THINK back to a memorable dream. Perhaps you were in the main street of your
town, except that it was more like a country road with cactuses as high as
houses growing alongside. You were talking to your brother, and halfway through
the conversation you realised he was a beautiful woman. Suddenly, you were on a
wooden ship en route to Timbuktu when a purple cloud engulfed the vessel and
lifted it up into the sky.
You never questioned any of it, you were too busy talking to the mixed-up
characters and being part of fast-changing scenes. What made you take it all so
very literally?
Allan Hobson of the Harvard Medical School has spent years
studying how the brain works during dreaming. He thinks he has an answer: the
dreaming brain, he says, is simply a different brain, a brain powered by a
chemistry set that is quite alien to our waking hours. This special 鈥渄ream
chemistry鈥� swings into action once every 90 minutes or so, during a phase of the
sleep cycle called rapid eye movement (REM) sleep. Cut off from signals from the
outside world, says Hobson, the higher, thinking regions of the brain are in
thrall instead to chemical signals from the primitive brainstem. These signals,
he argues, hold the secret to why the dreaming brain cooks up so much
nonsense鈥攁nd why it so credulously tries to make sense of that
nonsense.
Advertisement
Hobson鈥檚 is the most developed theory to offer a coherent biological
explanation for the origin of dreams in REM sleep. But naturally there are
dissenters, especially among those psychologists who are less impressed with
chemical and electrophysical theories. They challenge the orthodoxy that links
dreams chiefly with REM sleep. More than that, they say there鈥檚 not even
anything very special about the chemistry of a dreaming brain. We dream because
the mind is still the mind during sleep, says psychologist David Foulkes,
formerly of Emory University in Atlanta, Georgia, and a leading proponent of the
idea that our night-time reveries are, at root, no different from humdrum
daydreams.
Discarding garbage
Whatever researchers think about REM sleep and its link with dreaming, they
do at least agree that REM itself probably has an important biological function.
What function, though, nobody鈥擧obson included鈥攔eally knows. One
theory sees REM sleep mainly as a hangover from our days as a fetus and as a
newborn. Fetuses spend more than half their day in REM, and the sleep of
newborns is four times richer in REM activity than that of adults. Maybe fetuses
and newborns use REM activity to train the brain circuits they鈥檒l need to walk,
smile and eat, or maybe the electrical signals that dart around the brain during
REM help to wire up the nervous system.
But most theories focus on the role of REM sleep in adults. In some undefined
way, suggest some researchers, REM sleep may help the brain to process memories
and preserve skills. Unproven as it is, the idea is at least an enduring one.
We know that much because two prominent scientists have tried to turn the
thinking on its head, and failed.
In the early 1980s, Nobel prizewinner Francis Crick and Graeme Mitchison,
then both at the Salk Institute in La Jolla, California, argued that the purpose
of REM sleep is to eliminate patterns of unwanted or even harmful neural
connections that build up in the brain during the day. Our dreams are the
garbage we鈥檙e discarding, says this model. Crick and Mitchison even suggested
that trying to remember dreams should be discouraged because it could maintain
noxious thought patterns.
You could hardly imagine a bigger slap in the face for Freudian-style
psychoanalysis. But the idea never really caught on. One obvious problem with
it, says Hobson, is that people who remember their dreams seem no more
dysfunctional than people who don鈥檛.
Yet if Hobson doesn鈥檛 regard dreams as the mind鈥檚 garbage, neither does he
think they鈥檙e anything precious. 鈥淚 mean, why would dreams be so hard to
remember if they were high-priority as information?鈥�
And they are hard to remember. We have five REM periods a night, each an
opportunity for a dream. When sleepers are woken during REM and asked to recount
their thoughts, they report dreams about 70 per cent of the time鈥攄reams
they otherwise would forget. Yet when people are left to sleep, on average they
recall only one dream every week or two. A handful of people remember as many as
three to five a night, while others recall none in a whole lifetime. But there鈥檚
no evidence, says Hobson, 鈥渢hat poorer-than-average dream recall hurts you in
the slightest鈥�.
That鈥檚 because what matters biologically, in Hobson鈥檚 view, is not the dreams
themselves but the REM activity from which they emerge. 鈥淚 think we need to have
REM鈥攂ut I don鈥檛 think we need to dream,鈥� he says. And while Hobson can鈥檛
say what REM activity is for, he thinks he knows how dreams tumble out of
it.
His is a theory developed from thousands of gruelling hours recording
electrical signals from cat brains, from dripping chemicals into those brains to
see how they affect electrical activity during REM sleep, and from analysing
detailed dream reports. And the theory goes like this. Whether people are awake
or in REM sleep, the primitive brainstem acts like an on-switch, sending
electrical and chemical signals to certain parts of the cortex to stimulate
nerve cells to fire. But in REM sleep鈥攁nd only in REM鈥攁nother set of
signals, stronger and more rhythmic, akin to those produced during epileptic
fits, shoot up from the brainstem. Known as PGO waves, they stimulate the cortex
more widely and vigorously, especially the parts that process visual
information.
Normally, signals reaching these visual areas carry sensible information
about things we鈥檝e really seen. The PGO waves have nothing to do with such
mundane realities. But the cortex doesn鈥檛 know that. It assumes that the signals
are real and tries like crazy to make sense of them, to make them gel with
whatever other thoughts are zipping through its circuits at the time. 鈥淵ou don鈥檛
have the external world to orient you, so the brain does the best job it can,
running around after the fact trying to create coherence,鈥� says Hobson.
The result is a mix of things that were on our minds and things that are
thrown up more or less at random. It鈥檚 what Hobson calls an
鈥渁ctivation-synthesis鈥� model of dreaming. And he believes it explains the
bizarreness we experience in our dreams: characters that are composites of
several people, houses that are not-quite-houses-but-sort-of-merry-go-rounds,
and, when the contradictions become extreme, sudden shifts of venue.
None of this explains why we unthinkingly accept that bizarreness, or why we
find dreams so damnably hard to remember. For that, Hobson turns to something
else鈥攖he changes in brain chemistry that occur throughout our lives as the
brain cycles between REM sleep, deeper sleep and wakefulness.
Hobson鈥檚 group has identified two sets of nerve cells in the brainstem which
he thinks control the flip between REM activity and deeper sleep. One set, which
he calls the 鈥淩EM-on鈥� neurons, secretes the neurotransmitter acetylcholine. The
other set, the 鈥淩EM-off鈥� neurons, secrete two other chemicals, norepinephrine
and serotonin.
In Hobson鈥檚 model, the two nerve systems turn on and off in a 90-minute
cycle, which not only makes REM sleep come and go, but also controls the ups and
downs of the chemicals they secrete through the brain. In the thick of REM
sleep, for example, levels of norepinephrine and serotonin in the brain may be a
hundred times lower than during wakefulness, while acetylcholine is more
abundant.
And it鈥檚 these chemical 鈥渋mbalances鈥� that Hobson believes can explain why
critical thinking, insight and memory are all so woefully absent during dreams.
鈥淪omething鈥檚 been subtracted from the system. What鈥檚 been subtracted?
Norepinephrine and serotonin.鈥�
Is he right? Some of what鈥檚 known about the function of these chemicals
certainly fits the model. Many neuropsychologists have linked low serotonin
levels to lack of judgment and control over thoughts, and low norepinephrine
levels to poor concentration. If you can鈥檛 concentrate, you鈥檙e not going to
weigh up information, and you鈥檙e not going to remember it.
Support for the idea that the 鈥渄reaming brain鈥� is different from the 鈥渁wake
brain鈥� has come from a brain-imaging study that was published last year in
Nature. Pierre Maquet and his colleagues at the University of Li猫ge
in Belgium used PET scanning to discover which parts of the brain are more
active during REM sleep than when we are awake. What they found thrilled Hobson.
The amygdala鈥攁 part of the midbrain that triggers emotional
responses鈥攚as even more active in REM sleep than in wakefulness, and
dreams are often highly emotional. Anxiety crops up most often in dream reports,
with elation in second place and then anger.
By contrast, the Belgian team found that the front portion of the cortex was
less active in REM sleep than during wakefulness. Again, that鈥檚 good news for
Hobson鈥檚 model, since this part of the brain helps us to concentrate, think
critically and reflect on what鈥檚 going on.
It all sounds nice and neat. But Foulkes takes serious issue with Hobson鈥檚
model. First, he says, there鈥檚 no evidence to suppose that the seizure-like PGO
waves coming up from the brainstem introduce bizarreness into dreams. 鈥淭he
brainstem may turn the brain on, but I don鈥檛 think it determines what the brain
does,鈥� he says. 鈥淭he lights may be turned on in the cellar but what happens
upstairs is a function of who鈥檚 up there.鈥� Secondly, says Foulkes, the majority
of dreams that he and others have studied in sleep labs by waking people and
collecting their reports aren鈥檛 that bizarre at all. We only think that, he
argues, because the mundane ones are so forgettable they never make it into any
dream diary.
Foulkes believes night-time dreaming isn鈥檛 so different from day-dreaming. He
and John Antrobus, a cognitive neuropsychologist at the City University of New
York, have both left experimentees in darkened rooms and every so often asked
them, without warning, to report on what was running through their heads. 鈥淭hey,
too, reported mental activity that sounded dreamlike. It was hallucinatory, it
was dramatised, it was visual鈥攖hings as weird as `I was outside on the
wing of an airplane鈥�,鈥� Foulkes recalls.
There may be more workaday reasons for other aspects of dreaming, too. Take
all that anxiety that pervades our dreams. Shouldn鈥檛 we expect to feel anxious
as we wander about trying in vain to catch a train or struggling with the fact
that our house now looks like a merry-go-round?
And perhaps signals from the brain stem are not the explanation for all those
shifts of scene, either. Maybe these result instead from dreamers waking up
during their REM periods and then falling asleep again almost immediately. Each
of our dreams, in other words, might be several fragments stitched into a larger
story.
Foulkes鈥檚 wider point here is that there may be no great difference after all
between a dreaming brain and an awake brain. Left to its own devices, cut off
from the input we get from our eyes and ears and more, the brain may
automatically ruminate and wander. You don鈥檛 need a dearth of serotonin and
norepinephrine, or extra bizarreness triggered by the brainstem, thank you very
much. You don鈥檛 even require REM.
As evidence, Foulkes points to the fact that the link between dreaming and
REM sleep is not as clear-cut as scientists once thought. Some studies find that
virtually all dreams happen during REM. But others give numbers closer to 30 per
cent. It鈥檚 largely a matter of what one calls 鈥渁 dream鈥�. Most of the long,
complicated, fantasy-like stories we tend to remember take place during REM
sleep. Most deep sleep dreams tend to be short, fleeting scenes (鈥淚 was asking
directions to the supermarket鈥�) or more thought-like in nature (鈥淚 was thinking
about my mother鈥�).
But Hobson is adamant that the statistical distinctions make it clear that
REM dreaming is special. And he insists that anyone who thinks dream life
resembles daytime reveries has overlooked something important in their dream
analyses. 鈥淚n dreams, there are characters that are merged. I don鈥檛 have that in
my waking fantasies鈥攅ver.鈥�
Foulkes and Antrobus would argue that a much larger chunk of a dream is
concocted by the cortex, making questions about the content of dreams more
intriguing. But not even they can say whether there is a logic to all those
merged characters and jumbled scenes. It鈥檚 possible, as experiments show, to
influence the content of dreams鈥攂y spraying water on dreamers, for
instance. If the water ends up in a dream, it may be as rainfall, or perhaps as
a refreshing beverage. But how the brain makes such 鈥渄ecisions鈥�, nobody has the
slightest idea.
Perhaps, says Antrobus, and no, he isn鈥檛 joking, the only way to understand
dreams鈥攚hy this scene or that person crops up seemingly out of the
blue鈥攚ould be to get a better grip on how the brain works when it鈥檚 awake,
plug all the information into a supercomputer and then programme it to simulate
the effects of sensory deprivation.
To which, of course, it鈥檚 tempting to reply: in your dreams.
* * *
Dreamy gorillas
Dogs may whine, whimper and 鈥渞un鈥� in their sleep, but do they really dream.?
All mammals have REM sleep, which is strongly linked to dreaming. And other
evidence that animals dream comes from experiments on cats by French dream
researcher Michel Jouvet at the University of Lyons. Jouvet eliminated the
partial paralysis that usually occurs during REM sleep by severing certain
nerves in the cats鈥� brains. When asleep, the cats raised their heads, chased
nonexistent objects and showed aggression and fear鈥攁s if they were acting
our their dreams. This interpretation is backed up by observations at an animal
sanctuary in Woodside Caliifornia, where two gorillas, Koko and Michael, have
been taught to communicate using sign language. According to Penny Patterson,
the researcher who studies them, Koko or Michael will relate coherent 鈥渟tories鈥�
over breakfast that she knows never happened. Patterson believes these stories
could be accounts of the night鈥檚 dreams.