NEUROSCIENCE:
Getting
the Brain's Attention
Ingrid Wickelgren
While drug abuse continues to blight society, science is
getting a clearer view of the causes and effects of addiction, as
this special issue of Science shows. The following News
stories describe a new view of how the neurotransmitter dopamine may
contribute to addiction and how fetal cocaine exposure may affect
brain development. And a series of Articles beginning on page
45 covers areas ranging from basic mechanisms of addiction to
policies aimed at dealing with addictive drugs.
Rather than signaling pleasure as previously thought, the
neurotransmitter dopamine may be released by brain neurons to
highlight significant stimuli
A bite of moist chocolate cake, a romantic
kiss, or a puff of the day's first cigarette will send the hedonist
in many of us coming back for more. For 20 years, neuroscientists
thought that they understood why people succumb to these cravings,
as well as to the more dangerous ones for addictive drugs such as
heroin and cocaine. Their satisfaction triggers the release from
cells deep inside the brain of the chemical dopamine--a
neurotransmitter supposed to act on the brain's reward system to
produce feelings of pleasure. But recent evidence suggests a
different role for dopamine in producing these behaviors--one much
broader and more subtle than previously envisioned.
Dopamine is still thought to play a critical role in motivation
and reward, but many researchers no longer believe it acts by
directly producing feelings of pleasure or euphoria. Instead, new
data indicate that dopamine release within the brain highlights, or
draws attention to, certain significant or surprising events. These
include not only those the organism finds rewarding, such as
consuming a tasty morsel of food or engaging in sexual activity, but
also events that predict rewards, and stimuli, like loud noises and
flashing lights, that are simply startling. By underscoring such
events, say these researchers, the dopamine signal helps the animal
learn to recognize them--and in some cases, to repeat them.
Although not everyone agrees with this idea, it is winning some
prominent supporters. "These are very exciting years, because the
notion of what dopamine is doing is rapidly changing," says Kent
Berridge, a psychologist at the University of Michigan, Ann Arbor,
whose work in rats suggests dopamine is not a pleasure juice.
Indeed, even psychologist Roy Wise of Concordia University in
Montreal, the primary architect of the pleasure theory, has altered
his position. "I no longer believe that the amount of pleasure felt
is proportional to the amount of dopamine floating around in the
brain," he says. Facilitating learning may be one of its functions,
he agrees.
This new view of dopamine as an aid to learning rather than a
pleasure mediator may help explain why many addictive drugs, which
unleash massive surges of the neurotransmitter in the brain, can
drive continued use without producing pleasure--as when cocaine
addicts continue to take hits long after the euphoric effects of the
drug have worn off or when smokers smoke after cigarettes become
distasteful. It may also provide a better understanding of
schizophrenia and attention-deficit disorder (ADD), both of which
have been linked to abnormal dopamine transmission. People with ADD
and some schizophrenics become so distracted by sensory events that
most people's brains filter out--a passing car, say, or a bouncing
ball--that they can't work or study. The pleasure hypothesis can't
explain these symptoms, but the idea that dopamine works to heighten
attention to external events might.
The new work on dopamine may thus eventually lead to better drug
treatments that fix the attentional filter in ADD patients and
certain schizophrenics. It may also help experts develop new
therapies for drug addiction that work, for example, by blocking the
cravings addicts get when they see objects that remind them of a
drug--and that commonly cause relapse.
Producing pleasure
Before 1990, almost all of
the data on dopamine's role in addiction or reward suggested that
the transmitter signaled something akin to emotional value or
pleasure. In 1975, for example, Wise and Robert Yokel, also at
Concordia, found that rats given high doses of a dopamine-blocking
chemical called pimozide eventually stopped pressing a lever that
caused amphetamine to be delivered into their veins. Because the
rats first tried pressing the lever faster, ruling out the
possibility that their movements were impaired, the result indicates
that amphetamine was no longer rewarding to the animals.
Similarly, David Roberts and Hans Fibiger at the University of
British Columbia (UBC) in Vancouver and their colleagues showed in
the late 1970s that selectively destroying dopamine-producing cells
in a midbrain region called the ventral tegmental area (VTA) shut
down cocaine-seeking behavior by rats. Such experiments, says their
British Columbia colleague Anthony Phillips, "convincingly
implicated dopamine as a reward transmitter." Since then, other
researchers have produced a flood of data showing that a variety of
addictive drugs, from alcohol to heroin, unleash a surge of dopamine
from cells in the VTA. The dopamine is released into a part of the
brain, called the nucleus accumbens, that is known to be activated
by pleasurable behaviors.
Indeed, further work by Wise's team suggested that the dopamine
reward system plays a fundamental role in encouraging behaviors,
such as feeding, needed for life. They found, for example, that
dopamine-blocking agents blunt the reinforcing impact of food in
hungry rats, eventually making them stop running toward food.
Researchers logically concluded that the transmitter, when released,
causes the animals to feel good, and by setting off these positive
emotions, also teaches them to repeat pleasure-producing actions.
Skeptics noted, however, that the hypothesis that dopamine
produces pleasure was hard to reconcile with what's known about
schizophrenia, in which too much of the neurotransmitter seems in
many cases to cause a heightened state of arousal--not pleasure.
When Wise proposed the pleasure hypothesis, psychologist Richard
Katz, then of Johns Hopkins University in Baltimore, commented that
if Wise was right, then schizophrenics should be "inextinguishably
fat and happy. They are just the opposite ... as anyone who has
worked with them knows."
Beginning in the early 1990s, researchers began picking up hints
that even in pleasurable behaviors, dopamine may have a more subtle
role than the pleasure hypothesis suggests. Some of these hints came
from experiments in which they measured dopamine levels in the
nucleus accumbens and found that they surge not just in response to
food and sexual rewards, but also in response to events that predict
those rewards.
Predictive power
In 1993, for example,
Phillips's team in British Columbia found that after hungry rats
learned to associate the sound of a buzzer with food delivery, the
buzzer produced the same dopamine surge in the nucleus accumbens as
the food did. And in experiments reported just this year, Phillips's
team showed that dopamine levels in the accumbens surge in rats
during periods of sexual anticipation before mating. In this
experiment, dopamine levels rose 44% when male rats were exposed to
a female, who was kept out of reach behind a screen. When they were
then allowed to copulate with her, the levels went even higher, but
dropped off again as the males became sexually satiated. However,
the presence of a second female prompted another small dopamine
surge, which increased further, to 34% above baseline, as the male
copulated with her.
At about the same time, another line of work in which researchers
directly measured the activity of dopamine-producing nerve cells,
rather than brain concentrations of the neurotransmitter, added
support to the idea that the neurons' true role is to draw attention
to events that predict rewards rather than to trigger pleasure.
Wolfram Schultz, a neurophysiologist at the University of Fribourg
in Switzerland, and his colleagues were investigating the actions of
dopamine-producing nerve cells in a brain region called the
substantia nigra, which are thought to be important for controlling
movements. Such information would be helpful in understanding
Parkinson's disease, which is caused by a loss of those neurons, but
Schultz found the results of his experiments taking him in an
unexpected direction.
He and his colleagues began by recording the activities of
individual dopamine cells in the substantia nigra of monkeys who
were supposed to perform a series of learned movements. They
expected to see the neurons fire whenever the monkeys moved or
prepared to move. Instead, all the neurons remained silent. Then,
one day, the researchers gave the monkeys a piece of apple as a
reward during one of the experiments. This time, Schultz recalls,
"the neurons started going crazy. We couldn't believe it."
At first, the researchers assumed that the firing simply prepared
an animal to approach the apple morsel. But further work ruled out
that idea. For one thing, dopamine neurons fired just as much in
response to rewarding stimuli when the monkeys were not required to
move as when they were. Indeed, the Schultz team's results began to
point to a more interesting function for the dopamine-producing
neurons: providing incentives to act when something important is
going on. The researchers found, for example, that the cells would
respond only to "behaviorally significant" stimuli, such as rewards.
A piece of apple worked, but a bare wire that normally held the food
did not.
The cells might then simply signal the presence of a reward. But
in 1992 and 1993, Schultz's team showed that the neurons would fire
in response to stimuli that were not in themselves rewarding, such
as a light, if they predicted a food reward. Indeed, after monkeys
learned that a light predicted food, the neurons would stop firing
to the food itself. Because the food presumably remained more
pleasurable than the light, these data cast doubt on the idea that
dopamine is a pleasure transmitter, indicating instead that dopamine
teaches an animal to predict an upcoming reward correctly.
At the 1996 Society for Neuroscience meeting, Schultz and former
postdoc Jeffrey Hollerman presented perhaps the best evidence for
that idea: recordings of single dopamine neurons during an entire
learning episode. They first taught a monkey that touching one of a
pair of pictures would produce a reward. Then, while recording from
a dopamine cell, they switched to two novel pictures, so the monkey
no longer knew which to choose. When the monkey picked the correct
one by chance, the dopamine cell would fire as the monkey got the
reward, but after the monkey knew which one was correct, the cell
greeted the reward with silence. The researchers found the same
pattern in 45 neurons tested with separate picture pairs. The
results, says Schultz, "show that dopamine cells respond to reward
only when it occurs unpredictably--for example, during learning--and
this response, which reveals the difference between what is
predicted and what occurs, looks like a perfect teaching signal."
If dopamine does mark reward-predicting stimuli, that function
should be reflected in the responses of the neurons receiving its
signals in the brain's ventral striatum, a structure that includes
the nucleus accumbens. In 1996, a team led by neuropsychologist
Barry Richmond of the National Institutes of Health in Bethesda,
Maryland, garnered evidence that, indeed, striatal cells seem to
help an animal keep close tabs on its progress toward a reward.
The researchers trained monkeys in a task in which they received
a dollop of juice if they correctly performed a simple
action--letting go of a bar when a spot on a computer screen changed
from red to green--from one to three times. The animals could track
their progress by a second light, which brightened as the juice
reward neared. The monkeys acted as if they understood this light
cue: The closer they were to receiving a reward, the faster they
released the bar and the more accurate they were. Striatal neurons
reflected this knowledge. For example, many of the 138 cells sampled
in two monkeys fired more rapidly when a monkey knew a trial would
produce a reward. "You can figure out where you are in the sequence
of trials by listening to these neurons," Richmond says.
Now, a study in the September issue of Neuron suggests
an anticipatory role for the nucleus accumbens in drug abuse. A team
led by Hans Breiter and Bruce Rosen of Harvard Medical School in
Boston and Steven Hyman, who recently became director of the
National Institute of Mental Health, used functional magnetic
resonance imaging to scan the brains of cocaine addicts under the
drug's influence. This revealed the areas active when the subjects
felt a brief, euphoric "rush," and then those that remained active
after the euphoria wore off and the subjects were craving, or
anticipating, another hit. Despite its previous association with
pleasure, the nucleus accumbens remained active during the craving
stage, suggesting a tie between craving and the dopamine neurons
located there.
An attention-getting device?
Meanwhile, very
recent neurophysiological evidence suggests that the dopamine system
may have an even broader role in learning: It may highlight novel or
startling sensory events that attract an animal's attention but are
not related to reward. A team led by Jon Horvitz of Columbia
University in New York City and Princeton University's Barry Jacobs
exposed cats, at unpredictable times, to either loud clicks or
flashes of a bright light, and found that the neurons in the
animals' VTAs responded to both the clicks and the light flashes,
most commonly with a burst of activity of the type that unleashes a
surge of dopamine. The results suggest, Horvitz says, "that dopamine
neurons respond to salient events regardless of whether the salience
derives from conditioned reward properties or from physical
characteristics of the stimulus," such as a sudden onset.
Still, not all recent experiments have supported dopamine's
proposed role as an attention-getter. In 1995, for instance, a team
led by UBC's Fibiger repeatedly exposed a group of hungry rats to a
delicious liquid behind a wire mesh screen before lifting the screen
and allowing them to eat. They discovered a significant rise in
dopamine in the nucleus accumbens when the rats consumed the food,
but not while they eyed the food behind the screen. "My belief is
that changes in dopamine release are more reliably associated with
consumption of reward than anticipation of reward," Fibiger says.
Neurobiologists aren't sure why the results obtained by direct
recording from dopamine neurons don't always jibe with data--like
Fibiger's--that are based on dopamine levels. But the different time
scales of these two data types may offer an explanation. Because
dopamine levels are measured over minutes, they may miss the tiny,
momentary increases produced by dopamine-cell firing. "When the
smoke clears, I think dopamine will be playing a role in learning
the value of stimuli that are important in an animal's environment,"
Phillips predicts.
The smoke hasn't cleared quite yet, however. Other researchers
are still racking up correlations between dopamine and pleasure.
Earlier this year, an imaging study by Nora Volkow's team at
Brookhaven National Laboratory in Upton, New York, linked the
euphoria produced by cocaine to dopamine levels in a part of the
striatum next to the nucleus accumbens. The researchers gave
different doses of the drug to 17 people who were already cocaine
users. Then, to get an indication of dopamine levels in the
subjects' striata, they used positron emission tomography and
radioactive cocaine to detect cocaine binding to dopamine
transporter proteins. These proteins pick up dopamine released by
neurons and carry it back into the cells so it can be used again.
When cocaine binds to the transporters, it blocks them and results
in increased brain dopamine levels.
Volkow and her colleagues found that the intensity of the
subjects' highs increased in proportion to the percentage of
dopamine transporters blocked by the drug in the striatum--which,
they believe, reflect what's happening with the transporters in the
nucleus accumbens. "The ensuing rise in dopamine triggers the high,"
Volkow concludes.
Given the conflicting results so far, more work will be needed to
resolve whether dopamine has some role in conscious emotion, such as
pleasure, or whether it is mainly a subconscious signal in learning
and attention. One line of research that could settle the debate is
directed at dopamine's role, if any, in unpleasant events. Some data
suggest that dopamine helps animals learn and remember unpleasant
stimuli such as electric shocks. The evidence so far is
controversial, but if it holds up, it would buttress the idea that
dopamine functions primarily as a general purpose attention-getter.
Researchers could also bolster dopamine's proposed role in
learning by unraveling the brain mechanism through which the
neurotransmitter would exert this effect. One clue to the mechanism
may be the connections between dopamine cells in the VTA and the
frontal cortex, where short-term memories are held. Some scientists
think dopamine may cause frontal neurons to hold onto some temporary
memories for longer, increasing the chances that they will stick in
the mind. "I see dopamine as a possible modulatory cue that allows
short-term memories to hold their state for a while, such that
there's a better chance of remembering them the next day," says
Donald Woodward, a neurophysiologist who studies cocaine reward at
the Bowman Gray School of Medicine in Winston-Salem, North Carolina.
Ultimately, scientists hope their dopamine research will lead to
more general insights into how mammalian brains work, and equally
important, how they malfunction. If the dopamine signal serves to
draw attention to salient events of all sorts as Horvitz's work
suggests, then its overactivity in schizophrenia might explain some
of the disorder's symptoms. "It is very possible," says Horvitz,
"that the bizarre associations schizophrenics make, and the
significance they attach to stimuli that we would consider
irrelevant, are due to the fact that the stimuli are entering the
nervous system under a state of high dopamine activity."
Once scientists unravel the larger neural circuit in which
dopamine acts, they may find that as dopamine delivers its message,
other parts of the brain may reverberate with a conscious
emotion--pleasure, excitement, or even fear--depending on the
circumstances. But it is now doubtful that dopamine's cry directly
summons such feelings; instead, it may make a person look up in
surprise, like a schoolchild just called on by a teacher, to warn
that it's time to listen to the lesson.
Additional Reading
J. Altman et al., "The biological, social and clinical
bases of drug addiction: Commentary and debate," Offprint from
Psychopharmacology 125, 285 (1996).
E. M. Bowman et al., "Neural signals in the monkey ventral
striatum related to motivation for juice and cocaine rewards,"
Journal of Neurophysiology 75, 1061 (1996).
F. Fiorino, A. Coury, A. G. Phillips, "Dynamic changes in nucleus
accumbens dopamine efflux during the Coolidge effect in male rats,"
Journal of Neuroscience 17, 4849 (1997).
J. C. Horvitz et al., "Burst activity of ventral tegmental
dopamine neurons is elicited by sensory stimuli in the awake cat,"
Brain Research 759, 251 (1997).
W. Schultz, P. Dayan, P. R. Montague, "A neural substrate of
prediction and reward," Science 275, 1593
(1997).
N. D. Volkow et al., "Relationship between subjective
effects of cocaine and dopamine transporter occupancy,"
Nature 386, 827 (1997).
R. A. Wise, "Neurobiology of addiction," Current Opinion in
Neurobiology 6, 243 (1996).
