Professional Documents
Culture Documents
Linnaeus Centre HEAD, Swedish Institute for Disability Research, 2Department of Behavioural Sciences and Learning, Linköping University,
1
School of Health and Medical Sciences and Örebro University, 4Audiological Research Centre, Örebro University Hospital, Örebro, Sweden
3
Abstract
The aim of the following study was to examine the relationship between working memory capacity (WMC), executive
functions (EFs) and perceived effort (PE) after completing a work-related task in quiet and in noise in employees with
aided hearing impairment (HI) and normal hearing. The study sample consisted of 20 hearing-impaired and 20 normally
hearing participants. Measures of hearing ability, WMC and EFs were tested prior to performing a work-related task
in quiet and in simulated traffic noise. PE of the work-related task was also measured. Analysis of variance was used
to analyze within- and between-group differences in cognitive skills, performance on the work-related task and PE.
The presence of noise yielded a significantly higher PE for both groups. However, no significant group differences were
observed in WMC, EFs, PE and performance in the work-related task. Interestingly, significant negative correlations
were only found between PE in the noise condition and the ability to update information for both groups. In summary,
noise generates a significantly higher PE and brings explicit processing capacity into play, irrespective of hearing. This
suggest that increased PE involves other factors such as type of task that is to be performed, performance in the cognitive
skill required solving the task at hand and whether noise is present. We therefore suggest that special consideration in
hearing care should be made to the individual’s prerequisites on these factors in the labor market.
Keywords: Adverse conditions, cognitive skills, hearing impairment, labor market, normal hearing, perceived effort
they were currently using. If they had HI they were instructed were asked to type in (in a dialogue box appearing on the
to answer questions regarding etiology, duration of hearing screen) the last item from each of the four target categories.
loss, HA experience, uni- or bi-lateral fitting and current The target categories remained on the screen during the
HA model. whole of each trial. Each trial required three updates for
one of the categories (for example, four exemplars of this
Cognitive tests category were presented) and two, one and no update for the
All tasks were administered through a computer and the remaining three categories, respectively. By the presentation
instructions were presented in written form and complemented of the seventh word in each trial, at least one exemplar of
with oral instructions by the test leader. Three of the cognitive each of the four target categories had been presented (i.e.,
tests, lexical decision-making, rhyme-judgment and reading during the presentation of words 8-15, the participants were
span, are part of a cognitive test battery. The test battery has required to hold four items active in working memory). Two
previously been described in detail in Rönnberg et al.[16] and practice trials were followed by six main trials. The outcome
the tests included in the test-battery are all well-established measure was the proportion of correctly recalled items.
in the literature of cognitive psychology.[17-21] A condensed
description follows below: The sustained attention to response test (SART): SART
was used as a measure of inhibition.[23] One digit at a time
Lexical decision-making: The lexical decision-making test was presented at the center of the computer screen. The
was used as a measure of lexical access. The task was to task was to press the space bar as fast as possible when a digit
decide whether a combination of three letters constituted was detected. However, if the digit was “3”, the participants
a real word or not. Six items were initially presented for were to withhold their response and await the next stimuli.
practice and after that 40 items in total were presented and Each digit remained on the screen until a response had been
used for scoring. Half of the items were real words and they given or until 1000 ms had passed. Trials were separated by
were all familiar Swedish three-letter words. an interstimulus interval of 500 ms. Of the 120 presented
digits, 21 were the digit “3”. The outcome measure was the
Rhyme-judgment: A rhyme-judgment test was used as number of failures to withhold a response.
a measure of phonologic ability. The task was to decide
whether two simultaneously presented words rhymed or not. The number-letter task: The number-letter task, adapted from
Words were presented lexically on the computer screen. Four Miyake et al.,[12] was used as a measure of shifting ability.
items were used for practice and after that 32 items were In this task, pairs consisting of one digit and one letter (for
presented and used for scoring of the measurement. The example: 7G) were presented in one of the four corners of the
outcome measures of the lexical decision-making and rhyme- computer screen. The pairs were presented one at a time and
judgment were the proportion of correct responses. in a clockwise manner. The task was to decide whether the
digit was odd or even when a pair was presented in the upper
Reading span: The reading span test was used as a measure of half of the screen and whether the letter was a lower case or
WMC. The test used in the current study was a short version upper case when a pair was presented in the lower half of the
of the reading span test created by Rönnberg et al.[15] based screen. The participants responded by pressing one of four
on the original test created by Daneman and Carpenter.[22] buttons marked with the words “odd”, “even”, “lower” and
The task was to read 24 sentences and to decide whether the “upper”. Stimuli remained on the screen until a response had
sentences were absurd or not and after reading sets of three, been given or until 10 s had passed. Twelve practice trials
four or five sentences, to recall either the first or the final were followed by 38 main trials. Scoring was based upon the
words of the sentences in correct serial order. The sets were difference in reaction time between two consecutive trials:
always presented in the ascending order and the participants 1. When the present trial was a shift trial and the preceding
were cued to recall the first/last words post-stimulus trial was a no-shift trial and
presentation. The outcome measure was the proportion of 2. when the present trial was a no-shift trial and the preceding
correctly recalled items. trial was a shift trial. The outcome variable reported to
as “shifting” was thus the mean difference between trials
The keep track task: The keep track task, adapted from 1 and 2.
Miyake et al.,[12] was used as a measure of updating of
information. Each trial began with the presentation of four Work-related task and rating of PE
target categories at the top of the computer screen. After this, An information extraction task was used as a work-related
15 (mono- or disyllabic) words including 1-4 exemplars from task. In this task, tables of 15 numbered items belonging to a
each of six possible categories (countries, colors, metals, semantic category (for example fruits) and both categorical
fruits, relatives and animals) were presented one at a time. (for example place of origin) and continuous (for example
Words were presented for 3000 ms with an interstimulus price) information about the items were presented on a
interval of 500 ms. At the end of each trial, the participants computer screen. In each trial, the participants were asked
to identify one target item based on a question presented larger project, several other work-related tasks were included
below the table. The participants responded by typing the in the project in a counterbalanced order meaning that the
number (1-15) corresponding to the target item in a box extraction of information task could be performed either in
appearing at the bottom of the screen. The tables remained the first or second session. The task was always performed in
on the screen until a response had been given, or until quiet or in noise in the anechoic chamber and the group with
60 s had passed. The level of difficulty was manipulated HI always wore their HAs during this task.
by using one constraint in the easy condition (“Which fruit
has the highest price per kilogram?”) and three constraints Statistical analysis
in the difficult condition (“Which Portuguese fruit of which The SPSS software was used for statistical analysis.
at least 14,000 kg has been sold has the highest price per A Kolmogorov-Smirnov test for normality was first
kilogram?”). There were 16 easy and 16 difficult trials in conducted to ensure a normal distribution and the results
total and proportions of correct responses were used as of the test showed that the variables were non-significant
outcome measures of performance. (P > 0.05). Initially, a one-way analysis of variance
(ANOVA) was performed on the results of all the cognitive
After the extraction of information task was completed, the tasks between the two groups (normally hearing vs. hearing
participants were immediately asked to rate how effortful impaired). The reason for this was to check whether both
they found the task to be using the Borg CR-10 scale on study groups went into the quiet and noise conditions on
the computer.[24] This scale is a combination of ratio and similar cognitive level when performing the work-related
category scaling where verbal expressions and numbers are task. Secondly, Pearson’s correlation coefficients between
used congruently on a scale ranging from 0 (none at all) to each of the cognitive scores and demographic variables
10 (extremely great). The scale was given on a sheet of paper were calculated for the total sample. Thirdly, to analyze
next to the computer to help the participants with ratings and for within-group and between-groups differences in
the participants had to type in the number corresponding to performance on the work-related task and PE, two-way
the degree of effort perceived in the work-related task. One ANOVA with repeated measures was performed. This
question was asked for each level of difficulty of the working was followed with Bonferroni-adjusted post-hoc pairwise
task. The Borg CR-10 scale was used as a measure of PE comparisons. Finally, Pearson’s correlation coefficients
because this measure has shown to be sensitive enough to were used to examine the possible relationship between
identify differences between different background conditions cognitive abilities and PE after completing the work-related
in a previous study where it has been used (c.f., Larsby et al.[8]). task in quiet and in noise. All tests were two tailed and
conducted at a 5% significance level.
The noise used in the experiment was recorded at a crossroads
in Örebro during morning traffic using a microphone Ethics
employing the Ambisonics surround sound technique.[25] Participants received vouchers for cinema tickets or flowers
From the original recording, 19.6 min of traffic noise was for taking part in the study. The study was approved by the
extracted, edited and reproduced in an anechoic chamber Regional Ethics Committee in Uppsala (Dnr: 2010/072).
(5.5 m × 5.5 m × 4.5 m). Six loudspeakers were used to
reproduce the noise and they were placed in the circle around
the participants (sitting in the middle of the room with a
computer performing the work-related task) with a spread of
60° between each speaker. To ensure a realistic noise level, the
sound was reproduced with an equivalent A-weighted sound
pressure level of 73 dBA, matching the level of the original
recording location. In order to avoid dynamic variation of the
noise level, quiet parts of the recording file were removed
so that a consistent traffic noise was played in every test
trial. The final output of the noise was also looped and hence
that it could be played continuously until the work-related
task was completed by each participant. Figure 1 shows the
experimental set up in the anechoic chamber.
Procedure
Figure 1: The experimental set-up. The participants were seated
Data collection occurred on two separate sessions. Sessions in the middle of an anechoic chamber with a computer and
were completed within 4 weeks of each other. At the first there were six loudspeakers surrounding them in a circle. The
appointment, the demographic questionnaire, auditory and loudspeakers reproduced the traffic noise at 73 dBA and had a
cognitive tests were administrated. As the data is part of a spread of 60° between them
Table 3: Pearson correlation coefficients between demographic variables and cognitive skills in the total group of participants (n = 40)
Demographic variable Education PTA R PTA L Lexical Rhyme- Reading SART The keep The number- Shifting
match judgment span errors track task letter task
Age −0.21 0.39* 0.42** 0.25 −0.04 −0.34* −0.06 −0.18 −0.13 0.10
Education −0.22 −0.31 −0.04 0.63** 0.37* 0.14 0.44** −0.08 0.14
PTA R 0.95** −0.16 −0.13 −0.11 0.10 −0.30 0.04 −0.06
PTA L −0.14 −0.28 −0.14 0.06 −0.41** −0.02 −0.03
Lexical decision-making −0.16 −0.16 −0.22 −0.10 0.11 0.11
Rhyme-judgment 0.19 0.22 0.50** −0.07 0.06
Reading span 0.22 0.48** −0.08 0.09
SART errors 0.14 −0.10 0.06
The keep track task −0.29 0.13
The number-letter task −0.31*
*Correlation is significant at the 0.05 level, **Correlation is significant at the 0.01 level. PTA = Pure tone average, SART = Sustained attention to response test
(SD = 1.5) for the difficult task. A two-way repeated correlated with the easy task (r = −0.55, P = 0.02) for the
measures ANOVA revealed a statistical significant main group with normal hearing and with the difficult task (r =
effect of condition (F(3, 114) = 65.4, P = 0.001, ηp2 = 0.6). −0.46, P = 0.04) for the group with HI [Figure 4]. That is, the
No significant main effect between groups (F(1, 38) = 1.3, updating skill of both groups was significantly correlated with
P = 0.27) and interaction effect between condition and group PE in noise, where lower performance in the updating task
(F(3, 114) = 1.4, P = 0.25) were observed, meaning that no generated a higher PE. Interestingly, a difference between
significant differences in PE were observed between groups groups emerged where the updating skill was only correlated
in all conditions. Post-hoc analysis revealed however, that in the easy task for people with normal hearing and in the
both groups reported a statistical significantly higher PE difficult task for people with HI. Nonetheless, it is important
(P < 0.001) when noise was present when compared with to point out that PE in noise was trending toward statistical
the quiet condition and this effect was observed for both the significant correlation with updating for the group with HI
easy and difficult task. These findings demonstrate that the (r = −0.44, P = 0.057) in the easy task as well.
presence of noise generated a significantly higher PE for
both groups, regardless of task difficulty. In order to rule out the possibility that the participants rated
their performance instead of PE, an additional correlation
Correlations between cognitive tasks, performance in the analysis was performed between the performance of the work-
work-related task and PE related task and PE for both groups. The analysis showed
A series of Pearson product-moment correlation coefficient no significant correlations between performance and rated
tests were used to analyze the relationship between PE in the group with HI in quiet (easy: r = −0.18, P = 0.46;
performance on each of the cognitive tasks and PE in quiet difficult: r = −0.18, P = 0.45) and noise (easy: r = 0.07, P =
and in noise after performing the work-related task for each 0.76; difficult: r = −0.25, P = 0.29). Likewise, no significant
group [Table 4]. In general, no significant correlations were correlations were observed in the group with normal hearing
found between the cognitive tasks and PE in quiet and in in the quiet (easy: r = −0.13, P = 0.60; difficult: r = −0.12,
noise. Only updating was found to correlate significantly P = 0.62) and noise (easy: r = −0.10, P = 0.68; difficult:
with PE in noise for both groups, with PE being negatively r = −0.20, P = 0.41) condition.
Figure 2: Mean performance in the work-related task in quiet Figure 3: Means scores of perceived effort in quiet and in traffic
and in traffic noise for each group (error bars denote the 95% noise for each group (error bars denote the 95% confidence
confidence interval) interval)
Table 4: Pearson correlation coefficients between cognitive skills and perceived effort in the different conditions for both groups
Cognitive task Normally hearing (n = 20) Hearing impaired (n = 20)
Quiet Noise Quiet Noise
Easy Difficult Easy Difficult Easy Difficult Easy Difficult
Lexical decision-making −0.34 0.01 −0.11 −0.20 0.37 0.12 0.05 −0.07
Rhyme-judgment −0.12 0.17 −0.27 −0.22 −0.33 −0.03 −0.36 −0.32
Reading span −0.05 −0.22 −0.10 0.14 0.00 0.27 −0.11 0.11
SART errors −0.39 −0.44 −0.34 −0.08 −0.06 0.11 −0.02 −0.20
The keep track task −0.28 −0.24 −0.55** −0.19 −0.28 −0.14 −0.44 −0.46*
The number-letter task 0.13 0.15 0.20 0.22 −0.14 −0.34 0.18 −0.10
Shifting −0.16 −0.07 −0.20 −0.27 0.06 0.07 −0.07 0.03
*Correlation is significant at the 0.05 level, **Correlation is significant at the 0.01 level. SART = Sustained attention to response test
Correlations between PE in noise and updating HAs may still distort incoming signals which may lead
Statistical significant negative correlations were found, but to an even greater distraction for people with HI as more
not in quiet, between updating ability and PE of the work- cognitive resources are required to process the input. The
related task in noise for both groups. This demonstrates that authors further propose that recruitment of the HI may also
with lower ability to update new information, a higher PE lead to the perception of sharp onsets of sound that capture
in noise is generated regardless of hearing ability. Previous attention away from focal tasks. Therefore, for participants
studies have shown that in challenging situations, such as in with aided HI there could still be an additional demand of
noise, people with HI have to rely more on cognitive skills.[29,30] resisting attention capture that causes disruption or less
Further, both objective and subjective measures indicate that efficient focal task processing when noise is present. If this
background noise forces hearing-impaired individuals to argument was true, disruption of task processing should still
exert more effort in different listening tasks.[6,8,9] However, have been observed between PE in the easy task in noise and
it is important to remember that the current study employed updating, as the HAs were worn in both levels of difficulty.
a non-auditory extraction of information task and it is not In the present study, a trending, but non-significant,
correlation (r = −0.44, P = 0.057) was indeed observed in
apparent whether the detrimental effects from background
our results between the PE in the easy task in noise and
noise also applies to visual tasks for people with HI. The
updating for the group with HI. This means that participants
current findings suggest that there is a correlation between
with HI could still be distracted by the noise in the easy task
PE and cognitive skills in noise in both employees with HI
due to HA and that explicit processing capacity could still
and with normal hearing, even when the task is visual.
be involved in noise for this group even when performing
an easy task. The current statistical significant correlations
Interestingly, the current results showed that updating skill
should however be interpreted with caution as the strength
was only negatively correlated with the PE in the easy task
of the correlation was relatively small when judged against
for people with normal hearing. By itself, one might think
the standard criteria proposed by Cohen.[34] Future studies
that a more difficult work-related task would generate a
replicating the present findings would therefore be of
higher PE for the group with normal hearing. This might
interest to conduct.
not be the case according to Lavie’s[31] perceptual load
theory. This theory postulates that distractor perception
In addition, despite the multiple correlations, a Bonferroni
can be prevented when processing of task-relevant stimuli correction was not applied as it was felt that the increase in
involves high perceptual load. In other words, when high the risk of a Type 2 error occurring outweighed the potential
perceptual load engages full capacity in relevant processing benefits of reducing the risk of a Type 1 error. Exact P values
this would leave no spare capacity for perception of the are provided in order to enable the reader to evaluate whether
noise. Conversely, in situations of low perceptual load, applying a correction for multiple comparisons would have
any capacity not taken up in perception of task-relevant affected the outcome of the analyzes. We argue that the
stimuli would involuntarily “spill over” to the perception statistical significant correlations between PE in noise and
of task-relevant distractors. This means that the easy task updating skill are valid because: The task is highly based
left more room for processing of noise and based upon the upon updating skills and no other statistical significant
assumption that explicit processing capacity are brought correlations emerged with the different cognitive tasks and
into play in noise, updating skills were therefore involved the involvement of updating was only observed in the noise
in the easy task instead of the difficult task that yielded a conditions for both groups. In other words, regardless of
higher perceptual load. These predictions have previously hearing ability, the employees at work have to actively think
been tested in a series of experiments that assessed the effect about the task and devote cognitive resources to it during the
of varying perceptual load in task-relevant processing.[32,33] presence of noise, even if the task is non-auditory, which
If this theory were to apply to employees with aided HI, involves executive processes. The presence of noise does
a significant correlation should have emerged between PE not affect performance, further confirmed by the additional
in the easy task in noise and updating and not between the analysis where no significant correlations were observed
difficult task and updating observed in the current results. between PE and performance of the task, but it affected both
There might be some explanations as to why there was a groups differently due to the aided HI. More specifically, it
group difference in this finding and why the perceptual load affects how and when explicit processing capacity is engaged
theory does not apply on employees with aided HI. From to solve the task at hand and that a decreased performance
an audiological perspective, a HI may be advantageous relying on that specific process may lead to a greater PE
when performing visual tasks in noise, as the noise is not as for the individual in adverse conditions. Furthermore, our
intelligible for the group with HI. Moreover, all participants results are also in agreement with those Jahncke and Halin[7]
in our study were frequent HA users and they always wore reported regarding visual tasks for people with HI in noise.
their HAs when performing the work-related task, which An effect of noise was observed in their study where high
could have amplified the noise to a normal hearing level. noise affected the hearing-impaired participants’ recall
However, Jahncke and Halin[7] have proposed that the of semantic information and subjective effort negatively,
21. Shoben E. Semantic and lexical decisions. In: Puff CR, editor. Handbook assessments at 7 and 14 years of age. Appl Cogn Psychol 2004;18:1-16.
of Research Methods in Human Memory and Cognition. New York: 29. Pichora-Fuller MK. Perceptual effort and apparent cognitive decline:
Academic Press; 1982. Implications for audiologic rehabilitation. Semin Hear 2006;27:289-93.
22. Daneman M, Carpenter PA. Individual differences in working memory 30. Pichora-Fuller MK, Souza PE. Effects of aging on auditory processing
and reading. J Verbal Learning Verbal Behav 1980;19:450-66. of speech. Int J Audiol 2003;42 Suppl 2:2S11-6.
23. Manly T, Robertson IH, Galloway M, Hawkins K. The absent 31. Lavie N. Distracted and confused? Selective attention under load.
mind: Further investigations of sustained attention to response. Trends Cogn Sci 2005;9:75-82.
Neuropsychologia 1999;37:661-70. 32. Lavie N. Perceptual load as a necessary condition for selective attention.
24. Borg G. Psychophysical scaling with applications in physical work and J Exp Psychol Hum Percept Perform 1995;21:451-68.
the perception of exertion. Scand J Work Environ Health 1990;16 Suppl 33. Forster S, Lavie N. Harnessing the wandering mind: The role of
1:55-8. perceptual load. Cognition 2009;111:345-55.
25. Malham DG, Myatt A. 3-D sound spatialization using ambisonic 34. Cohen J. Statistical Power Analysis for the Behavioral Sciences. New
techniques. Comput Music J 1995;19:58-70. York: Academic Press; 1978.
26. Fisk JE, Sharp CA. Age-related impairment in executive functioning:
Updating, inhibition, shifting, and access. J Clin Exp Neuropsychol How to cite this article: Hua H, Emilsson M, Ellis R, Widén S, Möller C,
2004;26:874-90. Lyxell B. Cognitive skills and the effect of noise on perceived effort in
27. Salthouse TA. Aging and measures of processing speed. Biol Psychol employees with aided hearing impairment and normal hearing. Noise
2000;54:35-54. Health 2014;16:79-88.
28. Gathercole SE, Pickering SJ, Knight C, Stegmann Z. Working memory
skills and educational attainment: Evidence from national curriculum Source of Support: Nil, Conflict of Interest: None declared.