The cognitive determinants of behavioral distraction by deviant auditory stimuli: a review.

Numerous studies have demonstrated that rare and unexpected changes in an otherwise repetitive or structured sound sequence ineluctably break through selective attention and impact negatively on performance in an unrelated task. While the electrophysiological responses to unexpected sounds have been extensively studied, behavioral distraction has received relatively less attention until recently. In this paper, I review work examining the cognitive underpinnings of behavioral distraction by deviant sounds and highlight some of its key determinants. Evidence indicates that deviance distraction (1) derives from the time penalty associated with the involuntary orientation of attention to and away from the deviant sound and from resulting effects such as the reactivation of the relevant task set upon the presentation of the target stimulus; and (2) is mediated by a number of factors (some increasing distraction, such as aging or induced emotions; some decreasing it, such as a memory load or cognitive control). Contrary to the received view that deviants ineluctably elicit distraction, recent work demonstrates that it is contingent upon auditory distractors acting as unspecific warning signals in the service of goal-oriented behavior, and that deviants do not elicit distraction because they are rare but because they violate the cognitive system's predictions (which can be manipulated through implicit rule learning or explicit cueing). Evidence is also presented indicating that the capture of attention by spoken deviant sounds is followed by an involuntary evaluation of their semantic properties, the outcome of which can be robust enough to linger in working memory and interfere with subsequent behavior. Finally, I review studies suggesting that behavioral deviance distraction is not the mere byproduct of the mismatch negativity, P3a and re-orientation negativity electrophysiological responses and highlight a number of outstanding questions for future research.