Behavioral distraction by auditory novelty is not only about novelty: The role of the distracter’s informational value

Abstract

Unexpected events often distract us. In the laboratory, novel auditory stimuli have been shown to capture attention away from a focal visual task and yield specific electrophysiological responses as well as a behavioral cost to performance. Distraction is thought to follow ineluctably from the sound’s low probability of occurrence or, put more simply, its unexpected occurrence. Our study challenges this view with respect to behavioral distraction and argues that past research failed to identify the informational value of sound as a mediator of novelty distraction. We report an experiment showing that (1) behavioral novelty distraction is only observed when the sound announces the occurrence and timing of an upcoming visual target (as is the case in all past research); (2) that no such distraction is observed for deviant sounds conveying no such information; and that (3) deviant sounds can actually facilitate performance when these, but not the standards, convey information. We conclude that behavioral novelty distraction, as observed in oddball tasks, is observed in the presence of novel sounds but only when the cognitive system can take advantage of the auditory distracters to optimize performance.

Introduction

Unexpected novel events can capture our attention in an involuntary fashion, alerting us of a potential danger but, at the same time, potentially distracting us away from a task at hand. The registration of novel events is arguably a pivotal function of our central nervous system, one of strong clinical relevance (e.g., Fischer et al., 1999) offering a fertile avenue for fundamental research about brain functioning and cognition (e.g., Pullvermüller and Shtyrov, 2006, Schröger, 1996).

Novelty distraction has been observed in electrophysiological studies using, most often, auditory stimuli (e.g., Schröger, 2005) but also visual (e.g., Barceló et al., 2006, Polich and Comerchero, 2003) and tactile stimuli (Knight, 1996, Yamaguchi and Knight, 1991a, Yamaguchi and Knight, 1991b). Novel events trigger a pattern of responses marked by specific brain waves: an automatic novelty-detection response (mismatch negativity or MMN; e.g., Näätänen and Winkler, 1999, Picton et al., 2000), followed by an involuntary orientation response (P3a; e.g., Friedman et al., 2001, Grillon et al., 1990, Woods, 1992) and, when participants are engaged in some focal task, a re-orientation negativity (RON; e.g., Berti et al., 2004, Berti and Schröger, 2001, Schröger and Wolff, 1998). Of importance, novelty also results in measurable behavioral effects such as longer RTs and reduced response accuracy (e.g., Dawson et al., 1989, Grillon et al., 1990, Schröger, 1996, Schröger and Wolff, 1998, Woodward et al., 1991; see also Hughes et al., 2005, Hughes et al., 2007, for evidence of behavioral distraction due to auditory deviants in a visual serial recall task). It is this behavioral aspect of distraction that constitutes the focus of the present study.

Novelty distraction is also observed when target and distracter are temporally decoupled and presented in different sensory modalities, as observed in the cross-modal oddball task (e.g., Andrés et al., 2006, Escera et al., 1998), testifying of the fundamental nature of this phenomenon. In that task, participants categorize visual stimuli presented in sequence (e.g., digits) based on some criteria (e.g., parity). Each visual target is preceded by a task-irrelevant sound (standard on most trials, deviant or novel on rare trials) that participants are instructed to ignore. Just as in the auditory oddball task, auditory novels elicit a pattern of three brain responses (MMN, P3a, RON) and reduce task performance.

While most past research focused on electrophysiological responses to deviant or novel sounds, recent behavioral work suggests that (1) behavioral distraction reflects the time required for attention to switch to and from the novel sound rather than the slower processing of the target (Parmentier, Elford, Escera, Andrés, & San Miguel, 2008); and that (2) attention capture is followed by an involuntary semantic analysis of the novel’s content which can in turn impact on subsequent behavior (Parmentier, 2008, Parmentier et al., in press).

Past studies posit, somewhat peremptorily, that novelty distraction results from the novel’s low probability of occurrence and its corollary unpredictable nature. It is assumed to be about novelty per se. The objective of our study is to challenge this view and argue, instead, that (1) novelty is a necessary but not sufficient condition for novelty distraction; and that (2) an additional, yet so far unreported, important factor is the extent to which sound conveys information used by the cognitive system to optimize performance.

We suggest that past cross-modal oddball studies may not provide solid evidence that novel sounds ebb participants’ behavioral performance on the basis of their novelty per se, for the role of the sound as a warning signal has not been subject to punctilious attention. This role is evidenced by the fact that responses to a standard sound are significantly faster than in a silent condition (e.g., Andrés et al., 2006, Parmentier and Andrés, 2009), for a simple reason: in all past studies, sounds always harbingered the presentation of a target and its time of occurrence, for the interval between sound and target in this paradigm was fixed (see Table 1). Distraction, if purely reflecting the rarity of the novel’s occurrence, should be observed regardless of the sounds’ information value as a warning. To date, however, no study attempted to examine this issue.

That a stimulus preceding a target by a fixed interval speeds up response time is one of the oldest phenomena studied in psychology (e.g., Wundt, 1880; see Hackley, 2009, Niemi and Näätänen, 1981, for reviews), including across modalities (Bertelson, 1967, Davis and Green, 1969). Numerous studies have shown this effect in pure blocks where a warning always announces the upcoming presentation of a target and precedes it by a fixed amount of time, but not in mixed blocks where targets do not always follow warnings or do so after a temporal interval varying from trial to trial (e.g., Näätänen, 1970, Woodrow, 1914). This type of warning informs participants of the certainty of the target’s presentation and allows the orientation their attention to a particular point in time (Hackley, 2009).

We suggest that the notion that an event’s rarity is the sole determinant of behavioral distraction in the cross-modal task may be precarious, and that the potential role of the information conveyed by auditory distracters has not received sufficient attention. To address this issue, we asked participants to categorize visual digits preceded by standard and deviant sounds in three conditions. In the Informative Condition, the sound announced the upcoming of a target after a fixed temporal interval. In the Uninformative Condition, targets followed sounds with a probability of .5 and the interval separating them varied, thereby stripping the sound of its warning value. In the Informative Deviant Condition, the standard sound did not announce whether or when a target would be presented but the novel sound did. Two sets of predictions were opposed. According to the probabilistic view of distraction, deviants should distract participants as long as they are rare, therefore predicting similar levels of distraction in all three conditions. The informational view posits that the impact of novelty may be mediated by the informational content of sound, predicting distraction in the Informative Condition, no distraction in the Uninformative Condition, and facilitation in the Informative Deviant Condition.