Time plays a pivotal role in dreams. Much of the fantasy, incoherence, inaccuracies, and ambiguities of dream contents is owed largely to continuous weakening of the strengths of connections (synapses) between nerve cells of the brain over time. Considerations here relate mainly to this weakening influence. Emphasis is on the dreams of children, which are expressed in their purest form, minimally encumbered by the complicating influences that accumulate with age.
Dreams are the accompaniments or by-products of certain essential activities of the brain during sleep. They primarily occur as neocortical circuits (the most recently evolved brain circuits) become activated by spontaneous, self-generated, complex electrical oscillations. Superficially, these oscillations are expressed as the slow and fast scalp-waves of electroencephalograms (EEGs). A major role of dreams during sleep is the processing of phylogenetic and experiential memories, that is, inherited memories and memories of past waking events.
Memory Consolidation and Reinforcement
Regarding this processing, during nonrapid eye movement (NREM) sleep, recently acquired short-term memories stored temporarily in the hippocampus become converted into long-term memories stored in the neocortex by a process known as hippocampal replay. In addition, enormous numbers of already-stored fragments of memories in the neocortex become reinforced (strengthened). During rapid eye movement (REM) sleep, many already-stored, long-term memories in the neocortex also are reinforced to maintain their authenticity.
Because dreams usually are highly visual, the storage process for visual memories is used here for illustration, though essentially the same principles apply generally. Storage of memories in both the hippocampus and neocortex is sparse and distributed, sparse in the sense that only a fragment of the memory is represented at any one of the distributed physical locations. In the visual neocortex, sets of neurons having different response properties (fragments) for colors, textures, distances, orientation, positions, and so forth, are clustered together at the various distributed locations.
As an example of memory consolidation by hippocampal replay, consider the process for a new declarative memory, that is, for the conscious recollection or explicit remembering of a new scene or event. During replay, fragments of the memory, already sparse and distributed and stored in the hippocampus for the short term, become similarly established in the neocortex for the long term, by a repetitive interactive process. Such replay in humans might continue for as long as 3 years.
Spontaneous reinforcement of memories stored in neocortical circuits during sleep involves electrical activations by both fast and slow brain waves. Reinforcement often is accompanied by an “unconscious” awareness of the corresponding memories in the forms of static dreams, that is, isolated thoughts or perceptions. Narrative dreams are assembled from these thoughts or perceptions but, on any given night, only from a small fraction of them.
It is an intrinsic property of synapses of memory circuits that they need to be reinforced periodically. Otherwise the synapses weaken and the encoded memories deteriorate, in days, weeks, or little more than a month. Synaptic strengths (weights) persist for only limited periods, primarily because the macromolecules that are essential for synaptic function break down continuously (molecular turnover). Because sufficient numbers of these molecules are needed to preserve the specific synaptic strengths that encode given memories, if “lost” molecules were not replaced periodically, these strengths would gradually decline and the memories would deteriorate.
Dedicated Synaptic Strengths, Brain Waves, and Dream Contents
Dedicated (functional) values of synaptic strengths for a given memory become established and maintained by the memory’s use or periodic activation. But for memories that are used only infrequently during wakefulness, most synaptic reinforcement has to occur spontaneously during sleep. The functions of such spontaneously reinforced circuits that would disturb sleep usually are not triggered because of decreased or absent muscle tone or temporarily reduced behavioral responsiveness.
Circuits for dedicated functions remain labile to the extent that they are susceptible to being updated (reconsolidated) in response to related new experiences. Whereas the updating process drives existing synaptic strengths to new dedicated values, mere repetitious synaptic activity in the course of a dedicated function, or reinforcement during sleep, merely maintains existing strengths.
During REM sleep, spontaneous gamma oscillations (30-100 cycles per second) trigger a process known as temporal binding (binding by synchrony). In this process, neurons of distributed neocortical memory fragments are activated synchronously by gamma oscillations. The different embedded stimulus features of the fragments constituting a specific memory are brought together transiently in this way. By this process, first advanced by Christoph von der Malsburg, memories presumably are recalled. These include the thoughts and perceptions of waking as well as dream memories.
It seems likely that certain slow brain waves of REM sleep—the hippocampal theta waves (4-8 cycles per second)—trigger the serial linking of individual memories, recalled as described earlier, to form connected, narrative dreams. Reinforcement of memories during REM sleep and of enormous numbers of memory fragments during NREM sleep presumably occurs continuously. On any given night, however, only a small fraction of the reinforced circuit contents rise to the level of unconscious awareness, that is, enter dreams.
For memories, this limitation probably is because too many of them exist to be accommodated in one night’s sleep. For memory fragments, it is because the content of an isolated fragment is essentially meaningless. The existence of such limitations in the information conveyed by unbound fragments in the visual realm is illustrated by the “blind sight” of brain-damaged patients. Fragments of visual information are received and registered in the visual cortices of these patients, but because the fragments do not become bound, they are not accompanied by visual awareness.
For example, although certain patients cannot see a moving object in their visual field, they nonetheless become aware that movement is occurring there. In this connection, the activation of enormous numbers of memory fragments without temporal binding or awareness probably is how synaptic strengths in the vast majority of memory circuits become reinforced during NREM sleep.
Dream contents give clues to the brain’s priorities for circuit reinforcement during sleep. High priority goes to consolidation of recent occurrences, particularly significant events of the same and previous days. Highest priority is given to recent emotional events and/or actions and perceptions with significant survival value.
Illusory Contents of Dreams
From 85% to 95% of REM dreams are ordinary and mundane, with authentic, highly visual, and dynamic contents. Waking perceptual experiences reappear with remarkably lifelike details. G. W. Domhoff asserted that the waking mind and the dreaming mind seem to be one and the same. The main thesis of this entry is that many failures of our minds, as exemplified by illusory dreams, also have “one and the same” immediate cause— defective synaptic strengths.
Defective synaptic strengths have their origin in two broad categories of disruptive influences. The first category includes failures of intrinsic origin, owing simply to system complexity. Such failures are common even in much simpler cellular and subcellular systems. Also included is a normal weakening of synaptic strengths in neocortical networks with time (normal processes of decay). The second category of disruptive influences, which is not treated here, is due to pathologically altered brain waves.
It is most relevant, then, to ask what occurs when the strengths of a small fraction of the synapses—of the millions that may encode a given memory—are weakened slightly from their dedicated ranges. In answer to this key question, one would not expect a scene, for example, to be degraded beyond recognition as such. Rather, novel, unpredictable, probably relatively minor alterations would be expected—distortions, background or location ambiguities, altered identities, and so forth. A face, for example, still would be a face, but the alterations might make it unrecognizable.
Synaptic strengths in some memory circuits, even though adequately reinforced during sleep, probably accumulate chance errors in their records of dedicated strengths with time. This would place our general forgetting of our oldest memories over the years on a firm foundation, probably owing to the accumulation of these synaptic defects. Such an imperfection might even be favored by natural selection, as it would tend to eliminate useless memories of the distant past, freeing otherwise encumbered cortical tissue for the storage of new memories.
Static and Continuous Dreams
Most frequently, compatible circuits addressed by dream-producing mechanisms have overwhelmingly authentic contents. Temporal binding of small numbers of circuits with weakened synaptic strengths might lead merely to such discrepancies as unrecognized people and places, altered times, slightly altered thoughts, and so on. These dreams would not be considered illusory.
Illusory dreams, on the other hand, probably trace to the inclusion of older, variously incompetent memories and their fragments, often harking back to childhood experiences. These would contain greater numbers of defective synapses, leading to the incorporation of faulty thoughts and perceptions (beyond the mere minor discrepancies mentioned earlier). But such direct effects are unlikely to be the only cause of illusory dreams. Disruptive indirect effects probably also occur when flawed memories become serially linked, which inevitably would lead to incoherent connectivity.
When memories remain isolated in dreams (i.e., unlinked), the dreams are said to be “static” or “thoughtful.” Such memories typically occur in many dreams during adult NREM sleep, in children’s dreams up to about 5 years old, and in the dreamlike experiences that accompany some seizures and artificial brain stimulation. The means by which the unlinked memories of static dreams become serially linked to form dreams with narrative continuity can be dealt with only in broad outline.
Consider a dream that begins when gamma waves temporally bind certain memory fragments into an initial memory. Activation of a second memory would be expected not to be random but to be biased toward including other memories formerly associated with the first one and similarly for subsequent selections. Serial linkages between these memories, then, would possess narrative continuity.
One expects biases to exist in the activation of memories being bound and serially linked. In forming some “day residues” of dreams, in particular, the linkages conferring continuity would have been in effect only hours earlier, while awake. Accordingly, serial links must leave traces (temporally fortified connections) that guide subsequent binding and linking mechanisms.
The same reasoning would apply to old memories in long-term storage. Their ordered arrangement in narrative dreams implies that great numbers of much older, favored memory associations also persist. But any influence that tended to randomize the activation of fragments to be temporally bound, and the serial linking of the resulting memories, would be expected to favor the production of illusory dreams.
The circumstances in which temporal binding and serial linking mechanisms are subverted, leading to illusory dreams, occur only rarely during waking. This is to be expected, as few circuits employed during waking are for very old memories. Were it otherwise to any significant degree, dreamlike hallucinations, bizarre thoughts, and false memories—such as those that occur in certain pathological states—might not be uncommon in normal, awake individuals.
To further consider the basis for activation of memories whose fragments enter dreams, the process is treated as a three-step affair. In the first step, NREM slow waves (up to 14 cycles per second) address and activate circuits containing fragments of memories, thereby reinforcing the strengths of their internal synapses. However, the contents of the vast majority of these fragments do not necessarily enter dreams. Such nonentry is the probable primitive condition in animals that sleep. Dreaming takes place only if the second step occurs. In this step, proposed to be the more advanced condition in sleeping mammals and birds, REM gamma oscillations temporally bind the fragments together to yield memories.
Were the process to terminate with the second step, only a static dream, an isolated thought or perception, would occur, as is typical in NREM dreams. In the third step, the building blocks become serially linked by the actions of REM theta waves, thereby producing continuous, often narrative, and authentic dreams.
Vital Clues From Children’s Dreams
The conclusion that temporal binding can occur without serial linking has a compelling basis. Whereas bound memories always exist in dreams and dreamlike experiences, serial linking of them may fail. For example, up to the age of 5 years, the dreams of children consist largely of isolated occurrences, described as static thoughts and images (often of familiar animals). This finding suggests that serial linking mechanisms are not yet fully developed at 5 years. Because theta waves are present at earliest ages, some additional mechanism, perhaps functional completion of nerve-fiber myelinization, may be necessary for serial linking to occur.
Inasmuch as David Foulkes reported dreams by the typical child of 3 to 5 years old, in only 15% of awakenings from REM sleep and in none from NREM sleep, it appears that even the temporal-binding mechanism is only infrequently present at these ages during REM sleep and is absent during NREM sleep.
In a dramatic change, a “storylike” format in which characters move about and interact in a dynamic dream world closely modeled on the real world, often with the dreamer participating, begins to appear in children 5 years old. This format contrasts with the earlier static dream content, almost wholly lacking in social interactions and the presence of the dreamer. These changes accompany the well-known “5 to 7 shift” in children’s cognitive competence and functional completion of nerve myelinization—probably also when the serial memory-linking mechanisms mature. These also are roughly the ages when sufficient competence for formal schooling develops.
The other major time of prevalence of dreams without connectivity is during adult NREM sleep, when only 5% to 10% of dreams are indistinguishable from REM dreams. The other 90% to 95% of NREM dreams (themselves of infrequent occurrence) often are described as less visual, less vivid, less emotional, less bizarre, or as static or thought-like, and containing more day residues. The finding that only 5% to 10% of NREM dreams have connectivity presumably reflects the very low level of oscillations in the theta and gamma ranges.
Young children’s dreams portray activity of the same general kind that waking children perform or observe. Their dream maturation proceeds in an orderly manner, both reminiscent of and temporally associated with the unfolding of their other complex mental operations. With weakened or otherwise defective synapses accumulating with age, there should be relatively few weakened synapses in young children, whose dreams should reflect minimal or no memory-recall failures.
These deductions conform to Foulkes’s findings with children 3 to 15 years old, as they reported no illusory dreams. But dream distortions—dreams containing unrecognized people, animals, and places—were common. These did not begin to occur until 5 to 7 years of age. Earliest dreams, at 3 to 5 years old, were constrained and impoverished by cognitive immaturity. Almost without exception, however, dream contents at those ages were authentic, containing some few recognized family members and familiar settings, but mostly animals familiar from fairy tales and cartoons.
Dreams containing highly authentic memories at 3 to 5 years of age can be understood in terms of the existence of relatively few weakened synapses. These years also are times of intense synaptic pruning, as circuits become fine-tuned for dedicated functions. Mere dream distortions at 5 to 7 years reflect the expected minimal weakening of synaptic strengths. The frequency of occurrence of unknown people and places generally increases with age through (and beyond) adolescence, doubtless reflecting the accumulation of defective synapses.
Evidence of small numbers of weakened synapses also characterizes adult dreams, for which unknown people and places are a hallmark. Subjects typically describe objects or persons that have specific unidentifiable visual features. In one study, for example, people were recognized and recalled in only 20% of dreams—usually the familiar face of a relative, friend, or colleague.
Some fanciful explanations proposed by Freud and others for children’s dream distortions were discounted by Foulkes’s studies. However, his alternative explanation appears no less fanciful. He regarded the distortions as tracing to children’s increasing ability to imagine unfamiliar people and places as their cognitive processes mature. A more firmly grounded explanation would hinge simply on the presence of increasing numbers of defective synapses—the greater the numbers are, the older the child will be.
It follows that children experience less interference with the dream process than do adults because children have had less time for interfering influences to develop—the younger the child is, the less time and the less interference there will be. In consequence, time can be seen to play a crucial role in the authenticity of children’s dreams as compared to those of adults.
J. Lee Kavanau
See also Amnesia; Consciousness; Dreamtime, Aboriginal; Memory; Sleep; Time, Subjective Flow of
Buzsaki, G. (2006). Rhythms of the brain. New York: Oxford University Press.
Domhoff, G. W. (2006). Dream research in the mass media: Where journalists go wrong on dreams. Scientific Review of Mental Health Practice, 4(2), 74-78.
Foulkes, D. (1982). Children’s dreams: Longitudinal studies. New York: Wiley.
Kavanau, J. L. (2002). Dream contents and failing memories. Archives Italiennes de Biologie, 140, 109-127.