“Come again?” Unraveling degraded speech: an old trick!

Sadly, after the prime of life during adulthood, many of our cognitive abilities tend to decline when we grow older. Older people experience more problems with memory, become noticeably slower in cognitive tasks, and generally do less well on various aspects of fluid intelligence than younger people (Dalton, Cruickshanks & Klein, 2003). Moreover, as aging comes with physical decline in the ears and eyes, it is no wonder that older people typically have more difficulties in coping with complex situations in everyday life, such as trying to buy fruit at a busy market.

“the older we get, the more time we have to increase our knowledge of how the world works, and to polish and improve our language skills”

Nevertheless, as we continue to read and listen over the years, the older we get, the more time we have to increase our knowledge of how the world works, and to polish and improve our language skills. In other words, we become more experienced. In line with this, research has found that while fluid intelligence decreases with age, crystallized intelligence increases (Salthouse, 2004). Among other things, this is reflected by improved language skills and a larger vocabulary. This positive characteristic of aging might constitute a means of compensation to cope better with situations in which cognitive and sensory decline make it difficult to perceive and interact with the environment.

A way to study these compensation mechanisms in hearing research is by using the phonemic restoration paradigm (Warren, 1970), in which people are asked to listen to speech that is interrupted by periods of silence. This type of degraded speech destroys the continuity of the speech, removes critical speech information, and may lead people to hear misleading speech cues that put them on the wrong foot with respect to the interpretation of what is being said. To successfully restore and understand the meaning of such degraded sentences, people rely more on cognitive and linguistic skills, than they do with non-degraded speech (Stenfelt & Rönnberg, 2009). Of course, we don’t encounter speech interrupted by periods of silence on a daily basis. Therefore, in this paradigm, listeners are also presented with speech in which the silent periods are filled with loud, speech-like noise, which mimics the distortions caused by noisy everyday environments (e.g., the busy market). Inserting that noise brings back some of the continuity of speech and removes the spurious speech cues (such as sudden stops and starts), which improves understanding.

However, speech interrupted by periods of noise also leads to some ambiguity: Are parts of the speech indeed missing, or are they simply masked behind the noise? This ambiguity activates cognitive mechanisms that enable us to group together the audible speech segments and, by using our language skills and vocabulary, to fill in the missing segments. This way, people can often still extract the meaning of what is being said. As these so-called top-down restoration mechanisms are activated by the addition of the noise in the previously silent periods, the difference in intelligibility between silent and noisy conditions in the phonemic restoration paradigm reflects our ability to perceptually restore and compensate for degraded speech. We speak of a restoration effect when intelligibility is higher for speech interrupted with noise than for speech interrupted with silence.

In our study (Saija, Akyürek, Andringa, & Başkent, 2014) we tested whether older people are still able to rely on their preserved crystallized intelligence to restore degraded speech. This might be a tall order, because previous research showed that older people with hearing loss have a lower restoration effect than younger people without hearing loss (Başkent, Eiler, & Edwards, 2010). However, it is possible that restoration was impaired by hearing loss rather than age in this study. Hearing loss may degrade incoming speech in such a way that the restoration mechanisms cannot acquire enough information from the speech segments to group them and deduce the missing words in the silent periods. The addition of noise might then increase the continuity but not the intelligibility of the speech.

We used the phonemic restoration paradigm to test young (ranging from 19 to 26 years old) and older (ranging from 62 to 77 years) individuals with normal hearing. Our results showed that the restoration effect of the older participants was at least as large as that of the younger participants. Additionally, to test the effect of age-related decline of cognitive processing speed on restoration, in some conditions we presented speech that was slowed down with a factor of 2, without distorting the pitch of the speaker. Surprisingly, the older participants showed higher restoration effects than their younger counterparts in some of these conditions where speech was slowed. Most likely, the older participants used the extra processing time to recruit their superior language skills and vocabulary.

“older people are indeed able to […] extract meaning from degraded speech, by exploiting their life-long experience with language”

Based on these results, we can conclude that older people are indeed able to engage compensation mechanisms to extract meaning from degraded speech, by exploiting their life-long experience with language to make better use of the context of audible speech segments (Pichora-Fuller, 2008). The take-home message is that not everything goes downhill when we age: sometimes we are well able to compensate for cognitive decline by making successful use of our past experiences and accumulated knowledge of language and the world.


Relevant Publications

Başkent, D., Eiler, C.L., Edwards, B. (2010). Phonemic restoration by hearing-impaired listeners with mild to moderate sensorineural hearing loss. Hear Res 260:54–62.

Dalton, D.S., Cruickshanks, K.J., Klein, B.E.K. et al (2003). The impact of hearing loss on quality of life in older adults. Gerontologist 43:661–668.

Pichora-Fuller, M.K. (2008). Use of supportive context by younger and older adult listeners: balancing bottom-up and top-down information processing. Int J Audiol 47:S72–S82.

Saija, J.D., Akyürek, E.G., Andringa, T., Başkent, D., (2014). Perceptual restoration of degraded speech is preserved with advancing age. J Assoc Res Otolaryn 15, 139-148.

Salthouse, T.A. (2004). What and when of cognitive aging. Curr Dir Psychol Sci 13:140–144.

Stenfelt, S., Rönnberg, J. (2009). The signal–cognition interface: interactions between degraded auditory signals and cognitive processes. Scand J Psychol 50:385–393.

Warren, R.M. (1970). Perceptual restoration of missing speech sounds. Science 167:392–393.


Note: Image by Mario Mancuso, licenced under CC BY 2.0.

Jefta is a PhD-student in the Faculties of Medical Sciences, Behavioural and Social Sciences, and Mathematics and Natural Sciences. The focus of his research is on investigating links between auditory and visual domains, by mainly focusing on temporal integration and phonemic restoration. He finished his Bachelor’s degree in Artificial Intelligence in 2010, and his Master’s degree in Human Machine Communication in 2012.

You may also like

Leave a comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.