortical areas, eg. prefrontal cortex (PFC), reasoning and decision-making areas. This is supported by the association of REM sleep with synaptic homeostatic regulation (lower level hypothalamus, thalamus activities for example and specifically for procedural type memory (eg. memory reactivation carried out without conscious awareness such as motor movements) and emotional aspects of memory (REM disturbance contribute to poor daytime functioning in PTSD sufferers – Ogeil) which are reliant on the lower level brain areas such as the amygdala (Genzel). The situation may not be a case of a clear ´cut-off` between lower level and higher level, but a change in connectivity patterns. For example, there is strong activation reported in the amygdala, PFC and hippocampus in human REM sleep (Nir), but bidirectional changes in fear memory are selectively correlated with changes of theta connectivity between the amygdala, PFC and hippocampus during REM sleep. The same can be seen with spatial memory with theta wave disruption in REM sleep being associated with impairment of hippocampal dependent spatial memory (Varga, Boyce).
Therefore, referring to the ´dual processing hypothesis` NREM stages of non-lucid dreamers involve higher and lower levels of the brain with cortical type influence and conscious awareness diminishing and REM, with the lower levels having more influence and with no conscious awareness. However, this is not the case with lucid dreamers who report dreams in which they have conscious awareness and can perform designated actions. The differences in capabilities that the lucid dreamers have compared to non-lucid dreamers obviously must come from the physiological systems employed in the sleep cycle in the REM stage and there are two areas showing differences.
If we look at brain waves, we see that lucid dreaming is associated with increased beta brain waves in this REM phase. The increase appears to be restricted to the parietal areas and is attributed to the understanding of the realization by the lucid dreamer that they are in a ´dream` (use of the words – ´this is a dream`), the first person perspective (self-reflection, agency) as well as episodic memory recall. Although beta waves are also seen in the REM sleep of non-lucid dreamers as described above, this type of brain wave is associated with connectivity of the hippocampus to the PFC as well as feedforward firing of the thalamocortical areas and therefore, there is a difference between beta connectivity observed in the REM of non-lucid dreamers to that of lucid dreamers. This difference is supported by other brain wave oscillations and all point to an increased activity of the higher brain areas in lucid dreamers. The REM of lucid dreamers is associated with increased gamma brain waves in the brain`s temporal lobes and frontal lobes (Voss, Thomson) which is indicative of the waking state and higher level of control, eg. PFC. There are also reported reduced levels of delta pointing to upregulation of the delta-linked neuronal down states (Baird) and theta brain waves (Voss) with increased levels of coherence in these states, the largest of which are in the frontal and fronto-lateral areas. Although reports of alpha brain waves in lucid dreamers (Ogilvie) appears to be associated with a ´pre-lucid` state (thoughts pertaining to dreaming without becoming lucid), differences in EEG electrode placement (LaBerge) and individuality means that no reliable association between alpha brain waves and lucid dreaming can be made.
The brain wave oscillations give an indication of the neuronal firing connected networks taking place in the REM sleep phase and connectivity is also reported as being different in lucid dreamers in comparison to the non-lucid dreamers indicative of the higher cognitive areas alight. Whereas there is a decoupling between hippocampus and PFC in non-lucid dreamers, an increase in dorsolateral PFC activity, an area responsible in waking states for central executive functioning and reasoning (D`Esposito) is observed in lucid dreamers. There is also greater activity in the frontopolar cortices (eg anterior PFC responsible for metacognition and self-reflection – evaluation of one`s own thoughts and feelings). The connectivity and activity of the anterior PFC (aPFC – Brodmann area 10) is also found to be increased also in relation to the superior frontal gyrus, inferior/middle temporal gyri and occipital cortex (responsible for the ventral stream of visual processing involved in conscious visual perception), areas normally seen as reduced in the REM of non-lucid dreamers. Also, some see higher levels of connectivity between the aPFC and a network of areas involved in the frontoparietal control sub-network (Dixon) whereas most report high parietal activity (medial and lateral) eg. somatosensory cortex activity and precuneus (areas responsible for self-referential processing, episodic memory and experience of agency – Cavanna). In some areas, larger volumes are reported instead in lucid dreamers with increased gray matter being reported in the right anterior cingulate cortex, left supplementary motor area and bilateral hippocampus as well as other frontal pole areas (Baird). However, there is reported decreased functional connectivity between the left aPFC and bilateral insula (an area responsible for emotional feelings expression). Therefore, lucid dreaming is said to be associated with increased activity of the higher level brain areas whereas in non-lucid dreaming there is decoupling between the higher and lower areas.
If we assume that lucid dreamers maintain the activated states of the prefrontal and parietal areas and have conscious awareness, we have to ask why does it occur and does it have cognitive consequences? The first thing to say is the obvious observation that lucid dreams are likely to occur in a ´variation of REM`, not physiologically identical to the REM of non-lucid dreamers. Since spontaneous lucid dreaming occurs very infrequently in the general population, situations or conditions arise that make lucid dreams likely to occur, eg. the natural ´mental/emotional state of the individual, individual physiology favouring the SELF, natural sleep disruption or sleep fragmentation as seen as in conditions such as narcolepsy for example and even with some recreational drug use. The connection between sleep fragmentation and disruption is used to the advantage in experimental studies (eg. Gott and colleagues study) with researchers using the ´wake-back-to-bed` method of rapid alternating sequences of wake and sleep periods for their experimentation. This method shows an increased incidence of lucid dreaming (LaBerge), but lucid dreaming is also seen in a more natural setting of extended periods of wakefulness during the morning hours leading to a greater frequency of lucid dreaming in the subsequent sleep periods (LaBerge).
The correlation between sleep fragmentation and lucid dreaming is said to come from sleep fragmentation and short naps leading to a ´hybrid state` between wakefulness and REM sleep (Voss). This is a slightly different interpretation to our ´variation in REM` since this refers to a change in REM frequency and duration rather than a ´physiological change` in REM itself. The ´hybrid state` model is supported by sleep onset REM episodes with lucid dreaming occurring in healthy subjects particularly after periods of sleep interruption during the night (Fukuda), during daytime naps (Bishop), sleep fragmentation in narcolepsy and other intentional sleep disruptions in REM periods rapidly following wake periods and in recreational drug use where REM is suppressed but then goes through a re-bound (Roehrs). Our ´variation in REM` view instead is from a physiological perspective and relates to the brain wave and connectivity differences described above for lucid dreamers compared to non-lucid dreamers.
One view in this vein is the ´physiological model` described by Baird. This model describes the reason for lucid dreaming to be due to the activity in the NREM stages observed in the PFC being maintained through REM because of ´a process of inertia` (Hobson) with a higher number of wake-REM cycles reducing the likelihood of neural change. This implies that the excitatory firing of the higher brain levels is not extinguished and this is due to the failings of the ´switch-off` mechanism which takes the individual either back into the NREM stages as part of the BRAC rhythm or promotes awakening. There is support for this from the combination of the ´physiological and psychological` models suggested by Baird and others. For this combination model, the psychological model means top-down ´direction` where there is spontaneous activity of the higher levels producing the lucid dream due to sleep fragmentation. Disruption of sleep is described as causing ´failed lucidity` meaning that the entire brain is aroused and activated. It is uncertain as to how this works since the limiting factor is the activity in these higher brain areas and the model says that there is not enough to produce lucidity then the entire brain aroused and activated. If this is the case, then by other models this condition is likely to produce lucid dreaming since activation of the higher brain levels is not seen in this way in non-lucid dreamers. The ´physiological` side of the combination model is more in line with what brain waves and brain connectivity shows. Sleep fragmentation can lead to functional independence of areas that are part of the default mode network (DMN) which means that the bottom-up constraints on firing of higher levels are removed leading to higher levels of activation, beta brain waves and lucid dreaming occurrences, eg. precuneus firing linked to lucid dreaming, virtual representation and mind-wandering. Functional independence of the lower levels and DMN is possible and is likely through the normal hormonal and neurotransmitter influences of the brain stem and hypothalamus. For example via, the specific sleep centre in the hypothalamus and the specific wakefulness centre in the reticular activating system, brain stem plus lots of other areas. A switch from waking to sleep involves particularly the ventrolateral preoptic nucleus (VLPO) of the hypothalamus where neurons promote sleep by inhibition of activity in areas of the brainstem that maintain wakefulness, but is itself in waking hours inhibited by the active cortical areas of the brain and therefore, this one area where decoupling from the lower areas could occur. Further investigation would clarify this.
So, what do we think is happening? It is clear that lucid dreams occur in the REM stage of the sleep cycle and either occur naturally for some people or can be induced in some by disruption to the normal sleep cycle. In the case of lucid dreamers, the disruption of the sleep cycle leads to a maintenance of increased activity in the higher brain areas such as the PFC (such as dorsolateral), frontopolar cortices (aPFC) and parietal areas with a connectivity associated with beta, gamma and theta waves. This higher activated connectivity is linked to conscious awareness of the lucid dreamer, but not associated with language ability (the lucid dreamer is incapable of speech during the dream and reports the experience once awake). However, limited motor movements are possible as demonstrated by the LRLR eye signal capability used as experimental proof of lucid dreaming occurrence.
As far as the sleep cycle goes and disruption to it, the lucid dreamer probably is unable to exhibit the full BRACS rhythms in cyclical form and the REM stage lengths may be shorter than the non-lucid dreamer. This is seen in other cases of sleep disruption where in subsequent sleep episodes, the REM failing is remedied. In the lucid dreamer this shortened REM phase may be the reason why the likelihood of the lucid dream occurrence is increased and this brings on to another explanation of why increased activation of the higher brain areas is observed. Increased activation may occur because the individual knows that his sleep pattern is disturbed and hence, in the wake-back-to-bed method (or even unintentional methods with recreational drugs, anxiety etc.) ´alertness` must be maintained, eg. the individual knows he has a limited time to sleep, waits constantly for the alarm clock and knows that a report is required following the sleep. This type of ´alertness` is comparable to that seen with prospective memory where the cue for future action is stored and the individual performs unconscious processing of it until the cue is encountered in the external or internal environment. At this point the ´cue/idea` pops into the individual`s head and there is a shift from unconscious to conscious processing for a brief period of time with an inability to switch it off.
The link between lucid dreaming and prospective memory is not new with a proposal that lucid dreaming is accompanied by increased prospective memory capability (Baird) and the response to acetylcholine esterase inhibitors which is a core method for lucid dream induction (LaBerge). In both cases, the assumption is that the association to prospective memory comes from the individual having to remember to know that he is dreaming (Baird). From a neuronal perspective the association is stronger with prospective memory sharing similar characteristics to lucid dreaming. For example, prospective memory is top-down (McDaniel), there is conscious awareness as the cue hits and a similarity between activated brain areas. This includes multiple areas of the PFC (including the dorsolateral PFC) which in prospective memory is required for both event- and time-based memory and with the OFC (Brodmann area 11) being involved specifically with the projection of SELF into the future. Activation of the frontopolar cortices (Brodmann area 10) is also required for both lucid dreaming and prospective memory with deficits in this area associated with failure to follow instructions and to switch attention during tasks (Burgess). Frontal lobe activity is also required for both with activity associated with imagining (right side) and remembering future recall (left side) and medial areas keeping attention focused on the planned actions instead of other tasks. Parietal lobe activity is also required with links more to attentional monitoring (frontal gyri – temporal monitoring during time based prospective memory tasks- Harrington) rather than associations to visual processing and recognition of cues (not valid because of the sleep cycle phase). Prospective memory also shows a role for the cingulate cortex with lesions of the left leading to failure to recall (McCall). There appears not to be a specific role for the cingulate in lucid dreaming, but the area volume in frequent lucid dreamers is greater. In prospective memory, activity in this area provides an error signal for prospective memory failure which is not particularly applicable here in lucid dreaming where there is no ´error` just ´dream content`. However, the larger volumes in lucid dreaming could be linked to the disengagement of the ´error` signal in REM sleep in response to the abnormal cortical activation.
As indicated not all brain areas involved in prospective memory are also activated in lucid dreaming and this can be explained in some cases to the different conditions, eg. no incoming visual signal (such as V4 where visual cues match remembered required stimulus – Hayden), no attentional monitoring of external stimulus (Rampey, Rummel). Others also show discrepancy between the two with depletion of executive function having no effect on prospective memory performance (Cook), but is involved strongly in lucid dreaming, and the thalamus which is activated when intentional cues are prevented in prospective memory, but in lucid dreaming is not really linked to thalamus activity whereas REM in non-lucid dreamers is dependent on thalamo-cortical connectivity. The hippocampus is also strongly activated in prospective memory with links to memory recall searching for intended actions in the memory database and imagination of future events. It is also seen in non-lucid dreamers REM phases with decoupled connectivity to the PFC, whereas in lucid dreamers larger volumes are observed, but no distinctive hippocampal activity or connectivity. Hence, the larger volume may relate to general functions of the REM phase and general ´dream content` rather than specifically for lucid dreaming. Therefore, although a neurochemical similarity between prospective memory and lucid dreaming may exist it is not a definitive explanation for the physiological differences observed between lucid dreamers and non-lucid dreamers.
What this investigation shows is that lucid dreaming is likely to be a consequence of increased activation of higher brain areas, but does not cause it and that the REM phase in lucid dreamers is a ´variation` of the REM phase seen in non-lucid dreamers. The consequence of lucid dreaming is that it appears to have no effect on brain memory function and that the different physiology in the REM stage and conscious awareness of dreams and dream content have no effect on long-term memory consolidation and recall. This implies that that the effects required for successful long-term memory formation and consolidation occurs in the NREM stages or that there is no difference in the replay in the REM stage of both non-lucid dreamers and lucid dreamers independent of the physiological systems in play. Therefore, lucid dreaming does not follow the example of sleep deprivation where long term spatial memory is found to be impaired. To explain this difference, it could be suggested that the brain on permanent alert in the lucid dreamer with no ´shut down` leads to cognitive random firing, normal memory recall and reconsolidation repetition processes and therefore, the overall activity is still positive even if the dreams (the activated neural representations) are themselves bizarre. It is also possible that lucid dreamers compensate for their REM disruptions in the other earlier NREM stages and this is supported by increased sleep spindle compensation in Stages 2 and 3 and positive changes in the firing rate of the hippocampus (Mednick). Naps (equivalent to the wake-back-to-bed scenario) are found to be associated with increased spindles and significantly better verbal memory, perceptual learning deficits and no effect on motor learning (Mednick). Therefore, any effect on memory may be hidden if this compensation mechanism takes place. This may be possible, but unlikely since no differences in NREM are observed between non-lucid dreamers and dreamers.
We can conclude this investigation on a positive note that although we are prone to think that changes to sleep cycles are detrimental to memory, this view is not valid in all cases. Lucid dreaming with its ´variation of the REM` does not have negative effects on memory consolidation and replay even though higher brain area activation is maintained and there is conscious awareness when normally there is none. This implies that there is some flexibility in the physiological processes and compensatory or alternative mechanisms must exist even if not identified.
Since we`re talking about the topic……
….if Gott and colleagues experiments were repeated with the awareness signals being both the LRLR eye signal and a hand clench and imaging of the activities of brain areas measured would activities in the somatosensory cortex areas and motor cortical areas be observed as expected?
….if sleep spindle density can be increased by the administration of a short acting GABA A agonist such as zolpidem, but the amount of REM sleep decreased (Mednick), can we assume that in this case wake-back-to-bed experiments would lead to a decreased level of lucid dreams in lucid dreamers?
…. 5HT promotes wakefulness with decreasing REM. Therefore, can we assume that administration of a 5HT agonist could increase lucid dreaming in people susceptible to lucid dreaming when they finally are asleep? Would administration of orexin which regulates wakefulness, low levels of sleep and sleep instability with many short awakenings create the same conditions as the wake-back-to-bed experimental method and therefore, an increase in lucid dreaming frequency would be observed in those susceptible to this type of dream?
neurochitchat 