POETICKS

Mistakes, Dreams and Creativity

I propose that person’s dreams have two functions. Their primary “duty” is to relieve the person’s cortical neurons of stored energy that would otherwise make those cells excessively prone to out-of-context daytime activation that the person would experience as “mistakes.” I also contend that by causing a person to experience mixtures of highly incongruous data while he sleeps, dreams promote creativity. To account for these results, I postulate a mechanism that causes a portion of a person’s cortical neurons to become spontaneously active during REM sleep to produce the bizarre memories that, I claim, make up dreams.

Over the centuries, there has been much study of sleep, the state in which dreams normally occur. Many attempts have been made to assign some function to it. Most modern thinkers on the subject have suggested that sleep is the way the body conserves energy during times of low-activity, and gives the body, including the brain, time to repair or otherwise fine-tune itself, all of which makes perfect sense to me. There have also been numerous attempts (not reviewed here) to understand the nature of the dreams that have been shown to occur in birds and most mammals including man during a phase of sleep called rapid eye movement (REM) sleep. Biochemical replenishment, rearrangement of data, and the communication of “subconscious” messages have been most often cited as the function of those dreams.  This paper will sketch one more such attempt.

I go along with previous theorists, Francis Crick and Graeme Mitchison (Nature, vol. 304, 14 July 1983, pages 111 through 114), in believing that “in viviparous mammals the cortical system (the cerebral cortex and some of its associated subcortical structures) can be regarded as a network of interconnected cells which can support a great variety of modes of mutual excitation,” and that “Such a system is likely to be subject to unwanted . . . modes of behaviour, which arise as it is disturbed either by the growth of the brain or by the modifications produced by experience.” Like Crick and Mitchison, too, I postulate a mechanism other than conventional forgetting that is used by the brain to detect and counteract such unwanted modes of behavior—at least those resulting from modifications produced by experience. (I will not be concerned with unwanted modes of behavior caused by brain growth, like involuntary fits, which seem to me outside of normal human psychology.)

In what follows I first describe certain key assumptions of mine about the brain and memory. Next I postulate and describe the central mechanism of my theory, and how it differs from that proposed by Crick and Mitchison. Finally, after briefly discussing its testability, I trace some of the implications of my theory.

Preliminary Assumptions

First assumption: the existence in a person’s cortical system of “knowlecules”–neurons, or neuron-clusters, each of whose activation is experienced by the person as a discrete, unified image, idea, feeling or the like such as a visual image of a cat, some general idea of what a cat is, or the word, “cat” (an idea going back in experimental psychology to Penfield).

Second assumption: that each knowlecule can receive energy (the form of which is not relevant to this discussion) from sensory cells, other knowlecules, or itself, and that it stores such energy (in some form) until its supply reaches some pre-set threshold that causes it to become active. Third assumption: that remembering occurs when one knowlecule receives enough energy from other knowlecules (and/or itself) to become active (or re-active). The basic rule followed in this operation is simple: every active knowlecule divides a pulse of energy (which I call “k-energy,” which is short for “knowlecule-energy”) among all the knowlecules that have ever previously been active immediately after it.  So if knowlecule A is active during one “beat” of brain activity, and knowlecule B is active during the next “beat,” A’s later activation will cause it to transmit to B, and if it does so sufficiently strongly, or with sufficient help from other knowlecules, it will activate B as a memory.

There is a great deal more to the process than that, of course, particularly with regard to the manner in which context influences what percentage of its output of energy a given knowlecule will transmit to a second knowlecule (or itself). For the purposes of this paper, however, it is only necessary to know that a given active knowlecule transmits to a number of other knowlecules (and, possibly, itself) once active immediately after it.

Mistakes

If one grants my assumption that a knowlecule (or the equivalent) can store energy (in some form) with the potential to activate the knowlecule, it follows that an inactive knowlecule containing a great deal of stored energy can, upon receiving a very small amount of k-energy, become active. When, as sometimes must be the case, the activating k-energy is out-of-context (in a manner that should soon become clear), the resulting more or less inadvertant activation of the knowlecule will be experienced as a mistake. For instance, suppose one is asked, “In what city in Maryland is the U.S. Naval Academy located?” One might know very well that the answer is Annapolis, but what if one’s knowlecule for “Baltimore” has a nearly full store of energy? This might be the case if one had earlier read an ad for the Biltmore Hotel, say; and seen a picture of an oriole (if one is enough of a baseball fan to know of the Baltimore Orioles’ baseball team); then
talked to a friend named Al (if one has a friend named Al who lives in Baltimore, as I do). “Biltmore” might cause a little energy to go to the knowlecule representing the similar-sounding word, “Baltimore,” but not enough to activate it; ditto the oriole and the friend named Al.

In such a case, the small amount of energy the knowlecule for “Baltimore” might get from its association with “Maryland” when the person is asked where the Naval Academy is could be enough to activate it. If that happens, and at the same time—because, perhaps, the person is tired—the question doesn’t quite cause enough energy to go to the knowlecule representing “Annapolis” to activate it, the person might wrongly say that Baltimore is where the Naval Academy is.

Other psychological processes will no doubt quickly apprise the person of his mistake but they aren’t important here. What is, is that mistakes of a certain kind are sure to occur, given my assumption that knowlecules, or their equivalent, store energy without becoming active.  This idea of how mistakes come about, of course, is a speculative commonplace among cognitive scientists—though couched in sundry terminologies and unconfirmed by experiment. But it hasn’t been proved invalid, either, and it makes sense.

It also supports Crick and Mitchison’s model of dreaming, which hypothesizes that “the function of dream sleep . . . is to remove certain undesirable modes of interaction in networks of cells in the cerebral cortex,” by showing how those “undesirable modes of interaction” might arise, and why they would be considered undesirable. However, and this is the main point of this paper, I propose a mechanism of “knowlecule-flushing” different from Crick and Mitchison’s “reverse
learning mechanism,” for mine, among other things, does not result in the weakening of dream- traces, as theirs does.

Knowlecule-Flushing

The knowlecule-flushing activator (k-f activator) is my equivalent of the “dream-state generator” postulated by Hobson and McCarley, and on which the Crick/Mitchison model of dreaming is based. Hobson and McCarley place their mechanism somewhere in the pontine reticular formation from whence it causes both rapid eye movements and periodic dreaming. My similar mechanism also causes dreaming—but (probably) not rapid eye movements, which I believe dreams cause, by giving the eyes visual material they reflexively follow. In the course of
this paper I will offer no (new) empirical evidence to show that the k-f activator exists (as I define it) but hope through common sense arguments from long-established empirical data to make the possibility of its existence something worth serious consideration.

The k-f activator, in normal circumstances, can only operate during sleep. A person’s arousal center brings about that state when the person’s brain-activity reaches some pre-set low level. The person’s arousal center then slows his body down, to put it simply, and isolates most of his brain from the rest of his nervous system. That is, transmission of stimulation from the periphery to the cortex and vice versa is suppressed. The person completely relaxes, in the process shutting his eyes. Or so my theory has it, and I believe both common experience and the authorities in the field would agree.

Once asleep, the person will eventually experience REM sleep, in normal circumstances. This, I hypothesize, is caused by the person’s k-f activator, which joins every knowlecule in his brain. The k-f activator becomes active more or less reflexively, after a certain amount of sleep, I suspect—but with the length of time it takes to do so probably dependent on how full the person’s knowlecules are. Thereupon it causes all the knowlecules in the person’s brain that have more than some set amount of energy stored spontaneously to become active. The conditions thus set up (probably through dispersal of enzymes of some sort that increase knowlecule sensitivity to stored energy) also prime other knowlecules to become active whenever they, too, contain the new lowered activation threshold amount of energy.

The spontaneous flush of knowlecules by the k-f activator starts a dream, and the increased sensitivity to their stored energy of the rest of the brain’s knowlecules, as well as of the just- flushed ones, will keep the dream going. Its contents (as common experience and all previous research has shown) will be scrambled, weird, surrealistic—which follows from the knowlecules that initiate them dream’s being activated out of context. That is, they aren’t activated “logically,” but simply because of the amount of their stored energy.

They are therefore experienced as mistakes, many of them happening at once (in the safety of the periphery-isolation that prevents behavior based on them). Normal rationalizing behavior ensues, of course, as the person, in effect, tries to make sense of the data exploding in his mind. And his memory puts new matter into the dream taking place just as it puts various matter into his waking thoughts. That is, remembering occurs the same way in dreams that it does during waking hours. Just as a certain name heard at work might make one remember Cousin Jane, the same name heard in a dream might make one remember her. I won’t be discussing remembering further here, except to point out that there’s no need to hypothesize some kind of special remembering that a person uses only while dreaming.

Once the k-f activator sensitizes a person’s knowlecule to its stored energy, the knowlecule remains sensitized to the same degree until a k-f inhibitor that I also hypothesize turns it off when the person wakes up.  That doesn’t mean the person’s first dream of the night will last the entire night. On the contrary, just as common experience suggests, and dream experiments seem to verify, each dream, or dream-session, tails off and eventually ends within two hours at the most. The reason for this is simple. At first many knowlecules will become active due to their increased sensitivity to stored energy. They will transmit to other knowlecules to activate them, and those will in turn cause more activation. Eventually, however, no knowlecule will get enough energy from anywhere to become active, even with its activation-threshold reduced. Being isolated from the environment insures this.

A person’s first dream of the night won’t likely be his last, either.  According to researchers, people normally have more than one dream a night—five, on average. To explain this, I claim that a person’s k-f activator goes through a nightly cycle during which it five times enhances his knowlecules’ sensitivity to stored energy, each time making less energy able to activate the knowlecule storing it. Hence, the first dream-cycle might reduce the amount of stored energy capable of activating a knowlecule to 80% of what would have been needed to accomplish that during waking. The second, ninety minutes later, say, might reduce the activation-threshold to 60% of the daytime norm.  Later cycles might reduce it, respectively, to 40%, 20% and 2%.
(These, of course, are just guesses, no experimental data being available for any kind of precise estimate, or likely to be for a while.)

All this is based on the simple idea that, to avoid the build-up of mistake-potential, brain-cells (the components of knowlecules) need to be cleaned out, as in the Crick/Mitchison model. But because, unlike Crick and Mitchison, I believe that the energy flushed is re-distributed throughout the brain to other knowlecules (and, in some cases, back to the distributing knowlecules) rather than otherwise disposed of, the flushing I hypothesize cannot take place all at once—by an immediate reduction of knowlecules’ activation thresholds to 2%, say— because
the resulting dreams would be too dense. A person’s brain would be overloaded—so much so, in fact, that the person would probably wake up. And the “creative” juxtapositioning that I credit dreaming with making possible, and will describe later in this paper, would be overdone, and yield not creativity but confusion.

In any case, research indicates that dreaming generally goes through five stages much as I’ve described. Interestingly, the later dreams are generally described by those having them as more bizarre than earlier ones, which makes sense since more inappropriate data would be
brought into consciousness; that is, many knowlecules minimally ready for activation would contribute material to a person’s awareness during his last dreams.

If the Crick/Mitchison theory of energy dispersal rather than redistribution is accurate, by the way, it seems strange that (1) dreams last as long as they do—why couldn’t all the cells with stored energy be emptied all at once? and (2) why would we have more than one dream a night, many of them involving similar material—that is, cells one would expect an early dream to have cleaned out seem to participate in later dreams? I also wonder why we should experience dreams consciously at all, however sometimes fleetingly.

The Value of Dreams

Since Crick and Mitchison believe dream-traces are lost permanently unless the dreamer wakes up during a dream and thinks about it, dreams for them would seem to have no evolutionary advantage except as a way of getting rid of unneeded stored energy. This flies in the face of much cultural opinion, however unscientific, as to the value of dreams. I won’t get into that, but will try to present more hard-headed arguments for believing dream-traces are treated the same way that other memory-traces are. One of my arguments is that it would make no biological sense for a human being to evolve a system for getting rid of brain-energy if re-distribution of it through mechanisms already in place could accomplish the same thing—as it does in my model, in which “excess” stored energy in brain-cells is reduced to almost nothing, wit  to dump quickly, what to keep? And wouldn’t such a mechanism take up room comparable to the storage space required for simply storing the material? I say that it makes biological sense for a person to store everything, and let his remembering mechanisms decide what to return to according to what later becomes important rather than give him access only to what is initially thought to have the potential for importance.

The Crick/Mitchison concept of dream-forgetting goes against common experience, too, for all of us seem to remember dreams. Such memories are anecdotal evidence, to be sure, but vivid. I even remember seeing things in a dream and, while in the dream, recognizing their having been in other dreams—from days or months before.

Creativity

My main argument for our remembering dream-matter, however, is that it would allow for creativity-enhancement by allowing us to refer back to the arbitrary, “mistaken” connections we make in dreams and use them if they turn out to have some value, as any mistake can. Daytime mistakes are probably not as bizarre as dream-mistakes. Indeed, certain connections occurring in dreams would be close to impossible in daytime. At least in theory. A dream could easily allow a Kekule (who discovered the shape of benzene) to experience a snake-image at the same time he experiences an image of benzene if his knowlecule for “snake” happened to have the right amount of non-activating stored energy, at the same time his knowlecule for “benzene” also did. But nothing in his waking hours, unless he happened on a snake while thinking of benzene, would meld them. Even in the latter instance, he would think of benzene, then see the snake, rather than mentally experience them both at the same time.

I’m not saying creativity via a dream is what happened with Kekule, just that such an occurrence would be possible if my model of dreaming were accurate, and it wouldn’t be if the Crick/Mitchison model were. Since such juxtapositionings would seem to be of value, particularly if they were made undangerously, during sleep, then revisited more or less at liesure during waking hours, Nature should select for their storage as memories. And, I contend, has.

My theory’s Compatibility with Research and Other Theories

Like the theory of Crick and Mitchison, mine seems broadly compatible with a large amount of experimental data—and with everyday experience, as well. And it explains as effortlessly as Crick and Mitchison’s model both the need for dreaming in adult life and the large amount of it that occurs pre-natally (which I attribute to the propensity of the knowlecules of a developing brain for distributing k-energy willy-nilly and thus causing partial storage of k-energy to be
relatively wide-spread, it taking the brain time firmly to establish datapathways).

My theory, like theirs, is also compatible with the hallucinoid nature of dreams that all researchers, and non-researchers, remark on. Unlike theirs, my theory does not contradict Freud’s, but augments it, for it allows lessons to be recalled and thus learned from dreams, in keeping with Freudianism. It also permits repressed material to be popped into consciousness as Freud hypothesized, through the lowering of “repressed” knowlecules’ activation threshold. This agreement with Freudianism I mention only as an interesting feature of my model,
incidentally, not as an argument in its favor, Freudianism still not having been experimentally verified that I know of (and, in my view, 90% hogwash).

The effects of REM sleep deprivation certainly do not contradict my theory, though they don’t emphatically support it, either. That subjects of such deprivation sleep more when allowed to after their period of deprivation is what my theory would predict. That REM sleep deprivation in humans sometimes produces irritability would follow from my theory, too—because the mistakes a person makes as a result must irritate him once recognized, and because, as deflections from “right reasoning,” they will tend to strand him mentally, which would be conducive to irritation. This deflective property of mistakes would also explain the inability to concentrate experienced by some subjects of REM sleep deprivation, mistakes breaking their focus.

That feelings and wishes that he would ordinarily keep out of consciousness might intrude on a REM-sleep-deprived person’s thoughts, as some research indicates would happen, would be in
keeping with my model also, for knowlecules prevented from dreamtime activation would tend to build up stored energy until they had enough for waking arousal. Internal fantasizing should for similar reasons tend to increase among the REM-sleep- deprived, as has also been shown to be to a small degree the case. As for the possibility considered by some investigators that those deprived long enough of REM sleep would experience hallucinations, my theory is  noncommital.  According to it, REM sleep deprivation should yield just increased susceptibility to mistakes, as defined above—only, probably, after more than a few nights of deprivation, I might add.

My theory cannot account for the lack of any readily-observable detrimental effects from the complete blocking of REM sleep that certain monoamine oxidase inhibitors and other drugs cause. However, I believe that the drugs, which are anti-depressants, reduce people’s
anxiety about the mistakes they make, which makes those mistakes less noticeable. I believe also that the drugs energize those who take them, allowing them to power their way through their mistakes, before they multiply. A third factor would be the probably great length of time it
would take for any significant psychological deficits from any form of neurological deprivation to show up. The way the drugs involved no doubt interfere with normal distribution of k-energy must be a factor as well. My bottom line here, though, is the same as Crick and Mitchison’s concerning the same research: that it’s too small a factor to overthrow a theory so little contradicted elsewhere.

Testing My Theory

To prove or disprove the existence of my knowlecule-flushing activator would require much more knowledge of the brain, and much better brain -investigating technology now seems to be available. Crude tests of whether REM sleep deprivation will indeed increase a person’s
propensity to make mistakes, as I define them, or decrease his ability to come up with new ideas are perhaps possible but would not likely be very persuasive one way or the other. If we ever are able to pin down precisely what kinds of proteins or other substances are manufactured during the creation of memory traces, we might be able to compare the amounts of those substances produced during dreaming with the amounts produced during waking thought, and thus get a better idea of the likelihood of the data of dreams’ being stored or not.

Since my theory of dreams flows directly out of a (more or less) simple model of inter-cellular energy-flow in the brain, it could probably eventually be modeled by a computer program that could be used to check its plausibility. All my thinking on dreams is, in the final analysis, speculative, however. But since it is all based on a notion of the material make-up of brain-cells and auxiliary physiological mechanisms, it is all ultimately testable.

Possible Implications

If my theory of dreaming is close to being valid, it should help us understand and reduce (or increase, if desireable) the occurrence of mistakes, and appreciate and nourish creativity. It should provide some insight into the etiology and nature of certain kinds of psychoses, as well, some forms of schizophrenia being surely due to waking dreaming. Since in my model, dreams are accessible to remembering, the model’s validity would also suggest that perhaps dream-analysis in certain forms of psychotherapy might be of value, after all. It would certainly suggest that the high regard in folklore for dreams and what they say is not misplaced.

Viva dreams!                                         

.                                                               Bob Grumman

.             October 1997 (but based on my work in the early seventies)

26Apr14–38

2June14–64, a surprise

.