Effects of cognitive functions on a person's speech understanding and hearing aid fitting have received a lot of attention in recent years. Akeroyd examined 20 studies on the association between cognitive functions and speech reception/recognition in background noise (Int J Aud 2008;47[Suppl 2]:S53), and the results, when combined with those of Humes and colleagues, indicate that working memory and the degree of hearing loss were the most effective predictors for an individual's speech reception and recognition in noise. (J Speech Hear Res 1994;37:465; J Acoust Soc Am 2002;112[3 Pt 1]:1122; J Speech Lang Hear Res 2007;50:283.)
A person's working memory capacity generally can be measured in three ways:
- The reading span test uses a sentence presented word by word. The patient decides if the sentence makes sense at the end. He is asked to recall the first or last word of each sentence after reading several sentences. (J Verbal Learning Verbal Behav 1980;19:450.)
- Simon's memory test presents a patient with simple to complex sequences of colors and sounds, and he is asked to repeat them as accurately as possible. (Pisoni DB, Clearly M, Springer Handbook of Auditory Research; New York: Springer, 2004.)
- The visual letter or digit monitoring test presents a patient with a sequence of individual letters visually, and he is asked to judge if any of the consecutively presented letters would form a consonant-vowel-consonant word. Or the patient is presented with a sequence of individual digits, and is asked to identify an odd-even-odd sequence. (Int J Aud 2003;42[Suppl 2]:S77.)
Older adults are reported to have lower working memory capacities than younger adults (40% vs. 70%, Souza, 2012; audiologyonline.com). It is possible that at least part of the reduction in hearing abilities experienced by older adults, as reported by Gordon-Salante and colleagues over the years, can be attributed to reduced cognitive abilities. Older adults are found to have reduced speech recognition abilities when the speech signals are altered or degraded in some way with young adults (i.e., time compressed, frequency compressed, accented, or spectrally degraded). They also reported that older adults and those with hearing loss could benefit from slowing down speech components (e.g., time expansion of consonants, time expansion of temporal cues in speech) or increasing spectral contrast between consonants and vowels, both of which are supported indirectly by studies in amplification.
These results appear to bear a close link to recent findings between cognition and amplification. Gatehouse and colleagues (J Am Acad Audiol 2006;18:604) and Lunner and Sundewall-Thoren (Int J Aud 2007;45:153) found that adults with relatively lower cognitive functions could obtain higher speech recognition scores using hearing aids with slow compression release times (i.e., less temporal envelope alterations and greater temporal contrasts between consonants and vowels) than those with fast release time. Cox and Xu, on the other hand, showed that release times had minimal effects on hearing aid users with higher cognitive functions. (J Am Acad Audiol 2010;21:121.) Users with lower cognitive functions, however, benefited from shorter release time (increase in audibility of soft sounds) when speech was rich in contextual cues. They also performed better with longer release times (less distortions) when speech was low on semantic cues.
Working Memory Capacity May Influence Perceived Effort During Aided Speech Recognition in Noise
Rudner M, Lunner T, et al J Am Acad Audiol 2012;23(8):577
This article expands our current knowledge on the relationship among cognitive abilities, speech perception, and amplification. The authors determined in Experiment I the individual signal-to-noise ratio (SNR) for 16 Danish participants to understand 50%, 80%, 95%, 95%+1 ΔSNR, and 95%+2 ΔSNR, where ΔSNR was defined as the SNR difference between 80% and 95% speech recognition. The participants had mild to moderately severe sensorineural hearing loss and an average age of 63.5 years with a standard deviation of 12.1 years. They listened to sentences presented in a modulated and a steady state speech-shaped noise, and rated their listening effort using a visual analog scale. Their working memory capacities were measured using the visual letter monitoring task.
Participants reported greater listening effort in conditions with lower SNRs and in conditions with modulated background noise more than those with steady state noise. Higher working memory was associated with lower listening effort in all conditions with steady state noise but only in the two poorest SNR conditions with modulated noise.
SNRs for 50% or 80% correct sentence recognition of 30 Swedish hearing aid users with bilateral mild to moderate sensorineural hearing loss were tested in Experiment II. Average age was 70 years with a standard deviation of 7.8 years. Participants were randomly assigned to listen through hearing aid with fast or slow compression time constants. They then rated their listening effort at SNRs of -2, 4, and 10 dB in modulated and steady state background noises using the visual analog scale. Their working memory was tested using the reading span test.
Again, participants reported greater listening effort in conditions with lower SNR, regardless of the type of background noise or compression setting. Participants also obtained lower (better) average SNRs for 50% or 80% correct sentence recognition in modulated noise than steady state noise regardless of compression setting. No difference was found in working memory capacity between participants listening with hearing aid with slow time constants and those listening with fast time constants. The former, however, rated greater effort when listening to sentences in two of the three conditions with modulated background noise. This might be because hearing aids with slow compression time constants “preserved” the temporal fluctuations of the modulated noise more than hearing aids with fast time constants and resulted in greater listening effort.
The fact that some participants rated greater listening effort but obtained lower (better) SNR for 50% and 80% correct sentence recognition in modulated noise than steady state noise suggests that recognition performance does not necessarily reflect the listening effort spent on performing the task. Working memory decreased with participants' age when the reading span test was used. Higher working memory capacity was also associated with lower listening effort, indicating that listeners with higher working memory capacity generally require lower effort in any listening condition.
© 2013 Lippincott Williams & Wilkins, Inc.
- An abstract of the article, “Working Memory Capacity May Influence Perceived Effort During Aided Speech Recognition in Noise,” is available at http://1.usa.gov/TfF6Ha.
- Read past Journal Club columns in a special collection at http://bit.ly/JClubCollection.
- Click and Connect! Access the links in The Hearing Journal by reading this issue on our website or in our new iPad app, both available at thehearingjournal.com.
- Comments about this article? Write to HJ at HJ@wolterskluwer.com.
- Follow us on Twitter at twitter.com/hearingjournal and like us on Facebook at www.facebook.com/HearingJournal.