A. Sleep:

During sleep, we are essentially cut off from the sensory world. We do not hear, feel, taste, or smell, and we would not see if our eyes were pulled open. Everyone has different thresholds during sleep, though; enough of any stimulus will wake us. How does the brain manage to cut off sensory input, yet still let in the really important (or insistent) stimuli? The answer lies in the thalamus. Remember, no sensory information gets up to the cerebral cortex without first passing through the thalamus. If the gate of the thalamus is closed, then the cortex can shut out the world and go into sleep-mode.

B. The EEG:

Electrophysiologists define sleep in terms of the electrical activity of the brain. Just like the electrical activity of the heart can be measured with electrodes on the chest, depolarizations in the brain can be recorded at the scalp. This technique is the electro-encephalogram, or the EEG. An EEG is an average of all the electrical events going on in the brain, and in an awake person it would look something like this:

There is no apparent pattern to the activity, and the activity doesn't seem to be very strong. This is due to the fact that there are so many unrelated simultaneous events going on that many will cancel each other out, leading to a tracing which looks nearly flat. This type of EEG activity is called desynchronous, and is low amplitude but high frequency.

During sleep, the multitude of electrical events begin to fall into sync with one another, and the tracing takes on a new appearance:

Notice that the fluctuations are larger and slower than in the awake state. This pattern is called synchronous, and is high amplitude but low frequency.

There are four recognized stages of sleep, from the first and "lightest" stage to the deepest fourth stage. At night, you progress from stage 1 to stage 4 in the first hour of sleep, and spend the rest of the night cycling up and down between 1 and 4. The EEG gets progressively more synchronous with each deeper stage. Every time you return to stage 1, you enter REM sleep - a period of sleep characterized by rapid eye movements. REM sleep is also called paradoxical sleep, because the body appears to be more deeply asleep than in any other stage, but the EEG looks very much like the waking brain. The desynchronous activity of the brain may be due to dreaming, which also occurs during REM sleep.

C. The ascending reticular activating system:

Early on, someone noticed that if the midbrain of a cat was transected (at point A, below), the cat fell into a coma - their EEG became permanently synchronized. This finding alone is not too surprising. However, if the transection was made down in the medulla (at point B), the cat was only paralyzed, not comatose. The most surprising phenomenon was that if the rostral stump of the transected midbrain was electrically stimulated (blue arrow), the cat "woke up" - according to its EEG.

From these experiments it was concluded that there was some pathway originating in the pons or midbrain that ran forward into the cerebrum and stimulated wakefulness. The pathway was thought to originate with a group of neurons in the brainstem, the reticular formation. Therefore it was named the ascending reticular activating system.

They were close. In fact, the ascending pathway originates from a group of neurons around the fourth ventricle in the rostral pons (near midbrain). Most of these neurons are acetylcholinergic, and project to the thalamus, controlling whether the gate is open or closed. The key is in the action of acetylcholine. Acetylcholine cannot, by itself, activate or shut down the neurons of the thalamus. Instead it sensitizes them. By slightly depolarizing the thalamic neurons (it does this by closing a hyperpolarizing potassium channel), the ascending system can make the thalamus more sensitive to sensory input. This situation would correspond to an awake, alert state. Let's look at the whole system:

The acetylcholinergic neurons project both to the sensory areas of thalamus (such as VPL) and to the reticular nucleus, a layer of cells that wraps around the thalamus like the rind of an orange. This "reticular" has nothing to do with the "reticular" in the brainstem! The fact that the ascending reticular activating system targets the reticular nucleus is only maddening coincidence. The reticular nucleus of the thalamus has a general inhibitory effect on sensory thalamus. Now, although both areas are receiving acetylcholine, they have different receptors and respond in different ways. Sensory thalamus is sensitized by acetylcholine (or "facilitated") as described above, but the reticular nucleus is inhibited by acetylcholine. We can redraw the situation like this:

So what happens when the brain is awake? The cholinergic cells are active, so they facilitate sensory thalamus and inhibit the reticular nucleus. The inhibition of the reticular nucleus actually excites the sensory thalamus as well (negative x negative = positive). As a result the thalamus lets all sensory information through, and cortex is highly active and desynchronized dealing with all the input.

And when the brain is asleep? Now the ascending system is quiet, so sensory thalamus is not particularly sensitive. In addition, the reticular nucleus is freed from inhibition, so it can inhibit the sensory thalamus. The net effect is that thalamus is very insensitive to sensory stimuli, the gate is closed, and the cortex can rest. An interesting property of the thalamic neurons is that when hyperpolarized, they have slow intrinsic waves of activity, similar to the pacemaker of the heart. This activity may be the source of the slow synchronous pattern of the sleeping EEG. It also may function to keep the cortex in shape, ensuring that even when you are asleep the cortical neurons are active.

A summary:

ACh system active
ACh system inactive
sensory thalamus facilitated
sensory thalamus inhibited
reticular nucleus inhibited
reticular nucleus active
thalamocortical neurons active
thalamocortical neurons in slow rhythm
EEG desynchronous
EEG synchronous

Any student who regularly attends classes will recognize that these two states are merely the endpoints of a single continuum.

D. Dreaming

Dreaming occurs during REM sleep, the "paradoxical" sleep stage. Curiously, the ascending acetylcholinergic system actually turns on - it is as though the brain wakes up internally. Yet for some reason the person remains unconscious and unaware. Dreams generally do not make it to conscious memory unless the dreamer is awakened from the dream itself.

How is it that the cholinergic system can be on and the sleeper still unconscious? The answer probably lies in other neurotransmitters and nuclei of the rostral pons. The dorsal raphe nuclei, a cluster of serotonergic cells, and the locus ceruleus, a group of noradrenergic neurons, also play a role in sleep. They may help to keep consciousness suppressed during dreaming.

One of the striking things about REM sleep is the absolute stillness of the body. During most stages of sleep we toss and turn, but in REM sleep only the eye muscles twitch (and, for some unknown reason, the middle ear muscles!). This is due largely to a system of descending inhibition. Dreaming turns on a group of cells in the medulla that descend down the spinal cord and inhibit motor activity. Very specific lesions of these cells (a rare event) lead to a phenomenon called "violent sleeping", where the dreamer physically acts out his or her dreams. This is different from sleepwalking, which usually does not occur during REM sleep.

E. Language

Language, in most people, appears to be localized almost exclusively to the left cerebral hemisphere. Knowing what you know about the organization of the brain, where would be the most logical place to put a cortical area in charge of the production of language? First and foremost, you need fine control over the tongue and mouth. It would make sense to put your cortical area near the mouth section of motor cortex. Sure enough, just rostral to the motor-mouth area of the precentral gyrus is a small area that controls speech. It is called Broca's area, after the physician who discovered it in 1861. Broca had a patient who, after a stroke, completely lost the ability to speak. The patient could apparently understand language, but the only syllable he could produce was "tan", over and over again. After Tan's death, Broca performed an autopsy and determined the site of the stroke. Broca's area is shown below:

What about language comprehension? Where is the most logical place to put the comprehension area? First you must decide if language is primarily visual or auditory. When you read, do you "hear" the words in your head? When you listen, do you "see" the words as written? Which came first, written or spoken language? You probably agree that language is more of an auditory phenomenon than visual. As expected, the language comprehension area is just adjacent to auditory cortex, where the parietal lobe meets the temporal lobe. This area was discovered by Wernicke in 1874, by a studying patients with select comprehension deficits. These patients could not understand language, spoken or written, and could apparently produce copious flowing speech. Their speech, however, made absolutely no sense. Subjects and verbs would be strung together in a seemingly grammatical order, like fragments of real sentences, but the sentences seemed to bear little relation to what the patient was trying to express. Presumably the patients could not understand what came out of their own mouths. Wernicke's area is shown below:

But language is far too complex to be broken down into two discrete cortical areas. Obviously there are visual and manual components to language, for reading and writing. Where does sign language fit in? How do you explain a patient whose only deficit is an inability to name tools? He can describe the use of a hammer but not its name. How does a collection of syllables - a person's name - trigger the face, personality, birthdate, or voice of that person in your memory? Language is probably located all over the brain, with extensive crosstalk between areas. The discrete areas of Broca and Wernicke may be necessary for language, but they are certainly not sufficient.