Volume 106, Issue 1, January 2008, Pages 408-432
The cognitive locus of distraction by acoustic novelty in the cross-modal oddball task
a, b, , acFabrice B.R. Parmentier, Gregory Elford, Carles Escera, Pilar
acAndrés and Iria San Miguel
aSchool of Psychology, University of Plymouth, Drake Circus, Plymouth PL4 8AA, UK
bSchool of Psychology, University of Western Australia, Australia cCognitive Neuroscience Research Group, Department of Psychiatry and Clinical Psychobiology, University of Barcelona, Catalonia, Spain Received 24 July 2006; revised 11 March 2007; accepted 11 March 2007. Available online 18 April 2007.
Unexpected stimuli are often able to distract us away from a task at hand. The present study seeks to explore some of the mechanisms underpinning this phenomenon. Studies of involuntary attention capture using the oddball task have repeatedly shown that infrequent auditory changes in a series of otherwise repeating sounds trigger an automatic response to the novel or deviant stimulus. This attention capture has been shown to disrupt participants’ behavioral performance in a primary task, even when distractors and targets are asynchronous and presented in distinct sensory modalities. This distraction effect is generally considered as a by-product of the capture of attention by the novel or deviant stimulus, but the exact cognitive locus of this effect and the interplay between attention capture and target processing has remained relatively ignored. The present study reports three behavioral experiments using a
cross-modal oddball task to examine whether the distraction triggered by auditory novelty affects the processing of the target stimuli. Our results showed that variations in the demands placed on the visual analysis (Experiment 1) or categorical processing of the target (Experiment 2) did not impact on distraction. Instead, the cancellation of distraction by the presentation of an irrelevant visual stimulus presented immediately before the visual target (Experiment 3) suggested that distraction originated in the shifts of attention occurring between attention capture and the onset of the target processing. Possible accounts of these shifts are discussed.
Keywords: Novelty detection; Attention capture; Oddball task; Auditory distraction; Electrophysiology
2. Experiment 1
2.1.2. Stimuli, design and procedure
2.2.1. Response latencies
3. Experiment 2
3.1.2. Stimuli, design and procedure
3.2.1. Response latencies
4. Experiment 3
4.1.2. Stimuli, design and procedure
4.2.1. Response latencies
5. General discussion
6. Distraction and target processing
7. Mechanisms underpinning cognitive distraction
While efficient everyday functioning often requires the ability to selectively attend to some stimuli, it is equally important to maintain a certain degree of distractibility by task-irrelevant but otherwise potentially relevant events. Detecting an event violating the regularity of previous stimuli can for example warn of an imminent danger (e.g., an unexpected change in an aircraft engine’s noise may signal a malfunction).
In some circumstances, the propensity of transient environmental changes to capture our attention can have dramatic effects. For example, a report by the US National Transportation Safety Board examining 37 major accidents of US carriers from 1978 to 1990 revealed that nearly half these accidents involved lapses of attention associated with interruptions, distractions, or preoccupation with one task to the exclusion of another (Dornheim, 2000).
The distractive value of task-extraneous sound has been demonstrated in a variety of settings. For example, past research has shown that continuous irrelevant changing sounds can impair, for example, order memory [Jones et al., 1999], [Jones et al., 1995], [Jones and Macken, (e.g.,
1993], [Jones et al., 1992] and [Tremblay et al., 2001]) and various office
tasks ([Banbury and Berry, 1997] and [Banbury and Berry, 1998]). Single
auditory stimuli can also disrupt cognition by capturing attention away from ongoing cognitive processes, as shown in serial memory ([Hughes et
al., 2005] and [Lange, 2005]), arithmetic (Woodhead, 1964), visual
comparison (Woodhead, 1959), or motor pursuit tasks ([Jones and Broadbent,
1991] and [May and Rice, 1971]).
The present study examines the distraction occurring when stimuli in our environment deviate from the events expected by our cognitive system. As will become clear in the next pages, such distraction relates to core aspects of cognition, such as its tendency to build mental models of events to help deal with upcoming stimuli, and its propensity to orient attention towards events violating such models. The present study sought to examine the distraction yielded by such orientation responses. Specifically, it investigated whether attention-grabbing distractors impair ongoing cognitive performance due to competition for attentional resources or due to dynamic shifts of attention to and from distractors.
It is well established that infrequent auditory changes (so called oddball stimuli) in a train of repetitive stimuli capture attention in an obligatory fashion. From an electrophysiological standpoint, the brain response to auditory novelty is characterized by three specific responses, even when novel sounds are unrelated to the participants’ task. These responses, referred to as the ‘distraction potential’ (Escera & Corral,
2003) are: the mismatch negativity (MMN) and the enhancement of N1 generators when the distracter deviates a great deal from the repetitive background (Alho et al., 1998), P3a (sometimes referred to as novelty P3; see Friedman, Cycowicz, & Gaeta, 2001, for a review) and the
re-orientation negativity (RON; e.g., Schröger & Wolff, 1998b). The
mismatch negativity reflects the pre-attentive detection of a change (or even the illusion of a change, see Stekelenburg, Vroomen, & deGelder, 2004)
in the auditory context. The underlying mechanism eliciting this response is a comparison between a memory trace of the acoustic regularities of past sound events and the current auditory signal (see [Näätänen and
Winkler, 1999] and [Picton et al., 2000], for reviews). The output of this
detection system is an attentional interrupt hypothesized to involve frontal activations (Opitz, Rinne, Mecklinger, von Cramon, & Schröger, 2002) and resulting in the involuntary orientation of attention towards a novel sound, marked by the P3a response ([Friedman et al., 2001],
[Grillon et al., 1990] and [Woods, 1992]). When participants must perform
a primary task, a later deflection is also observed and interpreted as the re-orientation of attention towards that task ([Berti and Schröger,
2001], [Berti et al., 2004], [Escera et al., 2001] and [Schröger and Wolff,
While most studies on auditory attention capture focus on the sound-related electrophysiological responses, some researchers have used paradigms in which the impact of this capture can be measured behaviorally
from the participant’s performance in an attention-demanding task. For
example, Schröger (1996) found that MMN-eliciting deviants presented
among standards to the unattended ear delayed participants’ response to auditory targets in the other ear. In a slightly different task ([Schröger
and Wolff, 1998a] and [Schröger and Wolff, 1998b]), participants had to
discriminate between long and short sounds irrespective of their frequency (a repeated- or -frequency was presented in most trials standard
but replaced by a deviant on rare occasions). Even though frequency was irrelevant to the participants’ discrimination task and was to be ignored, response latencies in the primary task were significantly longer for frequency deviants relative to frequency standards. Distraction in this paradigm was shown to be independent of the stimulus-onset asynchrony (Roeber, Berti, & Schröger, 2003) but to decrease when participants
performed the discrimination task while concurrently holding a memory load (Berti & Schröger, 2003).
A large proportion of previous oddball studies have used the so called one-channel paradigm in which targets and deviants are presented auditorily, usually as distinct features of the same auditory object (e.g., frequency and duration). Remarkably, however, target and irrelevant stimuli do not need to be presented at the same time or in the same sensory modality for behavioral distraction to occur. This is important because it highlights the amodal nature of this phenomenon (that is, distraction does not merely occur due to the orientation of attention towards the same attended sensory modality as the target). Indeed, performance in a visual
discrimination task is also affected by auditory deviants or novels
(two-channel paradigm). This has been repeatedly demonstrated in studies using the cross-modal oddball task (e.g., [Andrés et al., 2006], [Escera
et al., 1998] and [Escera et al., 2002]). In this task, participants
categorize visual digits, presented in sequence, as odd or even. Each
digit is preceded by a task-irrelevant sound that participants are instructed to ignore. This sound is repeated on most trials (standard) but is replaced by a deviant or a novel on rare occasions (e.g., 20% of trials). In addition to the distraction potential elicited by novel or deviant sounds, response latencies in the categorization task are consistently longer for digits preceded by deviants or novels compared to standards.
While the oddball paradigm is one of the most widely used methods to study the neurophysiology of attention, there is currently no cognitive analysis of the distraction induced by novel stimuli. Of particular relevance in the present study, the cognitive locus of the effect of distraction observed in the cross-modal oddball task remains unexplored. Past studies have usually implicitly considered that the
electrophysiological response to a novel is responsible for the behavioral distraction in a subsequent and unrelated primary task. It is however not obvious how exactly the capture of attention by a novel sound should interplay with the processing of an unrelated and attended visual event. Why do participants take longer to report that, for example, “8”
is an even number when it is preceded by a novel sound compared to a standard? Up to date, it was unclear why the processing of a visual digit should be affected by the presentation of a sound that is (1) irrelevant to the participant’s task, (2) presented in a different modality, (3)
presented at a different time; and (4) that does not, arguably, afford any response or prime any mental representation that may possibly relate to the digit processing. This is the issue addressed in this study. Several positions can be entertained with respect to the potential ways in which behavioral performance in a primary visual task may be affected by auditory novels: through a competition for attentional resources
([Johnston and Heinz, 1978] and [Kahneman, 1973]) between the processing
of the target and the pre-attentive processing of the novel sound, and/or through an attention bottleneck ([Broadbent, 1958], [Pashler et al., 2001]
and [Welford, 1967]) delaying the onset of the processing of the target until attention switches from the sound analysis back to the visual task. Under the first hypothesis, the response to the visual target may be delayed because its processing is slowed by a relative depletion of attention resources. A previous study showing that visual ERPs (extraestriate N1) were attenuated following a deviant tone compared to a standard one (Alho, Escera, Dı?az, Yago, & Serra, 1997) would support
this hypothesis. Under the second hypothesis, however, responding to the visual target following a novel sound may be delayed because time is required for attention to shift to the novel sound, engage in its analysis, and return towards the primary task. The processing of the visual target itself would not be affected; its onset would. The two hypotheses predict different cognitive loci for the distraction observed in the cross-modal oddball task in which participants categorize visual digits (e.g., [Escera et al., 2000] and [Escera and Corral, 2003]). The resource
competition hypothesis places the locus of distraction at the digit processing stage of the task. The bottleneck hypothesis predicts that the processing of the digit is not the locus of distraction but that the shifts of attention to the novel and then back to the digit are.
In Experiments 1 and 2, we tested the hypothesis that distraction in the cross-modal oddball task may reflect the depletion of attention resources available to process the visual target when attention is captured by an irrelevant novel sound. Reporting that a digit is odd or even involves a number of processing steps, any of which potentially sensitive to distraction. In our effort to identify the possible cognitive locus of distraction among these processing steps, we started by dividing the
processing of the visual digit into two broad categories of processes: (1) the perceptual analysis of the visual input, from the registration of a pattern of light on the retina to the identification of the stimulus as a digit; and (2) the categorization and response selection. In Experiment 1, we manipulated the demands placed on the visual analysis stage and examined whether perceptually degraded visual stimuli would amplify the distraction induced by novel sounds. In Experiment 2, we applied the same logic to the manipulation of the categorization difficulty. In both cases, the rationale was the following: If the distraction entailed by the involuntary capture of attention by auditory novels results from the impaired processing of the target stimulus, then making this processing more demanding should amplify distraction. 2. Experiment 1
Twenty (14 females) undergraduates from the University of Plymouth took part in this experiment in exchange for a small honorarium. Participants were between 18 and 37 years of age (M = 22.8, SD = 4.47). All participants reported correct or corrected-to-normal vision and normal hearing.
2.1.2. Stimuli, design and procedure
Participants were presented with 4 blocks of 424 trials each. In each trial, participants had to categorize a digit (1–8) as odd or even. A fixation
box was visible at the centre of the screen throughout the task. The digits were presented in random order (different for every participant), but with equal probabilities across the task, at the centre of the presentation
box, sustaining a viewing angle of approximately 2.6? (participants were seated at approximately 50 cm of the screen). Each digit was presented for 200 ms.
In all trials, a 200 ms sound was presented 300 ms before the onset of the visual stimulus. Participants were told that the sound was a distracter and that they should ignore it. From the onset of a digit, participants had 1000 ms to respond. Following a further 100 ms, the next trial was automatically initiated. Participants used the keys Z and X on the computer keyboard to respond using two fingers of their dominant hand. The mapping between the response keys and the odd/even responses was counterbalanced across participants.
Two sound conditions were compared within each block. In the standard
(90% of trials), the sound was a 600 Hz sinewave tone of 200 ms condition
duration (10 ms of rise/fall times), hereafter referred to as the standard.
In the novel condition (10% of trials), we used 60 different environmental sounds adapted from Escera, Yago, Corral, Corbera, and Nuñez (2003),
hereafter referred to as novels. All novels had a 200 ms duration
(including 10 ms rise/fall times), were digitally recorded, and low-pass filtered at 10,000 Hz. Each novel sound was only used a maximum of three times across the experiment. Novel sounds were always preceded by at least one standard trial. All sounds were normalized and presented binaurally through headphones at approximately 75 dB. In each block, the first 24 trials, which only contained standards, were treated as warm-up trials and were not included in the data analysis. Across the 1600 test trials, 16 novel sounds were used within every successive group of 160 trials (proportion of novels = .1) to ensure a relatively even, yet unpredictable, distribution of novels across trials. The order of