A card player reaches for a card from one of four decks of playing cards, each of which is worth a different monetary value. But before making his pick, a bead of sweat drips down his forehead. He quickly changes his mind and chooses a card from a deck that has netted him the best returns in previous rounds of the research experiment he's participating in.

What prevented the participant from selecting the card that might have cost him money? His damp brow? Or was it the knowledge that the card wasn't the most valuable choice?

The predominant thinking was the former, until fourth-year Carnegie Mellon graduate student Tiago Maia and James McClelland, PhD, his adviser, co-authored a paper in Proceedings of the National Academy of Sciences (Vol. 101, No. 45, pages 16,075-16,080) challenging the popular presumption, known as the somatic marker hypothesis.

Proposed by neurologist Antonio Damasio, MD, in his book "Descartes' Error: Emotion, Reason and the Human Brain" (Quill, 1994), the somatic marker hypothesis suggests that bodily states can guide decision-making when conscious knowledge is not available, leading people to make advantageous decisions before they are conscious of which options are good and which are bad.

Specifically, the hypothesis maintains that when people choose between options, each possible outcome elicits a bodily state, or somatic marker, such as increased perspiration. That marker corresponds to an emotional reaction that influences decision-making, guiding the participant to make an advantageous choice even in the absence of conscious knowledge.

Maia's findings show that people in such situations may actually be making choices based on their knowledge of the situation in ways that researchers had previously overlooked. The findings have implications for future decision-making theories and understanding rational behavior, says McClelland.

"Researchers need to look more carefully at what people know when they make choices, but also at whether they actually use what they know when they do," McClelland explains. "To what extent can people actually make effective use of what they know in real- time situations?"


Much of the support for Damasio's hypothesis stems from a study involving the Iowa Gambling Task, a game in which participants select a card from one of four playing card decks. On every card, participants win play money. Two of the decks always award $100 and the others give $50. However, at random the $100 cards take money, in addition to awarding it.

In the long run, the decks that give $100 rewards produce net losses, while the decks that give $50 rewards produce net gains. Throughout the game, Damasio and his colleagues measured participants' skin conductance. After the first 20 trials of the task and then at 10-trial intervals, they asked participants questions such as, "Tell me all you know about what is going on in this game."

Damasio and his colleagues found that participants began making the correct selections in the game before they could state that those were the best selections. He also found that when participants were about to make a bad selection they exhibited a higher skin conductance response than when they were about to make a good selection.

From such findings, Damasio argued that the skin conductance responses reflected somatic markers that enabled participants to make advantageous selections before they demonstrated conscious knowledge.


However, Maia wondered if the questions Damasio asked in the study were sensitive enough to detect participants' knowledge.

"Damasio and his colleagues used broad, open-ended questions," Maia says. "And with open-ended questions you often get incomplete answers. As I read the study, I felt that Damasio's questions might have led to a very incomplete picture of participants' knowledge."

Maia also noticed that a series of recent research results suggested alternative interpretations for other evidence that had previously been used to support the somatic marker hypothesis, such as Damasio and his colleagues' findings that patients with damage to their ventromedial prefrontal cortex (VMPFC) continue selecting $100 decks in the Iowa Gambling Task after normally functioning participants have changed their strategy. While Damasio and his colleagues suggested that the VMPFC damage prevented the somatic marker effects that the normally functioning participants benefited from, Maia noticed that other research studies suggest that patients with VMPFC damage suffer from a reversal learning deficit in which they have trouble adapting to unforeseen developments.

With those findings in mind, Maia and his adviser, McClelland, set out to investigate whether it was indeed the case that, as Damasio and his colleagues argued, participants made advantageous decisions before figuring out the winning strategy in the Iowa Gambling Task.

In their study, Maia and McClelland restructured Damasio's open-ended questions into detailed, more specific questions that clarified participants' understanding of the game, such as "Suppose you were to select 10 cards from deck X. What would you expect your average net result to be?"

In doing so, they found that participants had explicit, reportable knowledge that could provide the basis for their judgments and behaviors. In fact, participants' selections were entirely in line with their knowledge of the goodness or badness of each deck, as well as their knowledge of the outcomes of each deck, Maia says.

According to Maia, the findings suggest that the participants displayed conscious knowledge when probed, thereby countering Damasio's claim that somatic markers influenced participants' decision-making. Yet Maia is careful to warn that his study does not end the debate.

"Our results do not prove that the somatic marker hypothesis is wrong," Maia says. Indeed, he notes, his and McClelland's findings did not explore the extent to which the participants actually based their behavior on conscious knowledge at the time of their decision.

Moreover, Damasio notes that Maia and McClelland didn't examine the presence or absence of emotion-related signals, the main tenet of the somatic marker hypothesis.

"Consciousness is not the issue with the somatic marker hypothesis," Damasio argues. "The main point is that somatic markers are emotion-related signals which are either conscious or unconscious. However, to be conscious of a somatic marker is very different than being conscious of knowledge of facts, options, outcomes and strategies involved in decision-making."

Despite these caveats, the fact that participants in Maia's study showed knowledge of the advantageous strategy may point to a more important role for conscious knowledge in advantageous decision-making than has recently been recognized, which is in line with the credo "think before you act," Maia notes.

Indeed, future research ought to explore whether a variety of judgment and decision-making tasks would yield results similar to Maia's findings, says McClelland, since the study only addressed a single task.


Maia's interest in decision-making stems from his post-undergraduate work. After graduating with a computer science and engineering licenciatura degree--a five-year degree in his native Portugal--Maia worked as a software engineer for three years, including one year at the European Space Agency, where he developed rovers capable of using artificial intelligence in order to operate autonomously during planetary exploration.

Then, after earning a master's degree in computer science from the University at Buffalo of the State University of New York , he decided to apply his computational background to psychology and neuroscience at Carnegie Mellon's psychology department. In addition to his recent work on the relationship between conscious and unconscious processing, Maia is examining the neuroprocessing of patients with obsessive-compulsive disorder as well as those with brain damage.

"I originally chose to pursue computer science because of its powerful set of tools for understanding and stimulating intelligence," he says. "However, I always kept an eye on the big prize: understanding the human mind."