From: Lee Corbin (lcorbin@tsoft.com)
Date: Sat Jun 14 2003 - 14:48:53 MDT
Hi Brett,
Sorry I mangled Antonio Damasio's name. Besides, one could
have reasonable read what I wrote as googlable, or being a
genuine reference.
Here a description of what I was refering to: Patients with
some kinds of brain damage behave exactly like normals on all
the usual tests. But one kind of test showed what was going
on, and moreover, illustrated how our emotions are *needed*
(at present) in our thinking. Here is an excerpt from my
book:
RISK TAKING: THE GAMBLING EXPERIMENTS Chapter in Damasio's
"Descartes Error" page 213-215.
Another approach we took to testing the somalic-marker hypothesis
made use of a task designed by my postdoctoral student.
Frustrated, as all researchers are, by the artificial nature of lost
experimental neuropsychological tasks, he wanted to develop as
lifelike a means as possible to assess decision-making performance.
The clever set of tasks that he devised, and further refined in
collaboration with Hanna Damasio and Steven Anderson, have come to be
known in our laboratory, predictably enough, as the "Gambling
Experiments." Overall, the setting for the experiments is colorful,
a far cry from the boring manipulations of most other such
situations. Normals and patients alike enjoy it, and the nature of
the investigation makes for amusing episodes. I recall the bulging
eyes and dropped jaw of a distinguished visitor who came to my office
after walking by the lab where an experiment was in progress.
"There're people gambling in here!" he informed me in a whisper.
In the basic experiments the subject, known as the "Player," sits in
one of four decks or cards labeled A, B, C, and D. The Player is
given a loan of $5,000 (play money but looking like the real thing)
and told that the goal of the game he is about to play is to lose as
little as possible of the loan and try to make as much extra money as
possible. Play consists of turning cards, one at a lime, from any of
the four leeks, until the experimenter says to stop. The Player thus
does not know the total number of turns required to end the game.
The Player is told also that turning any and every card will result
in earning a sum money, and that every now and then turning some
cards will result in both earning money and having to pay a sum of
money to the experimenter. Neither the amounts of gain or loss in
any card, nor the cards' connection to a specific deck, nor the order
of their appearance is disclosed at the outset. The amount to be
earned or paid with a given card is disclosed only after the card is
turned. No other instruction is provided. The tally of how much
has been earned or lost at any point is not disclosed, and the
subject is not allowed to keep written notes.
The turning of any card in decks A and B pays a handsome $100, while
the turning of any card in decks C and D only pays $50. Cards keep
being turned on any deck, and quite unpredictably, certain cards in
decks A and B (the $100-paying decks) require the Player to make a
sudden nigh payment, sometimes as much as $1,250. Likewise, certain
cards in decks C and D (the $50-paying decks) also require a payment,
but the sums are much smaller, less than $100 on the average. These
undisclosed rules are never changed. Unbeknownst to the Player, the
game will be terminated after 100 plays. There e is no way for the
Player to predict, at the outset, what will happen , and no way to
keep in mind a precise tally of gains and losses as the game
proceeds. Just as in life, where much of the knowledge by which we
live and by which we construct our adaptive future is doled out bit
by bit, as experience accrues, uncertainty reigns. Our
knowledge---and the Player's ---is shaped by both the world with
which we interact and by the biases inherent in our organism, for
example, our preferences for gain over loss, for reward over
punishment, for low risk over high risk.
What regular folds do in the experiment is interesting. They begin
by sampling from all four decks, in search of patterns and clues.
Then, more often than not, perhaps lured by the experience of high
reward from turning cards in the A and B decks, they show an early
preference for those decks. Gradually,k however, within the first
thirty moves, they switch the preference to decks C and D. In
general, they stick to this strategy until the end, although
self-professed high-risk players may resample decks A and B
occasionally, only to return to the apparently more prudent course of
action.
There is no way for players to carry out a precise calculation of
gains and losses. Rather, bit by bit, they develop a hunch that some
decks---namely, A and B---are more "dangerous" than others. One
might say they intuit that the lower penalties in decks C and D will
make them come out ahead in the long run, despite the smaller initial
gain. I suspect that before and beneath the conscious hunch there is
a nonconscious process gradually formulating a prediction for the
outcome of each move, and gradually telling the mindful player, at
first softly but then ever louder, that punishment or reward is about
to strike *if* a certain move is indeed carried out. In short, I
doubt that it is a matter of only fully conscious process, or only
fully nonconscious process. It seems to take both types of
processing for the well-tempered decision-making brain to operate.
The behavior of ventromedial frontal patients in this experiment was
most informative. What they did in the card game resembled what they
often have done in real life since they sustained their brain lesion,
and differed from what they would have done before the lesion.
Perhaps this can now be found on-line somewhere too. I'll look around,
because I consider this experiment absolutely fundamental.
Lee
This archive was generated by hypermail 2.1.5 : Sat Jun 14 2003 - 15:00:00 MDT