Perceived agency and the pre-feedback delay: a time-frequency analysis

Abstract

In everyday life, it is often the case that an individual must reflect on which, if any, of their actions led to a favorable or unfavorable outcome, especially if a sizeable period of time passed prior to knowing the outcome. If they attribute these outcomes to themselves, rather than to some external source, they would be said to have a "sense of agency." As of this paper, only event-related potentials and fMRI methods have been used to assess this phenomenon. In this experiment, we re-analyzed the data from a previous SPN study using a more powerful analytic method, time-frequency analysis, and tested a variety of hypotheses concerning previously established relationships between oscillatory activity and brain regions involved in feedback anticipation. Participants took part in a gambling task where their "sense of agency" and the valence of the expected outcome (potential loss versus potential gain) were manipulated. In choice trials, it was implied that participants' decisions influenced whether they won or lost money, while in no-choice trials, participants were correctly informed that the computer determined these outcomes. We found that pre-feedback alpha suppression in sites overlying right frontal and occipital areas was stronger on choice trials. Pre-feedback beta suppression in sites overlying left frontal sites was also greater on these trials. Both of these effects appeared to be modulated by contextual valence--effects were significantly larger in conditions with potential gain than those involving potential loss. These findings highlight the importance of selective attention and reward anticipation when the outcome of one's own action is awaited.

In everyday life, it is often the case that an individual must reflect on which, if any, of their actions led to a favorable or unfavorable outcome, especially if a sizeable period of time passed prior to knowing the outcome. If they attribute these outcomes to themselves, rather than to some external source, they would be said to have a "sense of agency." As of this paper, only event-related potentials and fMRI methods have been used to assess this phenomenon. In this experiment, we re-analyzed the data from a previous SPN study using a more powerful analytic method, time-frequency analysis, and tested a variety of hypotheses concerning previously established relationships between oscillatory activity and brain regions involved in feedback anticipation. Participants took part in a gambling task where their "sense of agency" and the valence of the expected outcome (potential loss versus potential gain) were manipulated. In choice trials, it was implied that participants' decisions influenced whether they won or lost money, while in no-choice trials, participants were correctly informed that the computer determined these outcomes. We found that pre-feedback alpha suppression in sites overlying right frontal and occipital areas was stronger on choice trials. Pre-feedback beta suppression in sites overlying left frontal sites was also greater on these trials. Both of these effects appeared to be modulated by contextual valence--effects were significantly larger in conditions with potential gain than those involving potential loss. These findings highlight the importance of selective attention and reward anticipation when the outcome of one's own action is awaited.