Thursday, October 01, 1998

Forbidden Zones: Consciousness in Abnormal States


By: J. Allan Hobson M.D.

Sigmund Freud move over. The search for a new model of consciousness is attracting psychologists and physiologists, artificial intelligence researchers and physicists. Prof. Hobson, an authority on sleep and dreaming, joins the quest with a hypothesis of brain-mind unity intended to address inadequacies that “impinge critically on brain-mind science.” In this excerpt, he applies his theory to abnormal states from coma to schizophrenia.


From Consciousness by J. Allan Hobson. To be published by Scientific American Library, © 1998. Used with permission of W. H. Freeman and Company


As new data on the brain become available, scientists and scholars of every stripe are searching for new insights into the relationship of brain and mind. “Never has the brain-mind problem been so hotly pursued as today,” writes Professor J. Allan Hobson. The result has been “a glorious variety of approaches and ideas now in ferment around the problem of consciousness.” At a recent conference in Tucson devoted to models of consciousness, more than 500 papers were presented, exploring the problem from the points of view of neuroscience, psychology, philosophy, artificial intelligence, philosophy, and mathematics—to name a few.

To see what is at stake, one need only consider the impact of an earlier model of consciousness—that of Sigmund Freud and his disciplines. As yet, no schema of remotely comparable impact has taken hold in our era, which may account for both the “glorious variety” of views and the fervor of the debate. The difference is that today’s architects of a model of consciousness must reckon with the data pouring forth from brain research. Prof. Hobson writes: “Brain science promises unanticipated revelations, revelations that will change our view of the mental world... Our future evolution as philosophers, psychologists, physiologists, spiritualists, moralists, and theists will be shaped by brain science in ways that we cannot yet imagine.”

With this introduction to his forthcoming book, Consciousness, Prof. Hobson begins his own ambitious construction of a new conceptual model of consciousness. For Prof. Hobson, the goal is “brain-mind unity in the study of conscious experience”: a “unifying framework” to aid in understanding both its normal and abnormal states of waking, sleeping, and dreaming.

In this model, normal conscious states result from the interplay of 

Activation: the energy, or processing rate, of consciousness—a concept as applicable to the rate of information processing as to the frequency of brain wave oscillation

Information: the input and output of information or signals and their source in the world or in ourselves

Modulation: the mode of information processing, such as attention, memory, thought, or volition.

Each of these factors, Prof. Hobson argues, can be given a clear physiological and psychological meaning—a “brain meaning” as well as a “mind meaning.” When they are pictured as three dimensions in space, our state of consciousness at any given time can be visualized as a point within that space, its position determined by the degree and kind of activation [A], information [I], and modulation [M].

Thus the full, waking consciousness of morning activities may involve high activation, external stimulation, and motor output, “the AIM conditions which society has chosen as optimal for the performance of what is called work.” In contrast, as we drift downward to the edge of sleep, there is a “sudden drop in external input-output traffic, A and I functions plummet...” Only with the onset of REM sleep and dreaming does A regain its waking intensity—but now with exclusively internal data.

Thus, writes Prof. Hobson, “the conscious state paradigm is now an effective set of working hypotheses about how the brain effects change in consciousness.”  In the excerpt below, from Chapter 8 (adapted) of Consciousness, Prof. Hobson puts his own model to the test, asking how well it applies to the multitude of states of consciousness that we describe as “abnormal.”


Woe be to those who by virtue of some involuntary genetic spin or some injudicious voluntary folly, find themselves in such forbidden zones as waking hallucination (as in schizophrenia or LSD psychosis) or delirium (as in alcohol or amphetamine withdrawal) or, worse yet, irreversible unconsciousness (as in the many forms of coma caused by brain disease and trauma).

Each of these conditions can be seen as an exaggeration or aberration of one or more of the state space axes.  Hallucination is an imbalance of Factor I (information) such that internal stimuli normally held in check during waking are released as they normally are only in dreaming. Delirium is an imbalance of Factor M (modulation) as a consequence of the ingestion of drugs that alter the normal metabolism of the brain’s neuromodulatory system.  This causes not only hallucinosis but also such dream consciousness components as disorientation, recent memory loss, and confabulation.  Coma is mainly the result of depressing Factor A (activation), the defective activation arising from a direct loss of neurons of either the thalamocortical end-organ of conscious experience or the brain stem systems that energize and modulate it.

The AIM state space model, of course, is too limited a construct to give even as summary an account of many of these pathological conditions as it does of normal dreaming.  But it does solve one important problem that no other formulation has even approached: the integration of apparently very disparate conditions into a unified conceptual framework. Several important advantages come with conceptual unification, and these are well worth the risk of oversimplification, a sin of which AIM is sure to be accused.

One clear advantage is functional continuity between the normal and abnormal states.

Of course, structural defects of either an intrinsic (genetic) or extrinsic (drugs) nature can be critical determinants of pathological conscious states. But they all act upon and through the brain systems that mediate the normal states that we have already described.  Furthermore, they act on those systems in ways that the normal vicissitudes of consciousness help us to recognize and understand.  And that understanding can be both technical and phenomenological.


If, as physicians—engineers of conscious states—we want to fix a given condition, like hallucinosis, we can look at the way the brain itself clamps internal stimuli and then simulate this clamping process.  (In fact, most treatments that work push Factor M in the waking direction.) If, instead, as psychotherapeutic empathizers, we wish to more convincingly identify with our fellow humans who are, say, delirious or otherwise psychotic, we have only to pay closer attention to our own dream consciousness to know, at first hand, how anxiety provoking and disorganizing is delirium.

We are all floating through the same state space with the same relative degree of helplessness. AIM provides a map that all of us lost souls can use. And, when lost, any map is better than none. 

There is a strong moral payoff in this approach, too, since it makes unconsciousness (as in deep sleep) and madness (as in dreaming) universal rather than special.  Once we get over the blow that this recognition deals to our pride, we can take satisfaction from the humble recognition that normal consciousness is—normally—subjected to deformations and distortions of pathological proportions.  Gone forever is the false comfort of the “we the sane, they the mad,” by which we have isolated, alienated, and confused our poor conscious state relatives for centuries.


In making this point I am not advocating the naive and irresponsible excesses of the deinstitutionalization movement.  Quite the contrary.  I am an advocate for more care, not less, and will even accept temporary restrictions of civil rights in the interest of helping to move people with disabled states of consciousness into functionally adaptive state space domains. Armed with the AIM model, I can approach my fellow man with a sense of true conscious state equity.  We are all floating through the same state space with the same relative degree of helplessness. AIM provides a map that all of us lost souls can use. And, when lost, any map is better than none.

A second advantage flows from the recognition that the consciousness state space is not only universally open to all, but universally continuous and universally dynamic. Just as the normal and abnormal states of consciousness are seamlessly linked to each other, so are the abnormal states more flexibly interchangeable than one- and two-dimensional models now in common use can accommodate.

Thus, rather than imagining psychopathology as an endless list of discrete conditions, each with its own particular cause and specific treatment, we envisage one single state space containing an infinitude of possible positions within it. And, just as no normal human is always in waking consciousness (or always in dreaming consciousness), but forever moving within and between such states, so are those who enter the forbidden zones always moving. Thus they may resemble one another more than themselves from one day (or psychotic episode) to another.

Here, again, a warning is in order.  In advancing a kind of psychopathological universalism I am not saying that there are no fixed points, no anchorages in the state space. Indeed, because there are preferred domains, diagnostic precision can be both valid and reliable.  But, as we say in New England, “if you don’t like the weather, wait a minute”!  And be careful: Today’s anchorage may be tomorrow’s shoal, especially since the chemical interventions chosen to alter consciousness may alter the brain as radically as the marine tide alters a shoreline.

A caveat to the reader of this formulation. The classification of coma as a disorder of activation, or depression as a disorder of modulation, is rather arbitrary.  This is because it is impossible to change one dimension of state space without changing others. The spirit of this discourse is to celebrate such dynamics and the diversity that they ensure rather than to enshrine a new orthodoxy with the same rigidity as the older structural models of consciousness.


A move to the left in the state space (signifying decreasing activation) theoretically impairs both waking and dreaming consciousness, although, in practice, it is impossible to have an account of dream consciousness without passing through the domain of waking.  This assumption that dreaming is lost is supported by evidence that abnormally deactivated states tend to resemble deep sleep in showing a high-voltage, slow EEG pattern more or less continuously. The most terrible and irreversible destructions of consciousness are the comas caused by structural damage to either the upper brain, by tumors or vascular insults, or to the brain stem (most commonly trauma). In both cases, the activation function is lost because the brain cells that mediate it are killed.  Thus comas can be prolonged, sometimes for years or even decades, without significant improvement.

The brain, so richly profligate in number of cells, so profuse in linking them into functional networks, does not replace killed cells with new ones as does almost every other tissue in the body.  Brain activation is also exquisitely sensitive to decreases in oxygen and blood sugar levels. And the brain has a fabulously elaborate vascular system to ensure the delivery of these vital fuels of consciousness. The breath-holding games of adolescence (which are designed to impair the activation of consciousness by voluntary anoxia) are usually reversible because the muscle atonia relieves the strangulation and oxygen levels then return to normal within seconds. And even the coma of diabetic hypoglycemia, which may last for hours, or for days, is fully reversible.  This sugar shock coma has, in fact, been voluntarily administered to human patients in a vain effort to alter the unwanted states of consciousness seen in schizophrenia.

Brain cells need be deprived of oxygen for only six minutes to be killed. The irreversibility of coma is caused by the anoxia of internal strangulation caused by strokes that close off blood vessels; by tumors that compress brain tissue; or by trauma that causes brain swelling.  To damage the thalamocortical system of the upper brain sufficiently to cause irreversible coma, the anoxia must be very widespread.  This occurs most easily when carbon monoxide, inhaled accidentally from unvented gas heaters, or in automobile fume suicide attempts, displaces oxygen from the blood-hemoglobin carrier.

Like carbon monoxide, encephalitis viruses are small enough and numerous enough to go everywhere in the brain, wiping out the cortex simply by usurping the cells’ DNA.  For strokes to have widespread effects many blood vessels must be closed off  and, even when this happens—in old age, for example—only some aspects of consciousness are lost. Because the brain stem is smaller and contained in a smaller space, it is easier to deal its neurons fatal blows than it is to kill the vast thalamocortical system that fans out overhead. This is why the twisting and tearing injuries of head trauma are particularly unwelcome in the brain stem. And they can be aggravated when damage to the upper brain—by tumor, stroke, or trauma—causes acute swelling.  Since the upper brain cannot expand upwards, it expands downward into the brainstem’s space, causing devastating damage to those smaller, critical state-control neurons that occupy it.

The disturbing theory that his behavior evoked in all who saw him was that he was conscious but unable to communicate with us. One reason for taking this idea seriously was that his upper brain showed the activated EEG pattern seen in normal waking consciousness.

Wherever and whatever the pathology, when the capacity to activate the organ of consciousness is lost, consciousness is altogether and unequivocally absent.  In some brain stem injuries, however, the results are not so clear cut. One of my high-school classmates, Sam Vesill, sustained a traumatic brain stem injury in an automobile accident in the late 1950s. Thereafter, Sam remained in a troubling state of suspended animation for 30 years.  He lay in his hospital bed immobile but reactive to events in his surroundings.  Because he made roving eye movements that often seemed to follow visual stimuli, his caretakers, his family, and his friends could not be sure he was not conscious. The disturbing theory that his behavior evoked in all who saw him was that he was conscious but unable to communicate with us. One reason for taking this idea seriously was that his upper brain showed the activated EEG pattern seen in normal waking consciousness.

This upsetting picture, called coma vigil (or akinetic autism) occurs when the central activating structures [A] are intact, but their links with the input-output system [I] are broken. As with babies and animals that cannot speak, we cannot know what kind of conscious experience they might have.  This uncertainty raises many moral dilemmas because consciousness is the attribute that we value above all others. Only when we feel sure that consciousness has been lost forever are we able even to consider terminating heroic and costly life-support programs.  The evidence that convinces some of us to abandon hope is the failure— ever—to show enough electrical activation of the brain to make the hypothesis of consciousness tenable. In these cases—when the brain waves disappear from the EEG—we refer to the individual as brain dead. Then we question the wisdom of keeping the unconscious body alive at whatever cost.  Since the evidence is not now and may never be adequate to convince everyone, it is essential for individuals to make their own wishes explicit, in advance, in a living will. It is equally important for society to determine democratically whether or not an individual has the right to die.


Our twin boys, Andrew and Matthew, now 15 months old, are as different as any two brothers because each was hatched from his own special egg fertilized by his own special sperm. Because they are growing up together in the same environment, it is easy to observe the marked differences in activation level that distinguish them. Andrew is calm and docile. He sits contentedly, focusing his emerging consciousness upon the objects and people in his small world; it is easy to imagine him in later life as a scholar satisfied to explore a narrow subject in depth. Matthew is active and demanding. He wants to move.  And he needs attention from others so much that he vocalizes, makes faces, and hurls objects constantly. It is easy to see him as a sportsman and adventurer later in life.  While Matthew now walks daringly unsupported, Andrew still prefers to play it safe and crawl.

These different behaviors are the presumptive signs of different brain-mind activation systems.  Time will tell where Andrew and Matthew lie along the continuum that distinguishes normal short sleepers (like their mother, who prefers to get six but can do with four hours of sleep) from normal long sleepers (like their father, who gets by with seven but feels better with eight or nine hours of sleep). Time will tell if the resulting five-hour difference in time spent in waking consciousness will be used for the contemplation and internal signal processing that their father prefers or for the external signal processing and interactive work upon which their mother thrives.

And time will tell whether or not they will exceed normal limits in the degree or kind of activated consciousness they experience.  The inability of some children, like Matthew, to sit still and to concentrate can evolve into the syndrome of hyperactivity called attention deficit disorder (ADD).  This disorder is mediated by hyperactivation of the perceptual and attentional modules of consciousness. Because in ADD every external signal has easy access to consciousness, none can be easily excluded and consciousness is a slave to the world and its data.

Hyperactive children thus have a disorder of the attention component of consciousness that current research of Michael Posner conceptualizes in a tripartite model. The localization of external stimuli is performed by networks in the posterior thalamocortical system, but the identification of those stimuli is made by more anterior circuits.  This “Where is it? What is it?” analytic sequence functions optimally when internal vigilance signals are reliably supplied to the upper brain by the brain stem. The vigilance signals are carried by norepinephrine, the aminergic neuromodulator made by the locus coeruleus, which works by toning down the thalamocortical system so that consciousness can be more selective in paying attention.

Already, the perspicacious reader will realize that this hyperactivation disorder [A] is the product of defective modulation [M] and that it is manifest by an over inclusion of external stimuli [I]. One implication is that both activation level and input level must be tempered by appropriate modulation to hold AIM in an optimal position within the waking subspace. In other words, it is not only the overall state of consciousness that is determined by AIM but its microscopic substrates, as well.  Another, more practical, implication is that the overactivation and over-inclusiveness of hyperactive conscious states can be counteracted by modulating consciousness artificially.  For instance, drugs like amphetamine, which mimic the action of norepinephrine and thus raise the flagging level of M, are the pharmacological treatment of choice for ADD.

A hyperactivation problem that is complementary to overinclusion of external stimuli during waking, as in ADD, is the failure to quell internal stimuli at the end of the day, when sleep is desired.  The result is inability to deactivate the thalamocortical system enough to allow it to go into the oscillating mode that pulls the brain-mind into the unconscious depths of NREM sleep.  The consciousness of insomniacs, over driven by internal emotional stimuli during the day, cannot shut down at night. Such individuals either take their trouble to bed with them or fail to go to bed at all because their internal stimulus processing is never done.  They know it will not stop even if they lie down and turn out the lights.

In insomnia, consciousness is overly activated [A], internally stimulated [I], and hyper modulated [M]. Remediation can be achieved by relaxation training.  This substitutes neutral information for arousing internal emotional stimuli, allowing normal deactivation [A] to occur. This can be abetted by sedative hypnotics that aminergically demodulate the brain [M].

But even when they have deactivated the brain enough for signs of EEG sleep to appear, they may not completely shut down the engines of internal emotional stimulation. The result is a dissociation of the normal coupling of activation [A] and input-output [I] mechanisms, so that their subjective experience is not peacefully oblivious but disturbed by the continuous cerebral noise of unproductive rumination and worry.  This condition has been called “pseudo insomnia,” as if the complaints of not being deeply or restfully asleep were invalidated by the evidence of EEG inactivation.  But since we know that low levels of conscious experience persist even in very deep NREM sleep, it is clear that the level of activation alone is not enough to decide the case. Instead of impugning subjective experience, we should admit the inadequacy of our one-dimensional model. By adopting the 3-D model we are in a better position to understand and respond to the problem of excessive consciousness during sleep.

Insomnia is a subtle and poorly understood disorder of consciousness that is often relegated to the clinical ash-heap. When the internal stimuli that invade consciousness are clearly of physiological origin, however, we are more easily able to fit the condition into classical models of disease. 


Insomnia is a subtle and poorly understood disorder of consciousness that is often relegated to the clinical ash-heap.  When the internal stimuli that invade consciousness are clearly of physiological origin, however, we are more easily able to fit the condition into classical models of disease. One such condition is epilepsy.  In epilepsy, consciousness is literally seized and held hostage more or less completely by uncontrolled hyperexcitability in one of the brain’s internal circuits.  Another condition is narcolepsy, in which the normal boundaries between waking and dreaming consciousness are dissolved. Like epilepsy, narcolepsy involves the release of normally inhibited neural circuits, which then generate internal stimuli and pre-empt motor output neural circuits.

Epilepsy has been called “Fire in the Brain.”  From the vantage point of the AIM model of conscious states, the brain fire consumes consciousness in one of two ways. If the focus arises in the thalamocortical system it can pre-empt consciousness completely by driving that system into a sleep-like oscillatory mode. This is generalized epilepsy.  It is called “grand mal” to indicate that it is big (grand) and bad (mal). In “petit mal” epilepsy, the seizure is more partial and short-lived. This leads to temporary lapses of consciousness, the so-called “absences” of petit mal (or not so bad) epilepsy.

If the seizure focus is elsewhere in the brain, in the temporal lobe for example, it is more likely to invade waking consciousness, contaminating it with internal emotional stimuli that, like the temporal lobe stimuli of dreams, are incorrectly assumed to come from the outside world.

Because they so clearly illustrate the basic assumptions of the AIM model, temporal lobe seizures are among the most instructive invaders of brain and consciousness.  Consider the following similarities and differences between normal dream consciousness and the abnormal consciousness of temporal lobe epilepsy:

  1. Both dreaming and temporal lobe epilepsy involve paroxysmal activation of the limbic brain. Both therefore are characterized by sudden, spontaneous emotions such as fear, elation, rage, and sexual excitement.
  2. The internally generated emotions are integrated by paralimbic cortical structures.  In the case of epilepsy they are integrated with ongoing reality (because the rest of the brain is still awake); in the case of dreaming they are integrated with recent and past memories (because all of the brain-mind is absorbed in the REM sleep process).
  3. Both dreaming and temporal lobe seizures are associated with a loss of insight, indicating that consciousness is held hostage (i.e., “seized”) in both conditions by such strong internally generated stimuli that the brain-mind uncritically projects the synthesis of the signals onto the outside world.
  4. Both the normal and abnormal states are associated with a suspension of memory for the experience itself, even though, in both states, the recall of past events may be present. Since both conditions share the paroxysmal takeover of the limbic lobe, this suggests that the mechanism of the amnesia for dreams may arise from Factor I as well as from Factor M.  In other words, the memory system of the limbic brain may be driven so powerfully into the playback mode by internal stimuli that it cannot store the newly synthesized product of consciousness in short-term memory.

Given these striking correspondences, it is difficult to decide whether dreaming results from normal temporal lobe seizures or temporal lobe epilepsy triggers abnormal dreaming.  One reason for taking the first hypothesis seriously is that the physiology of the internal stimulus generation, in driving normal dream consciousness to its hallucinatory and delusional experiences, is decidedly seizure-like.  PGO waves that arise in the pons and are conducted to the amygdala and visual brain during dreaming have precisely the same spike and wave form as the EEG stigmata of epilepsy.  And we know that at the cellular and molecular level both of these EEG signs of internal stimulus generation are produced by failures of inhibitory restraint.

As for the second hypothesis, the use of the term “dreamy state” to describe the consciousness of temporal lobe seizures speaks volumes about the deep correspondence between two conditions that we would prefer to believe are very different from one another. Subjects undergoing brain surgery for temporal lobe seizures also report dreamlike experiences when the neurosurgeon stimulates their temporal cortex electrically.  This means that remote memories, feelings, and sensations can be experimentally induced by simulating the internal stimuli of both dreams and temporal lobe seizures.  This classical experiment of clinical neurobiology takes on a new meaning when seen through the lens of the AIM model.


If normal dreaming were seriously considered the result of the epileptiform discharge of normal neurons, we might expect to see full-blown REM-sleep dreams suddenly replacing waking-consciousness. And this is exactly what we do see—in narcolepsy, a condition whose name suggests its seizure-like nature.  In narcolepsy, the pontine REM-sleep generator is inadequately inhibited. Thus it becomes possible for external stimuli to trigger attacks (or seizures) of dream consciousness during the daytime, when in most of us they are impossible.  Although the-fundamental mechanism is unknown, it appears to be a deficiency of dopamine—another one of the aminergic neuromodulators so crucial to maintaining waking consciousness.

While many narcoleptic attacks occur spontaneously, many are precipitated by external stimuli that evoke strong emotions, especially surprise, elation, humor, and—our old friend—sexual excitement.  Funny?  Yes, in a way; but not funny for the subject, who in addition to being plunged into the dream state, loses contact with the world with embarrassing and annoying abruptness.  Not only that, but muscle tone may melt away, adding the risks of bodily harm to those of social opprobrium. Why are these emotional stimuli so potent? Possibly because they activate the limbic circuits that normally participate in mediating emotion. These circuits may have intimate connections with the pontine REM generation via gates that are normally closed during waking. This AIM-derived hypothesis makes a three-way link between normal dreaming, temporal lobe seizures, and narcolepsy—all seen as the consequence of unchecked paroxysmal discharge of limbic lobe neurons, and all sharing many specific conscious state features.


Narcolepsy is instructive for a second reason. It exaggerates normally partial dissociations of wake and dream consciousness. Examples are visions occurring at sleep onset and offset (respectively called hypnogogic and hypnopompic hallucinations) and sleep paralysis (inability to move upon awakening from dreaming). All three of these dissociations are more disturbing to some subjects than the full-blown attacks because they are both more intrinsically frightening and because they raise the specter of mental illness. While having hallucinatory experiences within dreams is analogous to psychosis, having them while awake is identical to psychosis. 

Because the hypnogogic and hypnopompic hallucinations of narcolepsy are predominantly visual they are both gripping and terrifying. The subject may suddenly see people or animals in the bedroom, a place usually safe—even sacrosanct—and this vulnerability naturally increases their dread.  This panic anxiety is no less intense when, on awakening from a frightening dream, the subject finds himself unable to move, literally.  He is then literally—not just figuratively—frozen in fear.  The negative emotion associated with all of these symptoms of narcolepsy is again reminiscent of the conscious experience of temporal lobe seizures as is the insertion of the emotion into otherwise normal waking consciousness. The inevitable conclusion is that one (or more) part(s) of the brain is awake while the other(s) are dreaming.

In the pluralistic, multidimensional real world of conscious experience, a virtual infinitude of sometimes bizarre substates is the order of the day—and even more so of the night. It is this multiplicity of substates, their often hybrid aspect, and their dynamic interchangeableness that the AIM model addresses. The take-home message from this discussion is that the line between external stimulus “waking reality” and internal stimulus “dreaming fantasy” is always thin, and always shifting. On this view, the line between madness and sanity, however firm, may also be thin as our discussion of disorders of modulation shortly will reveal.  Consciousness is a many-splendored thing, but some of it splendors are both terrifying and disabling.


We now turn to the modulatory (M) dimension to try to explain the most debilitating disorders in neuropsychiatry, the psychoses.  Because the two major defining characteristics of psychosis— hallucination and delusion—are also prominent features of normal dreams, it is clear that there are powerful ways to disrupt what we might call the perceptual and cognitive components of normal conscious experience. Those ways are so commonplace—and so normal—as to suggest that psychosis is anything but the bizarre, recondite, and alien state of mind that folk wisdom has taken it to be.

We need to remember this as we consider the three most frequent and disabling conditions that can affect human beings.  There are two points to this consciousness-raising exercise. One is that the so-called mentally ill are not qualitatively different from the rest of us. The other is that neither are they so different from one another as our descriptions might suggest. From the vantage point of AIM, our patients are simply pulled toward and/or stuck in a part of state space that most of us glide through every night of our lives.  And they, too, can move in and out of these forbidden zones—or move from one to the other.

To understand these points, consider the following facts:

  1. The organic psychoses, in which delirium is prominent, are clearly caused by chemicals that raise havoc with the neuromodulatory brain systems that confine our dreams to sleep. Thus we can all be rendered psychotic simply by tampering with the M function of the AIM model.  And we don’t have to take drugs or alcohol to get crazy in this way. Even a commonplace manipulation such as sleep deprivation can produce this syndrome. And, of course, our dreams are properly considered to be organic psychoses caused by radical changes in M.
  2. The affective psychoses, otherwise known as bipolar or manic-depressive psychoses, are exaggerations of normal fluctuations in mood. Our mood goes up (toward mania) and down (toward depression) all the time. Most of us are lucky enough to experience these fluctuations within bearable bounds. But we should not miss the point that it is natural for mood to fluctuate fairly widely. While each of us may have a genetic predisposition set on the high side (mania), the low side (toward depression), or to oscillate widely (become manic-depressive), life events have an impressive power to push even our most stable moods further up or down than is comfortable. But even when depression of psychotic proportions is clearly triggered by overwhelming loss, are we correct in calling this “functional” as against “organic” psychosis? I think not. All psychosis, including that of our dreams, is functional and it is all organic.
  3. Schizophrenia, the condition rightly most dreaded, has a border with affective illness that is often blurred or invisible.  In schizophrenia, the classical conception holds that affect is flat and/or dissociated from cognition. But even this stigmatic feature is not constant. Many so-called schizophrenics clearly are depressed at certain phases of their lifelong trajectories through state space. And all of us so-called normals frequently find it advantageous to dissociate our thoughts from our feelings. Moreover, we are all capable of flattening our affect if it gets in our way or is unbearable.

The take-home message, again, is that consciousness is multifocal and dynamically regulated by brain processes that are subject to a myriad of controlling and disturbing forces.  In other words, even normal consciousness is a balancing act on the edge of a madness that is never far beneath the surface of our outward calm. Two insights emerge from this view of ourselves.  One is that to achieve the full glory of rational and creative human consciousness, nature has tinkered up a cluster of delicate brain mechanisms, each containing its own risk factor for downside dysfunction.  This being the case, the wonder is not that there is so much madness but so little. Sanity and madness are two sides of the same coin: heads, you are crazy only in sleep; tails, you are also crazy when awake.


We begin our state space analysis of psychosis with consideration of schizophrenia, the most clearly organic of the two so-called functional psychoses—and the one that is perhaps most different from normal dreaming.  These differences are worth noting:

  1. The hallucinations of schizophrenia are much more commonly auditory than visual (as in dreaming).  This means that while the I function may be similarly biased so that internal stimuli gain ascendance and produce fictive perceptions, those false percepts arise in the auditory networks of the temporal regions of the upper brain and not in the visual networks of the occipital region.
  2. The delusions of schizophrenia are far more commonly paranoid than are the delusions of dreams.  Dreams have this paranoid aspect surprisingly rarely, given the high level of anxiety and threat in dreams.  Why should this be? One possibility is that the hallucinated voices that plague the schizophrenic occur in the otherwise normal perceptual world of waking and are naturally projected onto it. The cognitive system needs an explanation for the sense of threat and anxiety, and an obvious one is a malevolent external agency that is assumed to be there even if it is invisible.  In this view, the paranoia is seen as a natural and inevitable response to the perception of threat and accusation in the voices.  In dreams, it is paradoxically easier for us to explain our anxiety and threat because we can visually hallucinate malevolent agencies.  Because our dream pursuers, attackers, and critics are visible we are not stuck for an explanation of the malevolence and our dreams are not paranoid. This may be the surprising benefit of our being more completely psychotic in our dreams than when schizophrenic in waking.

These two differences between dreaming and schizophrenic psychoses point up two important inadequacies of the AIM model:

  1. On the I dimension of the state space, no distinction is made between sensory modalities. A more complete model would have at least five I dimensions, one for each sense modality.
  2. The AIM model does not represent regional brain space with its important diversity, of which sensory modality is but one example. To deal with this deficiency, we would have to visualize an AIM state space in each functionally distinct brain region.  Our state space model, already far more complex than any other yet proposed, is actually grossly over simplified. 

We have considered this deficiency earlier, when we discussed temporal epilepsy as a disorder of the I function.  The locus of the internal stimulus is obviously an important determinant of its modality as well as of its emotional and cognitive impact.

Interestingly, there is mounting evidence that in schizophrenia the temporal lobe is dysfunctional in ways akin to temporal epilepsy. Like schizophrenics, epileptic patients may also be paranoid as they struggle to explain their strong internally generated perceptions in terms of the motives of people in their surroundings.

What about the modulatory mediation of schizophrenic symptoms?  In schizophrenia the leading culprit for the disordered M function is dopamine, an aminergic neuromodulator that does not seem to play a major role in triggering the normal psychoses of dreaming.  Dopamine is produced by neurons that are in the midbrain further forward than in the pons or medulla. But dopamine does share with norepinephrine (and with serotonin) a functional interaction with acetylcholine. This interaction subserves sensorimotor integration in the basal ganglia and related deep motor structures of the anterior brain. This could be a reason for the curious motor signs of schizophrenia such as catatonia and waxy flexibility that were seen before the advent of the antidopaminergic narcoleptic drugs and for the motoric side effects such as dystonia and rigidity.

The enormous benefit to schizophrenic consciousness of reducing the parasitic perceptions and paranoid cognition is often well worth these unfortunate side effects.  Following the introduction of the phenothiames in 1955 the mental hospitals were largely emptied, many were closed.  At the same time, the concept of an M function controlling the balance of internal and external stimuli became irresistible.  The modulatory neurons balance the neural nets so that they can catch real informational fish and let the fictive fish go free.

The different modulatory chemistry of schizophrenia could also help to explain the different conscious experience of its victims.  In particular, the sense of isolation and social distance from people is a disabling symptom that distinguishes schizophrenia from dreams. Schizophrenics are neither integrated as selves nor are they comfortably inserted in the social world. This split is still another variation on the theme of binding that recurs in any discussion of consciousness. Getting consciousness together and holding it together is not an easy task for everyone.

Recently, the M function theories of schizophrenia have had to be broadened, because the newer and sometimes dramatically effective anti-schizophrenic drugs not only counteract dopamine but bolster norepinephrine and serotonin. This suggests that schizophrenic consciousness, like normal dreaming, is a state of the brain-mind caused by generalized and widespread neuromodulatory dysfunction. Only when the unbalanced and deficient neuromodulatory systems are tuned by external chemicals with versatile and widespread effects are the core disintegrative symptoms resolved.  These deep effects upon the unity of conscious experience enable more unified, saner selves to emerge even after lifetimes of disintegration and disarray.


The M function of the AIM model has been used to define the important impact of neuromodulatory chemistry upon the mode of information processing, and especially on the brain-mind’s capacity to remember its own transactions. Thus, we have seen that when aminergic modulation fails (as in REM sleep) we cannot remember our conscious experience even though our consciousness is highly activated during REM. These effects are short-lived, lasting no more than an hour at a time before NREM sleep rescues and restores 50 percent of the lost M function. Six to eight hours later, waking restores 100 percent.

What happens if aminergic modulation is impaired in waking?  And over the long term? In this case, it is mood rather than memory that is the victim, as the affective disorders—better known as manic-depressive psychosis—make clear.

This story of the affective disorders and their relations to sleep and cognition offers the most persuasive argument in the casebook of abnormal conscious states for the state-space concept. The key to understanding the story is recognition that the M function profoundly affects the fundamental energy supply of the brain-mind as well as its mode of processing. In short, mode and mood go hand in hand.  As this concept may seem paradoxical, counterintuitive, and even contradictory to the activation

[A] function, we need to distinguish between the metabolic [M] levels of energy controlled by the aminergic modulators and the electrical levels [A] controlled (mostly) by other neurotransmitters (like glutamate and gaba).

To grasp this difference, imagine a remote cabin whose electrical power is generated by a gas engine. As long as there is adequate fuel, the generator can power a variety of devices including lights, heaters, and (in the case of modern day Henry David Thoreaus) even word processors.  But as the fuel supply dwindles, the lights dim, the house cools, and the computer crashes. In our brain, this happens if we run out of cerebral gas.  Our brightness, our warmth, and our ability to remember all ultimately fail if we become depressed. Depression occurs when the aminergic modulatory system is not topped up on a regular basis—as it is by sleep, especially REM sleep.  As this, too, is a paradoxical concept, we will come back to it when we discuss the surprisingly beneficial effects of REM deprivation on depressed mood later in this section.

In our modular dissection of consciousness, we had only one category for emotion and considered emotion only in reflexive terms. And certainly an emotion, like anxiety, may be triggered by stimuli and so heighten and/or disrupt conscious experience. But mood is pervasive emotion and, as such, relatively independent of stimuli—seeming to arise spontaneously from our depths and determine our response to stimuli rather than being determined by them. When we are high, the world is our oyster: Everything looks and feels good (even if it may be quite harmful or dangerous to us).  When we are down, we see and feel the world as forbidding: Gloom and doom color even vivid inputs brown.

Sometimes our mood may change overnight, but usually highs (as in hypomania and mania) and lows (as in neurasthenia and depression) take days, weeks, or months to develop and to clear.  The difference between emotion as an acute reflex response and mood as a chronic anticipatory emotional filter reflects the fact that each was engineered in a unique way to have a dramatically distinctive effect on consciousness.

To account for all of these differences it is helpful to recognize that modulatory chemicals like norepinephrine, serotonin, dopamine, and acetylcholine all have one (acute) effect at the level of the nerve cell membranes onto which they are secreted and another (chronic) effect on the metabolic machinery in the nucleus of that cell. The translation of the membrane signal into the metabolic message is effected by second-messenger molecules that are knocked off the inside of the membrane by M modulators and tell the nuclear DNA of the cell to make more gene product, thus cranking up one or another metabolic process within that cell. These second messengers are called cyclic nucleotides. The aminergic modulators norepinephrine, serotonin, and dopamine release cyclic AMP, which has energy-mobilizing effects on neurons.  The cholinergic modulator acetylcholine releases cyclic GMP, which has energy-conserving effects.  Too much cyclic AMP gives rise to hypomania, in which consciousness runs the risk of undue optimism. Too much cyclic GMP gives rise to depression, in which consciousness runs the risk of undue pessimism.

Each of us is born with a constitutionally embedded propensity to optimism and pessimism determined by where we sit on the M axis of the AIM model. This contributes to what we call personality. When colored by our moral sensibility and education, it constitutes what we call character. 

Each of us is born with a constitutionally embedded propensity to optimism and pessimism determined by where we sit on the M axis of the AIM model.  This contributes to what we call personality.  When colored by our moral sensibility and education, it constitutes what we call character.  High M people tend to be optimistic and to sleep little, while low M people tend to be pessimistic and feel rested only when they have lots of sleep.

Each extreme has both enviable and problematic aspects.  Short sleepers may get a lot done (because they have so much time in a state of energized, acquisitive, and inventive consciousness) but they are prone to over achievement, anxiety, and collapse into states of nervous exhaustion. Long sleepers may be more sensitively introspective, and hence more poetic in their response to the world (because they spend more time in inward, absorbed, and reflective consciousness) but they are prone to disillusionment, withdrawal, lugubrious rumination, and collapse into states of despair.


High M people can’t easily fall asleep; low M people can’t easily wake up.  Each may need cognitive, behavioral, and pharmacological interventions to move them into the part of the normal state space that chance and habit have conspired to avoid.  Indeed, it is in considering treatment of disorders of consciousness that we can appreciate the power of the AIM model and apply it to our analysis of the risks of living within the constraints of such a system.

Consider psychotic depression.  Of the three major classes of psychosis, it is arguably the least severe from a cognitive point of view. This is because the delusions (there is something wrong with my body) and hallucinations (I smell bad) are not so far from the mark.  But the retardation of energy, with its impairment of memory, can lead to states of cognitive imprisonment akin to dementia. And, worst of all, everything in black: the shame, guilt, low self-esteem, preoccupation with disease, death, dying, and loss cloud consciousness in a life-threatening way.  Consciousness may be so pained by its own state as to prompt the one definitive act of redemption: self-annihilation.

In such emergencies, it is helpful to pump up the aminergic modulators, tone down the cholinergic system, and thus jack up the M function so that consciousness can emerge from the slough of despond. This is exactly what the antidepressant drugs do.  They enhance the efficacy of noradrenergic and serotonergic modulatory molecules either by imitating them or giving them longer to act in the synaptic cleft. At the same time, they diminish the power of the cholinergic system by blocking the action of acetylcholine. And, true to the rule of acute synaptic versus chronic metabolic action, they have effects that are immediate and effects that are delayed.  One immediate effect is to correct the sleep disorder.

Since REM-sleep dreaming is triggered by acetylcholine and restrained by norepinephrine and serotonin, chronically depressed subjects tend to enter REM more quickly, have more intense REM, and stay in that state longer than normals. As a consequence, they have less deep NREM sleep and tend to wake up earlier feeling unrefreshed.  This contributes to their depressed state of consciousness. Anti-depressant drugs fix the sleep problem right away. When they do, the outlook for a delayed but positive effect on mood is brighter!

This two-step sequence implies that while the synaptic effects benefiting sleep are immediate, the metabolic effects benefiting mood are the delayed consequence of them.  In other words, we must sleep every night so that consciousness can be cognitively efficient the next day.  But we must sleep night after night so that consciousness can be effectively balanced over the long term.  Our nocturnal sojourn through the AIM state space thus has at least two linked but distinct goals: short-term cognition (and mood) and long-term energy regulation (and cognition).

One of the paradoxes of chronic depression is the apparently negative effect of REM sleep on mood. Once the aminergic system has failed, running the cholinergic system in REM makes depression worse—in the short term— while depriving of REM sleep makes depression better again—in the short term. So this gambit is instructive but impractical; there is no substitute for balancing both sides of the M system.

For the student of consciousness who seeks insights for day-to-day living, it should be clear that proper care and maintenance of the brain-mind begins with some simple principles of cognitive and behavioral psychology.  To give the AIM system maximal opportunities to enter the beneficial states of deep NREM and REM sleep, it is important to schedule (and program) one’s life so that somatic, cerebral, and conscious state mechanisms can interact positively and harmoniously.  In this great balancing act, the timing, amount, and quality of physical exercise, diet, intellectual exercise and content, and introspective self-reflection and guidance all play their part. With the luck of the genetic draw and skillful management of whatever constitutional hand we are dealt, we may even emerge from the casino of conscious life a winner —or at least with our losses cut to tolerable levels.

About Cerebrum

Bill Glovin, editor
Carolyn Asbury, Ph.D., consultant

Scientific Advisory Board
Joseph T. Coyle, M.D., Harvard Medical School
Kay Redfield Jamison, Ph.D., The Johns Hopkins University School of Medicine
Pierre J. Magistretti, M.D., Ph.D., University of Lausanne Medical School and Hospital
Robert Malenka, M.D., Ph.D., Stanford University School of Medicine
Bruce S. McEwen, Ph.D., The Rockefeller University
Donald Price, M.D., The Johns Hopkins University School of Medicine

Do you have a comment or question about something you've read in CerebrumContact Cerebrum Now.