8. BRAIN CONTROL OF SLEEPING AND DREAMING STATES

J. ALLAN HOBSON, MD

Neuroscience and Tibetan Buddhism have both devoted a great deal of attention to sleep and dream states, and there is a clear overlap of interests in this area. In fact, the topic generated so much interest on both sides that it was to become the focus of the fourth Mind and Life Conference.¹

Allan Hobson provides a general tour of current scientific knowledge on sleep and dream states, and the implications of these findings for our philosophical understanding of consciousness. Most significantly, consciousness is understood to be a natural condition of the activated brain, and most of the brain activity associated with various states, whether waking, dreaming, or in deep sleep, is generated internally by the brain itself, rather than being driven by sensory input.

Control of the regular cycles of sleeping, dreaming, and waking states is governed by reciprocal systems of neurotransmitters. Because of their role in controlling relaxation, these chemical systems may prove to play a role in meditation also.

Lucid dreaming, in which the subject is conscious of dreaming even as the dream occurs, is another area of particular interest, having only recently been recognized, let alone studied, in Western science. Tibetan Buddhism has a long tradition of dream yoga which includes training in lucid dreaming.

ALLAN HOBSON: I plan to identify some important themes in the science of sleep and dreams, and to orient the discussion toward the clinical and philosophical implications of this work.

Those implications include the following: I think we can now conclude that good sleep and good waking are reciprocally related, as Buddhist philosophy has long held. More particularly, I think we can identify within the brain some of the structural bases of these reciprocal relations. Further, I think we can help understand how some of the practices that the Buddhist tradition has developed actually work in relation to specific brain mechanisms. That is particularly exciting with respect to future experimental prospects.

Obviously, in the West we also believe that sleep is important to health, and to the degree that we can help people learn to sleep without pills, we will be in a much stronger position. One of the great afflictions of the West now is the overuse of sleeping pills. So behavioral techniques, such as those Buddhists have developed, that can be used to help people achieve peace of mind and good sleep are very welcome.

The second point about the clinical implications has to do with dreaming itself. I think we can now objectively identify the brain states associated with dreaming and help people, if they so wish, to have access to their dreams. In other words, we are now in a position more easily to control dreaming, if we want to. I will discuss dream control later.

There are a number of other interesting topics that are of great importance to philosophy and to clinical practice. Philosophically, these issues really have to do with the relationship of brain activity to consciousness. Perhaps the most important general conclusion coming from sleep research is that consciousness in all its varieties is specifically related to particular brain states. Consciousness seems to be a natural condition of the activated brain. Furthermore, most of the activity of consciousness is internally generated. In other words, the brain contains its own mechanisms for creating information, and most of brain activity is concerned with processing its own information, input from the outside world being relatively modest. Previously, in the West, models of brain function were largely input driven. Now, I think we see that the brain has highly organized spontaneous activity, such that the main rudiments of consciousness are inherent, self-organized, and built into the system.

So I hope these three themes—the clinical implications for sleep and its relation to waking, the nature of dreams and our capacity to access them, and the implications for theories of consciousness will be discussed following my presentation.

In such a brief time I can’t possibly cover the wealth of material now known about these issues, so I will review five major conclusions.

MEASURING SLEEP AND DREAM CYCLES

The first is that sleep can be objectively measured. We can, for example, measure activity of the brain, we can measure activity of muscles, we can measure activity of the eyes. When we make these three measurements, we can clearly distinguish three states: waking, sleeping, and sleeping with dreams. Furthermore, we can identify these three states in another way, simply by observing behavior, and using either photography or time-lapse video to keep track of changes in body posture that are associated with the underlying changes in brain state responsible for the shift from sleeping to dreaming. In time-lapse video, a picture is taken at regular intervals, typically every seven and a half minutes, using a video camera instead of a still camera to record a history of postural shifts throughout the night.

We have found that sleepers never make fewer than fourteen major body shifts during the night. They shift from right side to left side to right side to left side during six to eight hours of sleep. And when sleep becomes less sound, they make more movements. With respect to some of your traditions concerning sleeping on the right side,² the real issue is, can a trained subject suppress body movement shifting and still have normal sleep? This is a very simple experiment which could be performed with very simple techniques such as time-lapse video, which is easily portable.

Using measurements of the brain, eye, and muscles, we can identify three states: waking, sleeping without dreams, and sleeping with dreams. Nondreaming and dreaming sleep alternate in a regular cycle, which lasts about ninety minutes, with about the first sixty to seventy minutes being nondreaming sleep, the last fifteen or twenty minutes being dreaming. This means that in a night of six to eight hours of sleep, you are going to have four or five dream periods, each lasting fifteen to twenty minutes, or longer. This means that in any given night of sleep, we have as much as two hours of dreaming, distributed at regular intervals throughout the night. There is a lot of time spent dreaming.

DALAI LAMA: Are there any differences due to age?

ALLAN HOBSON: Yes. Sleep generally becomes shorter and shallower with age. And with aging, there is a slight decrease in the amount of time spent in the state of sleep associated with dreaming. The cycle duration remains fixed throughout the life span, once we are adult. Babies have a shorter cycle, and the cycle lengthens as the brain increases in size. The newborn infant has about four times as much dream-state sleeping as the adult does, which is a very interesting point to consider with respect to developmental issues.

DALAI LAMA: Is there any biological evidence to determine when the infant starts dreaming?

ALLAN HOBSON: Biological evidence could never prove when someone starts dreaming because dreaming is a psychological experience, but we know that REM sleep begins in utero. (Here we must be dualists with respect to language. There is an important distinction between language dualism and deep philosophical dualism.)

The three physiological phases that are associated with waking, sleeping without dreams, and dreaming, are characterized by measuring brain wave activity and that of the eyes and muscles. The stage of sleep associated with dreaming is called “rapid eye movement” or REM sleep because the eyes move around very dramatically. At the same time, muscle activity is blocked. Our muscles are paralyzed except for the eye muscles. And the brain is activated. This phase of sleep is clearly identifiable in the human fetus as early as thirty weeks of gestation, and it probably begins at least ten weeks earlier in a more primordial form. So as early as at twenty weeks of gestation, the brain has organized and differentiated sufficiently to generate this kind of alternation between brain states. But the dream state is much more prominent in the fetus even than in the newborn, so that at thirty weeks of gestation, the estimates are that this state occupies about 90 percent of all the time.

NEURONAL CONTROLS OF SLEEPING, DREAMING, AND WAKING

We know that sleep is organized as an alternating cycle. The next question for the Western scientist is: How is the cycle organized within the brain? We know that it is controlled by brain structures localized in part of the brain stem. The brain stem looks somewhat like the base of a flower. It connects the stem of the flower to the blossom just as in a lotus plant. The bulb of the flower, by this analogy, is the brain stem. This small but very important part of the brain, between the spinal cord and the rest of the forebrain, supports our conscious activities. Neuronal machinery which controls the alternation among sleep stages and wakefulness is located in a small region of the brain stem called the pons or bridge.

This location is obviously strategic. It can control inputs and outputs for the whole body. It can control activity throughout the whole upper brain, the forebrain, which we consider to be the organ of consciousness. Our third point, then, is that this regular alternation in sleep is controlled by the brain stem.

The obvious next question is: How does the brain stem do this? Within this pontine region are two populations of nerve cells that have distinctive chemical signatures. One is the neuronal population that supports the waking state, and we suppose it to be responsible for arousal and even anxiety. Its chemical signaling involves the release of amino acids, hence it is known as an aminergic system. When this system is very active, we are very alert, but we may also become too alert. We may become anxiously alert.

It is the proper regulation of this system that I think constitutes one of the goals of Buddhist meditative practices. And this same system is obviously of great significance also to Western medicine. This is so not only because this population of aminergic neurons controls arousal, wakefulness, alertness, and anxiety, but also because outputs of this system affect vital functions like breathing, blood pressure, and other visceral as well as cerebral contributions to our experiences.

This aminergic system in the pontine part of the brain stem is also involved in energy regulation and energy flow, and probably also with aggressive behavior. I think it is a key to understanding a number of very important aspects of human life. Moreover, it seems to play a significant role in many functions of prime importance in Buddhist thought and training.

DALAI LAMA: Are such emotions as aggression, love, and attachment also associated with that part of the brain?

ALLAN HOBSON: Not specifically, no. But, as part of the general continuum of activation, other forebrain structures will be engaged, which will then, according to external inputs, govern the emotional state of the individual. The emotional system is rather farther forward in the limbic part of the forebrain, including the hippocampus, which was discussed yesterday. This brain stem activation site is not a specific system for the control of emotions. It is a specific system for controlling the level of arousal of the individual as a whole, which thereby affects other systems, including emotional controls.

The aminergic system is one group of neurons in this critical region of the brain stem. The other group of nerve cells in this same region is called a cholinergic system because it’s chemical signature—its neurotransmitter—is acetylcholine. We can identify these two neuronal populations in the pontine brain stem, localize their cells precisely, determine their major connections, their chemical neurotransmitters, and record their patterns of electrical activity.

The cholinergic system is apparently held in restraint by the aminergic system. Thus, when the aminergic system is functioning at a high level, the cholinergic system is functioning at a reciprocally relatively low level. That is the situation in the waking state. As we go to sleep, the aminergic system decreases in its activity, and the cholinergic system becomes relatively more active. The cholinergic system becomes progressively more active throughout the period of deep sleep without dreams. Ultimately these two neuron populations become radically differentiated: the adrenergic system shuts off completely, and the cholinergic system reaches its highest level of activity just when you enter the dream state. Activation of the cholinergic cells generates signals that contribute to eye movements, to inhibition of muscle tone, and to activation of the forebrain.

These reciprocal shifts of functional states can probably also be influenced through meditative practice.

DALAI LAMA: Are you indicating that in the dreaming state you are even more relaxed than in the nondreaming state?

ALLAN HOBSON: It is a paradox, because the muscles are completely paralyzed. To speak of relaxation in this case is misleading. The muscles are actively suppressed, or inhibited. But the upper brain, the forebrain, is very active electrically. In contrast to the waking state, this electrically active brain in the dreaming state is chemically distinctly different because of the shift in the neurotransmitter ratios. The dreaming brain is very highly cholinergic, the waking brain is very highly aminergic, while in each of these states the forebrain is highly electrically activated. We believe that this is very important for understanding the differences between the waking state and the dreaming state.

So we now know that sleep is organized into a succession of states. We can identify and measure distinctive sleep states. We know that the brain stem controls the succession of waking/sleeping/dreaming states. And we know that the brain stem controls that succession of states by altering the production of specific neurotransmitters which are represented in two reciprocal systems of neuronal control.

The fifth and final point I want to make about the science of sleep is that we have tested this theory by making microinjections of very small amounts of chemicals into specific, localized regions of the brain stem of experimental animals. By this means, we can control the overall brain states of wakefulness and sleep. In other words, by imitating the activation by acetylcholine in very specific, localized parts of the brain stem, we can convert the whole brain from the waking state to REM sleep almost immediately and keep it there for many hours. If we put the same chemical, acetylcholine, into another part of the brain of experimental animals, we can produce waking. The differentiation of these control systems is specific and precise.

So we have obtained experimental control of the state of sleep in animals. To some extent, similar experiments have been replicated in humans. Obviously, we don’t inject chemicals directly into the brain stem in humans. We use human subjects to measure states of sleep and wakefulness and to obtain reports relating to conscious experiences. We use animal studies to investigate what’s going on neuronally within the brain stem during different sleep and wakefulness states. All mammals share identical organization of alternating states of wakefulness, deep sleep, and REM sleep (presumably dreaming) behavior. They all have obvious waking states complete with apparent awareness and interactive behavior with the environment. Such waking states alternate with slow-wave sleep, which lacks rapid eye movements and is associated with high-amplitude, low-frequency electrical activity throughout the brain. These slow-wave sleep states cycle regularly into the kind of sleep state associated with globally inhibited body movements except for rapid eye movements, specifically accompanied by low-amplitude, high-frequency electrical activity throughout the brain. In this latter state, all of the objective phenomena are equivalent to the state that in humans is identified by subjective testimony of dreaming.

In humans and other mammals we see a complete suppression of muscle tone during REM sleep, so the motor output is actively inhibited. Otherwise our dream states might be accompanied by our getting up and running around—still asleep—acting out our dreams. Our dreams are typically characterized by the hallucination of movements by ourselves and among other animate and dynamic things. That’s because the upper brain, the forebrain, is actually generating elaborate visual and motor patterns which are not allowed to be acted out by our muscles, perforce the general inhibitory control exerted by brain stem mechanisms. Only the eye muscles are permitted to express this internal sensorimotor dreaming state.

Meanwhile, during REM sleep, the brain is electrically activated, even more so than in quiet waking. The brain is intensely internally activated: hence we imagine that the dream arises because the manifestly activated brain is actively processing signals that would ordinarily be associated subjectively with direct, vivid experiences and outgoing behavior. We hallucinate the experiences and the inhibited behavior as if it were not inhibited. And that is our dream!

DALAI LAMA: What accounts for the rapid eye movements? The rest of the muscles of the body are paralyzed in the dreaming state, and yet the muscles associated with eye movements are not. Why is that?

ALLAN HOBSON: The answer is that the eye muscle system is a very different sort of motor system from the skeletal muscle system. Most of the skeletal muscle system is engaged in maintaining posture against gravity, and that system is obliged to use a lot of tonic inhibition. The eye muscles don’t have to do that. The eyes are essentially weightless, their specific gravity is about equivalent to their surroundings in the orbit. The activity of the eyes is to sweep about swiftly and relatively effortlessly. Because they work with straight beams of light, often from very remote objects, they have an enormous leverage and target speed as well as accuracy in relation to the visualized world. The eyes are never completely static and they don’t have to work against gravity.

Secondly, the eye motor nuclei which control eye movements lie upstream in the brain stem, forward of aminergic and cholinergic state control systems in the pontine brain stem. All other direct motor systems, for face and lower muscles, lie within the pons or below the pons along the lower brain stem and spinal cord. Hence, the eye motor nuclei are most anterior of all direct muscle control mechanisms. They are relatively so far forward of other direct muscle control systems that they are really a part of the forebrain systems that serve conscious experiences and higher mental life.

A third observation is that moving the eyes does not contribute to skeletal motions and other bodily effects that might have the consequence of jarring us awake.

ANTONIO DAMASIO: Accepting your idea that we have an active suppression of antigravity muscles in order to keep us from moving around, it is clear that the eye muscles would not do that. But do you ever see a lot of movement of the facial muscles, which are controlled only slightly lower, in the lower pontine and bulbar brain stem? They would not have much skeletal motion effects except by way of jaw movements. Moving muscles of the face presumably wouldn’t have the consequence of waking you up.

ALLAN HOBSON: You don’t see a lot of facial muscular movement, but you do see some. In the human infant, the facial expressions in REM sleep are particularly visible and dramatic, and quite charming. There is automatic smiling that is produced during rapid eye movement sleep. If you watch a mother nursing a baby, you will often see the baby, when he or she becomes satisfied with milk, begin to close the eyes as the sucking movements become very regular and rhythmic. Shortly, the eyes start to move actively, and the baby has gone into rapid eye movement (REM) sleep. Then you see dramatic, spontaneous facial expressions, especially smiling. And the mother believes the baby is sending her a message of contentment and happiness. You can observe this and talk to mothers about this, and they will uniformly interpret the baby’s behavior as meaningfully related to their generosity in giving the baby nourishment. That’s quite an important ethological concept, this intergenerational signaling system that is so adaptive and useful in contributing reinforcement to the mother for this uniquely mammalian, altruistic feeding behavior.

DALAI LAMA: Dogs make limb movements, and on occasion when a person has a rough dream, like a nightmare, the arms may flail about.

ALLAN HOBSON: Yes, but then it is not a nightmare. There are two kinds of frightening experiences that occur in sleep. Bad dreams, or nightmares, in which you imagine a scenario and frightening things happen to you may occur in REM sleep. They can erupt, which usually results in the subject’s spontaneous waking. There is another kind of night terror, which tends to occur in nondreaming sleep but not in REM sleep. This is a purely emotional experience, lacking the associated hallucinatory activity that accompanies dreaming. It is this night terror that may be accompanied by flailing limb movements.

There are some human subjects in whom the brain stem fails to inhibit the skeletal muscles. When they have REM sleep, they act out their dreams. This is a very dangerous brain stem defect.

Now let’s look at the way the sleep cycle is organized. First, as the night progresses, come phases of sleep associated with little or no mental activity. Next are periods of sleep associated with dreaming. The dream periods tend to become longer as the night progresses, lasting thirty, forty, even fifty minutes at a time. So the best time to obtain dream reports is in the early morning when these dreaming periods are quite long.

The cyclic alternation of non-REM and REM sleep is very regular. This suggests that the brain stem neuronal controls themselves constitute a clock, or that they observe an accurate clock, whereby they trigger activation of the dreaming state at regular intervals throughout sleep. This is an automatic, intrinsic process. It looks, therefore, as though this activation of the brain for dreaming purposes must be very important in some way. We need to understand this better.

WHAT IS THE PURPOSE OF DREAMING?

This is one of the great mysteries of the science of sleep at present. What is the function of the recurrent brain activation of dreaming during sleep? It is so prominent, almost dominant, in early life, that it attracts Western scientists to the idea that it may be important for the development of the brain. In later childhood and throughout adult life, although it is no longer so prominent, it nevertheless occupies a regular and a conspicuous part of one’s adult life, 10 percent or more. You go from an intrauterine experience of 90 percent dream time to adult experience of the reciprocal of that. We might suppose that dreaming is perhaps necessary for maintenance of the brain in some important respect. This is what we would like to study for the next ten years: Exactly what benefits does this forebrain activation, this dreaming state, confer? We suspect that it has something to do with the capacity to maintain attention during the waking state.

ROBERT LIVINGSTON: It might be that dreaming has also to do with problem solving. Among other dream scenarios, the subject matter of dreams could well include, directly or by analogy, aspects of frustrated, unfulfilled, perhaps failed experiences.

There is abundant evidence from among scientists, artists, and others engaged in creative activities, that many of them have picked up cogent novel ideas and new patterns of thinking and performance during a night’s sleep, and expressly during dreaming. Many of their most important challenges and frustrations have been cast in new light, opened to new strategies, or reformulated in entirely new and unexpected patterns for thought, action, and pursuit. These intellectually revolutionary solutions have emerged in consciousness during or promptly following dreams. Perhaps the dreaming state can try out alternative perceptions, judgments, and behavior without penalty of consequences.

ALLAN HOBSON: As yet we have no specific evidence bearing on that issue. We do not know whether cats dream, but they show brain waves indicative of activation associated with eye movements, and these occur periodically, at regular intervals, throughout their sleeping behavior. If we measure the electrical patterns of a cat’s brain when the cat goes from nondreaming sleep to REM sleep, we see that the high-voltage slow electrical waves give way to low-voltage fast electrical activity. At the same time the skeletal muscle tone is actively inhibited and the animal is totally relaxed. So the brain is turned on and motor output is turned off except for eye movements.

Cells localized to the brain stem, in the vicinity of the nuclei that control eye movements, send signals to the visual receiving areas of cortex that process visual data. The electrical activity occurs just prior to the execution of eye movements. This means that the brain has a way of keeping track of movements even before the movements occur. Signals recorded in the visual cortex faithfully encode the direction and distance of the eye movements, which are actually taking place in the dark. The brain stem is telling the visual perceptual system where and by how much the eyes are going to be displaced, and hence what comparable shifts must occur between subject and visual field during whatever scenario is playing in the dream. The brain thus has a highly specific information system operating throughout dreaming. This probably is the physiological basis for the dynamic visual experiences associated with dreaming. Dreams are richly visual because the visual system is stimulated along with activation of the eye muscles.

We now have the interesting paradox that our brain, in the dark, with eyes closed, can turn itself on and initiate messages that relate to vision and other aspects of our life experience. This is, indeed, what a dream is.

DALAI LAMA: It is impossible, isn’t it, for a person who has been blind from birth ever to see color in a dream state?

ALLAN HOBSON: It is impossible for them to realize this activation as visual because they have no categories of visual experience to bring to this synthetic task. This suggests that our visual system is entrained from exposure to visual events and that it is not simply intrinsic to that region of the brain. Blind patients must have had some previous visual experience in order to have visual imagery accompany their dreams. People with acquired blindness see in their dreams, and see only in their dreams.

Yesterday, Dr. Squire distinguished between declarative memory and procedural, or nondeclarative, memory. We might expect the forebrain, including the hippocampus and cortex, to be responsible for declarative sorts of memories, and other parts of the brain to be concerned with procedural memories associated with learned skilled behaviors. This part of the brain is constantly active throughout sleep and dramatically activated during REM sleep. In association with each eye movement, the activity of this part of the brain is altered, leading us to believe that one of the functions of REM sleep may be actually to rehearse brain programs for the benefit of behavior. This would provide procedural learning during dreaming. So the maintenance function theory of dreams now attains further specificity. What we may be doing every night in our dreams is rehearsing our basic motor skills, practicing to make perfect, if you like, making certain that the central programs for behavior are in good order. Although this is a speculative theory, it is provocative.

LUCID DREAMING

Dr. Hobson introduced the concept of lucid dreaming with an illustration of a man on a flying carpet. Flying is a favorite dream activity, and its obvious unreality provides a way of identifying the dream state.

ALLAN HOBSON: This man on a magic carpet, flying through the air, has taught himself to do this by taking advantage of some of the facts about dreaming. He knows that dreams are strange, that they have curious characteristics. So he tells himself before he goes to sleep that if he has a conscious experience that is strange, he will know that he is dreaming. He needs to do this for about three weeks, every night, for just a few seconds before going to sleep, and then he starts to gain consciousness of his dreams. He has created a new state in which part of his brain is acting as if it were awake while the other parts are dreaming. Western practitioners of this skill call themselves lucid dreamers. They can watch their dreams while their dreams are occurring.

DALAI LAMA: Is there special training that leads to that ability?

ALLAN HOBSON: It is simple. For three weeks you tell yourself before sleep that you ordinarily dream for about two hours, and that if you have a strange experience, you are going to recognize consciously that it is a dream. It helps to have a notebook at your bedside and to write down your dream experiences. You can induce lucid dreams by this automatic procedure, a split consciousness. In other words, the power of the mind is quite appreciable, capable of changing the state of the brain.

ANTONIO DAMASIO: Have lucid dreamers been studied over a long period of time?

ALLAN HOBSON: Lucid dreamers have not been studied over a long period of time in sleep labs. I suppose your question would be: What is the functional gain?

ANTONIO DAMASIO: What is the functional loss?

ALLAN HOBSON: Who knows? It could be that you are interrupting sleep in a deleterious way. All I can tell you is that their period of lucidity is fleeting. Even when they become highly skilled, the dream plot tends to slip away from them.

Until two years ago, there was no physiological evidence for lucid dreaming, and now there is. Subjects can be trained to make a prearranged sequence of eye movements so as to indicate that they are presently lucid, and these eye movements can of course be recorded electr­ophysiolog­ically.

During lucid dreaming, general body muscle tone is still inhibited. But of course they have eye motor control which is presumably under the control of the frontal eye fields in cortex of the frontal lobes. As a means of testing and documenting lucid dream states, the subjects are told to make three full-excursion movements to the left followed by three full-excursion movements to the right. The probability of this occurring spontaneously is essentially zero

DALAI LAMA: In the Buddhist practice of dream yoga, there is a particular practice in which the lucid dreamer, while in the dream state, is told to be conscious of the dream state and told to meditate on something specific.

ALLAN HOBSON: Yes, it could very well be that this would work, and it could be experimentally documented, but it has not yet been tried in the Western scientific tradition. It’s a doable experiment.

Now this finding has lots of important implications, and I would like to emphasize two. For Western science, it gives us a way of making time labels within the dream so that we can correlate physiological activity with mental activity more precisely. And it raises certain problems. It shows how suggestible dream contents can be.

What that means is that we can teach subjects to dream anything they want to dream about. Therefore, if the dream is taken as important evidence for a psychological or philosophical theory, we encounter the problem of a circular loop. The subject may be dreaming what he expects to dream about in order to prove the theory, and this does not constitute scientific evidence of anything.

DALAI LAMA: Are you implying that in psychology, where a dream may be considered of great importance for interpreting some of their theories, these experiments reveal that the psychological basis could be shaky?

ALLAN HOBSON: Yes, precisely. So that’s the bad news. The good news is that this lucid dreaming state is so plastic that it can be exploited for a variety of purposes.

DALAI LAMA: From a Buddhist point of view, one might be able to distinguish different states of dreaming. Generally speaking, a dream is a dream, something you can’t control. But for the highly advanced meditator, there could be possibilities for gaining certain insights through dreams.

ALLAN HOBSON: That’s possible. Even at that level you encounter the same problem of wondering whether this is evidence of anything other than purpose and expectation.

DALAI LAMA: I know some Tibetans who lived in Tibet prior to the 1959 uprising. Before their escape from Tibet, they did not know about the natural trails and passes by which to get over the Himalayas into India. Some of these people I met had very clear dreams of these tracks and, years later, when they actually had to follow the actual trails, they found that they were already familiar with them because of the very clear dreams they had had previously.

ALLAN HOBSON: This is a so-called precognitive dream, and there are many examples of this in the West as well. I would like to defer discussion of that until later, as it is an important question.

Dr. Hobson closed his presentation with another drawing based on a dream image: a man riding one of two bicycles which are connected by an elaborate magical device.

ALLAN HOBSON: The good news is that dreaming is an autocreative state. Its plasticity can be utilized for a number of different purposes. This picture shows another drawing of a dream which indicates the autocreative nature of dream experiences. What does it mean that there is no one sitting on the second bicycle? The dreamer is a bachelor who throughout his dream journal complains of the fact that he has no companion. Now this is an interpretation and probably can never be proved scientifically, but it is nonetheless poetically intriguing. Dreams are often poetically compelling and we should not lose sight of that important point.