“These neurons were not registering
whether something aversive was surprising,” Salzman says, suggesting they play a
specialized role in rewarding surprises.
Other amygdala neurons did the opposite — registering surprise when it was
aversive but not rewarding. By carrying
two different sets of neurons for surprise, the amygdala may be better able
to predict what’s coming in order to prepare appropriate actions, Salzman says.
“If you get something that’s highly
rewarding by surprise, you want to be
able to take actions in advance that
would help you acquire that rewarding
stimulus again,” he says. “By contrast, if
you receive a bad surprise you normally
would want to take some defensive
actions or avoidance actions to avert a
Ensuring a good hand
But the same stimulus, depending on the
circumstances, can be rewarding or not,
Salzman explains. He uses an example
from the card game blackjack.
The object of the game is to hold
cards with a total value as close as possible to 21, without going over. Players
are initially dealt two cards and then can
request additional cards.
Say a player is dealt two cards with
a total value of 11 and then gets a 10 of
hearts, worth 10 points, to reach a perfect
21. Now suppose that in the next hand,
after the cards are shuffled, the same
player is dealt two cards that total 15.
Cell-by-cell responses different nerve cells in the amygdala respond differently depending
on the stimulus. a study in monkeys found that while one set of cells (a) responded more robustly
to a reward, a sip of water, another (b) responded more to an annoying puff of air.
Neuron set A response to stimuli Neuron set b response to stimuli
Listening in on nerve cell chatter,
on the fly. “If you’re an animal out there
and you’ve just learned something about
either a harmful stimulus or a predator,
you may not get a second chance.”
Salzman says, evidence that the amyg-
dala’s emotional response to the same
sensory stimulus can be flexible.
Murray sees such flexibility among
monkeys making food choices in the lab.
Murray’s group is now using a simi-
lar method to see how values assigned
to various rewards can be linked with
actions. In a recent experiment, Sarah
Rhodes, also of the National Institute
of Mental Health, trained monkeys to
Given a choice between two
foods, monkeys (like most
people) will invariably reach
for their favorite. But after
gorging on a favored food
for a while, the reward value
of that treat diminishes, the
way the value of the 10 drops
when a blackjack player
already has 15 points in hand.
“I’m kind of
glad that the
brain isn’t as
simple as we
make it out
perform two actions, tap and
hold, to receive a reward. Six
rapid taps to a touch-sensi-tive screen were rewarded
with one kind of food, while
a continuous touch for two
seconds was rewarded with
At that point the monkeys
shift their choices and select a new food.
Monkeys without a working amygdala,
however, show little change in behavior
and continue to stuff themselves with
the first food choice, Murray’s studies
have found. These findings suggest that
the amygdala is needed to revise the
assessment of a reward’s value.
Rhodes, Murray and their
colleagues wanted to see if
the animals, once filled with
their favorite food, would shift their
actions along with their choices. Normal monkeys did, choosing a new action
once they got their fill of the first food.
Animals without an amygdala, however,
didn’t make the switch.
“What we think is happening is that
the amygdala is helping update the food
value, and sending that information
somewhere else — probably the orbitofrontal cortex,” Murray says. This same
circuitry may also help animals, including humans, automatically calculate and
update the value of goals or outcomes
Those findings, reported in November
in San Diego at the annual meeting of the
Society for Neuroscience, suggest that the
intact control animals engaged in goal-directed behavior, and that the amygdala
played a role, Murray says.
“It means the animals literally know
and somehow represent a goal. And that
if the value of the goal changes, they can
therefore change their behavior,” she
says. “We think this is what humans do,
so it’s really helpful to understand the
Another recent study shows that
people, like monkeys, do indeed use the
amygdala to judge the value of a particular item. Rick Jenison, who studies
neuroeconomics at the University of
Wisconsin–Madison, and his colleagues
recruited three volunteers who were
already undergoing a procedure that
required electrodes to be implanted in the
brain. Using the electrodes, the researchers eavesdropped on 51 amygdala nerve
cells as the volunteers put price tags on a
series of junk food items.
s.e. morrison and c.d. salzman/current opinion in neurobiology 2010