Why isn’t a lucid dream just a dream within a dream? Suppose I’m having a flying dream and I think, “I must be dreaming.” I’m in a dream state, so why I am not just dreaming that I’m dreaming? To put the question another way, if there’s a difference between knowing you’re dreaming and dreaming you’re dreaming, then what exactly is it?
Thirty years ago Daniel Dennett posed the following challenge:
“If someone gives a waking report of the lucid dream variety, there will be two hypotheses consistent with the report; the subject had, as she believes she had, a lucid dream; or the subject had an ordinary dream in which she was aware she was dreaming, decided to fly in her dream, etc. She wasn’t really aware she was dreaming, of course; she just dreamt she was aware she was dreaming. So subjective testimony . . . cannot establish that lucid dreams are anything other than a variety of ordinary dreams, viz., dreams of having lucid dreams” (Dennett 1979: 316).
Today two things can be said to meet this challenge. The first comes from phenomenology and the second comes from neuroscience. First, knowing you’re dreaming and dreaming you’re dreaming don’t feel the same and seem different to memory when you wake up. The reason, as we’ll see, is that knowing you’re dreaming involves a certain kind of attention that’s missing when you dream you’re dreaming. Second, we now have more evidence for lucid dreaming than just what people tell us when they wake up; we also have physiological and brain-imaging verifications of the lucid dream state.
Let’s begin with dream phenomenology. Normal waking consciousness is ego-structured, conditioned by the appearance of a sharp difference between a bounded self and an outside world. Dreams typically re-create this structure, for often one has a dream body in which one participates in the dream world. Even when one instead experiences oneself as an observational point of view, one still experiences oneself as a subject situated in relation to the dreamscape (see Jennifer Windt’s earlier posts at this blog). In this way, the phenomenal self/other structure characterizes the dream state.
The hypnagogic state, however, doesn’t fit this structure. There’s no hypnagogic world; the images occur on their own, one following another, with no ties to anything beyond them. A cow dissolves into the scattering of a flock of birds that transforms into a woman’s face, but none of these images participates in any larger universe. At the same time, there’s no hypnagogic ego in the sense of an I who acts as a participant within what’s being imagined. I can affect the images while watching them, but I don’t experience myself as immersed within the scene—if I do, then I’ve entered into full-blown dreaming (Windt 2010).
In the hypnagogic state, we look at visual patterns and they absorb us. When we dream, we experience being in the dream; more precisely, we experience being in the dream world. The experience of being a self in the world, which marks the waking state but diminishes in the hypnagogic state, reappears in dreams.
The great lucid dream pioneer, the Marquis d’Hervey de Saint-Denys, claimed he could observe, without any break in consciousness, how hypnagogic imagery turns into full-blown dreams. Here’s one of his dream reports:
“I close my eyes to go to sleep, thinking of some objects I noticed this evening in a shop in the rue de Rivoli. I remember the arcades of the street in question, and I catch the glimpse of something like luminous arcades forming repeatedly in the distance. Soon a serpent covered with phosphorescent scales appears before my mind’s eye, surrounded by innumerable ill-defined images. I am still in the period when things are confused. The images fade and re-form very quickly. The long fiery serpent has turned into a long dusty road burning under the summer sun. I immediately see myself travelling along it, and my memories of Spain are revived. I talk to a muleteer carrying a manta (blanket) on his shoulder; I hear the bells of the mules; I listen as he tells me a story. The countryside matches the central figure; at this moment the transition from waking to sleep is complete. I am completely taken up in the illusion of a clearly defined dream. I was offering the muleteer a knife, which he seemed to like, in exchange for a fine antique medallion he showed me when I was suddenly brought out of my sleep by an external cause. I had been asleep for some ten minutes, as far as the person who woke me could tell” (quoted in Mavromatis 1987: 91-92).
Here we see the transition from looking at hypnagogic imagery to being immersed in the dream world, with the third-person or observer view of the dream ego as an intermediate state between the two.
The hypnagogic-imagery technique for bringing about a lucid dream works with this kind of transition by taking advantage of the fact that waking awareness continues for a while in the hypnagogic state. The technique aims to lead you from the hypnagogic state into the dream state without losing the meta-awareness of your own conscious experience. Adopting the perspective of a detached observer, you watch the sleep-onset images arise, transform, and subside, without trying to control or hold onto them. In the words of lucid dream researcher Stephen LaBerge:
“When the imagery becomes a moving, vivid scenario, you should allow yourself to be passively drawn into the dream world. Do not try to actively enter the dream scene, but instead continue to take a detached interest in the imagery. Let your involvement with what is happening draw you into the dream. But be careful of too much involvement and too little attention. Don’t forget that you are dreaming now!” (LaBerge and Rheingold 1980: 90).
The experience of falling asleep is a human birthright, so it’s not surprising that many individuals from different times and cultures have hit upon this method of carrying meta-awareness into the dream state. Besides de Saint-Denys and LaBerge, Russian philosopher P. D. Ouspensky (1848–1947), American psychiatrist Nathan Rapport, and German lucid dream researcher Paul Tholey used this technique for lucid dreaming. But an older and more practiced form belongs to Tibetan Buddhist dream yoga, as described in a seventeenth-century Tibetan text:
“Before you fall into a deep sleep, there are so-called thoughts between falling asleep and dreaming. Before you actually fall asleep and you are still in the process of falling asleep, thoughts arise and sounds are faintly heard. You have a sense of the body’s becoming very torpid and a sense of becoming pressed into darkness. You also have a sense of the experience of deep breathing as you begin to relax. Right after that, there is a sensation of numbness at the point midway between the eyes. At that time, you will begin to feel vague impressions of people, animals, environments, or whatever your recent mental impressions are. These vague mental impressions are the cause for the dream. The dream you will have actually arises as the result of those impressions. If you recognize this, it is your chance to recognize the dream, like threading a needle right through the eye, and you will immediately enter the dream and apprehend it” (Gyatrul Rinpoche 1993: 109).
What makes this technique difficult is the spellbound quality of attention in the hypnagogic state. Too much absorption and you’re no longer able to direct your attention to the dreamlike quality of the images.
Here’s the crucial phenomenological point: The state of dreaming you’re dreaming lacks the mental clarity of being able to observe and direct your attention to your dreaming. As Andreas Mavromatis notes in his book Hypnagogia: “The experience of being conscious in one’s dream has a very different status from that of dreaming that one is conscious in one’s dream” (Mavromatis 1987: 106). The difference has to do precisely with attention and meta-awareness: to be conscious of your dream requires that the dreamlike quality of the state be directly available to your attention. Only if you can observe or witness the dream state this way can you know clearly that you’re dreaming.
In a strong lucid dream, you don’t merely think, “I’m dreaming.” You also directly experience your state as a dream by attending to the dreamlike quality of your state. This onset of lucidity typically increases the state’s sensory vividness without diminishing the clarity of the state awareness (the clear comprehension of being in the dream state). Finally, you don’t suffer from the cognitive deficiencies typical of nonlucid dreams (disorientation, poor working memory, inability to direct and sustain attention) but you have insight into the nature of your ongoing conscious state. In these ways, you now experience yourself as a dreaming subject. Your sense of self shifts as you experience the dream from the perspective of the dreamer instead of the perspective of the dream ego or the self within the dream.
This shift in your sense of self doesn’t necessarily mean you’re able to control the dream in the sense of actively and deliberately guiding the dream content. Such control comes in degrees and may be more or less present in any given lucid dream. So what defines a dream as being a lucid dream isn’t being able to control the dream content, but rather being able to direct your attention to the dream state as such. Of course, being able to direct your attention to the dreamlike quality of your state is a kind of cognitive control, and it does affect the dream content because what you dream depends on how you direct your attention in the dream. But one of the things that makes lucid dreaming so compelling is precisely the inability to control many dream features while recognizing that you’re dreaming them.
Critical readers will rightly point out that these phenomenological observations still aren’t enough to deflect Dennett’s challenge. For one thing, the observations are based on subjective testimony alone. For another, they’re based on waking recall, and so strictly speaking count as reports of the waking memories of lucid dreams.
But what if the lucid dreamer could signal to the outside world that she’s having a lucid dream? What if she could signal the presence of certain features or qualities of her lucid dream while she’s dreaming? And what if these signals could be correlated with behavioral and physiological measurements in the sleep lab?
One of the striking scientific findings about REM sleep—the finding that made possible the physiological verification of lucid dreaming—is that eye movements of the dream ego often correspond to physical eye movements of the dreamer (Dement and Kleitman 1957). The direction of your eye gaze within the dream often matches the direction of your eye movements during REM sleep: when your dream ego looks left, your physical eyes move left; when your dream ego looks right, your eyes move right.
In the late 1970s and early 1980s, two dream researchers—Keith Hearne in England and Stephen LaBerge in the United States—independently took advantage of this correlation in order to verify lucid dreaming in the sleep lab (Hearne 1978; LaBerge et al. 1981a, 1981b). They reasoned that if lucid dreamers can act voluntarily in the dream state, then they should be able to indicate when they become lucid by making prearranged eye movements within the dream that would be externally observable on a polygraph. Lucid dreamers were instructed to look right and left a specific number of times, and their eye movements were recorded. When they woke up, the subjects reported how many deliberate eye-movement signals they had made, and the number matched the number of recorded eye movements on the polygraph (see the figure below from LaBerge 2000).
These ingenious experiments were momentous. They showed that the subjects were genuinely asleep when they signaled their lucid dreams; they enabled subjects to report on their dreams while the dreams were happening, instead of just retrospectively; they effected a kind of transworld communication from the private dream world to the public waking world; and they opened up a new way to investigate consciousness whose prospects we’ve barely begun to tap.
Before these experiments, many scientists believed that lucid dreams were most likely experiences that happened during brief moments of awakening but were then later misremembered as having occurred during sleep. When the subjects signaled the onset of a lucid dream, however, the EEG showed brain waves indicating REM sleep. So, according to standard sleep-science criteria, the subjects were unequivocally asleep. This kind of lucid dream in the sleep lab is now known as a “signal-verified lucid dream.”
Further analysis by LaBerge and his colleagues showed that lucid dreaming typically happens during the most active period of REM sleep, called “phasic” REM sleep (LaBerge 2007). During both wakefulness and REM sleep, neurons fire at a high rate in many areas of the brain. In REM sleep this high level of activity can be broken down into two components—“tonic” or continuous background activity and “phasic” or transient activity. Phasic activity consists of short, periodic bursts of firing, which coincide with rapid eye movements, muscle twitches, and surges of activity in the middle ear muscles (middle ear muscle activity, or MIMA, being the ear equivalent of REM). In short, lucidity seems to occur when there are brief and rapid increases to the already high level of cortical activity in REM sleep.
Having verified lucid dreaming from the sleep-science perspective, these researchers could take the further step of asking people to report on features of their lucid dreams while they were dreaming. In this way, sleep science could go beyond relying only on the dream reports subjects make after they wake up.
Taking exactly this approach, LaBerge investigated time in lucid dreams (LaBerge 2007). Do dream actions take the same amount of time as waking actions, or do they take more or less time (as the movie Inception suggests)? Lucid dreamers estimated ten-second intervals by counting “one-thousand one,” “one-thousand two,” and so on, and marked the beginning and end of the intervals with eye-movement signals. In all cases, the time it took to count from one to ten during the lucid dreams corresponded closely to the time it took during wakefulness. Daniel Erlacher and Michael Schredl in Germany replicated these results (Erlacher and Schredl 2004). They also found, however, that performing squats (deep knee bends) took 44.5 percent more time in lucid dreams than in the waking state. Erlacher and Schredl speculate that longer tasks or motor tasks (such as squatting compared to counting) could require disproportionately more time in lucid dreams, but it’s not clear what accounts for this result.
In another recent study by Martin Dresler and his colleagues at the Max Planck Institute for Psychiatry in Munich, two experienced lucid dreamers were able to carry out the predefined motor task of clenching their dream hands after making eye movements to signal that they were in a lucid dream state (Dresler et al. 2011). These scientists compared the brain activity during lucid dream performance of this task with the brain activity when the same individuals clenched their real hands while awake. In both the lucid dream state and the waking state, their brains showed activations in the same sensorimotor regions, though the activations were weaker in the lucid dream state. This finding supports the idea that brain areas active when we carry out a motor action are also active when we imagine carrying out that action.
These studies rely on what I like to think of as a kind of transworld communication. A link is set up between the private dream world and the public waking world. The sender writes messages in the phenomenal medium of dream ego eye gazes; the receiver reads the messages in the physical medium of the sleeper’s eye movements. An even more amazing method uses Morse code. When the lucid dreamer clenches his left and right dream ego fists, the polygraph picks up a corresponding sequence of left and right forearm twitches. LaBerge (himself a lucid dreamer) sent a Morse code signal spelling the initials of his name, with left dream fist clenches corresponding to dots and right dream fist clenches corresponding to dashes (LaBerge 2007).
Can the communication go the other way, from the dream scientist to the dreamer? Can there be two-way communication between the waking world and the dream world?
The answer seems to be yes. In another German study, lucid dreamers were stimulated with repeated low tones interrupted by the occasional high tone. The subjects were asked to listen for the high tone and to signal hearing it with a single left-right eye movement. Three lucid dreamers were able to carry out this task, and the EEG showed a waveform known to indicate detection of this kind of target stimulus (see Erlacher and Schredl 2008 for a description of this unpublished study by J. Strelen). In the words of a recent article, “This is but one example of a simple way of communication between the ‘real world setting’ (research in the sleep lab) and the lucid dreamer in the dream world. Future exploration of two-way communication between the dreamed and waking world should provide very promising research opportunities” (Erlacher and Chapin 2010: 9).
What happens in the brain during a lucid dream? Since lucid dreamers can signal when they become lucid and scientists can monitor what’s going on in the brain at the same time, this question is now testable.
In a recent EEG study, German sleep scientist Ursula Voss and Harvard University psychiatrist and dream scientist Allan Hobson, compared brain activity during three states—waking, signal-verified lucid dreaming in REM sleep, and nonlucid REM sleep (Voss et al. 2009). They found that lucid dreaming resembles waking in showing increased activity in the gamma frequency range (36–45 Hz) in frontal and frontolateral areas, compared to the reduced activity in nonlucid REM sleep. They also found that the EEG “coherence”—a measure of the degree to which the brain waves at various frequencies are synchronized across areas—is much higher during lucid dreaming than during nonlucid REM sleep. This kind of “neural synchrony” is thought to be a way that the brain coordinates or integrates processes across multiple areas. In short, this study shows that during lucid dreaming the brain adopts a novel, coherent pattern of activity marked by features characteristic of alert waking consciousness, but without leaving REM sleep.
What’s the significance of the frontal and frontolateral regions where these electrical patterns predominate? An area at the front of the brain called the dorsolateral prefrontal cortex, which is crucial for working memory, volition, and reflective self-awareness, shows greatly reduced activity during REM sleep. We might hypothesize, therefore, that activity in the dorsolateral prefrontal cortex must increase in order for dream lucidity to occur. Hobson has long advocated this hypothesis. But although the increased gamma activity and coherence in frontal regions during lucid dreaming are consistent with this hypothesis, using EEG to measure electrical activity at the scalp can’t tell us whether the dorsolateral prefrontal cortex is the underlying source of these brain wave patterns.
More direct evidence comes from another recent study that used fMRI. In this study, conducted by Martin Dresler and Michael Czisch at the Max Planck Institute for Psychiatry in Munich, four lucid dreamers slept in an MRI scanner and were told to signal continuously their state of lucidity using quick left-right-left-right eye movements (Dresler et al. 2012). When the scientists compared lucid dreaming to nonlucid REM sleep, they found increased activity in a range of cortical regions, including the dorsolateral prefrontal cortex, during periods of lucidity (see the figure below from Dresler et al. 2012). These activated regions include frontal and parietal areas that are thought to form a control system for cognition and decision making. This “frontoparietal control system” links brain networks that support the ability to pay attention to the outside world with other networks supporting long-term memory and the ability to think about oneself—to remember one’s personal past and project oneself into an anticipated future. Thus lucid dreaming reactivates not just the dorsolateral prefrontal cortex but also a wide and distributed network of brain areas crucial for attention and our human sense of self and personal identity through time.
This finding supports the idea that lucid dreaming reframes experience from the perspective of the dreaming self rather than the dream ego, that is, from the perspective of the dreamer who can remember waking life and think about and direct attention to the dreamlike quality of experience.
In summary, it’s important to remember that Dennett didn’t deny that there may be a genuine difference between lucid dreaming and dreaming that you’re dreaming; he argued that subjective testimony alone would be insufficient to establish the difference between the two. The best way to take his argument is as a challenge to improve dream phenomenology and dream reporting and to validate these improvements with physiological and behavioral measurements. This approach is the basic strategy of neurophenomenology (Lutz and Thompson 2003; Fazelpour and Thompson 2015). My argument has been that neurophenomenology meets Dennett’s challenge and decisively shows that lucid dreams aren’t a variety of ordinary dreams. Rather, they’re a unique state of consciousness definable by distinct phenomenological, psychological, and neuroscientific criteria.
Dement, W. C. and Kleitman, N. (1957) “The Relation of Eye Movements During Sleep to Dream Activity: An Objective Method for the Study of Dreaming.” Journal of Experimental Psychology 53: 89–97.
Dennett, D.C. (1979) “The Onus Re Experiences: A Reply to Emmett.” Philosophical Studies 35: 315–318.
Dresler, M., et al. (2011) “Dreamed Movement Elicits Activation in the Sensorimotor Cortex.” Current Biology 21: 1–5.
Dresler, M. et al. (2013) “Neural Correlates of Dream Lucidity Obtained from Contrasting Lucid Versus Non-lucid REM Sleep: A Combined EEG/fMRI Study.” Sleep 35: 1017–1020.
Erlacher, D. and Chapin, H. (2010) “Lucid Dreaming: Neural Virtual Reality as a Mechanism for Performance Enhancement.” International Journal of Dream Research 3: 7–10.
Erlacher, D. and Schredl, M. (2008) “Cardiovascular Responses to Dreamed Physical Exercise During REM Lucid Dreaming.” Dreaming 18: 112–121.
Fazelpour, S. and Thompson, E. (2015) “The Kantian Brain: Brain Dynamics from a Neurophenomenological Perspective.” Current Opinion in Neurobiology 31: 223-229.
Gyatrul Rinpoche (1993). Meditation, Transformation, and Dream Yoga. Trans. B. Alan Wallace and Sangye Khandro. Ithaca, NY: Snow Lion.
Hearne, K.M.T. (1978) “Lucid Dreams: An Electrophysiological and Psychological Study.” Ph.D. diss. University of Liverpool.
LaBerge, S., et al. (1981a) “Lucid Deaming Verified by Volitional Communication During REM Sleep.” Perceptual and Motor Skills 52: 727–731.
LaBerge, S. (1981b) “Psychophysiological Correlates of the Initiation of Lucid Dreaming.” Sleep Research 10: 149.
LaBerge, S. (2000) “Lucid Dreaming: Evidence and Methodology,” Behavioral and Brain Sciences 23: 962-963.
LaBerge, S. (2007) “Lucid Dreaming.” In Patrick McNamara and Deirdre Barrett, eds., The New Science of Dreaming, Volume II: Content, Recall, and Personality Correlates of Dreams, 307–328. Westport, CT: Praeger.
Laberge, S., and Rheingold, H. (1990) Exploring the World of Lucid Dreaming. New York: Ballantine.
Lutz, A. and Thompson, E. (2003) “Neurophenomenology: Integrating Subjective Experience and Brain Dynamics in the Neuroscience of Consciousness.” Journal of Consciousness Studies 10: 31–52.
Mavromatis, A. (1987) Hypnagogia: The Unique State of Consciousness Between Wakefulnes and Sleep. London and New York: Routledge and Kegan Paul, 1987.
Voss, U., et al. (2009) “Lucid Dreaming: A State of Consciousness with Features of Both Waking Consciousness and Non-lucid Dreaming.” Sleep 32: 1191–1200.
Windt, J. M. (2010) “The Immersive Spatiotemporal Hallucination Model of Dreaming.” Phenomenology and the Cognitive Sciences 9: 295–316.