It's worse than you thought: The feedback negativity and violations of reward prediction in gambling tasks

Abstract

The reinforcement learning theory suggests that the feedback negativity should be larger when feedback is unexpected. Two recent studies found, however, that the feedback negativity was unaffected by outcome probability. To further examine this issue, participants in the present studies made reward predictions on each trial of a gambling task where objective reward probability was indicated by a cue. In Study 1, participants made reward predictions following the cue, but prior to their gambling choice; in Study 2, predictions were made following their gambling choice. Predicted and unpredicted outcomes were associated with equivalent feedback negativities in Study 1. In Study 2, however, the feedback negativity was larger for unpredicted outcomes. These data suggest that the magnitude of the feedback negativity is sensitive to violations of reward prediction, but that this effect may depend on the close coupling of prediction and outcome.

When decisions of others matter to me: an electrophysiological analysis

“…Because in the original Gehring and Willoughby study events were included in which a small loss was the most advantageous outcome of a trial (because the other response option would have resulted in a greater loss), it appears that the FRN responds to the gain/loss status of an event as opposed to whether or not the choice was erroneous or disadvantageous. In other studies, a strong influence of context on the FRN elicited by losses and negative events has been reported [16]. …”

Can event-related potentials serve as neural markers for wins, losses, and near-wins in a gambling task? A principal components analysis

“…While the negative deflection to unfavourable outcomes described above has been reported in a wide variety of circumstances, including simulated gambling (Hewig et al, 2007), guessing tasks Hajcak et al, 2005Hajcak et al, , 2007, time estimation tasks (Holroyd and Krigolsen, 2007;Miltner et al, 1997;Nieuwenhuis et al, 2005), and learning tasks (De Pascalis et al, 2010), the true nature of the FRN remains somewhat unclear, as this component is commonly superimposed on large amplitude P300 responses that occur immediately after it. It has been proposed that the reduced amplitude FRN observed following win outcomes may not be an actual attenuated response to these events, but is rather driven by larger P300 amplitudes following favourable outcomes (Yeung and Sanfey, 2004).…”

“…It has been proposed that the reduced amplitude FRN observed following win outcomes may not be an actual attenuated response to these events, but is rather driven by larger P300 amplitudes following favourable outcomes (Yeung and Sanfey, 2004). Furthermore, the relative contribution of negative and positive outcomes to the FRN remains unclear due to the fact that many studies have employed the computation of a difference waveform to measure FRN magnitude (e.g., Dunning and Hajcak, 2007;Foti and Hajcak, 2009;Hajcak et al, 2007;Holroyd et al, 2008;Miltner et al, 1997), and other recent research has suggested that, rather than a negative deflection to non-reward outcomes, the FRN is better conceptualised as a positive deflection that is greater following reward compared to non-reward outcomes (Foti et al, 2011;Holroyd et al, 2003;Holroyd et al, 2008).…”

“…Typically examined as a global component, the P300 (called the LPC in many studies), has also been shown to be sensitive to various aspects of incentive value on tasks that simulate gambling (Bellbaum et al, 2010;Hajcak et al, 2007). The inverse relationship between probability and P300 amplitude has been well established (Donchin and Coles, 1988), although the understanding of these results is subject to different interpretations (see Gonsalvez et al, 2007;Verleger, 1988).…”

“…The inverse relationship between probability and P300 amplitude has been well established (Donchin and Coles, 1988), although the understanding of these results is subject to different interpretations (see Gonsalvez et al, 2007;Verleger, 1988). Nevertheless, studies that have controlled for event probability have demonstrated that the P300 remains sensitive to manipulations of win and loss outcomes (e.g., Hajcak et al, 2007;Wu and Zhou, 2009;Yeung et al, 2005;Zhou et al, 2010), although the pattern of these results is somewhat variable. Some studies report a double dissociation between the FRN and P300, showing the FRN to be affected by valence but not reward magnitude, with the opposite pattern for the P300, regardless of whether the outcome is of positive or negative valence (Sato et al, 2005;Yeung and Sanfey, 2004).…”