You are on page 1of 11

J Am Acad Audiol 25:859-868 (2014)

The Impact of Different Background Noises: Effects


on Cognitive Performance and Perceived Disturbance
in Employees with Aided Hearing Impairment
and Normal Hearing
DOI: 10.3766/jaaa.25.9.8

Hakan Hua*t
Magnus Emilsson*f
Kim Kahari:j:
Stephen Widen*§
Claes Moller*§**
Bjorn Lyxell*f

Abstract

Background: Health care professionals frequently meet employees with hearing impairment (HI) who
experience difficulties at work. There are indications that the majority of these difficulties might be related
to the presence of background noise. Moreover, research has also shown that high-level noise has a
more detrimental effect on cognitive performance and self-rated disturbance in individuals with HI than
low-level noise.

Purpose: The purpose of this study was to examine the impact of different types of background noise on
cognitive performance and perceived disturbance (PD) in employees with aided HI and normal hearing.

Research Design: A mixed factorial design was conducted to examine the effect of noise in four exper­
imental conditions.

Study Sample: A total of 40 participants (21 men and 19 women) were recruited to take part in the study.
The study sample consisted of employees with HI (n = 20) and normal hearing (n = 20). The group with
Hi had a mild-moderate sensorineural HI, and they were all frequent hearing-aid users.

Intervention: The current study was conducted by using four general work-related tasks (mental arith­
metic, orthographic decoding, phonological decoding, and serial recall) in four different background con­
ditions: (1) quiet, (2) office noise at 56 dBA, (3) daycare noise at 73.5 dBA, and (4) traffic noise at 72.5
dBA. Reaction time and the proportion of correct answers in the working tasks were used as outcome
measures of cognitive performance. The Borg CR-10 scale was used to assess PD.

Data Collection and Analysis: Data collection occurred on two separate sessions, completed within
4 wk of each other. All tasks and experimental conditions were used in a counterbalanced order.
Two-way analysis of variance with repeated measures was performed to analyze the results. To examine
interaction effects, pairwise f-tests were used. Pearson correlation coefficients between reaction time
and proportion of correct answers, and cognitive performance and PD were also calculated to examine
the possible correlation between the different variables.

Results: No significant between-group or within-group differences in cognitive performance were


observed across the four background conditions. Ratings of PD showed that both groups rated PD
according to noise level, where higher noise level generated a higher PD. The present findings also dem­
onstrated that the group with HI was more disturbed by higher than lower levels of noise (i.e., traffic and
daycare setting compared with office setting). This pattern was observed consistently throughout four

"Linnaeus Centre HEAD, Swedish Institute for Disability Research, Linkoping University, Linkoping, Sweden; d epartm ent of Behavioural Sciences
and Learning, Linkoping University, Linkoping, Sweden; d epartm ent of Audiology, Sahlgrenska Academy, Gothenburg University, Gothenburg,
Sweden; §School of Health and Medical Sciences, Orebro University, Orebro, Sweden; **Audiological Research Centre, Orebro University Hospital,
Orebro, Sweden
Hakan Hua, Linkoping University, Department of Behavioral Sciences, SE-581 83 LINKOPING, Sweden; E-mail: huahakan@aol.com

S59
Journal o f the American Academy o f Audiology/Volume 25, Number 9, 2014

working tasks where the group with HI reported a significantly greater PD in the daycare and traffic set­
tings compared with office noise.

Conclusions: The present results demonstrate that background noise does not impair cognitive per­
formance in nonauditory tasks in employees with HI and normal hearing, but that PD is affected to a
greater extent in employees with HI during higher levels of background noise exposure. In addition, this
study also supports previous studies regarding the detrimental effects that high-level noise has on
employees with HI. Therefore, we emphasize the need of both self-rated and cognitive measurements
in hearing care and occupational health services for both employees with normal hearing and HI.

Key Words: Cognitive performance, labor market, hearing impairment, noise, perceived disturbance, work

Abbreviations: ANOVA = analysis of variance; HA = hearing aid; HI = hearing impairment; PD =


perceived disturbance; PTA = pure-tone average; RT = reaction time

IN T R O D U C T I O N effects on cognitive performance (Knez and Hygge,


2002; Schlittmeier et al, 2008; Ebissou et al, 2013).
earing ability is an im portant factor at the In a study by Schlittmeier et al (2008), three experi­

H workplace for understanding speech and iden­


tifying environmental sounds. Health care
professionals frequently meet employees with hearing
impairment (HI) who experience difficulties at work.
ments tested the impact of low background speech
(35 dBA) of both good and poor intelligibility, in compar­
ison to quiet and highly intelligible speech not lowered
in level (55 dBA). Mental arithmetic, verbal short-term
Researchers have reported disadvantages in both physi­ memory, and verbal-logical reasoning were used as
cal and psychosocial work conditions, increased sick leave measurements of cognitive performance, and self-rating
due to distress, and worse health among employees with scales of PD were also reported by the participants after
HI compared with their normal-hearing peers (Ringdahl completing each task in each experimental condition.
and Grimby, 2000; Danermark and Gellerstedt, 2004; The results showed that the performance was signifi­
Kramer et al, 2006). There are indications th at the cantly impaired by highly intelligible competing speech
majority of the negative effects of noise might be related (55 dBA) on the short-term memory and mental arith­
to the presence of background noise at work (Danermark metic tasks, but not in verbal-logical reasoning. However,
and Gellerstedt, 2004; Hua et al, 2013). Previous re­ ratings of PD were consistent over all three experiments
search has mainly focused on field studies with self- where quiet was rated least disturbing, and low back­
reported ratings of different working conditions and ground speech with poor speech intelligibility was rated
not so much on laboratory studies with controlled con­ less disturbing than good and highly intelligible speech
ditions and behavioral measures of cognitive perfor­ in 35 and 55 dBA, respectively.
mance. In the present study, we examine the impact Although such self-rated assessments and cognitive
of different types of background noise on cognitive per­ measures clearly speak in favor of reducing high-level
formance and perceived disturbance (PD) in employees noise in the workplace for employees with normal hear­
with aided HI and normal hearing. ing, very little empirical evidence exists regarding cor­
Noise annoyance is a fundamental concept in the area responding performance effects for employees with HI
of environmental effects of sound. It is perceived as a in the labor market. Moreover, although the negative
negative aspect of environmental conditions, but its effects of background noise on speech understanding
meanings are quite broad and diverse. Noise annoyance of persons with HI are well documented (Van Tasell,
is a multifaceted concept covering the effects of noise on 1993; Humes, 2002), most studies concerning individu­
behavior, such as PD or interference with intended activi­ als with HI focus mainly on processing of listening tasks
ties (Guski et al, 1999). Guski (1997) further states that and listening effort in noise (Larsby et al, 2005; Zekveld
the term annoyance can be used interchangeably with et al, 2011; Jahncke and Halin, 2012; Rudner et al,
subjective experiences, such as disturbing or irritating. 2012). In a study by Larsby et al (2005), different speech
Intelligible competing speech is usually considered a neg­ noises were found to interfere with cognitive processing
ative feature of the work environment, and the degree in individuals with HI and normal hearing. Four study
of intelligibility of competing speech has been found groups were included in their study, and the sample
to correlate positively with self-rating scales on PD. consisted of both young/elderly participants with nor­
As Venetjoki et al (2006) have demonstrated in a group mal sensitivity for pure tones and young/elderly partic­
of normal-hearing participants, the more intelligible ipants with HI. Cognitive tests of speech understanding
the competing speech is, the higher self-rating scales and self-rated effort were administered, respectively, in
of PD. Also, research on normal-hearing participants four background conditions: quiet and three types of
has demonstrated th at competing speech has negative speech or speech-like noise. The results demonstrated

860
E ffects of N oise on C ognitive P erfo rm a n ce a n d PD/Hua et al

th at the presence of noise compared with quiet had a viduals in the present study were participants in a
negative effect on accuracy, speed of performance in larger research project about individuals with HI in the
the speech-processing tasks, and self-rated effort for labor market, and were the same individuals as reported
the group with HI compared with the group with nor­ by Hua et al (2013). The hearing-impaired participants
mal hearing. were recruited from the Department of Audiology, Orebro
As stated earlier, few studies have examined the im­ University Hospital, Sweden. They were selected based
pact of different background noises on cognitive per­ on the following criteria: 18-65 yr old, mild-moderate
formance and PD among employees with HI, and it is binaural sensorineural hearing loss, history of aural
not clear whether the negative effects of background rehabilitation, and experienced hearing-aid (HA) users
noise also apply to nonauditory tasks for individuals (>3 mo). Exclusion criteria were retirement, long-term
with HI. Jahncke and Halin (2012) have previously sick leave, m oderate-severe tin n itu s, hyperacusis,
investigated whether open-plan office noise affects non­ psychiatric illnesses, dyslexia, and/or other significant
auditory performance and stress, and whether these diseases/disabilities. The two study groups were matched
effects differ between individuals with HI and normal according to age and educational level, and did not dif­
hearing. Twenty hearing-im paired and 18 norm al­ fer significantly (p > 0.05). The study was approved
hearing participants took part in their study, and cog­ by the Regional Ethics Committee in Uppsala (Dnr:
nitive performance, reported fatigue, and physiological 2010/072), and informed consent was obtained from
stress were measured during work in a simulated open- all participants.
plan office consisting of low and high noise levels at
30 and 60 dBA, respectively. The results showed that A udiological Tests
there was indeed a decreased performance in the group
with HI for nonauditory tasks th a t involved recall of Air- and bone-conduction thresholds were measured
semantic information in high-level office noise. Partici­ using standard audiological procedures in a sound-treated
pants with HI also reported higher levels of fatigue booth, and pure-tone averages (PTAs) were calculated
and tended to have higher stress hormone levels during for 0.5,1, 2, and 4 kHz for each ear. As part of the audio-
high noise exposure compared with the low noise level. logical assessment, speech recognition in noise was
Taken together, the presence of high-level noise could obtained for each participant and ear using standardized
have a detrimental effect on cognitive performance, and recordings of the Swedish phonemically balanced word
sparse literature exists on work performance involving lists (Magnusson, 1995). Each list was composed of 50
nonauditory tasks and rated disturbance among em­ monosyllabic words for recognition scoring and was pre­
ployees with HI. Furthermore, few studies have com­ sented with a carrier phrase produced by a male speaker
pared the impact of different work-related settings with a Swedish native accent. The participants were
consisting of both low- and high-level noise in this study instructed to repeat the word that followed the carrier
population. Therefore, the aim of this study was to com­ phrase, and performance was scored as a proportion of
pare the cognitive performance and PD between three correctly identified words. Table 1 gives the demographic
types of background noise varying in temporal and information of the participants.
spectral characteristics: simulating an office, daycare,
and traffic setting in comparison to a quiet condition, Working Tasks
in employees with aided HI and normal hearing.
A rith m etic Task
METHODS
The arithmetic task, in which participants were
Participants asked to solve addition problems, was a modified version
of a task used by Duverne and Lemaire (2004). On each
To estimate the required study sample for the current trial, one number and two to-be-added numbers (e.g., 16
study, The G*Power 3.1.7 power analysis package was and 9 + 5) were presented adjacently to each other on the
used. Data for estimation were based on our previous computer screen. The task was to indicate, by means of a
study using a general working task described by Hua key press, which one was the larger of the two (i.e., the
et al (2013). Assuming that the results will show an single number or the sum of the two to-be-added num ­
effect size of 0.9 (a = 0.05) between the two groups in bers). The addition problems remained on the screen
noise, it was estimated th at a minimum of 16 partici­ until a response had been given, or until 30 sec had
pants in each group were needed to reach statistical sig­ passed. Level of difficulty was manipulated by using
nificance with a power greater than 0.8. A total of single-digit operands in the easy condition (e.g., 9 + 5)
40 participants (21 men and 19 women) were recruited and three-digit operands in the difficult condition (e.g.,
to take part in the study and consisted of employees 588 + 192). To encourage exhaustive calculations in the
with HI (n = 20) and normal hearing (n = 20). The indi­ easy task, the difference between the sum of the operands

SB1
Journal of the American Academy of Audiology/Volume 25, Number 9, 2014

Table 1. Demographic Data for the Two Study Groups S eria l R ecall
Group Men/Women Mean (SD)
Normal Hearing (n = 20) 11/9
In the serial recall task, eight digits were presented
Age (yr) 40.5 (14.0) visually on the computer screen in each trial. The digits
PTA right ear 5.2 (6.3) were presented one at a time for 350 msec and were sep­
PTA left ear 7.0 (7.2) arated by an interstim ulus interval of 400 msec. After
Speech recognition in 84.1 (6.5) the presentation of the final digit, participants were
noise right ear (%) asked to type in (in a dialog box appearing on the
Speech recognition in noise 80.7 (4.1) screen) each of the eight digits in the order of presenta­
left ear (%) tion. Participants were instructed to always write down
Hearing Impaired (n = 20) 10/10
eight digits recalled in their correct position. In each
Age (yr) 48.0 (12.0)
condition, there were 10 trials. The outcome measure
PTA right ear 36.5 (6.0)
was the proportion of correct responses.
PTA left ear 36.0 (7.0)
Speech recognition in noise 62.3 (17.0)
right ear (%)
R ating of PD
Speech recognition in noise 61.2 (16.9)
left ear (%) After each working task was completed in each of the
four background conditions, the participants were imme­
diately instructed to rate how disturbing they found the
and the single number was never larger than ±2 (e.g.,
noise to be, using the Borg CR-10 scale (Borg, 1990). This
16 and 9 + 5), and in the difficult task, the single num ­
scale is a combination of ratio and category scaling where
ber was always approximately 1% (or 5-10 in absolute
numbers) from the sum of the operands. In all four con­ verbal expressions and numbers are used congruently
ditions, the same 16 easy and 16 difficult arithmetic on a scale ranging from 0 (no PD at all) to 10 (extremely
problems were used. To minimize recognition of prob­ great). An “absolute maximum” is located outside the
number scale to avoid a ceiling effect. The scale was
lems across conditions, four unique randomized orders,
given on a sheet of paper next to the computer to help
with easy and difficult problems interspersed, were
used. the participants with the ratings, and they were ins­
tructed to first choose the verbal expression th at best
O rth ographic D ecoding described their experience and then type in the number
corresponding to the degree of disturbance perceived for
In the orthographic decoding task, participants were each of the four conditions. One question was asked for
presented with two strings of letters at a time. The task each level of difficulty of the arithmetic task. The Borg
was to indicate, by means of a key press, which string CR-10 scale was chosen as a measurement of disturb­
constituted a correctly spelled Swedish word. The letter ance because it has been shown to be sensitive enough
strings remained on the screen until a response had to identify differences between different background
been given, or until 10 sec had passed. Four lists with conditions in previous studies in individuals with HI
different stimuli were used in the four conditions. Each and normal hearing (c.f. Hallgren et al, 2005; Larsby
list consisted of 24-25 trials. et al, 2005). Although this scale was originally devel­
oped to measure perceived effort, it is also a general
P h on ological D ecoding and common tool for measuring intensity of experiences
(Borg, 2008). This includes PD of different types of
In the phonological decoding task, participants were noise, and the scale has previously been used in several
presented with two strings of letters (real words versus investigations for this specific purpose (Ljungberg et al,
nonwords) at a time. The task was to indicate, by means 2004; Ljungberg and Neely, 2007).
of a key press, which string sounded like a proper Swedish
word. The letter strings remained on the computer Background Conditions
screen until a response had been given, or until 10 sec
had passed. Four lists with different stimuli were used All of the working tasks and measures of PD were
in the four experimental conditions. Each list consisted performed in the following four background conditions,
of 19 trials. each simulating a typical workplace setting with differ­
The outcome measures in the arithmetic, ortho­ ent sound level:
graphic decoding, and phonological decoding task was
mean reaction time (RT) (based on correct and incorrect 1. Quiet setting. No noise was simulated in this condition.
responses as well as omissions) and the proportion of 2. Office setting. The office noise used in the experiment
correct responses. Trials in all three tasks were separa­ was recorded in an office area, or more specifically in
ted by an interstim ulus interval of 500 msec. the reception/call center room, at Orebro University.

S6S
Effects of N oise on C ognitive Perform ance and PD/Hua et al

This space consists of three booths with two desks in group with HI always wore their HAs during all work­
each booth. This condition included a recording of ing tasks.
intelligible speech of normal level of the receptionist
answering phone calls. Statistical A nalysis
3. Daycare setting. The daycare noise was recorded in a
daycare environment in Orebro. Recording took The SPSS software (version 19) was used for statistical
place in the children’s playroom before the morning analysis. To control for normal distribution, raw scores of
assembly where the children played freely. The num­ the outcome measures and PD were plotted. This showed
ber of children in the playroom varied between 8 and that RT and PD were normally distributed, and the pro­
12 during the recording. In the middle of the room, portion of correct responses showed a positive skew in
there was an open-space area covered by a large car­ performance. Because the distribution of the partici­
pet. In one corner was a life-size toy castle, and in pants’ responses showed a positive skew, data were
another corner was playground decorated like a log transformed to obtain a more homogeneous error var­
small kitchen and, next to it, a couple of tables iance. To analyze for significant differences, a general lin­
and a few chairs. This condition introduced intelli­ ear model mixed two-way analysis of variance (ANOVA)
gible speech at normal to high levels and consisted was performed with four within-subject factors (PD in four
of the children talking, screaming, and shouting environmental conditions) and one between-subject factor
every now and then. (group). To examine interaction effects, pairwise /-tests
4. Traffic setting. The traffic noise used in the experi­ were used. Pearson correlation coefficients between RT
ment was recorded at a crossroads in Orebro during and accuracy of cognitive measures, and cognitive per­
morning traffic. The traffic noise consisted of sounds formance and self-rated disturbance were also calculated
from cars and trucks driving past the recording loca­ to examine the possible correlation between the different
tion at a normal to high level. No intelligible speech variables. All tests were conducted at a 5% significance
was introduced in this condition. level.

The three types of background noise were recorded RESULTS


by a microphone using Ambisonics surround sound
technique. From the original recordings, approximately C ognitive Perform ance
15-19 min of each noise was extracted, edited, and
reproduced in an anechoic chamber (5.5 X 5.5 X 4.5 As evident from Table 2, both groups obtained high
m). Six loudspeakers spaced 60° in a circle were used levels of performance and responded rapidly in the dif­
to reproduce the noise. The participants were seated ferent working tasks across all background conditions.
in the center of the loudspeaker array with a computer No large performance differences were observed among
performing the working tasks and rating of PD. To the four background conditions. As expected, the diffi­
ensure a realistic noise levels, the sound was repro­ cult arithmetic task also generated a slightly lower per­
duced with an equivalent A-weighted sound pressure formance level compared with the easy task for both
level of 56.6 dBA, 73.5 dBA, and 72.5 dBA for the office, groups (80-89% versus 92-97%), and serial recall showed
daycare, and traffic noise, respectively, matching the a performance score at approximately 60-70%. A mixed
level of the original recording locations. The final output two-way ANOVA conducted on the mean scores obtained
of the noise was also looped so th at it played continu­ for each working task showed no significant main effect
ously until the working tasks were completed by each of condition or group, and no significant interaction
participant. between condition and group (p > 0.05) (Table 3).

Procedure PD

Data collection occurred on two separate sessions. The average PD rating obtained for each task in the
Sessions were completed within 4 wk of each other, four conditions is summarized in Table 4. In the follow­
and each session lasted roughly 1.5-2.5 hr. Because ing results, only significant comparisons are discussed.
the data are based on a larger project, several other Generally, it seems as if the participants rated PD
tasks and tests were included in the project in a coun­ according to noise level across all tasks, where the quiet
terbalanced order, meaning th at the four work-related and office conditions were rated relatively low (0.3-3.8
tasks and the four experimental conditions could be per­ points), and the daycare and traffic noise were rated higher
formed either in the first and/or second session. How­ (2.0-5.5 points). A mixed two-way ANOVA was conducted
ever, the audiological tests were always performed in on the mean scores of each task, and this showed a signifi­
the first session before the participants performed the cant main effect of condition (p = 0.001), but no significant
required working tasks in the anechoic chamber. The main effect between groups (p > 0.05) (Table 5). An

863
J o u rn a l o f th e A m erica n A cad em y o f A u d iology /Volume 25, Number 9, 2014

Table 2. Mean Performance and SD in the Four Background Conditions for the Hearing-Impaired and Normal-Hearing
Groups
Quiet Office Daycare Traffic
Working Task Group Mean (SD) Mean (SD) Mean (SD) Mean (SD)
Arithm etic task perform ance - easy (%) NH 97 (4) 97 (5) 96 (5) 95 (11)
HI 93 (22) 94 (22) 94 (22) 92 (22)
Arithm etic task perform ance - difficult (%) NH 88 (10) 89 (10) 88 (7) 86 (15)
HI 82 (25) 80 (21) 81 (23) 80 (22)
O rthographic de co ding perform ance (%) NH 99 (2) 98 (2) 98 (3) 98 (2)
HI 98 (3) 9 9 (1 ) 97 (4) 99 (2)
Phonological decoding perform ance (%) NH 97 (5) 98 (4) 95 (5) 97 (4)
HI 94 (6) 93 (7) 94 (8) 92 (8)
Serial recall (%) NH 66 (14) 66 (14) 61 (18) 67 (15)
HI 62 (19) 58 (22) 61 (18) 60 (20)
Arithm etic task RT -E a sy (msec) NH 2705 (736) 2605 (729) 2597 (757) 2 5 8 2(9 2 1 )
HI 2553 (1089) 2695 (1224) 2631 (1282) 2 5 6 6(9 0 9 )
Arithm etic task RT -D iffic u lt (msec) NH 9209 (3575) 8737 (3916) 8 8 4 5(4 2 2 7 ) 8767 (3035)
HI 8341 (2859) 8997 (3735) 8871 (3291) 8 6 5 8(3 1 7 1 )
O rthographic decodingR T (msec) NH 1040 (244) 1044 (218) 1083 (228) 1028 (244)
HI 1185 (336) 1215 (297) 1177 (216) 1203 (327)
Phonological decodingR T (msec) NH 2031 (495) 2338 (553) 2401 (589) 2 2 6 5(5 0 2 )
HI 2756 (761) 2751 (889) 2784 (1054) 2703 (645)

interaction effect between condition and hearing was also ceived the daycare noise to be more disturbing than the
noted in all working tasks (p < 0.05), except in phonolog­ office noise in the serial recall task (t = 2.5, p = 0.003).
ical decoding (p = 0.06). Post hoc analysis revealed that
both groups reported th a t the three noise conditions C o rrelation s b e tw e e n C o g n itiv e P erform an ce
were significantly more disturbing than the quiet condi­ an d PD
tion (t = 5.2, p < 0.05) in all tasks, and that the group with
HI perceived the daycare and traffic noise to be more dis­ A series of Pearson product-moment correlation coeffi­
turbing than the office noise (t, = 5.3, p < 0.04). Further­ cient tests were used in order to examine whether the dif­
more, between-group differences were observed where the ferent outcome measures correlated with each other. As
hearing-impaired group reported a significantly higher there was no significant between-group difference in cog­
PD (t = 1.3—3.6, p = 0.001) in the daycare and traffic con­ nitive performance, but in self-rated PD, correlation ana­
dition compared with the normal-hearing group in the lyses were calculated separately for each group. However,
arithmetic tasks, orthographic decoding, and serial recall no significant correlations were observed between RT and
task. Although the group with normal hearing rated all the proportion of correct answers, and cognitive perfor­
three noise conditions quite similar consistently through­ mance and PD in the different conditions and work-related
out all tasks, post hoc analysis revealed that they per­ tasks (p > 0.05).
Table 3. Analysis of Cognitive Performance for All Working Tasks
Main effect Interaction effect between
F of condition F condition & hearing F Main effect of hearing
W orking task (df = 3,114) p-value ( t i p 2) (df = 3,114) p-value ( t i p 2 ) (df = 1,38) p-value (rip 2)
Arithm etic task 1.1 .34 (.03) 0.3 .80 (.01) 2.4 .13 (.06)
perform ance - easy
Arithm etic task 0.5 .67 (.01) 0.4 .73 (.01) 0.9 .33 (.03)
perform ance - difficult
O rtographic decoding 2.3 .08 (.06) 0.8 .51 (.02) 0.1 .93 (.00)
perform ance
Phonological decoding 0.5 .71 (.01) 0.8 .52 (.02) 1.9 .19 (.01)
perform ance
Serial recall 0.8 .50 (.02) 1.7 .17 (.01) 1.8 .18 (0 3 )
Arithm etic task RT - easy 0.3 .86 (.01) 0.8 .51 (.02) 0.4 .52 (.01)
Arithm etic task RT - difficult 0.3 .99 (.00) 0.6 .62 (.02) 0.1 .71 (.00)
O rtographic de co ding RT 1.9 .90 (.01) 0.8 .51 (.02) 3.6 .07 (.08)
Phonological de co ding RT 1.9 .14 (.05) 1.6 .19 (.04) 2.1 .17 (.04)

864
Effects of N oise on C ognitive Performance and PD/Hua et al

Table 4. Mean Score of PD and SD in the Four Background Conditions for the Hearing-Impaired and Normal-Hearing
Groups
Q u ie t O ffice D a yca re Traffic
W o rkin g Task G ro u p M ean (SD) M ean (SD) M ean (SD) M ean (SD )

A rith m e tic ta s k - e a s y NH 0 .7 (1.3) 2 .8 (1.7) 2 .9 ( 1 5 ) 3 .0 (1.9)


HI 0 .3 (0.6) 2.5 (1.8) 4.2 (4.8) 4.1 (2.0)
A rith m e tic ta s k - d iffic u lt NH 1.0 (1.9) 3 .8 (1.7) 3.8 (1.8) 3 .9 (2.1)
HI 0 .4 (0.8) 3 .2 (1.9) 4 .8 (2.3) 5 .0 (2.1)
O rth o g ra p h ic d e c o d in g NH 0 .3 (0.8) 1 7 (1 .1 ) 2 .0 (1.1) 2.0 (1.4)
HI 0 .2 (0.5) 1 .7 (1 9 ) 2 .8 (2.5) 3 .4 (2.7)
P h o n o lo g ic a l d e c o d in g NH 0 .5 (1.0) 2 .3 (1.0) 2 .8 (1.4) 2.9 (1.7)
HI 0 .3 (0.7) 2 .2 (2.1) 3.7 (2.5) 4 .0 (2 .5 )
S erial recall NH 0 .6 (1.0) 3 .8 (1.5) 4 .6 (2.3) 3 .7 (2.0)
HI 0 .3 (0.8) 3 .6 (2.0) 5 .0 (2.6) 5.5 (2.7)

N o te : HI = h e a rin g im p a ire d ; NH = norm a l he a rin g .

D IS C U S S IO N normal hearing, who rated all three noise conditions


equally similar in each working task, with the exception
n the present study, we examined the impact of dif­ of reported PD between the daycare and office settings
I ferent types of background noise on cognitive per­
formance and PD among some employees with aided
in the serial recall task. Lastly, correlation analyses
were performed separately for each group between RT
HI and some with normal hearing sensitivity. Summa­ and accuracy of the cognitive measures, and between
rizing the findings, the current results demonstrate cognitive performance and ratings of PD. However, no
th at both groups scored relatively high in accuracy significant correlations were observed between these dif­
and fast in RT in the different working tasks across ferent outcome measures.
all conditions. No significant between- or within-group The nonsignificant differences in cognitive perform­
differences in cognitive performance were observed across ance of the working tasks displayed an expected pattern
the four background conditions. Ratings of PD, on the because there was no indication as to why the hearing-
other hand, showed th a t both groups seemed to rate impaired group would perform significantly more poorly
PD according to noise level, where higher noise level than the group with normal hearing (c.f. Hallgren et al,
generated a higher PD. Generally, the lowest PD was 2001). Firstly, both groups were relatively young and
observed in the quiet condition, and both groups rated healthy with the main difference between the groups
PD of the three noise conditions significantly higher being mild-moderate aided HI. Secondly, the two
than the quiet condition. Interestingly, the present find­ groups were initially well matched in age and educa­
ings also indicate th at the group with HI was more dis­ tional level, and all participants were controlled for
turbed by higher than lower levels of noise (i.e., traffic physical/psychological diseases/disabilities before incep­
and daycare setting compared with the office setting). tion of the study. Previous studies have shown, however,
This pattern was observed consistently throughout four that performances are usually poorer in the group with
working tasks where the group with HI reported a sig­ HI when background noise is present (Hygge et al, 1992;
nificantly higher PD in the daycare and traffic settings Larsby et al, 2005). The difference between the current
compared with the office noise. In addition, this effect of study and previous studies conducted on individuals
the high-level noise was not observed in the group with with HI is that nonauditory working tasks were used.

Table 5. Analysis of PD for All Working Tasks


M ain e ffe c t In te ra ctio n e ffe c t b e tw e e n
F o f c o n d itio n F c o n d itio n & h e a rin g F M ain e ffe c t o f h e a rin g
W o rkin g ta s k (d f = 3 ,1 1 4 ) p -v a lu e (n p 2) (d f = 3 ,1 1 4 ) p -v a lu e (iqp2) (d f = 1,38) p -v a lu e ( t ip 2)

A rith m e tic 3 7 .4 .001 (.50) 4.0 .005 (.09) 0.2 .6 (.03)


ta s k - easy
A rith m e tic 3 8.5 .001 (.55) 3.8 .01 (.09) 0.3 .6 (.01)
ta s k - d iffic u lt
O rto g ra p h ic 3 0 .2 .001 (.4 4 ) 3.2 .02 (.08) 1.3 .3 (.03)
d e c o d in g
P h o n o lo g ica l 3 5 .2 .001 (.48) 2.5 .06 (.06) 0.9 .3 (.01)
d e c o d in g
S erial recall 70.2 .001 (.65) 4.2 .007 (.10) 0.6 .4 (.02)

865
Journal of the American Academy o f Audiology/Volume 25, Number 9, 2014

Our current findings clearly demonstrate that relatively PD in these two conditions was rated significantly
young, healthy, and well-matched employees with aided higher than the office noise, whereas the normal-hearing
HI do not differ in performance in visual working tasks group reported PD equally across all three noise condi­
compared with their normal-hearing peers, regardless tions, except in the serial recall task. A similar effect of
of whether or not background noise is present. F urther­ high-level noise (60 dBA) has been observed in the
more, previous studies have also reported th at specific hearing-impaired group in a study by Jahncke and
components within the phonological processing system Halin (2012), where the participants showed higher
could be deteriorating, indicating that working tasks self-rated fatigue and tended to have higher stress hor­
requiring phonological processing could be impaired in mone levels than in the low-level noise (30 dBA). There
the group with HI (Lyxell et al, 2003). This result is in might be a couple of explanations for these similar find­
contrast with our results because no significant difference ings in results, according to the authors. Firstly, the
was observed in the phonological decoding tasks; this greater sensitivity to a higher level of noise could be
m ight be due to the fact th a t the level of HI in pre­ due to recruitment, meaning th at the group with HI
vious studies was more severe-profound (>75 dB HL). perceived the high-level noises to be louder than what
Although the finding of nonsignificant differences in the normal-hearing group would perceive. Secondly,
cognitive performance was somewhat expected for the HAs were worn by the participants in all working tasks,
current sample, we cannot rule out the possibility th at which could have distorted incoming noise signals and
the measuring of cognitive performance is important consequently generating a higher disturbance for the
because previous studies have clearly demonstrated group with HI. As the daycare and traffic noise were
th at individuals with HI perform lower in certain cir­ around moderate sound levels (70 dBA), one would
cumstances (Hygge et al, 1992; Larsby et al, 2005; believe th at these two conditions were amplified to an
Lyxell et al, 2003). In addition to different study sam­ even higher sound level for the group with HI, and thus
ples, other reasons for the lack of difference in cognitive a higher PD was reported. In the current study sample,
performance m ight be th a t different tasks, noise all participants were fitted with different HAs from dif­
conditions, and study designs have been used in these ferent manufacturers, which could have affected the
studies. One could speculate that with different noise present results. However, before the study was under­
conditions and other types of visual working tasks, or taken, all participants had their HAs checked by an
even a greater difficulty for the current tasks, different audiologist, they were all frequent HA users, and a
results might have been revealed between groups. majority of the participants were using open-fit HAs
Therefore, we believe that it was a valid choice to include because of mild-moderate HI. This means that regard­
measures th at consider both objective and subjective less of the manufacturer, the study sample was relatively
dimensions of the effect of noise in the present study. homogeneous in regard to HA use and type, and they
Significant differences between and within groups received the required amplification based on their indi­
were observed in PD across the four conditions. Both vidual audiograms. Therefore, the current results should
groups rated the three noise conditions to be signifi­ be generalized on frequent HA users with mild-moderate
cantly more disturbing than the quiet condition, and HI, and it is possible that different results of PD would
this finding is consistent with w hat previous research have been observed if the study sample consisted of new
has shown among individuals with normal hearing sen­ HA users. We further argue that these two reasons are
sitivity (Schlittmeier et al, 2008; Ebissou et al, 2013). valid considering that the aided HI was the main differ­
Significant differences were also observed between ence between the two groups, and no difference in cogni­
groups where the hearing-impaired employees reported tive performance was observed.
a significantly higher PD than in participants with nor­ Previous research within the area of PD of noise and
mal hearing in the arithmetic easy and difficult task, mental performance at work has found th a t background
and orthographic decoding. No significant interaction speech has a detrimental effect on cognitive performance
effect was observed between conditions and groups in and PD in normal-hearing participants (Schlittmeier
phonological decoding. Group differences were noted et al, 2008). Irrelevant speech effect describes the signifi­
in the traffic setting in the serial recall task. Interest­ cant impairment of mental performance during irrele­
ingly, these group differences were only observed in the vant speech and some nonspeech sounds (Larsen et al,
two conditions with the highest levels (daycare: 73.5 2000). In our study, the office and daycare settings were
dBA; traffic: 72.5 dBA). Also, no significant differences the only two conditions that included speech, but no
were observed between groups in the quiet and office effect of irrelevant speech effect was observed in our
settings (56.6 dBA), meaning th a t a higher noise level results. This was apparent because no significant differ­
does indeed disturb employees with aided HI more than ences were observed between the traffic and daycare
their normal-hearing peers. When looking at the results noise within groups. However, it is important to empha­
within groups, we also observed a similar effect of the size that the speech sounds in the office noise were rel­
high-level noise conditions in the group with HI where atively low and took place in a rather quiet office space,

S 66
Effects of N oise on C ognitive Perform ance and PD/Hua et al

and the daycare noise consisted mostly of children talking hearing care and occupational health services for both
and screaming every now and then. Therefore, we do not employees with normal hearing and those with HI.
believe that this effect could have influenced the results.
As this study used different performance measures CONCLUSIONS
to examine the effects of different types of noise, we
wanted to find out if there was a correlation between he present results demonstrate that background
cognitive performance and PD, especially among the
participants with HI. Thus, additional correlation ana­
T noise does not impair cognitive performance in non­
auditory tasks in employees with HI and normal hearing,
lyses between variables were performed separately for but that PD is affected to a higher extent in the group
each group. No significant correlations were found with HI during higher levels of noise exposure. These
between RT and accuracy of the working tasks, and findings should only be generalized to individuals with
between cognitive performance and PD. This finding aided mild-moderate HI. We argue that self-rated mea­
indicates th at we cannot recommend one particular surements of PD are important and should be used
robust test to demonstrate the effect th at different types along with cognitive measurements when assessing noise
of background noise may have on individuals. Future annoyance in different working environments. In addition,
studies developing tests th at can measure both dimen­ this study also supported the results of studies regarding
sions, as an indicator of people’s perception of disturb­ the detrimental effect that high-level noise has on em­
ance, in a clinical setting would therefore be of interest. ployees with HI. Future research within this area should
Practical implications of where the present results focus on employees with severe-profound HI, interaction
are applicable include environments in daycare centers between different cognitive processes and work perfor­
and inventories where the background noises at the mance in relationship to different types of background
workplace tend to get high. In a recent study by Sjodin noise, and the impact on after-work activities.
et al (2012), stress-related health problems among 101
preschool employees with normal hearing were investi­ Acknowledgments. This study would not have been possi­
gated. The adverse effects of noise were analyzed using ble w ithout th e study participants who generously gave th eir
validated questionnaires and saliva cortisol samples. tim e to participate in th e cu rren t study. The authors also
Their results showed th a t stress and energy output th a n k Jonas Birkelof, Arvid Bjorndal, K ristina Ingvall, Jennie
were pronounced and th at 30% of the study sample H jaldahl, Johan n es Olsson, and Tobias A slund who carried
experienced burnout syndromes. Therefore, these find­ out some of th e tests and recorded the different sound environ­
ings along with the current results could indicate th at a ments; and G itte Keidser, William G. Beck, and two anonymous
higher risk for worse health could exist among employ­ reviewers for providing valuable suggestions on an earlier version
ees with HI, and this implication is in agreement with of the manuscript.
previous studies conducted on health and quality of life
showing worse health status in individuals with HI com­ REFERENCES
pared with that of their normal-hearing peers (Grimby
Borg G. (1990) Psychophysical scaling with applications in phys­
and Ringdahl, 2000; Danermark and Gellerstedt, ical work and the perception of exertion. Scand J Work Environ
2004; Kramer et al, 2006; Hua et al, 2013). The current Health 16(Suppl l):55-58.
findings are important because they demonstrate that
Borg G. (2008) A general scale to rate symptoms and feelings
although cognitive performance in nonauditory tasks related to problems of ergonomic and organizational importance.
are not affected in employees with aided HI, higher lev­ G Ital Med Lav Ergon 30(1, Suppl A):A8-A10.
els of noise are significantly more disturbing for this
Danerm ark B, Gellerstedt LC. (2004) Psychosocial work environ­
group compared with the group with normal hearing. This ment, hearing impairment and health. In t J Audiol 43(7):
finding could therefore indicate that employees with HI 383-389.
are more tired at the end of the working day, which
Duverne S, Lemaire P. (2004) Age-related differences in arith­
may cause participation restrictions in leisure activities metic problem-verification strategies. J Gerontol B Psychol Sci
outside work or even social isolation. If employees with Soc Sci 59(3):P135—P142.
HI were exposed to high-level noise for a longer period,
Ebissou A, Chevret P, Parizet E. (2013). Work Performance and
this might even be one possible reason for the reduced Mental Workload in Multiple Talker Environments. Paper pre­
physical health status and quality of life in this group. sented at the Proceedings of Meetings on Acoustics, Acoustical
In addition, no significant correlations were observed Society of America, Montreal, Canada, June 2-7.
between the different outcome measures. This result sug­ Grimby A, Ringdahl A. (2000) Does having a job improve the qual­
gests that alternative measurement of PD in conjunction ity of life among post-lingually deafened Swedish adults with
with measures of cognitive performance should be system­ severe-profound hearing impairment? Br J Audiol 34(3): 187-195.
atically used when assessing disturbance of noise in differ­
Guski R. (1997) Psychophysical methods for evaluating sound
ent working environments. Therefore, we emphasize the quality and assessing acoustic information. Acta Acust United
need of both self-rated and cognitive measurements in Acust 83(5):765-774.

857
Journal o f the American Academy of Audiology/Volume 25, Number 9, 2014

Guski R, Felscher-Suhr U, Schuemer R. (1999) The concept of Ljungberg J , Neely G, Lundstrom R. (2004) Cognitive performance
noise annoyance: How international experts see it. J Sound Vibrat and subjective experience during combined exposures to whole-body
223(41:513-527. vibration and noise. Int Arch Occup Environ Health 77(3):217-221.
Hallgren M, Larsby B, Lyxell B, Arlinger S. (2001) Evaluation of a Ljungberg JK, Neely G. (2007) Stress, subjective experience and
cognitive test battery in young and elderly normal-hearing and cognitive performance during exposure to noise and vibration.
hearing-impaired persons. J Am Acad Audiol 12(7):357-370. J Environ Psychol 27(l):44-54.
Hallgren M, Larsby B, Lyxell B, Arlinger S. (2005) Speech under­ Lyxell B, Andersson U, Borg E, Ohlsson IS. (2003) Working-
standing in quiet and noise, with and without hearing aids. Int J memory capacity and phonological processing in deafened adults
Audiol 44(101:574-583. and individuals with a severe hearing im pairm ent. In t J Audiol
42(Suppl 1):S86-S89.
Hua H, Karlsson J, Widen S, Moller C, Lyxell B. (2013) Quality of
life, effort and disturbance perceived in noise: a comparison Magnusson L. (1995) Reliable clinical determination of speech rec­
between employees with aided hearing impairment and normal ognition scores using Swedish PB words in speech-weighted noise.
hearing. Int J Audiol 52(9):642-649. Scand Audiol 24(4):217-223.

Humes LE. (2002) Factors underlying the speech-recognition per­ Ringdahl A, Grimby A. (2000) Severe-profound hearin g
formance of elderly hearing-aid wearers. J Acoust Soc Am 112(3 Pt im pairm ent and health -related quality of life among post-
1):1112-1132. lingual deafened Swedish adults. Scand A udiol 29(4):
266-275.
Hygge S, Ronnberg J, Larsby B, Arlinger S. (1992) Normal-hearing
and hearing-impaired subjects’ ability to just follow conversation in Rudner M, Lunner T, Behrens T, Thoren ES, Ronnberg J. (2012)
competing speech, reversed speech, and noise backgrounds. Working memory capacity may influence perceived effort during
J Speech Hear Res 35(1):208-215. aided speech recognition in noise. J A m Acad Audiol 23(8):
577-589.
Jahncke H, Halin N. (2012) Performance, fatigue and stress in open-
plan offices: the effects of noise and restoration on hearing impaired Schlittmeier SJ, Hellbriick J, Thaden R, Vorlander M. (2008) The
and normal hearing individuals. Noise Health 14(60):260-272. impact of background speech varying in intelligibility: effects on
cognitive performance and perceived disturbance. Ergonomics
Rnez I, Hygge S. (2002) Irrelevant speech and indoor lighting: 51(5):719-736.
effects on cognitive performance and self-reported affect. Appl
Cogn Psychol 16(6):709—718. Sjodin F, Kjellberg A, Knutsson A, Landstrom U, Lindberg L.
(2012) Noise and stress effects on preschool personnel. Noise
Kramer SE, Kapteyn TS, Houtgast T. (2006) Occupational per­ Health 14(59): 166-178.
formance: comparing normally-hearing and hearing-impaired
employees using the Amsterdam Checklist for Hearing and Work. Van Tasell DJ. (1993) Hearing loss, speech, and hearing aids.
In t J Audiol 45(9):503-512. J Speech Hear Res 36(2):228-244.

Larsby B, Hallgren M, Lyxell B, Arlinger S. (2005) Cognitive per­ Venetjoki N, Kaarlela-Tuomaala A, Keskinen E, Hongisto V.
formance and perceived effort in speech processing tasks: effects of (2006) The effect of speech and speech intelligibility on task per­
different noise backgrounds in normal-hearing and hearing- formance. Ergonomics 49(11):1068-1091.
impaired subjects. Int J Audiol 44(3):131-143.
Zekveld AA, K ram er SE, Festen JM. (2011) Cognitive load dur­
Larsen JD, Baddeley A, Andrade J. (2000) Phonological similarity ing speech perception in noise: the influence of age, hearing
and the irrelevant speech effect: implications for models of short­ loss, and cognition on the pupil response. Ear Hear 32(4):
term verbal memory. Memory 8(3): 145-157. 498-510.

8B8
Copyright of Journal of the American Academy of Audiology is the property of American
Academy of Audiology and its content may not be copied or emailed to multiple sites or
posted to a listserv without the copyright holder's express written permission. However, users
may print, download, or email articles for individual use.

You might also like