Background
The occurrence of an unexpected sound when performing a visual task (e.g., determining if two digits are the same or not) can be distracting. Although distraction may affect task performance, processing distractors is important for being aware of potentially important events outside the task domain.
The extent to which a distractor is processed and attended depends upon how it fits with an on-going model of the environment, i.e., when a distractor is novel it is more distracting [
1]. Also, distraction can be modulated by the degree to which a task is attended [
2,
3]. In this regard, it has been proposed that working memory (WM) acts to enhance and maintain the sensory processing required by a task, while at the same time reducing processing triggered by the presentation of a distracting stimulus [
3‐
6].
Evidence from younger adults that engaging WM reduces distractor processing and distraction comes from several studies [
3,
7]. For example, the SanMiguel et al. [
3] study used a cross-modal distraction paradigm in which participants performed either a working memory (W1) or no working memory (W0) visual matching task while auditory distractors were presented. It was found that the presentation of novel distractor sounds slowed responses in the W0 condition but not in the W1 condition. That is, there was less distraction in the WM task.
In addition to using behavioural measures of distraction (i.e., hit rate and response time), event-related potentials (ERPs) can also be used to determine the impact of working memory on the neural indices of distraction. ERPs provide a method to assess the brain function in response to sounds that has two major advantages compared to other neuroimaging procedures: excellent temporal resolution (in the order of milliseconds) and cost-effectiveness. The high temporal resolution ability of the ERPs have been used to investigate, low-level cognitive functions such as encoding of sounds [
8,
9], high-level functions such as selective attention, working memory, and language [
3,
10,
11] and functions intermediate between low and high cognitive functions such as sound discrimination, and involuntary attention [
12‐
14]. The ERPs and behavioural measures index different information regarding cognitive processing. While, ERPs provide processing information before, during and after a cognitive response, behavioural measures are usually related to the processing after the response.
Information processing related to distraction has been described by a model that posits three sequential processing stages with each stage reflected by a specific event-related potential (ERP), see Horváth, Winkler and Bendixen [
15]. According to this model, the initial stage of processing acts as a filter which adaptively adjusts the processing of sensory information to reduce the load on capacity-limited resources. This change detection stage is often associated with the mismatch negativity (MMN) ERP, a negative wave resulting from the subtraction of the ERP to a standard stimulus from that of a deviant one that peaks between 100 and 200 ms from the onset of the detected deviance. The MMN reflects a largely memory-based process that detects deviations from perceived regularities in the auditory input [
16]. There is, however a more basic process that is sensitive to stimulus onset and simple change detection, which is reflected by the N1 ERP, a negative potential that peaks fronto-centrally between 100 and 150 ms following a stimulus change [
17]. Given the similarity of the functions represented by the two event related components (to filter and highlight sensory events), it is often reported that this first stage of processing is represented jointly by N1/MNN, although Horváth et al. [
15] take care to separate these, which we do in our analysis.
The second stage in the model of distraction deals with the process of involuntary attention-switching mechanism towards the distractor once it has been detected as a change from some regular sensory context. Here, cognitive resources are allocated between voluntary attention to task-relevant events and involuntary attention to distractor events. This stage is marked by the P3a, that peaks between 200 and 300 ms after the distracting event [
18]. It has been suggested that the P3a represents an involuntary change in selective attention set that is invoked by the distracting event [
19]. That is, it is a cognitive orienting response that typically is generated to rare stimuli and may also be associated with arousal [
20].
The function of the third processing stage of the distractor model is to switch or refocus attentional resources back to task-relevant events [
21]. That is, provided the event that triggered a switch in attention does not require an on-going adaptive reorientation, the original task-related attentional set is restored. It has been proposed that this function is reflected by a modality-independent, fronto-central reorienting negativity (RON) component [
13,
22].
In their study, SanMiguel et al. [
3] found that for the working W1 task (as compared to the W0 task) there was no change in MMN amplitude; there was a reduction in the novelty-P3 amplitude (the P3a), and a larger RON in younger adults. The current study aimed to investigate the effects of working memory load on both the behavioural and neural indices of distraction processing in older adults.
There is a large and inconsistent research literature that has used both behavioural and ERP measures to examine age-related changes in auditory distraction. For example, some behavioural studies report that older adults’ control of auditory distraction is impaired [
23], while other studies find equivalent distraction levels between younger and older adults [
24,
25]. The ERP literature is similarly mixed. For example, in terms of age-related differences on MMN amplitude some studies report age-related reductions in MMN amplitude in response to deviant tones, suggesting older adults have deficits in encoding and retaining sensory information, reducing their ability to detect distractors within encoded sequences [
26,
27]. However, others find no age-related differences on MMN amplitudes or latencies, suggesting that older adults do not have deficits in automatic processing (involuntary, early attention) [
28,
29]. Research that has examined the later ERP components, P3a and RON has tended to paint a more consistent picture concerning age differences (although see Berti et al. [
30]). For example, it has generally been found that older adults have a smaller [
26,
31], and a later [
29] P3a compared to younger adults. This has been interpreted as indicating that older adults do not evaluate auditory distractors as efficiently as younger adults [
32]. Likewise, studies have indicated that the amplitude of the RON is also smaller (and occurs later) in older compared to younger adults, a finding suggesting that older adults are less efficient in re-focusing their attention to the task following distraction [
30,
32].
The different behavioural and ERP results outlined above highlight the difficulty in drawing conclusions from studies that have employed diverse experimental paradigms. In the current study we chose to extend the classic procedure used by SanMiguel et al. [
3] by examining the extent to which older adults (62–74 years) are protected from distraction when engaged in a working memory task (for comparison, younger adults aged 22–35 years were also tested). We selected the SanMiguel et al. procedure since it produced clear behavioural and ERP effects; used a simple cross-modal distraction paradigm and employed a categorical contrast between a working memory task and a similar task that did not involve working memory. In our view, when aiming to examine perceptual and cognitive effects on a different participant group it is an important first step to build on earlier studies by using their procedures.
SanMiguel et al. [
3] used an auditory oddball paradigm to determine the impact of working memory on the neural indices of distraction. The current study used the same paradigm to elicit neural indices of distraction processing in older and younger adults. In an oddball paradigm distractor sounds are the infrequent stimuli interspersed randomly but with a pre-determined probability in a series of repetitive frequent standard sounds. Before outlining the details of the current study, it is important to emphasize that although the results of SanMigel et al. [
3] clearly supported the claim that engaging working memory reduces the effects of a distractor, other studies found the opposite pattern, i.e., the involvement of working memory increased distraction [
33]. There are several reasons why engaging working memory may not always reduce the effect of a distractor and in the following section we outline some of these.
One factor that may influence the effect of working memory on distraction has to do with the relationship between the distractor and the task stimuli. For tasks where target and distractor stimuli compete for sensory processing and response selection (e.g. a unimodal task and distractors), increasing the working memory load tends to increase distraction [
34]. Whereas, for tasks where the distractors do not trigger response competition, engaging working memory typically reduces distraction [
3,
5].
Another factor that can modulate how working memory affects distraction concerns whether working memory is engaged as part of the task, or whether it is engaged in an unrelated task. That is, paradigms that load working memory with materials unrelated to the task more often report that working memory load increases distraction [
34].
Finally, there is the factor of how much working memory load is applied. That is, it appears that whether distraction can be suppressed in the early stages of processing or not depends on the resources available for executive cognitive control [
33]. The behavioural and ERP studies of SanMigel et al. [
3] and Lv et al. [
5] provide an example. In SanMigel et al., the working memory condition consisted of a simple visual 1-back task and performance in this condition was compared to that on a 0-back task, i.e., the contrast was with a non-working memory task. Two types of auditory distractors were used, a 600 Hz tone that occurred 80% of the time before the visual task (the standard); and novel environmental sounds that occurred 20% of the time. Overall, the presentation of novel compared to standard distractors reduced the hit rate and increased the response times to the visual task. This effect was driven by performance in the no working memory condition (0-back task), since in the working memory task there was no effect of distractor type (i.e., engaging working memory protected against distraction from the novel sounds). As mentioned above, SanMigel et al. [
3] also found that the novelty-P3 was attenuated in the working memory condition.
For the most part, this effect of working memory protecting from distraction was not found by Lv et al. [
5] who used the same cross-modal distraction paradigm but contrasted an easy and a difficult working memory task. That is, unlike SanMigel et al., Lv et al. found that engaging working memory (in their case, the high versus low memory condition) did not protect against being slower and less accurate on the visual task when presented with novel compared with standard distractors. Also, unlike SanMigel et al. who found no effect, Lv and colleagues found that an early ERP component (MMN) had a greater amplitude in the high load condition, suggesting that high memory load increased the salience of the novel distractor. Finally, and similar to SanMigel et al., Lv and colleagues did find later processing of the novel distractor (as indexed by P3a) was attenuated in the high WM versus the low WM condition. Similar findings also have been reported in auditory-only tasks. Evidence suggests that the amplitude of P3a and RON ERPs decreased in younger adults with increasing load in tasks, when WM load was modulated within the auditory domain only [
4,
35,
36]. On the other hand, the stage of distractor processing indexed by MMN, appears to be minimally modulated by task demands [
36] in younger adults.
As the current experiment aimed to determining whether engaging working memory will reduce distraction in older adults, we chose to use the cross-modal distraction paradigm employed by SanMiguel et al. that contrasted a working memory and a no working memory condition, since this design has produced the clearest effect of working memory on distraction in both the behavioral and ERP data. It is, however, an open question as to whether older adults will exhibit the same behavioral and neural effects of engaging working memory that younger adults showed in SanMiguel et al.
That is, two properties associated with cognitive aging could limit the extent to which working memory may protect against distraction. The first, is that older adults tend to be more distractible than younger adults, especially with cross-modal distraction (i.e., a visual task with an auditory distractor), see Leiva et al. [
37], and more generally the literature on the inhibitory-deficit hypothesis [
27,
38‐
41]. Older adults are more distractible than their younger counterparts in response to unexpected auditory events when engaged in a task that requires voluntary attention [
42]. Research using cross-modal (auditory-visual) oddball paradigms where participants engaged in a visual task while ignoring auditory distractors, found longer reaction times and low task accuracy in older adults compared to younger adults indicating increased distractibility [
23,
43]. Indeed, effects of increased distractibility in older adults have also been found in neural markers of distractor processing. Alain and Woods [
44] found that while there were no age-related changes in visual discrimination task performance following auditory distractors, older adults had a larger N1 and a smaller MMN to auditory distractors. These results were interpreted as age-related decline in the ability to suppress irrelevant auditory stimuli supporting inhibitory-deficits hypothesis. Previous research using unimodal auditory oddball paradigms have also reported age-related distraction effects either on behavioural or neural indices. Accuracy of an auditory duration discrimination task was reduced in middle-aged adults in the presence of irrelevant sounds, but no effect of age was found for reaction times [
29]. However, delayed P3a and RON latencies in middle aged adults indicated distractibility [
29]. Age-related greater behavioural distraction effects in older adults and low-performing older adults compared to younger adults were found in similar auditory duration discrimination tasks [
30,
38]. In a similar task, Horvath et al. [
34] reported delayed P3a and RON ERPs in older adults but no age-related behavioural distraction effects. There is, however, a dearth of research examining ageing effects on auditory distraction using auditory-visual paradigms.
The second property associated with cognitive ageing is that older adults may be less able to engage sufficient working memory resources overcome distraction due to age-related deterioration in the functioning of dorsolateral pre-frontal cortex [
45‐
48].