Iowa gambling task

From Wikipedia, the free encyclopedia

The Iowa gambling task is a psychological task thought to simulate real-life decision making. It was introduced by Bechara, Damasio, Tranel and Anderson (1994), then researchers at the University of Iowa. It has been brought to popular attention by Antonio Damasio, proponent of the Somatic markers hypothesis and author of Descartes' Error. The task is sometimes known as Bechara's Gambling Task, and is widely used in research of cognition and emotion.

Cost-benefit analysis of Iowa Gambling Task decks

Participants are presented with 4 virtual decks of cards on a computer screen. They are told that each time they choose a card they will win some game money. Every so often, however, choosing a card causes them to lose some money. The goal of the game is to win as much money as possible. Every card drawn will earn the participant a reward. Occasionally, a card will also have a penalty. Thus, some decks are "bad decks", and other decks are "good decks", because some will lead to losses over the long run, and others will lead to gains. The decks differ from each other in the number of trials over which the losses are distributed.

Screen shot of the Iowa Gambling Task
Screen shot of the Iowa Gambling Task

Most healthy participants sample cards from each deck, and after about 40 or 50 selections are fairly good at sticking to the good decks. Patients with orbitofrontal cortex (OFC) dysfunction, however, continue to perseverate with the bad decks, sometimes even though they know that they are losing money overall. Concurrent measurement of galvanic skin response shows that healthy participants show a "stress" reaction to hovering over the bad decks after only 10 trials, long before conscious sensation that the decks are bad. By contrast, patients with OFC dysfunction never develop this physiological reaction to impending punishment. Bechara and his colleagues explain this in terms of the somatic marker hypothesis. The Iowa gambling task is currently being used by a number of research groups using fMRI to investigate which brain regions are activated by the task in healthy volunteers as well as clinical groups with conditions such as schizophrenia and obsessive compulsive disorder.

[edit] Critiques of the Iowa Gambling Task

Although the IGT has achieved prominence, it is not without its critics. Criticisms have been raised over both its design and its interpretation. Published critiques include:

  • A paper by Dunn, Dalgliesh and Lawrence [1]
  • Research by Lin, Chiu, Lee and Hsieh [2], who argue that a common result (the “prominent deck B” phenomenon) argues against some of the interpretations that the IGT has been claimed to support.
  • Research by Chiu and Lin [3], the “sunken deck C” phenomenon was identified, which confirmed a serious confounding embedded in the original design of IGT, this confounding makes IGT serial studies misinterpret the effect of gain-loss frequency as final-outcome for Somatic marker hypothesis.
  • A research group in Taiwan utilized an IGT-modified and relatively symmetrical gamble for gain-loss frequency and long-term outcome, namely the Soochow Gambling Task (SGT) [4] demonstrated a reverse finding of Iowa Gambling Task. Normal decision makers in SGT were mostly occupied by the immediate perspective of gain-loss and inability to hunch the long-term outcome in the standard procedure of IGT (100 trials under uncertainty). Richard Peterson [5] [6]in his book, Inside the investor’s brain [7], considered the serial findings of SGT may be congruent with the Nassim Taleb’s [8] suggestion on some fooled choices in investment.

[edit] References

  • Bechara A, Damasio AR, Damasio H, Anderson SW (1994). "Insensitivity to future consequences following damage to human prefrontal cortex", Cognition 50: 7-15.

[edit] External links

  • A free implementation of the Iowa Gambling task is available as part of the PEBL Project [9]
  • Another, web based, implementation that will also run as a standalone application is available here.
  • An italian implementation is available here.
Languages