Measuring Wakefulness — Berger's EEG, the Reticular Formation and REM Sleep, 1924–1953

This is Part 2 of A History of the Science of Consciousness. Last time, psychology ran aground on disagreements between people's reports, and in 1913 Watson proposed dropping the word "consciousness" altogether. This time we follow an era that tried to measure consciousness without asking anyone anything. In 1924 the German psychiatrist Hans Berger began recording electrical fluctuations from the human head: the electroencephalogram, or EEG. Within a few decades the EEG could tell, from the trace alone, whether a person was awake, asleep, or had lost consciousness for a moment. What it never showed was what the person was seeing or imagining. And to learn what the new signs meant, researchers in the end had to wake sleepers up and ask them. Working from the original papers, this article sets out what could be measured and what could not.

The story of the mind-reading brain wave

The usual story goes like this. In 1929 Berger recorded the brain's electricity from outside the head and called it the electroencephalogram. From then on we could watch the brain at work from the outside. The EEG became a window into the mind.

The EEG did make it possible to tell waking from sleep without asking. But the first large wave to be found turned out mostly to signal that part of the brain had nothing to do. What a person was looking at, or picturing, did not appear in the waves. And what each new sign meant was settled, in the end, by checking it against what people said. This article runs from Berger's long hesitation to the discovery, in 1953, of rapid eye movements in the middle of sleep.

Recording the brain from outside the skull: Berger, 1924–1929

Hans Berger was a psychiatrist at the University of Jena. At eighteen, on army maneuvers, he came close to being killed. At about the same time his sister, far away at home, insisted that her brother was in danger, and their father sent a telegram to ask if he was all right. Berger took this as a mind reaching across a distance. Mental work, he reasoned, must involve a measurable energy, like heat or electricity, and he wanted to find it in the brain. He measured the brain's blood flow and temperature, without clear results.

He turned to the brain's electricity in 1910, using a string galvanometer: a very fine thread stretched between magnets, which twitches when a tiny current passes through it, its shadow magnified by a beam of light onto photographic paper. On 6 July 1924 he recorded electrical fluctuations from the surface of the brain of a 17-year-old patient whose skull had been opened by a surgeon. This is now counted as the first human EEG. With more sensitive instruments he was later able to pick up the waves through the scalp of people with intact skulls.

Even so, he hesitated for years. A diary entry from 1928 says, in effect, that he had worked in vain on the presumed EEG for many years and should abandon it. On 28 September the same year he wrote that he had found it. A record from October 1925 survives on which Berger traced over the photographed wave in white ink, and the wave he drew is far more regular than anything later photographs show. He may have been drawing not only what he saw but what he expected to see.

In 1929 he published "On the Electroencephalogram of Man", naming the recording the Elektrenkephalogramm. He called the larger, slower waves alpha waves and the smaller, faster ones beta waves, and reported that alpha waves vanish when the eyes are opened. Berger called the record a "brain mirror". But the paper was written in a difficult style and appeared in a psychiatric journal rather than a physiological one. The leading Berlin brain-research group regarded him as a technically naive amateur, and some there doubted that the alpha rhythm existed at all. Five years passed before anyone else confirmed it.

A wave that vanishes when the eyes open: Adrian and Matthews, 1934

In 1934 at Cambridge, Edgar Adrian, who had shared the 1932 Nobel Prize for work on nerves, and his engineer Bryan Matthews repeated Berger's experiments. They write that they started out skeptical, and that after a period of hesitation they were satisfied the waves came from the brain. They called it the Berger rhythm: an oscillation of about ten per second, largest at the back of the head, near the part of the brain that handles vision.

They mapped carefully what made it disappear. Opening the eyes in a lit room abolished it. In complete darkness it returned after a few minutes even with the eyes open, but vanished again as soon as the subject tried to make something out. With the eyes shut, a hard mental-arithmetic problem abolished it for as long as the subject struggled with it, and it came back the moment attention wandered. Yet counting aloud, reciting familiar poetry or making small talk barely disturbed it, and neither did squeezing a pair of pliers as hard as possible.

Visualizing shapes and figures left it unchanged. In Part 1, Titchener and the Würzburg psychologists gave opposite reports about whether thinking involves mental images. The EEG could not settle that question. Moreover, the pair's explanation of why closing the eyes favors the rhythm, that it withdraws attention from seeing, rested mainly on the introspection of one of them acting as his own subject. Some of the observations in the dark were made by one of them listening to his own rhythm through a loudspeaker and matching it against his sense of whether he was trying to see.

Their interpretation was that a patch of visual cortex, when it has nothing to do, falls into beating in unison, and that seeing or concentrating breaks up the unison. The rhythm thus showed, in their words, the negative rather than the positive side of brain activity: what an idle area does. Its rate in normal subjects stayed between about 9.5 and 10.5 per second, with little difference from person to person, although how readily it appeared varied a great deal. They called it disappointingly constant. In three patients who had been blind for years, they could find no trace of it.

When consciousness lapses, the trace changes shape: epilepsy and sleep, 1934–1937

In 1934 Hallowell Davis's laboratory at Harvard was also trying to confirm Berger. The first attempts, on two students' heads, found nothing; with the electrodes on Davis's own head, the rhythm appeared at once. The laboratory staff split roughly evenly into people who showed it clearly and people who did not.

In December that year Frederic Gibbs, Davis and William Lennox recorded two patients with what was then called petit mal epilepsy, in which consciousness lapses briefly while the body stays almost still. During the attacks the trace showed a regular pattern of a spike and a rounded wave repeating about three times a second. Their 1935 paper, covering twelve children, was titled "The electro-encephalogram in epilepsy and in conditions of impaired consciousness". A lapse of consciousness had become visible as a shape on paper, with no need to ask the patient anything.

Sleep became measurable too. From 1935 the physicist Alfred Loomis and colleagues, working in a private laboratory outside New York, made long recordings of sleeping people. They described short bursts of waves at about fourteen per second, the sleep spindles, and built the first scheme for grading the depth of sleep from the EEG alone. The depth of a sleeper's sleep could now be read off without waking them. What the EEG showed so far was not the content of consciousness but its level: how awake a person was.

What keeps the brain awake: from Bremer's cuts to the reticular formation, 1935–1949

Where in the brain is that level set? In 1935 Frédéric Bremer in Brussels cut through the base of the cat brain in two ways. Cats cut where the brain meets the spinal cord showed, in their EEG and their eyes, signs of both sleep and waking. Cats cut higher up, at the level of the midbrain, seemed to stay continuously asleep, even though the pathways from the ears, eyes and nose were still intact. The brainstem, the stalk at the base of the brain, evidently mattered for the level of consciousness.

In 1948 Giuseppe Moruzzi of Pisa came to work with Horace Magoun at Northwestern University near Chicago. They set out to study a different question: how the cerebellum influences the motor cortex. In their first experiment, in December 1948, stimulating part of the lower brainstem made the EEG of the motor cortex go flat, and they wrote in the protocol book that motor-cortex activity was completely inhibited. Recording from other areas that same day, they realized that the note gave a distorted picture.

What they had been stimulating was the reticular formation, the core of the brainstem, where nerve cells and fibers interlace like a net. In cats prepared without anesthesia, stimulating it replaced large, slow, synchronized waves with small, fast ones, a change that looked the same as the EEG of an animal woken by a natural stimulus such as a sound or a touch. In the first volume of the new journal Electroencephalography and Clinical Neurophysiology in 1949, they reported that this effect did not travel along any of the known sensory or motor pathways. Looking back, they recalled that a steady background of less intense activity in this upward-running system might help maintain the waking state.

Here too, what was measured was a cat's EEG and behaviour. Cats report nothing. "Awake" meant that the EEG became small and fast and the animal responded to sounds.

Asleep, but the eyes are darting: Aserinsky and Kleitman, 1953

At the University of Chicago the sleep researcher Nathaniel Kleitman gave his graduate student Eugene Aserinsky a task: find out whether blinking stops abruptly or gradually at the onset of sleep. Aserinsky could not settle on a workable definition of a blink, so he switched to lid and eye movements and recorded them all night. He later described the approach as a credo of desperation.

The eyeball carries a small voltage between its front and back, so when it moves, the electrical potential of the surrounding skin shifts. A record of this is called an electro-oculogram. Aserinsky placed electrodes above and below and on either side of one eye, and recorded from the forehead at the same time, comparing the two traces to tell eye movements apart from brain waves and muscle activity. He also watched sleepers directly under dim light, and filmed two of them without waking them.

In twenty healthy adults, periods appeared during sleep in which both eyes moved together in rapid jerks. In eleven people left to sleep undisturbed, the first such period began on average 3 hours 14 minutes after going to bed. The movements came and went, and the period lasted about 20 minutes on average. A second period began on average 2 hours 16 minutes after the start of the first; with longer sleep there was a third, and occasionally a fourth. During these periods the EEG was low and irregular; Aserinsky later called it a waking-type EEG. Breathing quickened from an average of 13.4 to 16.9 breaths a minute.

Suspecting that this was when people dream, they woke sleepers and asked: ten people over fourteen experiments. Of 27 awakenings during the eye movements, 20 produced detailed dreams, usually with visual imagery. Of 23 awakenings while the eyes were still, 19 produced no recall at all. The quiet-eye awakenings were timed by EEG to match the depth of sleep as closely as possible. Because the eye movements, the EEG pattern and the breathing changes went together, the paper concluded that they, and probably dreaming, were manifestations of a particular level of cortical activity in normal sleep, and that the method gave a way to count how often, and for how long, people dream.

At a scientific meeting that same year Aserinsky introduced the acronym REM, for rapid eye movement, and this phase later came to be called REM sleep. By his own account, the finding contradicted the view, held by Pavlov and by his own supervisor Kleitman, that the cortex rests during sleep.

What could be measured at this point, and what could not

What could be measured was the level of consciousness. Awake, drowsy, deeply asleep, briefly absent: each could now be read from the shape of a trace without asking the person anything. Animal experiments showed that the system holding up that level runs through the reticular formation of the brainstem. And the periods of sleep in which dreams were likely now had an outward sign, the movement of the eyes.

What could not be measured was the content of consciousness. Picturing shapes left the alpha rhythm unchanged, and counting aloud did not disturb it. The most conspicuous wave mostly showed a patch of cortex with nothing to do. The content of a dream could be learned only by waking the dreamer and asking. Even then, 7 of the 27 rapid-eye-movement awakenings yielded no detailed dream, and 4 of the 23 quiet-eye awakenings yielded at least a sense of having dreamed. The very reading of rapid-eye-movement periods as dreaming periods had been fixed by matching them against what woken sleepers said. And a small, fine EEG marked waking in Moruzzi and Magoun's cats, yet appeared in Aserinsky's subjects in the middle of sleep, during the periods when they were likely dreaming. The shape of the wave alone could not say which state it was. Behaviorism had set reports aside. The EEG gave a quantity that could be measured without them, but to fix what that quantity meant, reports came back in another form.

Next time: one head, two reports. In patients whose corpus callosum, the bridge between the two halves of the brain, had been cut to treat epilepsy, asking the right half and the left half of the brain separately produced different answers. We follow split-brain research, 1940–1981.