4 Chapter 4: Unconscious Dungeons
The Essential Neuroscience of Human Consciousness
The Essential Neuroscience of Human Consciousness
Amedeo D’Angiulli
Chapter 4: Unconscious Dungeons
4.1. Autonomous Continuous Brain Activity
4.2. Sleep & Dreaming
4.2.1 Stages and Sleep-Wakefulness Cycle
4.2.2 Neurofunctional Pathways of Wakefulness and Sleep
4.2.3 Spontaneous EEG and Metabolic Activity During Sleep
4.2.4 Phenomenological Consciousness and Dreaming During Sleep
4.3. Sleep-related Dissociated States
4.3.1. Sleepwalking
4.4. Anesthesia
4.4.1. Phenomenology and Behavior Observed During Anesthesia
Chapter 4: Unconscious Dungeons
4.1. Autonomous Continuous Brain Activity
The brain is a funny organ. Despite accounting for just over 2% of the entire body’s weight, it represents 20% of its energy consumption. Metabolically speaking, the brain is a very expensive commodity. Things get even more interesting when you look at the brain’s own energy budget. Up to 80% (average range 60-80%) of the brain’s metabolic expenditure goes to neuronal signaling for basic functions, and less than 5% of that functional budget is devoted to conscious stimulus-response or performance-evoked changes in the brain. Such allocation is constant over time and rarely changes the overall budget distribution of the brain.
If you think about it, the brain’s energy distribution suggests a clear relationship between conscious and unconscious processes in terms of energy expenditure. Consciousness is the tip of the iceberg, supported by vast structural and functional “underground” machinery that consumes most of the system’s resources. The repertoire of possible conscious responses has a much smaller but dedicated resource allocation. This is the foundational idea behind a current big picture framework known as the connectome. The backbone of consciousness is probably made up of extensive underground anatomical and functional networks that define relationships among brain areas. One hypothesis spurred by fMRI research is that the connectome more or less coincides with a large neurofunctional system known as the Default Mode Network (DMN). The DMN is associated with ongoing natural idling or a neuronal resting state, which is reflected in the autonomous (or intrinsic) continuous activity of the whole brain.
The DMN was discovered as a shortcoming of the subtraction method. About a decade after fMRI had become widely used in neuroimaging, many researchers around the world started to notice odd results. Subtraction of fMRI signals between task and control (task minus control) showed in some cortical areas task-related increases in BOLD (activation) for some attention experiments, following the principle that increases in BOLD reflect added mental operations relative to the control, a bit like subtracting activity to detecting a red lightbulb would be assumed to change and increase BOLD signal relative to simply detecting the light from the uncolored light bulb (the pure insertion assumption, see Chapter 2). However, in other cortical areas, the fMRI signal subtraction resulted in signal decreases (). To make it more obvious that it was a decrease in activation, fMRI researchers used reverse subtraction, control minus task, which had negative activation or task-specific deactivation. That is, if the areas were not involved in processes underlying the task one would have just expected no difference in activation with the control. But the results paradoxically showed an insertion of added or different contributing mental operation by the control, which would be paradoxical (since it is the main task that adds operations by definition).
Introducing an alternative design study and using “resting” states, in other experiments, researchers were able to show an opposite pattern in the same brain structures in task-control versus task-resting fMRI signals. Results showed increased BOLD for control-resting states subtractive analysis, with decreased BOLD for the control-task subtractive analysis (i.e., more brain and cortical activity at rest!). In different studies, random fluctuation of activity (increased absolute activation) at rest in these areas was also correlated with decreased attention during the tasks. DMN activity would be continuous during a resting condition and changed when a task was introduced, with an increase in BOLD in task-related areas. The deactivation in resting states is not random, but consistent and replicated within the same subject in the same neural networks. Although there are variations in the neuronal networks activated from condition to condition, and bearing in mind individual differences, three structures always seem to be involved: the precuneus, posterior cingulate, dorsal and ventral media prefrontal cortex.
An important question that emerges, however, is what exactly defines a “resting” condition. Passive control conditions used in many experiments, like visual fixation or eyes closed, cause deactivation of that neural network. This is important for two reasons: first, the task-dependent deactivation may reflect a decrease from initial activation that goes on at rest. If this hypothesis is correct, logically, task-dependent activation is tracked by increases and decreases in fMRI signals. (A similar debate occurs over whether the “work” done by neural tissues during cognitive processing happens where the blood is supplied to, where it departs from or even in both sites).
To better understand what fMRI changes mean in the DMN, researchers decided to turn to PET. The physiological basis of BOLD signal is the oxygen extraction fraction (OEF). Recall that as blood supply increases, the ratio of oxygen consumed to oxygen delivered becomes smaller. In other words, when blood flow increases locally in association with a task, activation can be defined as a momentary local decrease in oxygen extraction or, similarly, a momentary increase in oxygen availability. Thus, activation increases when OEF gets smaller. OEF is uniform across the brain except (slightly) in the visual cortex. PET can be used to precisely quantify OEF across the brain. When researchers assessed whether OEF was different in the DMN than other areas, they discovered that it was identical! The conclusion that many came to was that the deactivation in fMRI indicates a change in ongoing sustained intrinsic functionality (hence the term default functions) in the DMN areas that is attenuated by the introduction of a task.
Ongoing random fluctuations (noise) in the BOLD signal and resting state EEG show a remarkable level of coherence and correlation in the DMN. This means that fluctuations of spontaneous resting activity over time are synchronized in those brain areas. More specifically, a repeated pattern of spatial coherence in activation and deactivation around a zero baseline (as referred to % BOLD change) is identifiable in DMN areas. The decrease of activity over time during a task shows a virtually identical correlation with random “noise”. It’s very important to note that “noise” of the default networks is inversely correlated with the dorsal attentional systems, also known as task-positive networks. This is because these other attentional centers show increased activation during tasks that require attentional engagement. Thus, intriguingly, this “noise” has a deeper meaning in terms of ongoing intrinsic activity that always runs in the background . Harking back to the connectome theory, it’s possible that most energy is used in background processes with some minute amount allocated to “emotional or cognitive moments”. These would require consciousness, be it in terms of effort, planning, monitoring or even evaluation of the task at hand. Comparing these moments of consciousness to “The tip of the iceberg” seems particularly apt.
There are still many issues to iron out before knowing what default autonomous ongoing activity in the brain really means. Spontaneous spatially coherent patterns of ongoing noise persist in the DMN during tasks and at any level of consciousness. This leads to questions regarding the development of connectivity and plasticity during people’s lives and refers to fundamental questions about the nature of neurophysiological signals we record through fMRI and EEG. Does the DMN reflect balance between excitation and inhibition of neuronal pools? And is it equivalent to the spontaneous variations in cortical excitability (up-/down-states) as seen in direct current electroencephalography during sleep? Research suggests that there is similarity in that recordings in rest and sleep conditions show highly coherent brain activity patterns, spatially and temporally. However, this type of brain activity is always present or, in the case of the DMN, only slightly attenuated during tasks that show temporal stationarity. This contrasts with non-stationary state-dependent recordings of other coherent brain activity in other areas, which in contrast are markers of functional connectivity during actual cognitive deployment. At present we do not have good grasp of what stationary and resting states are in the brain. In normal conditions, the brain seems to be always on to some extent, and so is consciousness. To better appreciate the last point, we need to delve into the most mysterious known intrinsic brain connectome of all, which is the complex system that regulates wakefulness and sleep.
4.2. Sleep & Dreaming
In the early stages of brain development, between birth and 2-3 years of age, humans sleep on average 50-60% of the day. That amount of sleep declines quickly and steadily to less than half by mid-life (45 to 90 years old) but even at its minimum, sleep amounts to a significant 30-40% of our daily lives. The connectome theory and the brain’s energy consumption budget seems even more relevant for this part of the brain and the states of consciousness experienced during this period of our life. The sleep-wake cycle encompasses various forms of altered states of consciousness, including paradoxical ones that seem to preserve some aspects of full consciousness, such as dreaming.
As originally proposed, the idea of a human connectome (Sporns, Tononi & Kotter, 2006) implied an anatomical continuum from the cellular to the highest systems level of organization. The methods used in sleep research follow a similar principle. Sleep is studied at membrane potential intracellular level, spiking activity at the extracellular level, EEG activity, electromyographic activity (EMG), electrooculographic activity (EOG), neurotransmitter release and behavior/phenomenology. Sleep research involves the integration of observations and measurements from all these sources of information, which is known as polysomnography.
4.2.1 Stages and Sleep-Wakefulness Cycle
A sequence of consistent and coherent changes observed for these measures defines the stages of sleep-wakefulness cycle. During typical resting wakefulness, we can observe fast EEG waves or low-voltage activity, alpha waves (8-13 Hz), in the occipital region. EOG activity reflects a mix of frequent voluntary movements and eye blinks with EMG patterns associated with tonic and phasic voluntary muscle activity. In this stage, there is an ordinary increase or release of all main neurotransmitters, such as for example (but this list is not exhaustive): Acetylcholine (ACh), Norepinephrine (NE), Serotonin (5-HT), Histamine, Hypocretin/Orexin and Dopamine (DA).
Falling asleep, “the state of transition to sleep”, is called Non-REM stage 1 or just N1. We start to observe important changes in our measurements: the EEG has a lot less alpha waves and a lot more slow theta waves (3-7 Hz); the EOG shows slow, rolling eye movements; the EMG still looks normal; motor activity is in the form of sudden jerks and sensory awareness decreases. There is rapid return to wakefulness when the individual is called by his/her name or gently shaken, and most people often report brief dream-like experiences. Microsleep episodes (5-10 sec) are also observed in this phase.
In Non-REM stage 2 (N2), we start observing EEG signatures typical of real sleep: K-Complexes and Sleep spindles. K-complexes are a sharp, high amplitude, negative wave followed by a positive slow wave. Sleep spindles are high frequency bursts of 12-15Hz lasting for ~1 second. This is consistent with intermittent cellular membrane potential activity and alternating extracellular spiking. Arousal threshold is significantly increased, making it more difficult to wake up; eye movements and muscle tone are much reduced; there is a reduction of DA and other neurotransmitters, especially hypocretin. Dopamine fluctuates around these parameters during other phases but does not vary predictably due to movements during sleep. This is real sleep – if interrupted, people will report they were sleeping.
Afterwards, in Non-REM stage 3 (N3), a new signature appears: EEG slow waves, which are mostly made of very high amplitude delta frequencies (> 75microVolt, <2Hz). Intra and extra-cellular activity goes dark. The arousal threshold goes further down, eye movements cease and muscle tone is relaxed. This is deep sleep – if interrupted, people remain confused for a bit before reaching wakefulness. Most people report some dreaming but in the form of thoughts or inner conversation/monologue, involving words and linguistic content.
Finally, during rapid eye movements (REM), we observe cellular membrane activity similar to what is observed during wakefulness. Tonic EEG activity shows alpha- and theta-like waves, similar to N1, or a mix of the two. ACh is increased and while the release of all other neurotransmitters is depressed, NE and 5-HT are especially at a low minimum. A loss of muscle tone is observed except in extraocular muscles and the diaphragm. There is also some phasic activity in the form of REM bursts and muscle twitches. As for behaviour, the arousal threshold is just as low as during deep sleep. 90% of people’s reports in this stage mention dreams, usually vivid and with visual or sensory modalities.
The succession of NREM stages followed by an episode of REM sleep is called the sleep cycle. Each sleep cycle lasts for about 1.5-2hrs (90-110 min). Typically, people experience a total of 4-5 sleep cycles every night. Slow wave sleep occurs earlier in the night and holds an inverse relationship with REM: as slow wave sleep decreases, REM increases. The time spent in each phase of the sleep cycle in healthy individuals averages 5% in N1, 50% in N2, 20-25% in N3 and 20-25% in REM.
4.2.2 Neurofunctional Pathways of Wakefulness and Sleep
As for pathways regulating the sleep/wake cycles, the same brain centers are involved in antagonistic fashion to determine each state with relative weighting of excitatory and inhibitory input from major neurotransmitters. During wakefulness, cortical activation is produced by keeping neurons depolarized and on the brink of firing; this is more or less a global state across the brain referred to as up-state. Pathways with high firing during wakefulness and REM but next to none during NREM are cholinergic. These two major pathways have a dorsal (reticular activating system (RAS) > Pons > Thalamus > Cortex) and a ventral stream (RAS > Dorsal Brainstem > Basal Forebrain > Cortex). Other pathways mainly active during wakefulness are cholinergic neurons in the hypothalamus, which feed on histaminergic neurons and then the cortex; glutamate inhibitory pathways that project from the RAS throughout the cortex; and 5-HT/Noradrenalin-secreting nuclei. Unaffected pathways are dopaminergic neurons targeting the frontal cortex. As mentioned, this is due to reduced motor activity and planning in rest and sleep.
In sleep, cells are hyperpolarized and in a global down-state. The anterior hypothalamus is a major player in inhibiting (through GABA) wake-promoting centers, so that during N1-2 activity is largely attenuated but mostly distributed as it is in wakefulness. Excitatory activity is reduced to a minimum during N3, but REM marks an intense recovery of focused and limited excitatory activity streams (“creeks”) in targeted areas of the thalamus and cortex. REM-generating pathways are somewhat anatomically distinct from those involved in wakefulness. Though they also involve cholinergic cells, these originate in the pons, use ACh and have as first-relay targets the thalamus and basal forebrain ventral network. From there on, activity goes to the limbic system, which then feeds the cortex.
4.2.3 Spontaneous EEG and Metabolic Activity During Sleep
Wakefulness and REM EEGs are remarkably similar since they include a mix of fast low voltage waves, have background alpha and beta waves, and a small proportion of gamma waves. During wakefulness, neurons are depolarized and predisposed to firing; during REM, depolarization and firing rates are even higher (paradoxical sleep). Thus, if one were to just consider EEG, REM would look like active wakefulness.
EEGs exhibit a coherent sequence of activity patterns during NREM. This sequence can be described as a progression of three synchronized states across cortical neurons:
- Up state, with neuron depolarization across the cortex similar to or higher than wakefulness
- Down state, with neuron hyperpolarization across the cortex relaying slow waves
- Silence, without synaptic activity.
The transition from one stage to another is associated with the ordered sequence of global sweeps in the cortex of up state, down state and then silence periods. This follows a top-down path, from frontal to occipital lobes, which becomes progressively more frequent as one goes from N2 to N3. K-complexes are produced at the onset of these sweeps when the RAS triggers them. Following this, the temporally close appearance of sleep spindles reflects a dynamic loop, the crosstalk between cortex and thalamus during the upstate: Cortex > Thalamus > Cortex. Spindles are recorded when the signal comes back to the cortex from the thalamus (i.e., Thalamus > Cortex). They seem to indicate the intervals when the thalamus is temporarily unblocked.
During NREM, metabolism and blood flow can be as low as 40% of normal rates. This indicates widespread reduced synaptic activity and reduced excitation of most cortical areas (except the primary sensory areas). During REM, metabolism and blood flow have similar levels to wakefulness, but activity in some brain areas varies compared to wakefulness. Limbic system & amygdala-related areas (anterior cingulate, parietal lobule and extrastriate visual areas) show more activity than during wakefulness; in contrast, there is also less activity in the parietal, posterior cingulate and dorsolateral prefrontal cortex.
4.2.4 Phenomenological Consciousness and Dreaming During Sleep
The phenomenological reports of what sleepers subjectively and consciously perceive bear some correspondence to the neurophysiological stages they go through. From N1 to REM, reports are consistent with a progressive shut out from environmental inputs, with a progressively higher arousal threshold. During NREM, thalamocortical hyperpolarization acts like a firewall, partially shutting off the “thalamic gate” to the cortex. Thus, although primary sensory areas remain active, the input is not forwarded to higher cortical areas (breakdown of effective connectivity). The REM-generating mechanisms are still largely unknown. One of the best hypotheses is that during that stage, input from the sensory areas goes through the thalamic firewall. However, there is a disconnection of attention due to deactivation of frontal and parietal areas (external input comes in but is not attended to).
Contents of consciousness also vary consistently with the neurophysiological states that define sleep stages. During N1, 80-90% of sleepers report hypnagogic hallucinations (snapshots or still frames) when woken up. During NREM, sleepers’ reports do not indicate images but rather “thought-like” content, especially in early sleep. These contents later become more dream-like. During REM, most reports are of typical dreams rich in modality-specific vivid images, which are mostly visual but also include other senses.
One of the best theories to understand the link between neurophysiological changes and states of consciousness during sleep has been provided by the Information integration theory of consciousness (Tononi, 2004). According to this model, we acquire consciousness when information is shared globally by a number of neural networks across the brain. This process is one of increased functional connectivity and reduced segregation (narrow localization). Thus, consciousness is based on
- Information availability: Available in relatively segregated local neural activity states which contain information
- Information integration: Activity from local subsystems must be integrated.
The brain can always achieve 1 and 2 in wakefulness, while there is little interference of suppression of integration in N1 and REM. In NREM, however, the sweep of slow waves in local downstate disrupts both availability (with local disconnections) and integration (with global down state).
This theory has also led to a neurocognitive model of dreaming which is able to match the phenomenology of dreaming to brain activity patterns. Accordingly, dreaming is generated by a complex brain pattern:
- Mostly inactivated primary sensory cortices (partial disconnect from external world)
- Dorsolateral PFC mostly inactivated (reduced decision making and thinking)
- Limbic system and temporo-parieto-occipital (TPO) junction sufficiently activated (Important: this is where sensory images are stored)
The brain pattern associated with dreaming explains typical recurring characteristics of dreaming phenomenology reported by most, if not all, of us. Disconnection from the environment reflects reduced input, attention and availability of primary sensory and perceptual areas; internally generated world-analogue of its own sometimes bizarre and crazy rules; reduced voluntary control and reduced reflective thought, that show deactivation of frontal and prefrontal areas; amnesia, likely due to local and global disconnections; and finally, hyper-emotionality, which reflects substantial involvement of the limbic system.
4.3. Sleep-related Dissociated States
A number of “in-between” states called dissociated states have been identified and are currently subjected to intense study. They have some characteristics of waking consciousness and some of consciousness in sleep. Daydreaming is a very common mental state which can be defined as dream-like waking images and thoughts independent of the current task. Although neurocognitive and neurophysiological correlates are unknown, some research suggests that daydreaming and mind-wandering are closely related and correlated to the DMN. However, current studies show that daydreaming likely involves a different pattern of brain activation than REM sleep.
In another, related form of dissociation, lucid dreaming has sleepers become aware of their dreaming, seemingly even able to guide and control the activity. Lucid dreaming has been linked to REM sleep but there is no clear explanation; the most plausible account is that it may consist of some REM episodes in which the usual deactivation of the PFC does not occur.
A few other interesting conditions have been identified. REM sleep behavior disorder consists of episodes of dream enactment during REM. Usually found in elderly males, this condition occurs with extremely vivid dream images and complete recall of the episode and of the dream that is enacted. This disorder is presumably due to an impairment of some regions of the pons, which is responsible for inhibiting muscle tone and motor activity during REM. Narcolepsy is characterized by daytime sleepiness attacks, with variable cataplexy (muscle weakness attacks), hypnagogic hallucinations and sleep paralysis – all symptoms that can happen independently. People who suffer from this condition go directly into REM. Causes are not related to lack of sleep but rather an inability to stay awake and keep muscle tone, presumably due to inactivation of brainstem mechanisms during REM. Narcolepsy is linked with a genetic mutation and depletion of hypocretin.
4.3.1. Sleepwalking
The best known and studied sleep-related dissociated state is sleepwalking. Sleepwalking consists of complex motor behaviours that interrupt deep sleep. It is initiated during sudden arousal from slow wave sleep and culminates in a deambulatory activity with an altered state of consciousness and judgment. Paradoxically, this is not a disorder of sleep but a disorder of arousal, since it is essentially a state of increased and incomplete arousability from slow wave sleep. The incidence of this condition is very rare in adults (2-4%) but less so in children (10-15%, mostly 8-12-year-olds). It tends to occur once per night and lasts anywhere from 1 to 10 minutes. Clinical features of this syndrome are the following: During sleepwalking, patients seem to wake up, sit, appear confused and start walking. Their eyes are open and staring, and patients can respond and talk but rarely make sense. It is difficult to wake them up and when they do, there is confusion and amnesia of the episodes.
Episodes are usually preceded by a lot of slow EEG waves with a very peculiar signature: rhythmic slow delta over frontal central scalp sites (Hypersynchronous Delta waves, HSD). Heart rate accelerates abruptly during (but not before) sudden arousal and polysomnography reveals atypical fragmentation of N3 sleep. Much more arousals per night per cycle are present in these patients. Note that HSD are linked with sleep deprivation.
There are a number of predisposing and triggering factors; a genetic predisposition for sleepwalking has been found a few times. There is a high family history incidence, 10 times higher in first degree relatives than in other individuals. In one of the best and largest twin studies (the Finnish cohort study), although the highest correlation of sleepwalking symptoms was found among monozygotic twins, there was also a strong correlation among dizygotic twins. Research synthesis suggests that genetics account for an estimated 57-66% of variability in children and 36-80% in adults. Common triggers that act on predispositions are sleep fragmentation (caused by breathing disorders, such as sleep apnea) and excessive slow wave sleep (caused by a rebound after sleep deprivation). Fever, alcohol, medication & stress are also possible links. Interestingly, pregnancy is associated with a reduction of sleepwalking episodes.
Theoretical models of sleepwalking have focused on explaining its two main pathophysiological aspects: complex motor behavior (deambulation) and impaired consciousness. Complex motor behavior could be the result of subcortical hyperarousability, while impaired consciousness has been linked with the deactivation of prefrontal areas (hypoarousability). Indeed, confirming these hypotheses, neuroimaging shows the hypothesized increased blood flow in the cerebello-thalamo-cingulo-cortical pathway, which is linked with motor behavior, but a decreased CBF in the thalamo-frontal pathway. Furthermore, some studies have found a link between sleepwalking, migraines and serotonin regulation.
Due to the rare and elusive nature of sleepwalking episodes, one may wonder about the issues that might arise in differential diagnosis of sleepwalking from other, more common nocturnal disorders. In this case, there are two essential pieces of evidence the researcher or clinician cannot do without. The first is the individual’s clinical history, while the second is video evidence where the complex motor behavior can be caught in the act. This is usually done with videography and video-polysomnography, i.e. video documentation of the entire sleep-wake period during the patient’s visit. Conditions with similar episodes to sleepwalking include REM disorder, NREM parasomnias, nocturnal epilepsy, confusional arousals/sleep drunkenness and nocturnal wandering/sundowning in dementia patients.
Finally, one of the most notable aspects of sleepwalking is the complications that may manifest with its occurrence, such as self-injury, violence, sexual behavior and compulsive eating. A famous example of this is the case of Kenneth Parks. One night, this man drove half an hour across Toronto to repeatedly stab his in-laws, killing his sleeping mother-in-law and seriously injuring his father-in-law. He then, in a confusional state, turned himself in at a police station as he realized he was covered in blood and had the vague impression of having done something terrible he could not remember. Parks was eventually acquitted thanks to an exemplar forensic diagnosis that demonstrated without a doubt he was in a severe sleepwalking state when he committed the attacks. As in the case of Parks, forensic diagnoses for violent sleepwalking need to be based on a lot of converging evidence. First of all, personal and familial clinical history must be collected, in addition to sleep tests with full video polysomnography. The duration of episodes needs to be established. The nature of the behavior is often meaningless, and the emotional reaction after the episode tends to be surprise in the form of perplexity and horror. As I have mentioned, a key clue is complete amnesia of the episode. Finally, a very detailed history of documented sleep deprivation and stress is another necessary condition for the diagnosis of violent sleepwalking.
4.4. Anesthesia
If one should assign an award to acknowledge the best experimental tool to tackle consciousness, anesthesia would win without a doubt. Anesthesia allows us to chemically induce temporary, reversible unconsciousness and to chemically manipulate levels of consciousness. In addition to this, there is a practical aspect that makes anesthesia invaluable: 1 in 1000-2000 people that undergo surgery have intraoperative awareness. That is, they appear completely anesthetized during surgery, yet are fully conscious and aware for the entire procedure.
There is a straightforward relationship between anesthesia and levels of unconsciousness. The level of unconsciousness depends on the amount of anesthesia, which has a dose-dependent effects on consciousness. There is a minimal dose to induce a first stage of unconsciousness but also a dose-dependent function, according to which loss of consciousness becomes deeper and deeper. These vary from person to person but only slightly. Loss of consciousness in the operating room is defined by loss of the ability to respond upon verbal request, and failure to move in response to a rousing shake. Anesthesia plus neuroscience methods such as neuroimaging give us strong tools to achieve a greater understanding and knowledge of the neurobiology of consciousness.
There are two primary categories of chemical substances or agents that can act as anesthetics, which are defined according to delivery method: intravenous or inhalational. Intravenous delivery (induction) is rapid and continuous; it generally involves barbiturates (sodium thiopental, methohexital, etomidate, propofol) and in some cases sedatives (benzodiazepines: midazolam, clonidine, dexmedetomidine, or opiate analgesics). In contrast, inhalation delivery is long-lasting but much slower. It uses gases at room temperature (ex: nitrous oxide or xenon) or the vapors of volatile liquids (isoflurane, sevoflurane, desflurane). These agents are administered using calibrated vaporizers that deliver an exact dosage, a percentage of the chosen anesthetic agent plus oxygen). Anesthesiologists can give exact dosages by turning the calibrated knob on the vaporizer to a certain point.
The doses of inhaled agents can be characterized in terms of their relative potency by referring to the critical amount needed to prevent movement during surgery. The minimum alveolar concentration (MAC) of an inhaled agent that prevents movement in 50% of patients in response to painful surgical stimulation is defined as 1 MAC. To be certain movement will be prevented during surgery, dosage is increased to a bit above 1 MAC. MAC-awake is the point at which there is no response to verbal request in 50% of patients. This is typically considered the point of loss of consciousness, typically at 0.3-0.4 MAC.
Although nobody is exactly sure of the cellular mechanisms, anesthesia interferes with cellular protein channels that control synapses. The hypothesized mechanism reduces excitation or is inhibitory, leading to hyperpolarization through GABA, Glutamate and ACh channels. Anesthesia becomes very interesting if we look at brain imaging. Virtually all anesthetics decrease global cerebral metabolism in a dose-dependent fashion (reduction of glucose metabolic rate) with the important exception of Ketamine, which increases global cerebral metabolism. If we look at the global metabolic rate of glucose, brain activity could be reduced to 30-60% during anesthesia. Some brain regions are more suppressed than others, such as the thalamus, posterior cingulate and medial parietal cortex.
A lot of evidence has accumulated showing that anesthetics interfere with the normal tonic firing pattern of cells in the thalamic networks (similarly to deep, NREM3 sleep). This has been most clearly shown in experiments where in the somatosensory portion of the thalamus, neural responses to external stimulation decrease as a function of increased anesthetic dosage. However, that is reversible and responses increase when dosage decreases. Thalamic nicotinic mechanisms in the central medial thalamus reverse the action of anesthesia, turning on consciousness with microinjections of nicotine. In spite of the evidence supporting a thalamic site of action, most of it shows that anesthetics act first on the cortex and then on the thalamus, with GABA pathways majorly involved in the process. Direct evidence of this was provided by studies of Parkinson’s patients that monitored the time-course of anesthesia via EEG from the cortex and thalamus. Cortical EEG changed right after consciousness was lost while thalamic EEG changed 5-10 minutes later. These changes did not show a sudden drop in EEG activity concurrent with the introduction of the anesthetic agent, rather tapering off much more gradually. During surgery, the thalamus shows another response the cortex does not, which is a change in some parameters of EEG activity due to intense somatosensory stimulation but unrelated to loss of consciousness. In other words, the thalamus shows reliable activity to stimulation even when patients stop moving in response to intense somatosensory stimulation. Overall evidence seems to suggest that anesthetics may act via cortico-thalamic feedback, i.e. feedback from the cortex to the thalamus until loss of consciousness. When the cortex stops responding, the site of action shifts to the thalamus. The most plausible account for the effect of anesthesia on motor output is that anesthesia interferes with effective connectivity in the ascending excitatory pathways from cerebellum to cortex via thalamus.
A number of issues have been raised regarding specificity versus interaction of anesthetic effects and such systems. One such issue is whether there are major overlaps between the mechanisms of sleep and anesthesia. Research shows that they are quite small and their occurrence always involves segments of the RAS, presumably around the thalamus. However, the overlaps and interactions most likely occur at lower levels of unconsciousness during sedation, not loss of consciousness. Another issue is that evidence of activity suppression in the posterior cingulate and parietal medial cortex during anesthesia suggest a link to the default networks. Nevertheless, default network fluctuating activity seems to be unchanged by complete unconsciousness induced by anesthesia (at least in monkeys). Overall, current evidence shows that anesthesia involves a much larger number of network and more complex connectivity across them than other systems to which it is compared.
Surprisingly, research using goat models shows that arousal from an external stimuli decreases the amplitude of EEGs when < 1 MAC. When > 1 MAC, EEG remains the same as before, indicating no change to arousal state. This research also shows that unconscious responses to painful stimulation, as shown by EEG changes at 1.3-1.4 MAC, could shift down subcortically or even be mediated at the level of the spinal cord. This new line of research shows that anesthesia’s effect may be a progressive suppression (mediated by feedback loops) of the brain’s ability to arouse itself into consciousness; this is known as the arousal blockade hypothesis.
4.4.1. Phenomenology and Behavior Observed During Anesthesia
Assuming induction, in the first 30 minutes of anesthesia, progressively increasing MAC from 0.1 to 0.3 produces initially higher sensitivity (hyperalgesia) then decreased sensitivity to pain (analgesia). The other major change is impairment in declarative/explicit memory. Increasing MAC to 0.6 intensifies the analgesic state and memory impairment, which now includes procedural memory.
Increasing to 0.8 MAC in an additional 30 minutes produces the following phenomenological reports: feelings of intoxication (with visual disintegration), disconnection from the environment and/or intense focus on a very narrow visual spot, drunken-like behavior, numbness and tingling of hands and feet, tiredness and drowsiness. When 0.8 is reached, one feels like falling asleep (N1) and there is still movement in reaction to painful stimuli. Full arousal is still possible at this stage, but the individual seems in a state of fast transition or fluctuation between sleep and consciousness.
Once MAC 1.0 is reached, disconnection from the environment is much more intense, arousal threshold much higher, the individual starts closing eyes and by MAC 1.1, it appears as if the individual falls asleep. Snoring is observed, there are no rapid transitions between sleep and consciousness, arousal seems blocked and there is no movement in reaction to painful stimuli. This is the externally manipulated dungeon of deep unconsciousness.