Starkey Research & Clinical Blog

Cognitive Benefits of Digital Noise Reduction

Desjardins, J. & Doherty, K. (2014). The effect of hearing aid noise reduction on listening effort in hearing-impaired adults. Ear and Hearing 35 (6), 600-610.

This editorial discusses the clinical implications of an independent research study and does not represent the opinions of the original authors.

Understanding speech in noise can be a challenge for anyone with hearing loss, but it is especially difficult for older listeners (Plomp, 1978; Duquesnoy, 1983; Dubno et al., 1984; Helfer & Freyman, 2008).  Older individuals also experience increased listening effort in noisy conditions as compared to younger listeners (Desjardins & Doherty, 2013).  Listening effort is often measured in a dual-task paradigm during which subjects perform a second task while simultaneously repeating speech in noise. Increased listening effort is reflected by the allocation of cognitive resources away from the secondary task, resulting in poorer performance. When this effect is considered with reference to everyday situations, increased listening effort could have repercussions for elderly individuals beyond communication, affecting their ability to multi-task which could have associated safety concerns.  Listening effort could also affect the risk of social isolation as elderly, hearing-impaired individuals may feel reluctant to exert the energy and effort required to interact with others in group situations.

In a review of the literature on speech recognition and cognitive abilities, Akeroyd (2008) found that hearing was the primary predictor of speech recognition performance, but working memory capacity was the second best predictor. Because speech perception is affected by peripheral auditory processing as well as cognitive functions like working memory and speed of processing, the measurement of speech recognition scores alone may not be sufficient to evaluate the potential benefits of noise reduction in hearing aids. To this end, some studies have examined the effects of noise reduction on listening effort and the allocation of cognitive resources. Ng et al. (2013) found that noise reduction improved working memory function for some subjects and Sarampalis et al. (2009) found that the use of noise reduction reduced listening effort, resulting in quicker visual reaction times and better word recall on secondary tasks.

The current study investigated the relationships among noise reduction, listening effort and speech recognition in middle-aged to older adults with hearing loss. A dual-task paradigm, with speech recognition in noise as the primary task and visual tracking as the secondary task, was used to evaluate listening effort. Working memory tends to decline with advancing age (Salthouse, 1994), as does the speed of perceptual processing (Salthouse, 1985; Wingfield et al., 1985), so measures of working memory and processing speed were also examined. Twelve subjects participated in the study, ranging in age from 50 to 74 years.  All had symmetrical, sensorineural hearing loss and were experienced hearing aid users. For the purpose of the study, subjects were fitted with behind-the-ear hearing aids and disposable canal earmolds, fitted to DSL v5 targets (Scollie et al., 2005). Hearing aids were programmed with two memories: in the first, all special signal processing features were disabled and in the second, noise reduction (Voice iQ2) was set to maximum.

Speech recognition in noise was examined using the R-SPIN Test (Bilger et al., 1984), which consists of eight lists of 50 sentences each. Half of the sentences on each list are high context and half are low-context.  The sentences were presented in two-talker babble (TTB), composed of two female talkers reciting nonsense sentences. TTB has significantly affected listening effort for older subjects in previous research (Desjardins & Doherty, 2013).  Prior to the dual task procedure, subjects performed a sentence recognition test in noise to determine the SNR required to achieve 76% performance and 50% performance.  Not surprisingly, subjects required more favorable SNRs to achieve 76% performance, with average SNRs of 4.4dB for 76% performance and 1.8dB for 50% performance. The individual SNRs were used later to derive four listening conditions for the dual task procedure:

1.              Moderate listening condition (more favorable SNR) with noise reduction

2.              Moderate listening condition without noise reduction

3.              Difficult listening condition (less favorable SNR) with noise reduction

4.              Difficult listening condition without noise reduction

Following the sentence blocks in the dual-task procedure, listeners were asked to rate their ease of listening. They were asked to rate how easy it was to listen to the sentences on a scale from 0-100, with 0 being “very, very difficult” and 100 being “very, very easy” (Geller & Margolis, 1984; Feuerstein, 1992).

Results for the R-SPIN speech recognition task revealed significantly higher scores in the moderate listening condition compared to the difficult listening condition. There was no main effect of noise reduction or any interaction between listening condition and noise reduction, indicating that noise reduction did not have a significant impact on speech recognition ability in noise. Scores for high-context sentences were significantly better than for low-context sentences, indicating that listeners used context to help them understand the sentences better. There was no interaction between context and listening condition.

Listening effort was measured with the dual-tasks of speech recognition in noise and visual tracking. Poorer performance on the visual tracking task indicated higher listening effort or the allocation of more cognitive resources toward speech recognition. Sentence recognition scores were not significantly different based on the four test conditions, but listening effort was affected by test condition.  Without noise reduction, listening effort was significantly higher for the difficult listening condition than for the moderate condition. With noise reduction, there was no significant difference in listening effort between the moderate and difficult listening conditions. In other words, noise reduction reduced listening effort in the difficult listening condition, but did not have a significant effect in the moderate condition. Perhaps surprisingly, listening effort did not vary depending on sentence context. Secondary task performance remained consistent during high and low context sentences, despite the fact that speech recognition scores were significantly better for high context sentences.

Self-perceived ease of listening ratings showed that subjects rated the moderate listening condition as significantly easier than the difficult condition. There was no significant difference in their ease of listening ratings based on noise reduction, nor was there an interaction between noise reduction and listening condition. These results indicated that the subjects did not feel that noise reduction made listening easier, despite the fact that their measured listening effort was significantly reduced with noise reduction, at least in the difficult listening condition.

There were no significant effects between working memory or processing speed on listening effort, with or without noise reduction. There was, however, a significant trend for subjects with faster processing speeds to show reduced listening effort with noise reduction activated, but this occurred only for the difficult listening condition.

The results of this study are in agreement with prior reports, in that noise reduction did not significantly improve significant speech recognition scores in noise. Noise reduction did, however, reduce listening effort in the difficult listening condition with the poorer signal-to-noise ratio. This is in agreement with Sarampalis et al. (2009) who reported listening effort reduction only in the more difficult listening condition of -6dB SNR.  Though working memory and processing speed did not significantly affect listening effort in this study, there was a trend showing subjects with faster processing were able to derive more benefit from noise reduction in the difficult listening situation. Prior studies have shown relationships between working memory, processing speed and listening effort, so this is an area that requires further study.

Desjardins and Doherty’s study provides further evidence that tests of listening effort may be a reasonable tool for evaluating the benefits of noise reduction. The authors point out that the ability of noise reduction to reduce listening effort in noisy conditions could have implications for multi-tasking, which could in turn affect safety in some scenarios.  For instance, an older hearing-impaired person, driving a car while talking to a passenger, may unwittingly divert cognitive resources toward the task of understanding conversation, thus potentially reducing their ability to respond to other stimuli related to driving. If noise reduction helps reduce the cognitive demand required for speech recognition in this scenario, in theory there would be more cognitive resources available for driving and attending to surrounding activity.

Concerns like this are particularly important when considering hearing aid fittings for older patients.  Hearing loss increases with advancing age, while working memory and processing speed are known to decline with age.  Therefore, older individuals are more likely to be challenged by the cognitive demands of multi-tasking while at the same time facing additional obstacles like poorer hearing and speech recognition ability and impaired vision.  Though the maximum noise reduction settings used in this study are not usually used in clinical fittings, it may be appropriate to use higher noise reduction settings for older hearing aid users, especially those with known cognitive processing deficits. Directional microphones and automatic noise programs with low-frequency gain reduction may provide additional benefit.  In counseling sessions, clinicians should discuss potential multi-tasking difficulties and the need to reduce distractions in order to optimize speech recognition ability. The importance of properly fitted hearing instruments to facilitate speech communication during simultaneous tasks should also be emphasized, especially at early appointments when hearing loss is diagnosed and hearing aids are being considered. This is particularly relevant to situations in which older, hearing-impaired listeners may be executing other tasks like driving, walking on stairs or through parking lots or streets, in which safety is a concern.

Social isolation is a concern for older adults with hearing loss. The National Council on Aging (www.ncoa.org) reports that hearing loss can reduce participation in social activities, reduce self-confidence, cause depression and strain relationships with family and friends. Hearing loss is also known to increase mental fatigue, making hearing-impaired listeners feel “exhausted” and fatigued from conversational interaction (Hornsby, 2013). Older adults who treat their hearing loss with hearing aids are less likely to experience these negative effects. It is plausible to assert that factors affecting listening effort could also affect the risk of social isolation, as the increased effort and fatigue required in group interactions could be too daunting for some individuals.  The beneficial effect of noise reduction on listening effort is encouraging and should be considered for hearing aid fittings with older adults at risk of social isolation and depression.

 

References

Akeroyd, M. (2008). Are individual differences in speech reception related to individual differences in cognitive ability? A survey of twenty experimental studies with normal and hearing-impaired adults. International Journal of Audiology 47 (Suppl 2), S53-S71.

Bilger, R., Nuetzel, M., Rabinowitz, W. & Rzeczkowski, C. (1984). Standardization of a test of speech perception in noise. Speech and Hearing Research 27, 32-48.

Desjardins, J. & Doherty, K. (2013). Age-related changes in listening effort for various types of masker noises. Ear and Hearing 34, 261-272.

Desjardins, J. & Doherty, K. (2014). The effect of hearing aid noise reduction on listening effort in hearing-impaired adults. Ear and Hearing 35 (6), 600-610.

Dubno, J., Dirks, D. & Morgan, D. (1984). Effects of age and mild hearing loss on speech recognition in noise. Journal of the Acoustical Society of America 76, 87-96.

Duquesnoy, J. (1983). The intelligibility of sentences in quiet and noise in aged listeners. Journal of the Acoustical Society of America 74, 1136-1144.

Feuerstein, J. (1992). Monaural versus binaural hearing: Ease of listening, word recognition and attentional effort. Ear and Hearing 13, 80-86.

Geller, D. & Margolis, R. (1984). Magnitude estimation of loudness I: Application to hearing aid selection. Journal of Speech and Hearing Research 27, 20-27.

Helfer, K. & Freyman, R. (2008).  Aging and speech on speech masking. Ear and Hearing 29, 87-98.

Hornsby, B. (2013). The effects of hearing aid use on listening effort and mental fatigue associated with sustained speech processing demands. Ear and Hearing 34(5), 523-534.

Ng, E., Rudner, M. & Lunner, T. (2013). Effects of noise and working memory capacity on memory processing of speech for hearing aid users. International Journal of Audiology 52, 433-441.

Plomp, R. (1978). Auditory handicap of hearing impairment and the limited benefit of hearing aids. Journal of the Acoustical Society of America 68, 1616-1621.

Salthouse, T. (1994). The aging of working memory. Neuropsychology 8(4), 535-543.

Salthouse, T. (1985). A Theory of Cognitive Aging. New York, NY: North-Holland.

Sarampalis, A., Kalluri, S. & Edwards, B. (2009). Objective measures of listening effort: Effects of background noise and noise reduction. Journal of Speech, Language and Hearing Research 52, 1230-1240.

Scollie, S., Seewald, R. & Cornelisse, L. (2005).  The desired sensation level multistage input/output algorithm. Trends in Amplification 9, 159-197.

Wingfield, A., Poon, L. & Lombardi, L. (1985). Speed of processing in normal aging: Effects of speech rate, linguistic structure and processing time. Journal of Gerontology 40, 579-585.

Woods, W., Nooraei, N. & Galster, J. (2010). Real-world listening preference for an optimized digital noise reduction algorithm. Hearing Review 17(9), 38-43.