Deciding Advantageously Before Knowing the Advantageous Strategy

See allHide authors and affiliations

Science  28 Feb 1997:
Vol. 275, Issue 5304, pp. 1293-1295
DOI: 10.1126/science.275.5304.1293


Deciding advantageously in a complex situation is thought to require overt reasoning on declarative knowledge, namely, on facts pertaining to premises, options for action, and outcomes of actions that embody the pertinent previous experience. An alternative possibility was investigated: that overt reasoning is preceded by a nonconscious biasing step that uses neural systems other than those that support declarative knowledge. Normal participants and patients with prefrontal damage and decision-making defects performed a gambling task in which behavioral, psychophysiological, and self-account measures were obtained in parallel. Normals began to choose advantageously before they realized which strategy worked best, whereas prefrontal patients continued to choose disadvantageously even after they knew the correct strategy. Moreover, normals began to generate anticipatory skin conductance responses (SCRs) whenever they pondered a choice that turned out to be risky, before they knew explicitly that it was a risky choice, whereas patients never developed anticipatory SCRs, although some eventually realized which choices were risky. The results suggest that, in normal individuals, nonconscious biases guide behavior before conscious knowledge does. Without the help of such biases, overt knowledge may be insufficient to ensure advantageous behavior.

In a gambling task that simulates real-life decision-making in the way it factors uncertainty, rewards, and penalties, the players are given four decks of cards, a loan of $2000 facsimile U.S. bills, and asked to play so that they can lose the least amount of money and win the most (1). Turning each card carries an immediate reward ($100 in decks A and B and $50 in decks C and D). Unpredictably, however, the turning of some cards also carries a penalty (which is large in decks A and B and small in decks C and D). Playing mostly from the disadvantageous decks (A and B) leads to an overall loss. Playing from the advantageous decks (C and D) leads to an overall gain. The players have no way of predicting when a penalty will arise in a given deck, no way to calculate with precision the net gain or loss from each deck, and no knowledge of how many cards they must turn to end the game (the game is stopped after 100 card selections). After encountering a few losses, normal participants begin to generate SCRs before selecting a card from the bad decks (2) and also begin to avoid the decks with large losses (1). Patients with bilateral damage to the ventromedial prefrontal cortices do neither (1, 2).

To investigate whether subjects choose correctly only after or before conceptualizing the nature of the game and reasoning over the pertinent knowledge, we continuously assessed, during their performance of the task, three lines of processing in 10 normal participants and in 6 patients (3) with bilateral damage of the ventromedial sector of the prefrontal cortex and decision-making defects. These included (i) behavioral performance, that is, the number of cards selected from the good decks versus the bad decks; (ii) SCRs generated before the selection of each card (2); and (iii) the subject's account of how they conceptualized the game and of the strategy they were using. The latter was assessed by interrupting the game briefly after each subject had made 20 card turns and had already encountered penalties, and asking the subject two questions: (i) “Tell me all you know about what is going on in this game.” (ii) “Tell me how you feel about this game.” The questions were repeated at 10-card intervals and the responses audiotaped.

After sampling all four decks, and before encountering any losses, subjects preferred decks A and B and did not generate significant anticipatory SCRs. We called this period pre-punishment. After encountering a few losses in decks A or B (usually by card 10), normal participants began to generate anticipatory SCRs to decks A and B. Yet by card 20, all indicated that they did not have a clue about what was going on. We called this period pre-hunch (Fig. 1). By about card 50, all normal participants began to express a “hunch” that decks A and B were riskier and all generated anticipatory SCRs whenever they pondered a choice from deck A or B. We called this period hunch. None of the patients generated anticipatory SCRs or expressed a “hunch” (Fig. 1). By card 80, many normal participants expressed knowledge about why, in the long run, decks A and B were bad and decks C and D were good. We called this period conceptual. Seven of the 10 normal participants reached the conceptual period, during which they continued to avoid the bad decks, and continued to generate SCRs whenever they considered sampling again from the bad decks. Remarkably, the three normal participants who did not reach the conceptual period still made advantageous choices (4). Just as remarkably, the three patients with prefrontal damage who reached the conceptual period and correctly described which were the bad and good decks chose disadvantageously. None of the patients generated anticipatory SCRs (Fig. 1). Thus, despite an accurate account of the task and of the correct strategy, these patients failed to generate autonomic responses and continued to select cards from the bad decks. The patients failed to act according to their correct conceptual knowledge.

Fig. 1.

Presentation of the four periods in terms of average numbers of cards selected from the bad decks (A and B) versus the good decks (C and D), and the mean magnitudes of anticipatory SCRs associated with the same cards. The pre-punishment period covered the start of the game when subjects sampled the decks and before they encountered the first loss (that is, up to about the 10th card selection). The pre-hunch period consisted of the next series of cards when subjects continued to choose cards from various decks, but professed no notion of what was happening in the game (on average, between the 10th (range: 7 to 13) and the 50th card (range: 30 to 60) in normals, or between the 9th (3 to 10) and the 80th card (60 to 90) in patients. The hunch period (never reached in patients) corresponded to the period when subjects reported “liking” or “disliking” certain decks, and “guessed” which decks were risky or safe, but were not sure of their answers [on average, between the 50th (30 to 60) and 80th card (60 to 90) in normals]. The conceptual period corresponded to the period when subjects were able to articulate accurately the nature of the task and tell for certain which were the good and bad decks, and why they were good or bad [on average, after the 80th card (60 to 90) in both normals and patients]. (Top panels) Bars represent means (±SEM) of the mean magnitude of anticipatory SCRs generated before the selection of cards from the bad decks versus the good decks. Anticipatory SCRs are generated in the time window before turning a card from any given deck, that is, during the time the subject ponders from which deck to choose (2). SCRs in association with the good and bad decks from normal controls or patients were not significantly different during the pre-punishment (baseline) period. However, there was a significant increase in the magnitude of these SCRs during the pre-hunchperiod, but only for normal controls. During the next two periods, SCR activity in normal subjects was sustained in the case of the bad decks, but it began to subside in the case of the good decks (8). (Bottom panels) Bars in the “Behavioral responses” plots represent means (±SEM) of the mean number of cards selected from the bad decks versus those selected from the good decks. Normal controls selected more cards from the good decks during the pre-hunch, hunch, and conceptual periods. In contrast, prefrontal patients selected more cards from the bad decks during these periods (9).

On the basis of these results, we suggest that the sensory representation of a situation that requires a decision leads to two largely parallel but interacting chains of events (Fig. 2). In one, either the sensory representation of the situation or of the facts evoked by it activate neural systems that hold nondeclarative dispositional knowledge related to the individual's previous emotional experience of similar situations (5). The ventromedial frontal cortices are among the structures that we suspect hold such dispositional knowledge, the activation of which, in turn, activates autonomic and neurotransmitter nuclei (such as those that deliver dopamine to selected cortical and subcortical forebrain regions), among other regions. The ensuing nonconscious signals then act as covert biases on the circuits that support processes of cognitive evaluation and reasoning (6). In the other chain of events, the representation of the situation generates (i) the overt recall of pertinent facts, for example, various response options and future outcomes pertaining to a given course of action; and (ii) the application of reasoning strategies to facts and options. Our experiment indicates that in normal participants, the activation of covert biases preceded overt reasoning on the available facts. Subsequently, the covert biases may have assisted the reasoning process in cooperative manner, that is, biases would not decide per se, but rather facilitate the efficient processing of knowledge and logic necessary for conscious decisions (7). We suspect that the autonomic responses we detected are evidence for a complex process of nonconscious signaling, which reflects access to records of previous individual experience—specifically, of records shaped by reward, punishment, and the emotional state that attends them. In this light, damage to ventromedial cortices acts by precluding access to a particular kind of record of previous and related individual experience.

Fig. 2.

Diagram of the proposed steps involved in decision-making.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
View Abstract

Stay Connected to Science

Editor's Blog

Navigate This Article