| Article Access Statistics|
| Viewed||15280 |
| Printed||442 |
| Emailed||13 |
| PDF Downloaded||359 |
| Comments ||[Add] |
|Year : 2001
: 4 | Issue : 13 | Page
|Low frequency noise "pollution" interferes with performance
Kerstin Persson Waye1, Johanna Bengtsson1, Anders Kjellberg2, Stephen Benton3
1 Department of Environmental Medicine, Göteburg University, Sweden
2 National Institute for Working Life, Stockholm, Sweden
3 Department of Psychology, University of Westminster, London, United Kingdom
Click here for correspondence address
To study the possible interference of low frequency noise on performance and annoyance, subjects categorised as having a high- or low sensitivity to noise in general and low frequency noise in particular worked with different performance tasks in a noise environment with predominantly low frequency content or flat frequency content (reference noise), both at a level of 40 dBA. The effects were evaluated in terms of changes in performance and subjective reactions. The results showed that there was a larger improvement of response time over time, during work with a verbal grammatical reasoning task in the reference noise, as compared to the low frequency noise condition. The results further indicated that low frequency noise interfered with a proof-reading task by lowering the number of marks made per line read. The subjects reported a higher degree of annoyance and impaired working capacity when working under conditions of low frequency noise. The effects were more pronounced for subjects rated as high-sensitive to low frequency noise, while partly different results were obtained for subjects rated as high-sensitive to noise in general. The results suggest that the quality of work performance and perceived annoyance may be influenced by a continuous exposure to low frequency noise at commonly occurring noise levels. Subjects categorised as high-sensitive to low frequency noise may be at highest risk.
Keywords: Low frequency noise, performance, annoyance, noise sensitivity
|How to cite this article:|
Waye KP, Bengtsson J, Kjellberg A, Benton S. Low frequency noise "pollution" interferes with performance. Noise Health 2001;4:33-49
| Introduction|| |
The introduction of modern technology and computerised machinery in industry has reduced the occurence of high noise exposure situations but introduced other types of occupational noise of more moderate noise levels. In many cases, the change to moderate noise levels has been achieved by building insulated control rooms from which industrial processes are supervised. The noise in such control rooms is often dominated by noise in the frequency range of 20 to 200 Hz (low frequency noise) caused by ventilation and air conditioning systems as well as by the lower attenuation of the low frequencies by the walls, floors and ceilings. Other occupational environments, such as office areas, house a number of noise sources that generate low frequency noise at moderate levels.
Major examples of such sources are network installations, ventilation, heating and air-conditioning systems.
There is a growing body of data showing that low frequency noise has effect characteristics that are different from other environmental noises of comparable levels [Persson Waye 1995; Berglund et al. 1994]. Symptoms that have been reported in connection with annoyance caused by low frequency noise and which may also reduce the working capacity are fatigue, headaches and irritation [Tokita 1980; Nagai et al. 1989; Persson Waye and Rylander 2001]. Although the importance of low frequency noise has been acknowledged in the WHO document on community noise [Berglund et al. 2000], the effects are less well explored compared to noises of higher frequencies and the specific regulations for control in the occupational environment are unsatisfactory.
Many occupational tasks have stringent demands on the employee as concerns knowledge, learning, flexibility, attention, and productivity. Computerised equipment determines the working speed of many employees, and the tasks demand concentration and rapid decisions. Supervising equipment on instrument boards also requires constant attention and, in the case of error messages, rapid remedial actions.
Information on how low frequency noise influences performance in these types of work situations is scarce. An experimental pilot study indicated that low frequency noise from ventilation equipment at a level of 42 dB LAeq could increase the time taken to respond in a verbal grammatical reasoning task, compared to a ventilation noise of equal level but not dominated by low frequencies (Persson Waye et al. 1997). Similarly, Kjellberg and Wide (1988) found a slower learning rate in this task when it was performed during exposure to simulated ventilation noise. Persson Waye et al. (1997) further showed that the subjects' self-rated performance and mood were affected to a higher degree by the low frequency noise than by noise not containing low frequency components.
It has also been shown that infrasound and low frequencies at 42 Hz may lower the level of wakefulness [Landstrom 1987; Landstrom et al. 1985; Landstrom et al. 1983]. This effect indicates that performance on monotonous, machine-paced tasks such as signal-monitoring tasks may be sensitive to low frequency noise exposure. Performance in these kind of tasks has consistently been found to be most sensitive to changes in wakefulness [Hockey 1986].
Reading comprehension and other verbal tasks have often been found to be more sensitive to noise than other tasks [Jones 1990]. However, it remains to be demonstrated if low frequency noise affects the performance of such tasks.
Generally, the research into the effects of noise on performance presents a rather inconsistent picture [Kjellberg and Landstrom 1994; Smith and Jones 1992]. One reason for this may be the large individual differences in noise sensitivity in general, and possibly, specifically for low frequency noise. It was previously found that subjects high-sensitive to noise in general, as measured by a sensitivity scale [Weinstein 1978], had the lowest performance accuracy under conditions of exposure to traffic noise [Belojevic et al. 1992]. Similarly, Jelinkova (1988) found that noise sensitive persons had a reduced working ability and attention when exposed to recorded traffic noise at 75 dB LAeq, compared to persons tolerant to noise.
Previous experience recorded from subjects disturbed by low frequency noise in their homes has shown that persons sensitive to low frequency noise are not necessarily sensitive to noise in general as measured by general noise sensitivity scales [Persson Waye 1995]. It is therefore important to categorise subjects not only in terms of sensitivity to noise in general but also with respect specifically to low frequency noise. However, the relation between self-rated sensitivity and performance effects is not clear.
The present study was undertaken to further elucidate the influence of low frequency noise on performance and attempted to answer the following questions:
- Can low frequency noise at a level normally present in control rooms and office areas influence performance and subjective well-being?
- What kind of performance tasks are affected by low frequency noise?
- How is the performance affected by duration of exposure?
- What is the relation between self-rated noise sensitivity and noise effects?
| Material and methods|| |
The subjects performed a series of performance tasks during exposure to a low frequency noise or a reference noise. Based upon responses to questionnaires, the subjects were categorised as having a high- or low sensitivity to noise in general and low frequency noise in particular. Their subjective reactions to the test session were recorded using questionnaires. To assess stress, saliva samples were taken and the amount of cortisol was determined. After each saliva sample, the subjects answered a questionnaire evaluating their perceived stress and energy [Kjellberg et al. 1989]. These latter data will be reported elsewhere [Persson Waye et al. 2001].
The exposure noises were two ventilation noises, one of a predominantly flat frequency character (reference noise) and the other of a predominantly low frequency character (low frequency noise). The reference noise was recorded from a ventilation installation. To obtain the low frequency noise, sound pressure levels in the frequency region of 31.5 to 125 Hz were increased using a digital sound processor system [Aladdin interactive workbench, Nyvalla DSP Stockholm, Sweden]. Furthermore, the third octave band centred at 31.5 Hz was amplitude-modulated with an amplitude frequency of 2 Hz. Both noises had a level of 40 dBA.
[Figure - 1] shows the equivalent third octave band sound pressure levels for the two noises, measured at the position of the subjects' head.
For the study, 19 female and 13 male (n=32) subjects with an average age of 23.3 (Sd= 2.58) were recruited by advertising. Each person underwent a hearing test [SA 201 II Audiometer, Entomed, Malmo, Sweden] and only persons with normal hearing (<20 dB HL) were allowed to participate. The subjects were given financial compensation for their participation.
Subjective sensitivity to noise
To assess sensitivity to low frequency noise and sensitivity to noise in general, two questionnaires were answered after the last test session. On the basis of the subjects' scores on two of the questions in the questionnaires, subjects were categorised as highly sensitive (high-sensitive) or less sensitive (low-sensitive) to low frequency noise. The first question "are you sensitive to low frequency noise" had five response alternatives ranging from "not at all sensitive" to "extremely sensitive". The second item, "I am sensitive to rumbling noise from ventilation systems", had six response alternatives ranging from "do not agree at all" to "agree completely". The subjects were also categorised as highly sensitive (high-sensitive) or less sensitive (low-sensitive) to noise in general, using the question "are you sensitive to noise in general" (with five response alternatives ranging from "not at all" to "extremely sensitive") and using the total number of points scored in a noise sensitivity evaluation questionnaire [Weinstein 1978]. The questionnaire had a total of 120 points; the higher the point scores, the higher sensitivity to noise. The subjects' answers ranged between 40 and 114 points, with an average of 70.8.
The categories of sensitivity were elicited through a principal component analysis with direct oblimin rotation of the four sensitivity questions. Two correlating factors, which explained 85% of the variance, were extracted. In the first factor, the two questions on low frequency noise had a loading of 0.9 and the questions on sensitivity to noise in general had a load below 0.5 see [Table - 1]. The other factor showed the opposite load pattern. The correlation between the two factors was 0.46. Factor scores were calculated for both the two factors, and the four categories (highsensitive/low-sensitive to low frequency noise and to noise in general) with the medians as cutoff point.
In the group, 15 females and three males were high-sensitive to low frequency noise, and four females and ten males were low-sensitive to low frequency noise. Eleven females and five males were categorised as high-sensitive to noise in general, and eight females and eight males were low-sensitive to noise in general.
The two categorisations were virtually independent (chi 2 =0.508, p=0.473). Eight of the 16 subjects categorised as low-sensitive to noise in general, belonged to the group assessed to be high-sensitive to low frequency noise. Six of the subjects in the group high-sensitive to noise in general did not belong to the group assessed to be high-sensitive to low frequency noise.
The experiment was performed in a 24 m 2 room, furnished as an office with a desk, computer and bookshelf. Behind the subject was a window with a closed Venetian blind so the person could be observed during performance. The sound was produced by four loudspeakers, hidden behind curtains and placed in each corner of the room. To amplify the low frequency noise, there was a subwoofer (ace-bass B2-50) which can reproduce frequencies down to 20 Hz. The background noise from the test chamber ventilation was less than 22 dBA, and the sound pressure levels for frequencies below 160 Hz were below the threshold of normal hearing [ISO 389-7:1996].
In the experiment, four performance tasks were used. Tasks I, II and IV involved working with a computer and task III involved working with pen and paper. The tasks were chosen in order to involve different levels of mental processing. A high workload was generated by instructing the subjects to work as rapidly and accurately as possible. All performance tasks were carried out twice in each test session, once in phase A and once in phase B see [Table - 2].
Task I was a simple reaction-time task and is part of the SPES computer test battery [Gamberale et al. 1989]. The subject was told to press a button as quickly as possible when a red square appeared on a black screen. Mean response times for the five, one-minute periods were recorded.
Task II was a short-term memory task. A set of numbers, e.g. 1 2 5 4, was shown on the computer screen. This set was followed by one number, e.g. 7. The subject was to respond, by yes or no, to whether that number was also present among the set of numbers shown earlier. The total response time and total number of correct and false answers were recorded.
Task II was carried out together with a secondary task, the bulb-task, previously used by Persson Waye et al. (1997). This task consisted of four differently coloured light bulbs, placed at four different positions on an arch at the periphery of the subject's visual field. Each of the four bulbs was illuminated at random intervals and in random sequence. The subjects' task was to respond only when a yellow bulb was illuminated, after which the subject was instructed to, as quickly as possible, push a response button that matched the colour (red, green or blue) of the light bulb that was illuminated prior to the yellow light bulb. The set-up used for task II with a primary and secondary object was designed to require the subject's full attention and concentration. The total response time and number of correct and erroneous responses were recorded.
Task III was a proof-reading task [Landstrom et al. 1997]. The subject read a text, printed on paper, for exactly ten minutes, and the task was to mark errors in the text. The number of lines read, correct marks, erroneous corrections and the total number of marks were recorded and related to the number of lines read for each subject; correct marks per line, erroneous corrections per line and total number of marks per line.
Task IV was a computerised verbal grammatical reasoning task, translated into Swedish from the original version [Baddeley 1968]. The task is based on grammatical transformation of sentences that have various passive, active, negative and positive structures. The subject was instructed to respond to whether a sentence is false or true in relation to a letter combination following the sentence. For example:
The set-up used for task IV was designed to impose a high mental workload. In total, the task consisted of eight blocks of 32 sentences. The mean response time for the eight blocks and the number of correct and false answers were recorded.
Following tasks II, III and IV, a questionnaire was administered to evaluate how much effort the subjects judged had used in order to perform each task. The subject could choose between five response alternatives ranging from "none at all" to "extremely".
A questionnaire evaluating mood [Sjoberg et al. 1979] was completed before and after the test session. The questionnaire consisted of 71 adjectives describing feelings of different kinds, and the adjectives formed the following six mood dimensions: social orientation, pleasantness, activation, extraversion, calmness and control. The subject could choose between four response alternatives: "I agree completely", "I somewhat agree ", "I do not agree" and "I certainly do not agree".
When the test session was completed, the subject completed a questionnaire evaluating selfreported estimates of annoyance due to noise, presence of symptoms experienced during or after the experiment and, also, questions were asked about whether the subject judged that the capacity to work had improved or been impaired due to noise, temperature or light during the tasks. Regarding impaired working capacity, the subject could choose between seven response alternatives ranging from one to seven: "major improvement", "rather much improvement", "some improvement", "neither improvement or impairment", "some impairment", "rather much impairment" and "major impairment". The alternatives for annoyance, ranging from one to five, were "not at all annoyed", "a little annoyed", "rather annoyed", "very annoyed" and "extremely annoyed". For the questions on presence of symptoms experienced during or after the experiment, questions were posed concerning headaches, pressure over the eardrum or head, occurrence of nausea, lack of concentration, irritation, tiredness, dizziness, irritation in eyes or throat or a sensation of unpleasant taste. The subject could choose between five response alternatives ranging from "not at all" to "extremely".
Experimental design and procedure
The experiment had a 2 (noises) × 2 (phases) × 2 (sensitivity groups) factorial design with repeated measures in the first two factors with independent groups representing the sensitivity factor. In the analyses of the simple reactiontime task and the verbal grammatical reasoning task, a fourth factor, time blocks within the task, was added.
On a separate occasion before the main test session, the subjects learned the procedures and practised on short versions of the performance tasks for about one hour with the reference noise at 35 dBA. Before each task, both written and verbal instructions were given to emphasise the need to "work as rapidly and accurately" as possible. The subjects were also informed that, if needed, they could communicate with the research director through a microphone on the desk.
In the study, each subject took part in two test sessions, on separate days and always in the afternoon. The total exposure time was on average 2 hours and 10 minutes with a variation of ±9 min. The variation was due to the difference in the individuals' performance time carrying out task IV during phase B.
Of the 64 test sessions, 37 started at 12.30, and 27 started at 15.00. The proportion of subjects starting at 12.30 and 15.00 for the two noise conditions was similar, 18/14 for the low frequency noise condition and 19/13 for the reference noise condition. During each test session, the subjects worked with four performance tasks and were exposed to the reference noise or the low frequency noise. A detailed plan of the experimental set-up is found in [Table - 2]. Half of the subjects started with the reference noise and the other half with the low frequency noise. To minimise subjective influence caused by the attitude to noise, motivation and the individual's level of expectations before the test sessions, the written and verbal information about the experiment did not explicitly refer to noise exposure.
Analysis and statistical methods
Analyses of variance, ANOVA, were performed to evaluate the influence of noise exposure, time, subjective sensitivity and their interactions on the different performance tasks and subjective ratings. The p-values are based on degrees of freedom corrected with Greenhouse-Geisser epsilon, when appropriate. To evaluate the difference of means for specific periods, a Student's t-test for dependent data was applied. Correlations between subjective data and performance were done using Pearson's correlation analysis. All tests were two-tailed, and a p-value of <0.05 was considered statistically significant, while a p-value up to 0.10 is reported as a tendency.
The statistical analyses employed SPSS [SPSS base 10.0 for Windows].
| Results|| |
No significant interaction of noise and gender was found for the subjective estimations or for any of the performance tasks.
No significant main effect of noise condition on reaction-time in the simple reaction-time task was found (F(1,29)=1.952, p=0.173).
A tendency to a two-way interaction in reactiontime was found between noise and sensitivity to noise in general (F(1,29)=4.141, p=0.051). Subjects high-sensitive to noise in general had a somewhat longer reaction-time during the low frequency noise condition compared to the reference noise condition, while the lowsensitive subjects had a similar reaction-time during both noise conditions.
The results from the short-term memory task and the bulb-task are shown in [Table - 3],[Table - 4] respectively.
No significant difference in total response time was found between noise exposures for the short-term memory task (F(1,31)= 0.561, p=0.46) or the bulb-task (F(1,31)= 0.304, p=0.585). In the short-term memory task, the total number of errors made in phase A and B was small and did not differ between noise exposures.
In the short-term memory task, a significant three-way interaction in response time was found between noise, phase and low frequency noise sensitivity (F(1,30)= 4.949, p<0.05). As can be seen in [Table - 3], subjects high-sensitive to low frequency noise decreased their response time considerably from phase A to B during reference noise, while their response time decreased only slightly during the low frequency noise. Subjects low-sensitive to low frequency noise, in contrast, decreased their response time during low frequency noise, while their response time decreased only slightly during the reference noise.
Among subjects categorised according to sensitivity to noise in general, a significant threeway interaction was found between noise, phase and sensitivity to noise in general (F(1,30)= 6.576, p<0.05). Subjects high-sensitive to noise in general decreased their response time from phase A to B during low frequency noise, but not during reference noise. Subjects low-sensitive to noise in general, on the contrary, only decreased their response time during reference noise.
During the bulb-task, [Table - 4], subjects highsensitive to low frequency noise had, regardless of noise exposure, a longer response time than low-sensitive subjects (2674 ms compared with 2150 ms, F(1,30)=7.545, p<0.01). No significant difference was found for subjects categorised according to general noise sensitivity.
The results of the proof-reading task are given in [Table - 5].
No significant effect of the different noise exposures was found on number of correct marks per line.
A significant two-way interaction between noise and phase, (F(1,31)= 10.069, p<0.005), was found for the number of erroneous corrections per line. The number of erroneous corrections was lower during phase B in low frequency noise, but not during reference noise. A two-way interaction between noise and phase was also found for total marks (correct and erroneous) per line, (F(1,31)=7.018, p<0.05).
Regardless of noise exposure, subjects highsensitive to low frequency noise made slightly fewer total marks per line than low-sensitive subjects. This tendency was the same in both noise conditions. Subjects high-sensitive to low frequency noise also read fewer lines than lowsensitive subjects (on average 129 lines compared with 142 lines). There was a significant three-way interaction between noise phase and low frequency noise sensitivity (F(1,30)=5.306, p<0.05). Subjects high-sensitive to low frequency noise read a larger number of lines in phase B as compared to phase A during reference noise conditions and a lower number of lines in phase B compared to phase A during low frequency noise conditions. The reverse was seen for low-sensitive subjects.
The same analysis with subjects categorised according to general noise sensitivity showed partly different results. The difference in numbers of lines read, was not present (134 lines for high-sensitive subjects as compared with 136 among low-sensitive). The interaction between noise, phase and sensitivity was significant F(1,30)=7.976, p<0.01), but the pattern of a lower number of lines read for subjects highsensitive to low frequency noise was not seen for subjects high-sensitive to noise in general.
The results of the verbal grammatical reasoning task are shown in [Figure - 2],[Figure - 3].
The total response times for phases A and B were 3.8 s for the low frequency noise condition and 3.6 s for the reference noise condition. [Figure - 2] demonstrates that no difference in total response time was found between noise conditions in phase A. The mean response time was shorter during phase B as compared to phase A in both noise conditions (3.7 s versus 3.9 s, F(1,31)=9.014, p<0.01), but the decrease in response time in phase B was less pronounced during the low frequency noise conditions. This two-way interaction between noise and phase was significant (F(1,31)=5.750, p<0.05).
Subjects high-sensitive to low frequency noise had on average a similar response time between noises in phase A. [Figure - 3] shows that the difference in response time during low frequency noise and reference noise conditions was larger in phase B, and a tendency to a three-way interaction between low frequency noise sensitivity, noise and phase was found (F(1,30)=3.319, p=0.078). For subjects categorised as high-sensitive to noise in general, no difference between the noise conditions was detected.
In summary, the main results from the performance tasks were that during work with the proof reading task a lower number of erroneous marks as well as total marks were made during low frequency noise. During work with the verbal grammatical task subjects showed a greater improvement over time during reference noise exposure compared to low frequency noise exposure.
The average rated effort for the short-term memory task, the proof-reading task and the verbal grammatical reasoning task is given in [Table - 6].
As can be seen, the rated effort was lowest for the proof-reading task and equivalent to "some effort" to "moderate effort". The highest rating was given for the verbal grammatical reasoning task, which scored closest to the category of "rather much effort". No significant difference was found between the noise conditions on rated effort during the proof-reading task or the verbal grammatical reasoning task. For the short-term memory task, however, a two-way interaction between noise and phase was found (F(1,31)=4.307, p<0.05). The interaction was due to the task being rated as more effort demanding in phase B as compared to phase A during the reference noise condition, while the subjects rated the same effort in both phases during the low frequency noise.
Subjects high-sensitive to low frequency noise rated, regardless of noise, more effort than lowsensitive subjects on the proof-reading task and the verbal grammatical reasoning task (3.1 versus 2.4; F(1,30)=4.371, p<0.05 and 4.1 versus 3.4; F(1,30)=5.886, p<0.05).
The low frequency noise was on average rated as more annoying than the reference noise (2.5 versus 2.0; F(1,31)=9.922, p<0.005). Furthermore, there was a two-way interaction between low frequency noise sensitivity and noise (F(1,30)=6.534, p<0.05). Subjects highsensitive to low frequency noise were more annoyed by the low frequency noise than by the reference noise (3.1 compared to 2.3), while low-sensitive subjects reported on average the same annoyance after both noises (1.6). No significant difference between noises was found
when the same analysis was done with subjects categorised according to general noise sensitivity.
Low frequency noise was on average considered to impair the working capacity more than the reference noise (5.2 versus 4.8; F(1,31)=6.808, p<0.05). When the data was subdivided into the two noise sensitivity groups, no significant effect due to noise condition could be detected.
No significant main effect of noise condition was found for the mood dimensions. There was, however, a significant three-way interaction between noise, phase and low frequency noise sensitivity (F(1,29)=4.352, p<0.05) for perception of "being in control" [Figure - 4]. The figure demonstrates that the high-sensitive subjects' perception of being in control was lower after (3.3), as compared to before (3.4) the exposure to low frequency noise. The opposite results were found for low-sensitive subjects (3.5 after compared to 3.4 before). A tendency to the same three-way interaction, between noise, phase and low frequency noise sensitivity, was present for "activation" (F(1,29)=3.837, p=0.06). The interaction showed a lower value for perception of activation during both noise conditions for subjects high-sensitive and subjects low-sensitive to low frequency noise, but the decrease was greater for high-sensitive subjects during the low frequency noise condition. However, when the analysis on control and activation was conducted with subjects categorised according to general noise sensitivity, these effects were not present.
No significant main effect of noise condition was found for the different symptoms.
In summary, the main results from the subjective estimations were that the low frequency noise was rated as more annoying and also considered to impair working capacity more than the reference noise. No direct effects of noise condition for symptoms were found.
Relations between performance and subjective estimations
Impaired working capacity due to reference noise exposure was negatively correlated to number of lines read in phase A (rxy -0.495, p<0.005).
A significant correlation was also found between rated tiredness and response time in the verbal grammatical reasoning task in phase B during low frequency noise (rxy=0.524, p<0.005). For the reference noise, there was a correlation between response time in the simple reactiontime task in phase B and headaches (rxy=0.517, p<0.005).
Impaired working capacity due to low frequency noise exposure was significantly correlated to lack of concentration (rxy=0.507, p<0.005), nausea (rxy=0.460, p<0.01), tiredness (rxy=0.471, p<0.01) and a feeling of pressure on the head (rxy=0.494, p<0.005). No significant correlation between noise impairment due to reference noise and symptoms was found.
Annoyance due to low frequency noise was correlated to subjective estimation of the following symptoms: a feeling of pressure on the head (rxy=0.664, p<0.001), tiredness (rxy=0.519, p<0.005), dizziness (rxy=0.519, p<0.005), and lack of concentration (rxy = 0.537, p<0.005). Reference noise annoyance was correlated only to nausea (rxy=0.522, p<0.005).
In summary, relationships between annoyance respectively impaired performance and several symptoms were found after work in low frequency noise, while a relationship between annoyance and nausea was found after work in reference noise.
| Discussion|| |
The experiment was designed to test the effects of low frequency noise in a situation requiring an increased level of attention and awareness for a fairly prolonged time period. As the experiment was performed under laboratory conditions, the relevance of the results for normal working conditions must be evaluated with care. Alterations in performance found under experimental conditions could incorporate a bias induced by the experimental situation and particularly by the acute exposure conditions [Rylander and Persson Waye 1997]. On the other hand, tiredness and decrease in performance induced by a particular environmental stimulus, in this case low frequency noise, would probably have a low level of adaptation and transfer into effects that could be registered in real life after long-term exposure.
The results indicate that low frequency noise, at levels normally occurring in office and control rooms, could negatively influence performance in more demanding verbal tasks, while the effects on the routine tasks were less clear. Decreases in performance on verbal tasks and on tasks that put high demands on information processing have previously been reported in other studies using other types of noise exposures and at higher noise levels [e.g. reviews Smith 1989; Smith and Jones 1992; Kjellberg and Landstrom 1994]. Importantly, this study used lower noise levels. This was done in order to address the impact of less intense, but more widespread noise. In spite of the comparatively low noise levels, significant differences in performance could be detected and related to the content of low frequencies in the noise. This supports the previous hypothesis [Persson Waye 1995] that different mechanisms mediate the effects on performance under conditions of exposure to low frequency noise as compared to higher frequency noise.
The decrease in response-time over time during work with the verbal grammatical reasoning task was larger in reference noise, indicating a higher learning effect in this noise condition. This together with the results from the proof-reading task, where subjects made fewer total marks over time during exposure to low frequency noise, give some support for the hypothesis that low frequency noise is more difficult to ignore or to habituate to [Benton 1997a; Benton 1997b]. The larger decrease in response time during the verbal grammatical reasoning task in phase B during exposure to reference noise indicates a higher learning effect during the reference noise condition as compared with the low frequency noise condition. Less habituation to low frequency noise may reduce the available information processing resources, and may lead to higher competition between available resources, which would interfere with cognitive processing abilities. The observation that the effects appeared in the second phase of the experiment supports this hypothesis, as the effort to cope in low frequency noise would develop over time and thus be more strenuous over time.
In the proof-reading task, fewer total marks per line read and fewer erroneous corrections per line read were found during the low frequency noise condition in phase B. The tendency to make fewer total marks per line read could be a result of a less thorough treatment of the text material. Such coping strategies for contextual errors per line read have previously been reported by Weinstein [1974; 1977], while he did not find an effect on the number of noncontextual errors as a result of noise exposure. In the study presented here, the total number of marks was rather few and it was thus not meaningful to subdivide the analysis into contextual and non-contextual errors.
The comparatively longer response time over time seen for the verbal grammatical reasoning task during the low frequency noise condition is in agreement with previous findings [Persson Waye et al. 1997]. In that study, a tendency towards longer response time over time was found in the low frequency noise condition, using the same performance task and exposure noises as used in this experiment but involving a smaller number of test subjects. The simple reaction-time task has previously been found to be sensitive to tiredness [Kjellberg et al. 1998]. In this study, a difference between noise conditions for this task was only found for subjects high-sensitive to noise in general. This moderate effect is comprehensible as the design of the study aimed to generate a high and correct work load and the exposure time was limited to two hours. To evaluate tiredness, further studies should be carried out involving subjects working at their own pace and during a longer exposure time.
Few other studies have previously investigated performance after exposure to low frequency noise. Benton and Leventhall (1986) found that exposure to pure tones (centred at 40 Hz and 100 Hz, modulated at 1 Hz and a narrow band centred at 70 Hz at a level of 25 dB above the individual threshold) gave rise to more errors as compared with exposure to traffic noise at 90 dB lin, or silence. The effects were especially pronounced during the last 10 minutes of the total 30-minutes exposure. Some support for impaired performance caused by low frequency noise was also given by Benton and Robinson (1993). Previous studies are thus in agreement with the findings presented here, but further studies need to be carried out to evaluate more specifically how low frequency noise affects performance and which tasks or situations that are most vulnerable for noise interference.
The results do not give direct support for the hypothesis that low frequency noise would induce different symptoms that could impair performance. No direct effects of noise condition on symptoms, or clear relationships between symptoms and performance effects, were found. However, the relationships between symptoms and annoyance respectively, symptoms and impaired performance, were particularly frequent after work in the low frequency noise condition, while for the reference noise a relationship was found only between annoyance and nausea. Although the study is not able to predict whether symptoms impair performance or whether the strain of performing during the low frequency noise condition could lead to a development of symptoms, the findings support a link between symptoms and the experience of impaired performance.
The reasons for choosing the specific low frequency noise used in this study was to achieve a noise that resembled a realistic ventilation noise, which often includes a tonal component and a modulation characteristic [Broner 1994]. The effects observed after low frequency noise could be related to specific acoustical characteristics such as amplitude modulation and the tonal character at 31.5 Hz. In one study, the presence of modulations was found to lead to increased sleepiness [Persson Waye et al. 1997], but the influence of a tonal character in the low frequency range has been shown to be of little or no importance for annoyance, reduced wakefulness or performance [Landstrom et al. 1991; Landstrom et al. 1995; Holmberg et al. 1993]. While the presence of amplitude modulations thus could have increased the effects, the tonal character was of less importance.
Subjects high-sensitive to low frequency noise generally performed less well and also reported the highest annoyance due to low frequency noise. In other studies, subjects high-sensitive to noise in general have been found to have the lowest performance accuracy during exposure to traffic noise [Belojevic et al. 1992]. Interestingly, this study also indicate that the response between the two categorisations of sensitivity to low frequency noise and sensitivity to noise in general were partly different. Some of these differences were found regardless of noise exposure, such as the difference in response time in the simple reaction-time task found using the categorisation of sensitivity to low frequency noise, while this difference was not found using the categorisation of sensitivity to noise in general. Other differences were related to noise exposure, such as the longer response time found in phase B during low frequency noise on the verbal grammatical reasoning task, for subjects high-sensitive to low frequency noise, while no difference between noise conditions was found using the categorisation according to sensitivity to noise in general. Differences related to noise exposure were also found for some of the subjective responses such as a higher rating of annoyance and lower perception of control among subjects high-sensitive to low frequency noise, while this difference was not found using the categorisation according to sensitivity to noise in general.
While the results from the study show that subjects categorised as high-sensitivity to noise in general or to low frequency noise generally gave a higher subjective rating of annoyance and impaired working capacity, the difference caused by noise exposure upon performance and subjective estimations was most obvious among subjects categorised with regard to sensitivity to low frequency noise. This agrees with previous observations that low frequency noise sensitivity is a specific issue. The validity and practical relevance of these characteristics should be further evaluated for effects caused by low frequency noise.
In conclusion, the study supports a hypothesis that low frequency noise at levels normally occurring in office-like environments may influence work performance and subjective perception of annoyance and lead to work impairment. The study also points to the importance of including factors related to individual sensitivity to noise when evaluating effects. According to the results obtained here, subjects categorised as high-sensitive to low frequency noise seem to be at highest risk.
| Acknowledgements|| |
The project was supported by funds from the Swedish Council for Work Life research (grant nr 1998-06-08). The project is also part of a network-program for occupational health research funded by the National Institute for Working Life. We gratefully acknowledge the valuable comments by Prof. Ragnar Rylander and the skilful research assistance of Agneta Agge and technical assistance of Martin Bjorkman.
| References|| |
|1.||Aladdin interactive work-bench, Nyvalla DSP, Stockholm, Sweden. |
|2.||Baddeley A.D. (1968) A 3 min reasoning task based on grammatical transformation. Psychon. Sci. 10: 341-342. |
|3.||Belojevic G., Ohrstrom E. and Rylander R. (1992) Effects of noise on mental performance with regard to subjective noise sensitivity. Int. Arch. Occup. Environ. Health. 64: 293-301. |
|4.||Benton S. and Leventhall H.G. (1986) Experiments into the impact of low level, low frequency noise upon human behaviour. J. Low Freq. Noise Vibr. 5: 143-162. |
|5.||Benton S. and Robinson G. (1993) The effects of noise on text problem solving for the word processor user (WPU). Noise and man 1993. Proceedings of the 6th International Congress on Noise as a Public Health Problem, Nice, France. pp 539-541. |
|6.||Benton S. (1997a) Measurement challenges in assessing the annoying characteristics of noise. Is low frequency noise a special case? J. Low Freq. Noise Vibr. 16: 13-24. |
|7.||Benton S. (1997b) Low frequency noise and the impact upon individual's quality of life: Case study reports. J. Low Freq. Noise Vibr. 16: 203-208. |
|8.||Berglund B., Hassmen P. and Soames Job R.F. (1994) Health effects induced by low-frequency noise: A critical review. Archives of the Center for Sensory Research. 1: 128. |
|9.||Berglund B., Lindvall T., Schwela D. and Goh K-T. (2000) Guidelines for Community Noise. World Health Organization. ISBN: 9971-88-770-3. |
|10.||Broner N. (1994) Determination of the relationship between low frequency HVAC noise and comfort in occupied spaces - objective phase. ASHRAE 714-RP, Vipac Report 38114. |
|11.||Gamberale F., Iregren A. and Kjellberg A. (1989) SPES: The computerized Swedish Performance Evaluation System. Background, critical issues, empirical data and users manual. Arbete och Halsa, National Board of Occupational Safety and Health, Solna, Sweden. |
|12.||Hockey G.R.J. (1986) Changes in operator efficiency as a function of environmental stress, fatigue, and circadian rhythms. In: Boff KR, Kaufman L & Thomas JP eds. Handbook of perception and performance. Vol. II. Cognitive process and performance, Wiley, New York, chapter 44: pp 1-49. |
|13.||Holmberg K., Landstrom U. and Kjellberg A. (1993) Effects of ventilation noise due to frequency characteristics and sound level. J. Low Freq. Noise Vibr. 12: 115-122. |
|14.||ISO 389-7:1996. Acoustics Reference zero for the calibration of audiometric equipment Part 7: Reference threshold of hearing under free-field and diffuse-field listening conditions. |
|15.||Jelinkova Z. (1988) Coping with noise in noise sensitive subjects. Proceedings from the 5th International Congress on Noise as a Public Health problem, Stockholm, Sweden, August 21-25, pp 27-30. |
|16.||Jones D. (1990) Recent advances in the study of human performance in noise. Environ. Int. 16: 447-458. |
|17.||Kjellberg A. and Iwanowski S. (1989) The stress/energy questionnaire - Development of a method to evaluate estimations of mood at work. National Institute of Occuaptional Health. Report 1989:26 (In Swedish). |
|18.||Kjellberg A. and Landstrom U. (1994) Noise in the office: Part II - The scientific basis (knowledge base) for the guide. Int. J. Indust. Ergonomics. 14: 93-118. |
|19.||Kjellberg A. and Wide P. (1988) Effects of simulated ventilation noise on performance of a grammatical reasoning task. Proceedings of the 5th International Congress on Noise as a Public Health Problem, Stockholm, Sweden, August 21-25, pp 31-36. |
|20.||Kjellberg A., Muhr P. and Skoldstrom B. (1998) Fatigue after work in noise - an epidemiological survey study and three quasi-experimental field studies. Noise & Health, 1: 47-55. |
|21.||Landstrom U., Bystrom M., Kjellberg A. and Nordstrom B. (1997) Performance and annoyance during exposure for amplitude-modulated noise at different noise levels. National Institute of Occupational Health. Report 1997:15 (In Swedish). |
|22.||Landstrom U. (1987) Laboratory and field studies on infrasound and its effects on humans. J. Low Freq. Noise Vibr. 6: 29-33. |
|23.||Landstrom U., Bystrom M. and Nordstrom B. (1985) Changes in wakefulness during exposure to noise at 42 Hz, 1000 Hz and individual EEG frequencies. J. Low Freq. Noise Vibr. 4: 27-33. |
|24.||Landstrom U., Lundstrom R. and Bystrom M. (1983) Exposure to infrasound- perception and changes in wakefulness. J. Low Freq. Noise Vibr. 2: 1-11. |
|25.||Landstrom U., Kjellberg A., Soderberg L. and Nordstrom B. (1991) The effects of broadband, tonal and masked ventilation noise on performance, wakefulness and annoyance. J. Low Freq. Noise Vibr. 10: 112-122. |
|26.||Landstrom U., Akerlund E., Kjellberg A. and Tezars M. (1995) Exposure levels, tonal component and noise annoyance in working environment. Environment International, 21: 265-275. |
|27.||Nagai N., Matsumoto M., Yamasumi Y., Shiraishi T., Nishimura K., Matsumoto K., Miyashita K. and Takeda S. (1989) Process and emergence on the effects of infrasonic and low frequency noise on inhabitants. J. Low Freq. Noise Vibr. 8: 87-99. |
|28.||Persson Waye K. (1995) Doctoral Thesis, Goteborgs University, ISBN 91-628-1516-4. On the effects of environmental low frequency noise. |
|29.||Persson Waye K. and Rylander R. (2001) The prevalence of annoyance and effects after long term exposure to low frequency noise. J. Sound Vib. 240: 483-497. |
|30.||Persson Waye K., Rylander R., Benton S. and Leventhall H.G. (1997) Effects on performance and work quality due to low frequency ventilation noise. J. Sound Vib. 205: 467474. |
|31.||Persson Waye K., Bjorkman M., Hellstrom P-A. and Rylander R. Acute effects of low frequency noise in relation to susceptibility to low frequency noise. (Submitted) J. Sound Vib 1997. |
|32.||Persson Waye K., Bengtsson J., Rylander R., Hucklebridge F., Evans P. and Clow A. Low frequency noise enhances cortisol among noise sensitive subjects during work performance. (In press) Life Science 2001. |
|33.||Rylander R. and Persson Waye K. (1997) Human perception of environmental noise in acute and chronic exposure conditions. Proceedings of Inter Noise, Budapest, Hungary 25-27 August, pp 971-976. |
|34.||Sjoberg L., Svensson E. and Persson L.O. (1979) The measurement of mood. Scand. J. Psychol. 20: 1-18. |
|35.||Smith AP. (1989) A review of the effects of noise on human performance. Scand. J. Psychol. 30: 185-206. |
|36.||Smith AP. and Jones D.M. (1992) Noise and performance. In D.M. Jones and AP Smith eds. Handbook of Human Performance, Vol 1: The Physical Environment. Academic Press, London, pp 1-28. |
|37.||SPSS base 10.0 for Windows. |
|38.||Tokita Y. (1980) Low frequency noise pollution problems in Japan. Proceedings of the Conference Low frequency noise and hearing, Aalborg, Denmark, pp 189-196. |
|39.||Weinstein N.D. (1974) Effects of noise on intellectual performance. J. Appl. Psychol. 59: 548-554. |
|40.||Weinstein N.D. (1977) Noise and intellectual performance: A confirmation and extention. J. Appl. Psychol. 62: 104-107. |
|41.||Weinstein N.D. (1978) Individual differences in reactions to noise: A longitudinal study in a college dormitory. J. Appl. Psychol. 63: 458-466. |
Kerstin Persson Waye
Department of Environmental Medicine, Göteburg University, Box 414, Göteburg 405 30
Source of Support: None, Conflict of Interest: None
[Figure - 1], [Figure - 2], [Figure - 3], [Figure - 4]
[Table - 1], [Table - 2], [Table - 3], [Table - 4], [Table - 5], [Table - 6]