by Kristen Silveira
BA, Queen’s University, 2014
A Thesis Submitted in Partial Fulfillment of the Requirements for the Degree of
MASTER OF SCIENCE in the Department of Psychology
© Kristen Silveira, 2016 University of Victoria
All rights reserved. This thesis may not be reproduced in whole or in part, by photocopy or other means, without the permission of the author.
Supervisory Committee
The Impact of Musical Affect and Arousal on Older Adults’ Attention by
Kristen Silveira
BA, Queen’s University, 2014
Supervisory Committee
Dr. Colette Smart (Department of Psychology) Supervisor
Dr. Catherine Mateer (Department of Psychology) Departmental Member
Abstract
Supervisory Committee
Dr. Colette Smart (Department of Psychology) Supervisor
Dr. Catherine Mateer (Department of Psychology) Departmental Member
Selective attention is a specific area of executive control that declines in older adulthood and may be amenable to cognitive rehabilitation. This study explored background music as an accessible and typically enjoyable tool that may exogenously facilitate attention. Two particular properties of a musical piece – (1) mode (i.e., major, minor, or atonal), and (2) tempo (i.e., stimulative or sedative) – influence affect, arousal, and cognitive function, ultimately enhancing or hindering cognitive performance on attention-demanding tasks. Six musical pieces were selected to represent different combinations of mode and tempo. Older adults (i.e., 65-80 years-old; n=16) were recruited from Victoria, BC. Participants completed the Multi-Source Interference Task (MSIT) assessing selective attention at baseline and under the six counterbalanced musical conditions. In each condition, participants reported motivation and task-difficulty, as well as affect and arousal on the Activation-Deactivation Checklist (AD ACL). Musical affect impacted reaction times on MSIT control and interference trials for the first block, but had no influence during the last block. Musical arousal did not
significantly impact attention. AD ACL responses, as well as task-difficulty and motivation to succeed on the task did not vary as a function of the music. The results illuminate older adults’ allocation of executive resources between competing goals of regulating musical affect and succeeding on an attention task. Implications are discussed for selecting music specifically to facilitate older adults’ attention in everyday life.
Table of Contents
Supervisory Committee ... ii!
Abstract ... iii!
Table of Contents ... iv!
List of Tables ... vi!
List of Figures ... vii!
Acknowledgments ... viii!
Dedication ... ix!
Introduction ... 1!
Selective Attention ... 3!
Perceptual Mechanism of Selective Attention ... 5!
Executive Mechanism of Selective Attention ... 7!
Selective Attention Deficits in Older Adults ... 9!
Emotion Regulation in Older Adults ... 12!
Arousal and Attention ... 20!
Musical Affect and Arousal ... 21!
Study Aims and Hypotheses ... 24!
Method ... 26! Participants ... 26! Measures ... 26! Procedure ... 35! Results ... 38! Descriptive Statistics ... 38! Primary Analyses ... 46! Secondary Analyses ... 50! Discussion ... 53!
Musical Affect and Selective Attention ... 54!
Musical Arousal and Attention ... 60!
The Role of Motivation and Task Difficulty ... 62!
Implications ... 65!
Conclusion and Future Directions ... 67!
References ... 70!
Appendix A ... 81!
Appendix B Routine Arousal Questionnaire ... 82!
Appendix C ... 84!
Appendix D ... 85!
Appendix F Geriatric Anxiety Scale ... 89! Appendix G Post-Task Questionnaire ... 91! Appendix H Multi-Source Interference Task Trials ... 93!
List of Tables
Table 1. Characteristics of musical conditions ... 34!
Table 2. Demographic, musical, and arousal-related characteristics of the sample ... 39!
Table 3. Percentile rankings a by subtest on the Test of Everyday Attention (TEA) ... 40!
Table 4. Enjoyment of background silence and music. ... 41!
Table 5. Familiarity and reminiscence triggered by silence and music ... 41!
Table 6. Mean positive and negative affect scores on the PANAS at baseline ... 42!
Table 7. Mean positive and negative affect scores on the AD ACL during baseline silence and musical conditions ... 42!
Table 8. Mean Energetic Arousal and Tense Arousal scores on the AD ACL during baseline silence and musical conditions ... 43!
Table 9. Frequency of emotions across silence and musical conditions ... 44!
Table 10. MSIT mean accuracy scores as percentages across silence and musical conditions ... 45!
Table 11. Mean first-block and last-block interference RT across silence and musical conditions ... 47!
Table 12. Mean first-Block control and interference RTs across silence and musical conditions ... 49!
Table 13. Summary of linear regressions with silence last-block interference RT as predictor ... 51!
Table 14. Mean self-reported task difficulty and motivation scores across silence and musical conditions ... 52!
List of Figures
Figure 1. Experimental protocol for participant randomly assigned to complete high-arousal musical conditions in first session and low-high-arousal conditions in second session.. ... 37!
Acknowledgments
Thank you to my supervisor, Dr. Colette Smart, for all of her support throughout my Master’s thesis, including equal doses of expert guidance and inspirational freedom – to ask research questions that are truly important to me and to design my own study to explore their answers. I would also like to thank my committee member, Dr. Catherine Mateer. It is an honour to have her expertise and insights from the field of cognitive rehabilitation, and I am grateful for her kindness and time in serving on my committee.
Thank you to my fellow “Smarties” in the SmartLab, as well as my friends in the program and cohort, for their moral and academic support, and for the laughs. A special thanks to Jacob Koudys for his careful and thoughtful research assistance in preparing the database for analyses. Thank you also to the friendly faces at the Centre on Aging for their endless help with recruiting participants for the study, and thanks to all older adults in the community who responded to the call to participate. I would also like to
acknowledge my undergraduate research supervisor at Queen’s, Dr. Lola Cuddy, for introducing me to the world of music cognition and continuing to support my academic pursuits.
I find such strength, support, and inspiration in my friendships across the country and here in Victoria – thank you. Finally, a humungous thank you to my family for valuing and supporting my academic endeavours, and cheering me on until and beyond this point.
Dedication
To my Gammy, whose beauty, strength, and delight in life always captivate my attention.
Introduction
Popular theories of cognitive aging focus on normative declines in executive functions in healthy older adults (e.g., Hascher & Zacks, 1988; Rush, Barch, & Braver, 2006). Relatedly, musical cognitive interventions for older adults have largely aimed to improve working memory and other areas of executive functions (e.g., Thompson, Moulin, Hayre, & Jones, 2005). Executive functions refer to a set of interrelated complex cognitive abilities that enable people to modify thoughts, emotions, and behaviours (Schmeichel, 2007). These higher-order abilities are essential for organizing thoughts and behaviour in a goal-directed nature amidst a changing environment (Jurado and Rosselli, 2007). The importance of executive functions for appropriate social and self-serving behaviour is emphasized by Lezak, Howieson, and Loring’s (2004) claim that intact executive functions can enable a person to live independently and productively, even with considerable cognitive loss in other domains. Declines in executive functions are predictive of disruption in older adults’ instrumental daily activities above and beyond other cognitive declines, and improving executive functioning would have positive
implications for older adults’ activities in the home and community, including completing housework, taking medications, managing money, shopping, and using the telephone, other technology, and transportation (Cahn-Weiner, Malloy, Boyle, Marran, & Salloway, 2000; Royall, Palmer, Chiodo, & Polk, 2004; Shallice & Burgess, 1991).
Deterioration of both posterior (i.e., perceptual) and anterior (i.e., executive) cortices is commonly associated with surpassing 65 years of age, but the greatest
proportion of cell loss occurs in the frontal cortex (Kramer, Humphrey, Larish, Logan, & Strayer, 1994). In terms of cellular activity, neuroimaging studies have shown increased
activity in the prefrontal cortex and decreased activity in visual processing regions (Grady, 2012). This activity pattern in older adults relative to younger adults is called the “posterior to anterior shift with aging”. Increased anterior activity can indicate
compensatory mechanisms to account for cell or strategy loss, or it may represent less efficient use of executive resources, and it is not related to enhanced performance on executive tasks.
Research on executive function declines in healthy older adults is paralleled by the recent research focus on “attentional control” declines in seemingly healthy older adults who are actually in the preclinical stages of Alzheimer’s disease (e.g., Belleville, Chertkow, & Gauthier, 2007). Attentional control is a comprehensive construct that bridges to executive functions such as inhibition, updating, and switching. Tracking and rehabilitating attentional control deficits is pertinent to older adults at preclinical stages of neurodegenerative conditions such as Alzheimer’s disease; likewise, research on
attentional control deficits may also be warranted for healthy older adults. Older adults’ normative deficits in working memory and executive control intersect, overlap, and correlate with attentional control, or more specifically with selective attention (i.e., an area of attention involving inhibition; McDowd, 2007). Thus, selective attention may be a particularly relevant construct for healthy older adults as well as preclinical individuals. Notably, there is ongoing debate regarding the separability and unity of executive functions that overlap with working memory and attention – specifically, inhibition, updating, and switching (e.g., Miyake, Friedman, Emerson, Witzki, & Howerter, 2000). Due to the relevance of selective attention for rehabilitation, it is treated in this study as a separable construct that encompasses the executive function of inhibition.
Selective attention may be most easily described as freedom from distractibility (Sohlberg & Mateer, 2001). It encompasses effortful processes, including facilitative processing of relevant information as well as inhibitory processing of distracting information. A real-world example could be an older adult on public transportation reading the upcoming bus stops and street signs while ignoring the conversations of other passengers on the bus. If it were shown to effectively enhance selective attention,
listening to background music would be an ideal cognitive intervention because it is non-invasive, accessible, affordable, and typically enjoyable for community-dwelling older adults. Music is a complex and varied stimulus, and this thesis study is a novel
exploration of different types of music that may enhance or hinder older adults’ selective attention.
Selective Attention
Executive functions are highly interrelated with attention in their functional significance as well as their neurocircuitry, as highlighted by Sohlberg and Mateer (2001) in their outline of predominant attention models. For example, the experimental model by Posner and Petersen (1990) and Petersen and Posner (2012) proposes three networks of attention. The first network – “orienting” – involves posterior brain areas and is a primitive form of attention that is not usually relevant to cognitive intervention. The second network – “alerting” – involves right prefrontal regions and is used to maintain vigilance in the absence of salient novel external stimuli. The third network – “executive control” – is located in anterior brain areas and involves executive functions such as shifting and alternating attention. In Petersen and Posner’s model, the alerting and executive control networks of attention involve processes and brain regions associated
with executive functions. Sohlberg and Mateer (2001) also outline a clinical model of five components of attention that are particularly relevant to cognitive rehabilitation – focused, sustained, selective, alternating, and divided attention. Focused and sustained attention would likely correspond with Petersen and Posner’s (2012) alerting network, whereas selective, alternating, and divided attention seem to overlap with the executive control network. Based on the experimental and clinical attention models by Petersen and Posner and by Sohlberg and Mateer (2001), selective attention appears to be a specific element of attention that is related to executive functions and may be enhanced by cognitive rehabilitation.
In order to effectively rehabilitate selective attention, we must understand its underlying mechanisms. Broadbent’s (1958) seminal ‘early selection’ model of attention was based on the premise that perceptual capacity is limited and protected by an
attentional filter. This filter was thought to exclude irrelevant information as soon as its simple perceptual features had been processed. The early selection model accounted for earlier observations of participants being unable to report information presented to their unattended ear during dichotic listening tasks (e.g., Cherry, 1953) However, it could not explain why the participants could sometimes respond to the same information when it was preceded by a special mention of the participant’s own name (Moray, 1959). Participants’ recognition of their name demonstrated that they had processed the unattended information on a semantic level, beyond the earlier level of perceptual
processing. A rival “late selection” model by Deutsch and Deutsch (1963) was born from these findings; it proposed that all information is fully perceived and irrelevant
information is filtered from post-perceptual processes such as memory and behavioural responses.
Perceptual Mechanism of Selective Attention
After decades of research and discussion surrounding early versus late selection theories, Lavie, Hirst, Fockert, and Viding (2004) presented a promising theory to resolve the debate. Lavie and colleagues’ (2004) load theory proposed two mechanisms of
selective attention: the perceptual mechanism and the executive mechanism. Information is first processed through the perceptual mechanism. This initial stage is automatic and passive, depending solely on perceptual capacity and perceptual load. Perceptual capacity may vary across age groups, presumably as posterior cortices and sensory abilities such as vision and hearing develop and subsequently deteriorate across the lifespan. Perceptual load varies depending on the number of stimuli items or complexity of perceptual
processing demands. For example, load may be increased by including more task-relevant stimuli (e.g., adding more dots to a target group of dots). Load may also be increased by increasing the perceptual processing requirements for the stimuli, for example by changing the simple task of detecting whether or not the stimuli is present to the more complex task of discriminating the stimuli’s shape and texture. When perceptual load is increased and perceptual capacity is reached, irrelevant information and
distractors are filtered and excluded from processing at the perceptual level (i.e., early selection occurs). When perceptual load is low, the remaining perceptual capacity is automatically allocated to processing irrelevant information and distractors, and this information is not filtered until it reaches the post-perceptual executive mechanism (i.e., late selection occurs).
Empirical evidence suggests that age-related impairment on a selective attention task may be improved by manipulating the perceptual mechanism (e.g., Lavie et al., 2004). In one experiment, older adults completed the Eriksen flanker task under various levels of perceptual load. Under the lowest level of perceptual load, one target and one distractor were presented, and under higher levels the number of target-distractor pairs increased. Distractibility on the flanker task referred to slower reaction times to the target letter when the target letter(s) was presented with the incompatible distractor letter(s) (i.e., Y when the target was X) than when the target letter(s) was presented with the compatible distractor letter(s) (i.e., X when the target was X). Older adults demonstrated that they had a lower perceptual capacity because they required a smaller increase in perceptual load than younger adults required to decrease distractibility on the flanker task. This experiment revealed one avenue for enhancing older adults’ selective attention – increasing the perceptual load of the task. For everyday real-world attention-demanding tasks, however, increasing task-relevant stimuli or perceptual processing requirements of task-relevant stimuli is not necessarily possible, practical, or appropriate.
Considering older adults have a lower perceptual capacity, we would expect them to be less distractible than younger adults on a task with low perceptual load, as there would be less perceptual room available for distractors to be processed. However, this was not the case – In the experiment described above, Lavie and colleagues (2004) found that, when completing the flanker task under the lowest level of perceptual load, older adults were more distractible than were younger adults. Their distractibility could not be attributed to impairment at the perceptual level, because an impaired or lowered
and therefore would have reduced potential for distraction. More likely, older adults’ distractibility may be traced to impairment at the executive level. In other words, when the target stimulus was perceptually simple, there was room at the perceptual level for both the target and distractor stimuli to be processed and to reach the executive level. Then, at the executive level, older adults were required to exert inhibitory control, and were ill equipped at this level to ignore the distractor stimuli. Interestingly, Lavie et al. (2004) note that the response profiles of older adults on the flanker task resemble those of patients with frontal lobe damage (e.g., Shalice & Burgess, 1991). Given the involvement of frontal regions in executive functioning, this comparison further supports the executive nature of older adults’ impaired performance on the selective attention task.
Executive Mechanism of Selective Attention
The executive mechanism of Lavie and colleagues’ (2004) theory is exercised when there is low perceptual load and distractors are perceived. In contrast to the passive and automatic perceptual mechanism, the executive mechanism is a controlled and active process that depends on higher cognitive functions, such as working memory, planning, and inhibitory control. The executive mechanism works the opposite way that the perceptual mechanism works; whereas high perceptual load decreases distractor processing, when there is high executive load, the capacity of executive processes to control and reject distractors is drained, and selective attention suffers. This idea that executive resources are limited is corroborated by resource allocation theory, as well as other research (e.g., Schmeichel, 2007). The basic principle is that tasks that use
executive resources can deplete these resources and hinder one’s performance on concurrent or subsequent tasks that also require executive resources.
In support of the executive mechanism, Lavie et al. (2004) reference studies that correlate working memory capacity with performance on a selective attention task. Individuals with low working memory capacity tend to perform worse on the Stroop task than do individuals with higher working memory capacity. The measures of working memory capacity that show this negative correlation with selective attention performance are those that involve executive functions such as task-switching and task-dividing. Lavie et al. (2004) suggest that these executive components of working memory capacity overlap with attention. In fact, their proposed executive mechanism inherently suggests that attention and executive functions share overlapping processes – when executive resources are occupied, selective attention is impaired.
To summarize the construct of selective attention, it involves initial perceptual processing as well as deeper executive-level processing. Older adults are technically impaired at the perceptual level because they have lower perceptual capacity related to posterior cortical loss and accompanying sensory difficulties with hearing and vision. This impairment could actually serve to enhance attention if the task’s perceptual load can be increased to fill older adults’ perceptual capacity, leaving less room for distractors to be processed. When tasks have low perceptual load, distractors reach the executive level of processing. Resources are limited at the executive level, and appear to be particularly impaired in older adults; efforts to enhance older adults’ selective attention should ideally not drain their limited executive resources. This consideration is relevant for the current exploration of musical intervention, and will be discussed in later sections in the context of music-induced emotion and the draining nature of emotion regulation on executive resources.
Selective Attention Deficits in Older Adults
Hasher and Zacks (1988) have proposed that older adults’ selective attention deficits are rooted in inhibitory deficits. Inhibitory mechanisms ideally prevent entrance of task-irrelevant information to working memory. However, a person’s goals ultimately guide these inhibitory mechanisms, and older adults are thought to place increased importance on personal goals, values, and experiences. Thus, task-irrelevant information that nevertheless aligns with an older adult’s goals is likely to enter his or her working memory and affect selective attention. Interestingly, this claim seems to be compatible with Mather and Carstensen’s (2005) more recent Socioemotional Selectivity Theory. In brief, Socioemotional Selectivity Theory says that older adults focus on their emotional goals, and are thus more likely to pay attention to emotional stimuli and to allocate their executive resources to emotion regulation.
Hasher and Zacks’s (1988) theory has been empirically supported by older adults’ impaired performance on a variety of inhibitory tasks (i.e., directed forgetting paradigms, the garden path sentence paradigm, Stroop tasks, negative priming, response
compatibility, the stop-signal paradigm, and the Wisconsin Card Sorting test). Multiple researchers have corroborated the finding that older adults exhibit deficits on some of these tasks, specifically on those traditionally used to assess selective attention. For example, Rush, Barch, and Braver (2006) found age-related deficits on the Stroop task and on the stop-signal task, but not on the garden path sentence task. In this case, the former task is traditionally used to assess inhibition as it applies to selective attention, whereas the latter task assesses inhibition as it applies to implicit memory. Hasher and Zacks’s (1988) theory supports the idea that the construct of selective attention overlaps
with inhibitory control, and Rush et al.’s (2006) findings show that this area of overlap is affected by aging.
Rush and colleagues (2006) have additionally suggested that executive control deficits in the area of context processing underlie changes in older adults’ attention. Context processing is considered to be an aspect of executive control because it involves internally representing, maintaining, and updating environmental cues in order to use these cues to control thoughts and behaviour. Rush and colleagues (2006) outline how context-processing deficits account for older adults’ performance patterns on the AX-CPT, which is a modified continuous performance task designed to assess context-processing. The AX-CPT is arguably also assessing selective attention, as the participant must focus and initiate a response when the target X appears on the screen, and ignore and inhibit the response when the distractor Y appears on-screen. The task further assesses the ability of context processing to help guide the appropriate response, as it requires a response to the target X when it appears in a pair with the context letter A, but not when it is paired with the context letter B. Inhibitory control would still be required here, because the participant must ignore and inhibit a response towards the BX pair. Note that the AX-CPT is a context-processing task rather than a true selective attention task, and likely taps into constructs of attention besides selective attention. For example, the task appears to require divided attention, as the participant must attend to whether the target is X or Y, as well as whether the context letter is A or B.
In comparison to younger adults, on the AX-CPT older adults made fewer errors on AY trials, during which the context letter A suggests a response but the distractor letter Y suggests a non-response (Rush et al., 2006). Older adults also showed
disproportionately slower responses on BX trials, during which the context letter B suggests a non-response but the target letter X suggests a response. These results suggest there is a qualitative age difference in context processing. Rush et al.’s (2006) sample of older adults also showed significant impairment on the Stroop and stop-signal tasks assessing inhibitory function. Furthermore, they exhibited slower reaction times, as would be predicted by Salthouse’s (1996) processing speed theory of aging. Still, Rush and colleagues (2006) dismiss inhibitory and processing speed accounts of cognitive aging and argue that context processing deficits fully accounted for older adults’ impairment on the AX-CPT. Their reasoning is that older adults’ impairment on
inhibitory tasks was of a smaller magnitude than their impairment on the AX-CPT, and that processing speed could not fully account for the pattern of results on the AX-CPT. An alternate interpretation of their results could be that multiple executive function deficits (i.e., both context-processing deficits and inhibitory deficits) contributed to older adults’ impairment in selective attention performance on the AX-CPT. Tasks that could effectively parse out these different executive domains would be needed to ascertain the unique and shared contributions of these factors to older adults’ performance.
The Multi-Source Interference Task. The current thesis study employed the Multi-Source Interference Task (MSIT; Bush & Shin, 2006, Bush, Shin, Holmes, Rosen, & Vogt, 2003) to assess older adults’ selective attention at baseline and during musical intervention. The MSIT assesses participants’ ability to attend to task-relevant stimuli and ignore task-irrelevant or interfering stimuli. The MSIT incorporates multiple sources of interference by combining elements of the Stroop, Eriksen Flanker, and Simon Effect tasks. Furthermore, the MSIT is strongly supported by behavioural and neuroimaging
data to reliably and robustly activate the cingulo-frontal-parietal attention network (CFP network). The data show sufficient test-retest reliability, making the task ideal for assessing changes in selective attention with musical intervention. Notably, the MSIT was designed to maximally activate the dorsal anterior midcingulate cortex (daMCC) of the CFP network, which is strongly implicated in executive control.
Salami, Reickmann, Fischer, and Backman (2014) have used the MSIT to explore age differences in the neural and behavioural correlates of selective attention. They found that older adults’ reaction times were slower in comparison to younger adults during both interference and control trials of the MSIT, however their performance was only less accurate than younger adults during the interference trials. Thus, the interference effect was stronger for older adults than for younger adults. Salami et al. used neuroimaging data and functional connectivity analyses to determine that during interference trials, older adults demonstrated greater activation of the dorsolateral prefrontal cortex and anterior cingulate cortex, as well as posterior regions. Their findings suggest that (1) older adults are less efficient at recruiting an “interference network”, and (2) older adults attempt to compensate by activating a different posterior network that facilitates conflict resolution less effectively. Salami and colleagues’ (2014) research clearly shows that the MSIT is a well-suited task for assessing older adults’ selective attention.
Emotion Regulation in Older Adults
Music serves various purposes in people’s everyday lives. A person’s intention when listening to music may be to enjoy or appreciate the music, to have a background stimulus that facilitates completion of either mundane activities or complex tasks, or to process emotions and memories induced by the music. Many or most of the executive
and attention tasks previously described pertain to “cold” cognitive functions, i.e., those with no particular emotional valence. Considering the relevance of executive functions to selective attention as well as emotion regulation, one might wonder how older adults’ executive resources would be allocated in the context of an emotionally-neutral attention task presented simultaneously with an emotionally-charged stimulus such as music.
Emotion regulation refers to the process of strengthening, weakening, or
maintaining the intensity of positive or negative emotions according to one’s goals (Gross & Thompson, 2007). Emotion regulation is related to executive functions, as it is a
complex process requiring attention to affective cues, detection of cognitive conflict between one’s present affective state and one’s desired affective state, and allocation of executive resources towards the goal of altering the affective state or inhibiting responses associated with the affective state (Teper, Segal, & Inzlicht, 2013). Interestingly,
executive functions that are highlighted as being relevant to emotion regulation, such as inhibition, are also particularly relevant for selective attention. Empirical support for the involvement of executive control in emotion regulation is found in several studies. For example, Mather and Knight (2005) showed that older adults who performed poorly on tests of executive control or were distracted were less likely to regulate emotions. Similarly, in other studies when participants completed an executive control task
composed of completing mathematical equations while also encoding and recalling target words, they later performed poorly on the emotion regulation task of watching a
negative-affect inducing video of animal slaughter while following instructions to not show any facial expression of emotion.
Knight et al. (2007) discuss how age-related declines in executive processes of working memory and selective attention are ironically contrasted with improvements in the executive process of emotion regulation. This contrast is supported by
neuropsychological task data demonstrating that older adults’ performance deficits are minimal on affective-processing tasks (Peters, 2007). Furthermore, neuroimaging data shows less of a change in older adults’ cortical areas associated with affective-processing (e.g., the amygdala and ventromedial prefrontal cortex), relative to their non-affective areas (e.g., the dorsolateral prefrontal cortex). Thus,we can expect older adults’ executive resources to be occupied by their enhanced emotion regulation performance.
According to Mather and Carstensen’s (2005) Socioemotional Selectivity Theory, older adults are motivated to allocate their executive resources to emotion regulation. Whereas younger adults perceive time as open-ended and focus on preparatory goals such as gathering information, novel experiences, and knowledge, older adults view their time horizons as limited and their goals focus on regulating current feeling states to optimize well-being and promote positive affect. Although Socioemotional Selectivity Theory includes language surrounding “motivation”, this motivation is a presumed drive that is not explicitly verified; if this motivation is present, it is likely more of an implicit process seen in perceptual biases than an explicit or conscious drive that older adults would verbalize. That said, there are several studies showing how older adults’ attention may be impacted by their (implicit) motivation to regulate emotions. For example, in one study researchers presented younger and older adults with a negative and neutral picture simultaneously, and tracked their eye movements (Rosler et al., 2005). Both age groups glanced initially at the negative picture, however the younger adults sustained attention
longer to the negative picture than did the older adults. Thus, there is no difference in the age groups’ passive perceptual process of detecting and orienting to negative stimuli. However, there appears to be an age-related difference in the active executive process of inhibiting attention to negative stimuli; that is, only older adults inhibited their attention to the negative picture. In terms of positive stimuli, both younger and older adults showed greater sustained attention to positive stimuli than to neutral stimuli. This sustained attention to positive stimuli would also likely involve executive resources in order to actively ignore the neutral stimuli while directing and sustaining attention to the positive stimuli. Petersen and Posner’s (2012) model supports the involvement of executive resources in sustained attention. In the context of the current study, then, Rosler et al.’s (2005) research suggests that older adults’ executive resources may be occupied by both negative and positive musical stimuli, albeit for converse reasons, more so than neutral stimuli. By extension, then, one might predict that emotionally neutral music would have more of a facilitative effect on attention-demanding tasks than music that is affectively valenced, either positively or negatively.
Mood congruence. Mood congruence occurs when an individual attends to stimuli with a similar emotional valence to his or her mood; it is an automatic or passive lower-level process associated with the amygdala rather than a controlled executive-level process relying on frontal regions (e.g., Knight, Maines, & Robinson, 2002; Suslow et al., 2009). Despite some stereotypes to the contrary, it is well recognized that older adults are typically in a more positive and less negative mood than younger adults (e.g., Peters, Hess, Vastfjall, & Auman, 2007). Thus, older adults may be more inclined to attend to
positive stimuli and less inclined to attend to negative stimuli simply because their mood automatically facilitates this attention.
In the context of the present study, though, it is important to note that older adults’ diverting of attention away from negative stimuli is more likely an executive process of emotion regulation than an automatic mood congruence process. Inhibition and switching are two executive functions in particular that appear to be required for this attentional shift. For example, in Rosler and colleagues’ (2005) research, older adults who were presented with both neutral and negative stimuli initially attended to the negative stimulus, and then presumably used the executive functions of inhibition and switching to divert attention from the negative stimulus and towards the neutral stimulus. The use of executive functions in diverting attention from negative stimuli is notable because negative music would be predicted to drain older adults’ executive resources as they divert attention away from the music, leaving little resources available for the actual attention task.
In terms of attention towards positive stimuli in the above study by Rosler and colleagues (2005), it is not clear whether older adults’ greater sustained attention to positive rather than neutral stimuli relied on executive functions or mood congruence. Based on Petersen and Posner’s (2012) model, though, orienting attention is an automatic process, whereas sustaining attention requires executive resources. Thus, a plausible interpretation is that older adults’ positive mood facilitated initial orienting towards the positive stimulus, and executive functions were used to sustain this attention. Following this interpretation, positive music in the current study would also be expected to use older adults’ executive resources and consequently hinder performance on the attention task.
Motivation and effort in the context of succeeding on a cognitive task. From the perspective of Mather and Carstensen’s (2005) Socioemotional Selectivity Theory, motivation to fulfill emotional goals underlies older adults’ mobilization of effort towards regulating emotions. Existing literature emphasizes the role of motivation and effort in successful performance on a cognitive task, indicating that older adults would likely perform poorly on a task when simultaneously presented with emotional music, as effort would be directed towards emotional regulation. In their research, Gendolla, Abele, and Krusken (2001) emphasize task-demand appraisals in the mobilization of effort. These researchers induced positive and negative affective states in younger adults using either music or autobiographical recollection. Then, participants completed a letter-cancellation task and reported task-demand appraisals by rating task difficulty, subjective ability, and amount of effort required on seven-point scales. Effort was also assessed by the
autonomic measure of blood pressure. Negative affect was associated with higher demand appraisals and stronger effort than was positive affect. Furthermore, there was a weak but significant relationship between effort and task performance. Gendolla and colleagues (2001) suggest that the higher mental effort and performance associated with negative affect was due to the higher task-demand appraisal associated with negative affect. In this view, negative affect made participants think the task difficulty exceeded their mental ability, and this is why participants engaged mental effort.
Sarter, Gehring, and Kozak (2006) present a model that may further elaborate on the role of motivation in the study by Gendolla et al. (2001). According to Sarter et al.’s (2006) top-down model of interacting cortical, mesolimbic, and cholinergic systems, both task-demands and motivation are integral to attentional effort. Task demands are a
stimulus for potentially adjusting effort, however the actual adjustment depends on the individual’s motivation to perform. In this view, attentional effort may or may not be mobilized following increased task demands such as distractors, prolonged time-on-task, circadian phase shifts, stress, or sickness. Effort will be mobilized only if there is
cognitive incentive from internal or external motivational forces to counter performance decline. Empirical support for this view comes from Sarter and colleagues’ research, in which participants’ performance on a sustained attention task was superior when they were paid to perform well.
Prior literature has suggested that cognitive and emotional regulation become competing tasks in older adulthood, and that according at least to Socioemotional Selectivity Theory, older adults would be motivated to direct more resources toward the latter than the former (Peters et al., 2007). As such, on a demanding (i.e., executive) selective attention task, rather than motivating older adults to direct effort towards performance on the task, positive and negative music-induced affect would likely motivate older adults to direct resources toward regulating emotions. Older adults typically regulate emotions through strategies of response-focused coping and positive reappraisal (Carstensen, Fung, & Charles, 2003). As an example of these respective strategies, negative affect would likely cause them to direct mental effort towards changing the negative affective response or towards changing the negative appraisal of the music to a positive appraisal. Younger adults may be motivated to regulate emotions through different strategies than those that are used by older adults. For instance,
Schwarz’s (1990) theory suggests that positive or negative induced affect would inform the younger adult whether or not the person-environment relationship is sound, with
negative affect motivating him or her to engage in high-effort analytical information processing on the attention task to solve the person-environment problem.
Previous research has considered music to be a useful tool for affecting task demands and consequently impacting attentional effort and performance. For example, in work by Unal, Steg, and Epstude (2012), music was found to enhance younger adults’ performance on an attention-demanding real-world driving task. Their participants drove via computer simulation in regular and conflict traffic situations in one of two conditions: music or silence. In the music condition, participants selected music they would usually listen to while driving. They verbally reported a number between 0 and 150 to indicate effort on the Rating Scale Mental Effort at 13 points throughout the driving task. The main finding was that music drivers indicated higher levels of mental effort than silence drivers. Music enhanced performance on two of the conflict traffic situations, and mental effort mediated this enhancement in the car-following situation involving executive attention. Unal and colleagues (2012) explained their results in the framework of
Hockey’s (1997) compensatory theory of mental effort, which says that when mental load is increased, for example because of the distracting nature of music, drivers compensate by adjusting their effort to complete the task. Thus, Unal and colleagues (2012) reached similar conclusions as did Gendolla et al. (2001) about increased task demands providing a stimulus for the person to adjust mental effort to succeed on the task. Notably, both of the aforementioned studies pertained to a younger-adult sample, and older adults would not necessarily exhibit enhanced performance, especially when the factor increasing task demands is emotionally-charged music. The body of literature disentangling roles of motivation, effort, and task demands in performance on a cognitive task informed the
measurement of affect, motivation, and task-demand appraisals associated with the music and task in the present research.
There are, however, several additional considerations for applying findings from this literature to the research area of using music to enhance older adults’ selective attention. In noting limitations of their research, Unal and colleagues (2012) suggest that future research examines different types or properties of music, specifically in the population of older adults whose declining cognitive resources may make them more susceptible to the task demands induced by music. Moreover, instrumental music may not even pose task demands to participants; in Unal et al.’s (2012) study, participants self-selected their favourite music, which may have involved distracting lyrics. Various studies have shown that lyrical music selections that are presented in their original form with lyrics or in their “karaoke” form without lyrics may impair performance on
cognitive tasks, presumably because they trigger overt or subvocal rehearsal of the lyrics, as conceptualized by Baddeley’s phonological loop of working memory (e.g., Alley & Greene, 2008; Iwanaga & Ito, 2002; Pring & Walker, 1994). For this very reason, in the current thesis study, instrumental music was selected in order to avoid the distraction imposed by lyrics. Furthermore, the current study attempted to parse out the role of affective and arousal properties of music.
Arousal and Attention
The potential for music to induce physiological arousal is a pertinent reason for why music was explored as a rehabilitative tool for attention in this study. For this study’s purposes, arousal will refer to physiological symptoms of activation, such as increased heart rate. Thayer, Hansen, Saus-Rose, and Johnsen’s (2009) neurovisceral
model explains why different types of music may differentially impact attentional performance on an executive-level task. Specifically, a musical stimulus that decreases arousal or increases heart-rate variability (HRV) would be predicted to enhance
performance on an executive task. This neurovisceral model is based on an inhibitory subcortical pathway by which increased HRV is associated with activation in the prefrontal cortex (i.e., the area associated with executive functions). This pathway is supported by pharmacological and neuroimaging data. In empirical investigations, Thayer et al. (2009) measured HRV’s association with attention during the continuous performance task, which includes executive tasks, such as detecting identical stimuli, as well non-executive tasks obtaining simple reaction-time measurements of attention. Individuals with high HRV performed better on the executive tasks.
Affect, arousal, and attention. Arousal is a construct that is related to and may moderate the impact of affect on attention. An example of this interaction between affect and arousal comes from responses on the Stroop task in a study by Wurm, Labouvie-Vief, Aycock, Rebucal, and Koch (2004). Stroop task responses are commonly slower for positive and negative emotion words than for neutral words; further to this impact of affect, Wurm et al. (2004) found that older adults took longer to name the colour of high-arousal positive and negative emotion words than low-high-arousal positive and negative emotion words. Thus, an emotional stimulus may be more disruptive to attention when it is more arousing.
Musical Affect and Arousal
Husain, Thompson, and Schellenberg (2002) have proposed the “arousal-mood” hypothesis, which supports arousal and mood as mediators in the relationship between
music-listening and cognitive performance. They defined mood as relatively long-lasting emotions, self-reported as “sad, happy, discouraged, gloomy”; however one should note that their measurements more likely represent affect, as they used adapted versions of the Affect Grid by Russell, Weiss, and Mendelsohn (1989) as well as participants’ self-reported ratings of how they were feeling at the time of testing. Affect is a momentary emotional state as compared to longer-lasting moods. Arousal refers to a degree of physiological activation, self-reported as “vigor, activity, wakefulness”. There appear to be two properties of music that impact the listener’s affect and arousal: The first is its major, minor or atonal quality, which primarily impacts affect, and the second is its stimulative or sedative quality, which primarily impacts arousal.
In the context of music, one can consider how different properties might promote arousal. Gaston (1951) provided original descriptions of stimulative and sedative music. Stimulative music is characterized by staccato notes (i.e., short, detached, percussive), whereas sedative music consists of legato notes (i.e., long, sustained, smooth).
Stimulative music builds up physical energy, whereas sedative music triggers
contemplative and dreamlike states. Gaston (1951) had initially associated sedative music with depression and sorrow in noting its relaxing effect on muscles, indicating that the music heightened low arousal states on the spectrum of negative affect. Subsequent research by Smith and Morris (1976) showed that sedative music also decreased nervousness, panic, tension, and uneasiness; in other words, the music dampened high arousal states on the spectrum of negative affect. Therefore, sedative music does not necessarily increase negative affect, but rather it decreases arousal. The impact of the sedative or stimulative quality of music on arousal is supported by results from Husain et
al.’s (2002) research dissecting the “Mozart effect”. Husain and colleagues (2002) manipulated a Mozart sonata to produce four variations in tempo and mode: slow minor, slow major, fast minor, and fast major. The tempo of the music would correspond with its sedative or stimulative quality. Fast-tempo or stimulative music increased young adults’ arousal, whereas slow-tempo or sedative music decreased their arousal. At the same time, when the music was divided into these elements it became clear that the major versus minor modality of the music, rather than its stimulative versus sedative quality, impacted affect. Specifically, listening to major music increased positive affect, whereas listening to minor music increased negative affect.
In research assessing stimulative music’s impact on arousal, Hirowaka (2004) measured arousal with Thayer’s Activation-Deactivation Adjective Checklist (AD ACL) before and after music exposure. The AD ACL is a multidimensional measure, eliciting physiological and psychological self-reports of Energetic Arousal, with the subscales energy and tiredness, and Tense Arousal, with the subscales tension and calmness. Hirowaka found that stimulative music increased older adults’ energy. The work by Husain and colleagues (2002) and Hirowaka (2004) on the relationship between music and arousal corresponds to research uncovering the impact of stimulative versus sedative music on heart-rate variability (HRV). For example, Iwanaga, Kobayashi, and Kawasaki (2005) found that stimulative music decreased HRV and was associated with lower perceived relaxation and higher perceived tension than sedative music. Thus, stimulative and sedative properties of music may be a valid way of manipulating arousal and
associated HRV, ultimately impacting attention via Thayer and colleague’s (2009) neurovisceral pathway.
Study Aims and Hypotheses
Selective attention is a worthy construct for cognitive rehabilitation among older adults. Older adults show normative age-related declines in their executive processing of task-relevant and task-irrelevant stimuli. Musical affect and arousal are two musical properties that may impact older adults’ efficient use of their limited executive resources. In regards to musical affect, older adults would be motivated to use their executive resources to regulate affect (i.e., ignore negative musical stimuli and attend to positive musical stimuli) rather than to meet the goals of an attention task. In this case, neutral music would likely be the ideal stimulus for optimizing older adults’ arousal while not occupying executive resources. At the same time, musical arousal may also impact older adults’ executive functioning, specifically via bottom-up neuro-visceral pathways. Although heart-rate variability is not measured in the current study, one can understand the potential impact of musical arousal on attention through understanding the construct of heart-rate variability. Specifically, high-arousal music would hinder executive
functioning, presumably via the impact of decreased heart-rate variability on the
prefrontal cortex. Low-arousal music, on the other hand, would be predicted to increase heart-rate variability and enhance executive functioning.
The aim of this thesis study was to determine the impact of musical affect and arousal on older adults’ selective attention.In this study, older adults completed the Multi-Source Interference Task (MSIT) while under exposure to six musical conditions, each condition representing a specific combination of musical affect and arousal (i.e., high arousal & negative, high arousal & positive, high arousal & neutral, low arousal & negative, low arousal & positive, and low arousal & neutral). Older adults’ MSIT
performance was compared across the three affective conditions (i.e., negative, positive, and neutral), across the two arousal conditions (i.e., high and low), and within each of the six combinations of affect and arousal.
Hypotheses about older adults’ MSIT performance are presented in relation to musical affect, musical arousal, and the interaction between these two musical properties.
1. Musical affect: Older adults would perform better on MSIT interference trials while listening to neutral music, in comparison to negative and positive music. 2. Musical arousal: Older adults would perform better on MSIT interference trials
while listening to low-arousal music, as opposed to high-arousal music. 3. Interaction between musical affect and arousal: Finally, there was an expected
interaction of the above main effects, such that older adults’ best performance on MSIT interference trials would occur while listening to low-arousal & neutral music.
Method Participants
The target population for this study was healthy older adults between the ages of 65 and 80 years. Participants were recruited from the Victoria and surrounding
community via flyer distribution at several senior centres and gyms, a presentation at the Embrace Aging Event at the Yakimovich Wellness Centre, and emails to the University of Victoria Retirees Association and University of Victoria Alumni Association. Flyers were also posted in the “Gettin’ Higher Choir” newsletter and on websites for the SmartLab and the Institute on Aging & Lifelong Health. Participants were not compensated for their participation, although they were offered reimbursement for parking or bus fare when traveling to campus for the two study sessions.
Exclusion criteria for participants were (1) self-reported uncorrected hearing or vision loss, and (2) a history of dementia, stroke, major head injury, or any other
neurologic diagnosis. Participants completed the Lawton Instrumental Activities of Daily Living Scale to screen for functional impairment in their daily lives that could be
indicative of clinically significant cognitive impairment. All participants attained the maximum score of 8, indicating relative functional independence.
Measures
Self-report measures. Questionnaires assessed participants’ levels of affect, arousal, depression, and anxiety. Furthermore, participants reported their perceptions of the attention task, the silence, and each musical piece.
Demographic information. The demographic questionnaire, provided in Appendix A, pertained to various characteristics, including: age, gender, education, musical training, and favourite genres of music.
Daily arousal-related activities. Questions referred to the quality, frequency, or length of typical daily activities, such as music exposure, caffeine and alcohol
consumption, smoking behaviour, drug use, sleep patterns, and physical activity. The Routine Arousal Questionnaire is included in Appendix B.
Positive and Negative Affective Schedule (PANAS). The PANAS is a highly reliable and valid scale assessing two factors: positive and negative affect (Watson, Clark, & Tellegen, 1988; see Appendix C). The participant rates the extent to which he or she feels 20 different emotions in the present moment. Ratings are made on a five-point scale (1= “Very Slightly or Not at All”; 5= “Extremely”). Ten emotion words (e.g., “interested”, “strong”, and “alert”) count towards a positive affect score and 10 emotion words (e.g., “distressed”, “guilty”, and “nervous”) count towards a negative affect score. The minimum possible score for each emotional valence is 10, and the maximum is 50.
Activation Deactivation Adjective Checklist (AD ACL). The AD ACL is a multidimensional assessment of momentary arousal (Thayer 1967, 1978, 1986, 1989; see Appendix D). The Energetic Arousal dimension incorporates the subscales Energy and Tiredness, and the Tense Arousal dimension includes the subscales Tension and Calmness. Furthermore, Energy and Calmness correspond to high and low arousal dimensions of positive affect, and Tension and Tiredness represent high and low arousal dimensions of negative affect. The AD ACL is a 20-item form with five items for each of the four subscales. Some examples of adjective items, along with their corresponding
subscales, are “vigorous” (Energy), “drowsy” (Tiredness), “jittery” (Tension), and “placid” (Calmness). A minimum score of five and a maximum score of 20 may be obtained on each subscale, as participants rate the extent to which they feel each item in the present moment, on a four-point scale with verbal anchors (no= “definitely do not feel” ; ? = “cannot decide”; v= “feel slightly”; vv= “definitely feel”). Although its verbal rating scale is unconventional, the AD ACL is a psychometrically sound measure. Self-reports on the AD ACL correlate well with affective states, physiological disturbances, sleep patterns, exercise, cognitive and information processing functions, and stress.
Geriatric Depression Scale (GDS). Considering the relevance of depression to affect and arousal, participants were assessed for depression with the GDS (Yesavage et al., 1983; see Appendix E). The GDS is a reliable and valid scale that was designed specifically for the elderly population. For instance, its items do not refer to somatic depressive symptoms because non-depressed healthy older adults typically experience somatic ailments (i.e., sleep disruption, sexual dysfunction, digestive difficulties, and pain). Furthermore, its format is simple and easily understood, requiring a yes or no response to all 30 items. Some example items are “Are you in good spirits most of the time?” and “Is it easy for you to make decisions?”. In the GDS validation study, older adults over the age of 55 were classified as “normal”, “mildly depressed”, or “severely depressed” according to the Research Diagnostic Criteria for major affective disorder. Average GDS scores were 5.75 (SD= 4.324) for the normal older adults, 15.05 (SD= 6.50) for the mildly depressed older adults, and 22.85 (SD = 5.07) for the severely
depressed older adults. A score of 10 or higher is the cut-off for clinical depression, with a possible range of one (minimum) to 30 (maximum).
Geriatric Anxiety Scale Version 2.0 (GAS). The GAS is a brief 25-item self-report measure that assessed participants’ anxiety, as anxiety is a construct relating to affect and arousal (Segal, June, Payne, Coolidge, & Yochim, 2010; see Appendix F). The GAS has shown sound psychometric properties in a sample of community-dwelling older adults as well as a sample of older adults receiving outpatient mental health care. In both samples, there was strong internal reliability, construct validity, and convergent validity with measures such as the Geriatric Depression Scale, State-Trait Anxiety Inventory, Beck Anxiety Inventory, and Adult Manifest Anxiety Scale-Elderly Version (AMAS-E). The GAS items are based on symptoms for anxiety disorders listed in the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV-TR). Three GAS subscales cover domains of somatic, cognitive, and affective symptoms. It is beneficial that the current study’s anxiety measure emphasizes somatic anxiety symptoms such as heart racing and shortness of breath, as these physiological signs of arousal are particularly relevant to attention. Participants indicate how often they have experienced each symptom over the past week. Responses are made on a four-point scale, which is ideal for detecting subtle variations in anxiety symptoms in comparison to other elder-specific anxiety assessment measures which require a true or false answer (e.g., the AMAS-E and the Geriatric Anxiety Inventory). Obtained scores include a total score, with a possible range from zero (minimum) to 75 (maximum), as well as individual subscale scores: Somatic subscale scores range from zero (minimum) to 36 (maximum), and Cognitive and Affective subscales each range from zero (minimum) to 24 (maximum). In the clinical sample of older adults seeking outpatient treatment for caregiver-related concerns and various depressive and anxiety symptoms, the total mean score was 20.75 (SD= 10.73),
the somatic score was 8.59 (SD= 4.28), the cognitive score was 5.94 (SD= 3.89), and the affective score was 6.21 (SD= 3.45).
Post-task questionnaire. For the silence baseline and each musical condition, a post-task questionnaire assessed participants’ perceptions of the silence or musical piece and of the attention task (See Appendix G). Questions referred to participants’ emotions, memories, enjoyment, and sense of familiarity triggered by the music or silence, as well as their perception of task difficulty and motivation to succeed on the task, all factors that may have influenced objective performance on the task.
Standardized clinical-neuropsychological measures. The Test of Everyday Attention (TEA; Robertson, Ward, Ridgeway, & Nimmo-Smith, 1996) is an ecologically valid test that was used in this study to characterize participants by their attentional performance on tasks mimicking everyday real-world scenarios. The TEA was developed in accordance with Posner and Petersen’s (1990) theoretical attention networks, and the executive control network may be assessed by six of its subtests: Map Search, Telephone Search, Telephone Search while Counting, Visual Elevator, Elevator Counting with Distraction, and Elevator Counting with Reversal. These six executive subtests were selected for the current study. All subtests load onto one of four factors: visual selective attention/speed, attentional switching, sustained attention, and auditory-verbal working memory. Map Search and Telephone Search assess visual selective attention. Visual Elevator measures the ability to switch attention. Elevator Counting with Distraction and Elevator Counting with Reversal assess auditory-verbal working memory and likely also taps into auditory selective attention. Telephone Search while Counting loads onto the sustained attention factor, although Robertson and colleagues (1996) note that it was
designed to assess divided attention and is sensitive to the ability to handle complex everyday tasks. The TEA was normed on a sample of 154 healthy volunteers, stratified into four age bands, with the oldest group ranging from 65 to 80 years of age.
Experimental measures. Several experimental measures were used as the primary independent variables in the current study.
Multi-Source Interference Task (MSIT; Bush & Shin, 2006; Bush et al., 2003). Each stimulus trial of the MSIT incorporates cognitive interference elements from the Stroop, Eriksen, and Simon tasks, as well as other elements known to activate the
dorsolateral anterior mid cingulate cortex (i.e., decision-making, target detection, novelty detection, error detection, response selection, stimulus/response competition, and task difficulty). In each trial, participants see three numbers on-screen and must indicate, via key press, the one (target) number that differs from the other two (distractor) numbers. During control trials, the target number 1, 2, or 3, is placed congruently in the first, second, or third position on the screen, and the two distractors are zeros. During
interference trials, the on-screen placement of the target number 1, 2, or 3 is incongruent with its numerical value, and the distractors are other numbers rather than zeros.
Appendix H includes pictorial examples of control and interference trials.
Participants are explicitly instructed that the on-screen numbers will change approximately every 2 s, and that they should “answer as quickly as possible, but since getting the correct answer is important, do not sacrifice accuracy for speed” (Bush & Shin, 2006). After instructions are reviewed, participants complete a practice session consisting of a 12-trial control block followed by a 12-trial interference block. If the participant is found to press the key corresponding to the position rather than the number
of the target, the experimenter should review this error with the participant and have them repeat the practice session correctly. The actual task is comprised of four 24-trial control blocks that alternate with four 24-trial interference blocks. Half-way through the task (i.e., after two control blocks alternated with two interference blocks), participants are given feedback on their performance on the first set of trials, and are instructed to press a button to continue the next set of trials.
Accuracy scores reported by Bush and Shin (2006) for 25 healthy adult volunteers were 99.4% (SD= 1.3) for control trials, and 97.4% (SD= 2.0) for interference trials. Accuracy and RT differences between the two trial types were both highly significant (i.e., p< .00001). The difference between RTs in interference and control trials is known as the cognitive interference effect, typically starting at 300-400 ms during the first 5 min, and then stabilizing, although with extensive practice it still remains above 200 ms.
Musical pieces. Six non-lyrical instrumental pieces were selected to form the six musical conditions for this study (see Table 1). Three music students from the University of Victoria’s School of Music helped to select these pieces, by suggesting possible pieces and confirming the tempo (i.e., stimulative or sedative) and mode (i.e., major, minor, or atonal) of the final selections. Each piece represents a different combination of tempo and mode, and is intended to evoke a particular combination of affect and arousal.
Stimulative and sedative music correspond to high and low arousal, respectively (e.g., Hirowaka, 2004; Husain et al., 2002). Gaston’s (1951) guidelines of stimulative and sedative music state that stimulative music is characterized by faster-tempo staccato notes (i.e., short, detached, percussive), whereas sedative music consists of slower legato notes (i.e., long, sustained, smooth).
Major, minor, and atonal music are related to positive, negative, and neutral affect, respectively (e.g., Daynes, 2010; Husain et al., 2002). Atonality in particular is a subjective psychoacoustic property referring to a lack of intelligible key (i.e., the listener cannot decipher whether the music is major or minor). Atonality is commonly found in relatively modern genres, such as New Age and Free Jazz. All pieces were modern, from genres that best represent the specific combination of arousal and affect. Modern music was thought to be ideal for this experiment because it would likely be less familiar to older adults, and therefore less likely to evoke autobiographical memories that would occupy executive resources and lead to distraction from the attention task (Bernsten, 1998; Janata, Rakowsk, & Tomic, 2007). Furthermore, modern music would be typically accessible to older adults in the community (i.e., via radio, background music at locations such as grocery stores and banks).
Table 1.
Characteristics of musical conditions.
Condition Tempo Mode Artist Title Date Length Genre
High-arousal positive
Stimulative Major Todd Day Band “People” 2014 2:40 Jazz Fusion
High-arousal negative
Stimulative Minor Animals as Leaders
“Nephele” 2014 4:31 Progressive
Metal High-arousal
neutral
Stimulative Atonal Sounds of Isha “The Leap” 2004 12:57 World
Low-arousal positive
Sedative Major Enya “And Winter Came” 2008 2:22 New Age
Low-arousal negative
Sedative Minor Ludovico Einaudi
“Walk” 2013 3:27 Contemporary
Classical Low-arousal
neutral
Sedative Atonal Morton Feldman “For Samuel Beckett” 2006 18:51 Avant-Garde (Experimental) Classical
Procedure
Baseline assessments. Participants were invited to the lab for two separate morning sessions, each session spanning approximately two hours. Older adults’
circadian rhythms tend to peak in the morning, and this is the time of the day when their cognitive performance is optimal, especially on interference tasks (Borella, Ludwig, Dirk, & Ribaupierre, 2011). For the first session, upon entering the lab and consenting to the study, participants completed a battery of self-report questionnaires (i.e., demographic questionnaire, Routine Arousal Questionnaire, GDS, GAS, and PANAS). For the
beginning of the second session, participants renewed consent and responded again to the Routine Arousal Questionnaire and PANAS. The TEA was administered at the end of the second session.
Experimental protocol. In this within-subjects study, all participants completed the MSIT seven times over the course of two sessions. At the beginning of the first experimental session, they completed the MSIT in a baseline silence condition. Following this baseline, participants were randomly assigned to complete the MSIT under either (1) the three counterbalanced low-arousal musical conditions during the first session, then the three counterbalanced high-arousal musical conditions during the second session, or (2) the three counterbalanced high-arousal musical conditions during the first session, followed by the three counterbalanced low-arousal musical conditions during the second session. Figure 1 displays the experimental protocol for a participant who was randomly assigned to the latter ordering of conditions.
First experimental session. Each participant sat in an isolated room in the lab that is free of visual distractors such as windows and posters. First, the participant sat alone in
silence for three minutes. At this point, the experimenter entered the room and gave the participant the AD ACL. The participant was left alone again for two minutes to
complete the AD ACL, and then the experimenter returned to present the MSIT on a laptop. The participant read the on-screen instructions, and completed the practice session. The experimenter monitored participants’ performance during the practice session to ensure that the participant understood the requirement to press the key corresponding to the number of the target rather than the position of the target. If a participant pressed the key corresponding to the target’s position on multiple practice trials and did not identify these responses as errors, the experimenter reviewed the error and had the participant repeat the practice session once or twice more, as necessary. The number of practice sessions was recorded for each participant. The participant was left alone to complete the MSIT. After informing the experimenter of task completion, the participant responded to the post-task questionnaire.
This protocol was repeated, without the practice block, for the three musical conditions. For each musical condition, the representative piece was played from an iPhone over portable speakers. Participants were allowed to adjust the volume to a comfortable listening level. The participant sat alone listening to the piece for three minutes, and the piece continued to play on repeat while the participant complete the AD ACL, MSIT, and post-task questionnaire.
Second experimental session. Participants completed the AD ACL, MSIT, and post-task questionnaire under the remaining three musical conditions. The MSIT practice block was administered to refresh participants’ understanding of the task for the first musical condition of this session.
