Hearing and Speech Processing in Midlife

Abstract

Middle-aged adults often report a decline in their ability to understand speech in adverse listening situations. However, there has been relatively little research devoted to identifying how early aging affects speech processing, as the majority of investigations into senescent changes in speech understanding compare performance in groups of young and older adults. This paper provides an overview of research on hearing and speech perception in middle-aged adults. Topics covered include both objective and subjective (self-perceived) hearing and speech understanding, listening effort, and audiovisual speech perception. This review ends with justification for future research needed to define the nature, consequences, and remediation of hearing problems in middle-aged adults.

1. Introduction

Problems with hearing and understanding speech in noisy environments become increasingly common as people age, and for the most part these problems do not come on suddenly. Rather, these difficulties often begin to emerge during middle age. Yet most research on aging compares performance of younger adults to that of older adults (> 60 – 65 years of age) with little consideration of what happens in the interim age span. This approach leaves researchers and practitioners with only a limited understanding of the problems faced by adults in this critical age range and an incomplete picture of the time frame for changes in various aspects of auditory functioning and speech understanding. This article provides a review of what is known about changes in hearing and speech understanding in the midlife period (from approximately 45 years of age to 64 years of age) and concludes with clinical and research implications.

2. Hearing sensitivity in middle-aged adults

Most middle-aged adults do not have hearing sensitivity like they had when they were younger. By the time they reach middle age, a substantial proportion of adults have measurable hearing loss. Results of the Beaver Dam Offspring Study demonstrate that hearing loss is relatively common among middle-aged adults, with prevalence of hearing loss (defined as thresholds in either ear between 500 Hz and 4 kHz greater than 25 dB HL) of about 11% of adults 45–54 years of age and 26% of individuals 55–64 years of age (Nash et al., 2011). Wang et al. (2019) reported results of a cross-sectional study of hearing in adults between 30 years and 59 years of age. Of their participants in the 50–54 year range, 9% had thresholds > 25 dB HL bilaterally (mean of thresholds for 1 kHz, 2 kHz, and 4 kHz pure tones); in the 55–59-year age range, prevalence increased to 22%. When considering high frequency hearing (the average of 4 kHz and 8 kHz thresholds), prevalence was even higher, at 14% in the 50–54 group and 44% in the 55–59 group. Likewise, the Busselton Baby Boomer Cohort found that 21% of middle-aged adults (45–66 years) had high-frequency hearing loss (Bucks et al., 2016). Results of large epidemiological studies that use automated speech-in-noise tests instead of pure-tone threshold assessment also suggest that age-related changes in these measures begin by middle age in many individuals (Dawes et al., 2014; Nachtegaal et al., 2009). Taken as a whole, these epidemiological studies point to the fact that hearing begins to change relatively early in the aging process in many adults.

As discussed below, a proportion of individuals who have what would be considered “normal” audiograms report considerable difficulty understanding speech in adverse listening situations. One reason this might be the case is that clinical audiometric pure-tone testing generally only extends up to 8 kHz. It has long been known that frequencies above 8 kHz are vulnerable to both noise exposure and early aging (e.g., Jilek, Suta and Syka, 2014; Yeend et al., 2017). Several studies have demonstrated that pure-tone thresholds above 8 kHz, often referred to as extended high frequency (EHF) thresholds, can contribute to explaining a portion of inter-subject variability in speech understanding in noise. Badri et al. (2011) compared the hearing thresholds of two groups of adults between 18 years and 50 years of age who either did or did not self-report speech perception problems in noisy situations. Pure-tone thresholds at conventional audiometric frequencies did not differ significantly between these two groups, but the group with self-reported hearing problems had higher EHF thresholds. These results suggest that elevated EHF thresholds could contribute to people’s perceptions of difficulty understanding speech in noise. Other studies measuring objective speech perception have demonstrated the value of EHFs for listening to one talker in the presence of competing talkers (Levy et al., 2015; Monson et al., 2019; Moore et al., 2010). Accordingly, Yeend et al. (2019) found that among adults 30–57 years, EHF thresholds together with working memory explained 41% of total variance in a composite speech-in-noise metric.

Why should EHF thresholds affect speech understanding, given that the spectral information that is most important for speech perception is primarily in lower frequencies, and that there is only a limited amount of acoustic energy that is audible in the EHF range when speech is presented at typical conversational levels? EHF energy in consonants can be helpful for perceiving those sounds when information at lower frequencies is not available (e.g. Lippmann, 1996) or when speech is embedded in noise (Apoux and Bacon, 2004). It has been proposed that information in the EHF range serves as a grouping cue to help individuals segregate target voice information from background sounds (Monson et al. 2019). Moreover, providing energy in an extended high frequency range contributes to speech being perceived as more natural (Moore and Tan, 2003). Another consideration is that elevated EHF thresholds maybe a sign of cochlear synaptopathy or “hidden hearing loss” (Liberman et al., 2016) and so may be a marker for people whose speech understanding in noise is compromised by the presence of that condition. The reasons why EHF thresholds are associated with speech understanding are yet to be determined, but what is known so far points to the idea that evaluating hearing sensitivity in this frequency range may help to explain why some middle-aged adults who show little hearing loss on routine audiometric testing report difficulty understanding speech in adverse listening situations.

As the research discussed above indicates, many middle-aged adults have at least some degree of elevation of auditory thresholds. Although measuring pure-tone hearing is of obvious clinical value, it gives us only limited insight into how adequately people can understand speech in real-world situations. Next we focus on research that directly examines subjective and objective speech understanding in middle-aged adults.

3. Speech understanding in middle-aged adults

Pure-tone hearing loss is an inadequate indicator of how much difficulty middle-aged adults believe they experience in adverse acoustic environments. Demeester et al. (2012) found that middle-aged (55–65 years) people with normal hearing sensitivity reported significantly more problems in challenging listening situations than younger adults. In an analysis of NHANES data, Kamil et al. (2014) determined that people in their 50s tended to overestimate their hearing problems, while older adults tended to underestimate them. Similar findings were noted by Bainbridge and Wallhagen (2015), where the prevalence of middle-aged adults reporting subjective hearing difficulty was greater than the prevalence of objective pure-tone hearing loss in this population. This pattern was reversed in older adults. Data from our lab supports this disconnect between self-perceived hearing difficulty and hearing thresholds: Speech and Spatial Questionnaire (SSQ: Gatehouse and Noble, 2004) scores aggregated across several studies showed little difference between older and middle-aged adults in how adequately individuals felt they can function in adverse listening conditions, even though the older participants had substantially more hearing loss and so would be expected to experience more problems in these situations (Helfer, Merchant, and Wasiuk, 2017).

What are the reasons behind this discrepancy between objective and subjective hearing ability in middle-aged adults? Certainly, there are auditory abilities that may be impacted earlier in the aging process (and/or to a greater degree) than pure-tone hearing sensitivity. Tasks that have a substantial temporal component appear to be most sensitive to early aging, including measures of temporal fine structure processing (Füllgrabe, 2013; Grose and Mamo, 2010) and gap detection or discrimination (e.g., Grose, Hall, and Buss 2006; Humes et al. 2010; Leigh-Paffenroth and Elangovan 2011; Lister, Besing, and Koehnke 2002; Ozermal et al., 2016). Changes in the way the auditory system processes fundamental aspects of sounds likely contribute to problems segregating sound sources (e.g., Füllgrabe, Moore and Stone, 2015; Neher et al., 2012; Summers and Leek, 1998), an ability that is of considerable importance when trying to listen to one voice in the presence of other voices. The underlying reasons contributing to a decline in psychoacoustic abilities by middle age have not yet been established, but it should be noted that even very mild peripheral hearing loss may disrupt performance on these (and other) auditory tasks (e.g., Bernstein and Trahiotis 2019; Giroud et al. 2018).

The problems reported by middle-aged people often revolve around understanding speech in complex acoustic environments with multiple talkers. These subjective complaints have been confirmed in research studies demonstrating that middle-aged adults have particular difficulty understanding speech in the presence of speech maskers (e.g., Baskent, van Engelshoven, and Galvin 2014; Cameron, Glyde, and Dillon 2011; Füllgrabe, Moore, and Stone 2015; Glyde et al., 2013; Hannula et al. 2011; Shinn-Cunningham et al., 2013; Tremblay et al. 2015; Wiley et al. 1998). Data from our labs has consistently shown that middle-aged adults’ performance is similar to that of younger adults when the masker is steady-state or fluctuating noise, but closer to that of older listeners when the masker is understandable speech (Helfer and Freyman, 2014, 2016; Helfer, Freyman, and Merchant, 2018; Helfer and Jesse, 2015; Helfer and Vargo, 2009; Jesse and Helfer, 2019).

These functional changes in speech understanding are likely based on a complex interplay of age-related alterations in peripheral auditory, central auditory, and cognitive processing. Increasing evidence also shows anatomical changes in the central auditory system beginning relatively early in the aging process, especially in individuals who have elevated pure-tone thresholds. For example, Neuschwander et al. (2019) recently documented that age-related hearing loss can lead to reduced thickness in auditory cortex in middle-aged individuals. The interaction of hearing loss and cognition is discussed later in this paper.

3.1. Speech perception in face-to-face communication

Many of our conversations take place face-to-face. In these face-to-face conversations, listeners process and combine the information obtained from hearing and seeing a speaker talk in order to determine what was said. Access to audiovisual speech renders recognition more robust (for an overview, see Massaro and Jesse, 2008). Listeners of all ages are more likely to recognize correctly what the speaker said if they have access to both auditory and visual information than when they only receive auditory information (e.g., Jesse and Janse, 2012; Jesse et al., 2000; MacLeod and Summerfield, 1987; Massaro et al., 1986; Ross et al., 2011; Tye-Murray et al., 2016). This audiovisual gain typically exceeds what would be predicted from taking an unweighted sum of the performance scores of auditory-only and visual-only conditions. Its size is in part determined by the relative distribution of information across the modalities, in that audiovisual gains are larger to the degree that the two sources complement each other in the information they provide rather than providing duplicate information (Grant and Walden, 1996; Jesse and Massaro, 2010; Walden et al., 1974).

Not surprisingly, access to visual speech becomes more important when hearing is impaired. In a group consisting of middle-aged and older adults, audiovisual benefits for the detection of target words in a continuous speech stream accompanied by multitalker babble noise tended to be larger for those individuals with a greater degree of high-frequency hearing loss (Puschmann et al., 2019). However, this positive association was largely due to a negative correlation between hearing loss and performance in the auditory-only presentation condition. This latter relationship is to be expected as even in young adults with normal hearing, lower recognition scores for auditory-only presentations tend to lead to larger audiovisual gains. For example, a further analysis of data from Jesse et al. (2000), displayed in Figure 1, shows a negative correlation between the recognition of CID (Central Institute for the Deaf) sentences from auditory-only presentations (presented with speech noise) and the observed audiovisual gain, measured as the difference in recognition scores between audiovisual and auditory conditions (r = −.82, t(69) = 11.73, p < .00001). Information obtained from seeing a speaker can, however, help overcome deficits in speech perception due to age-related high-frequency hearing loss. For example, Puschmann et al. (2019) found that hearing loss only predicted performance in an auditory speech-in-noise condition (despite individually adjusting the speech-to-noise ratio) but not in an audiovisual condition in their group of middle-aged and older adults.

Figure 1.

Re-analysis of data from Jesse et al. (2000) showing the association between word recognition scores for auditory-only presentations in noise and audiovisual gain (AV-A) for 72 participants. Shading represents the 95% confidence intervals.

The ability to extract and process information from visual speech seems to gradually decline across the adult lifespan (e.g., Sommers et al., 2005; Tye-Murray et al., 2016). These age-related changes in speechreading become more apparent in difficult processing situations, such as for the speechreading of sentences rather than of simpler consonant-vowel syllables (Shoop and Binnie, 1979; Walden et al., 1993), especially when the latter only consists of a very small set (Sekiyama et al., 2014). This reduction in the ability to speechread seems to have an early onset, apparent already at mid-life (Shoop and Binnie, 1979; Tye-Murray et al., 2016). Gender differences in speechreading found in young adults can persist in midlife (Alm and Behne, 2015). Both young and middle-aged women seem to be better at speechreading than men. Furthermore, middle-aged women were more influenced by visual speech than their male peers, as measured by proportion of fusion responses in a McGurk effect task. Results showed that women were more likely than men to respond with an alveolar consonant (e.g., /ta/) when presented with incongruent audiovisual stimuli consisting of auditory presentation of a bilabial consonant (e.g. /pa/) while seeing the production of a velar consonant (e.g. /ka/). This gender difference was not linked to hearing loss and was not observed in younger adults.

The ability to recognize speech from seeing a talker’s face thus weakens across the adult lifespan, but individuals who are better at speechreading earlier in their life, such as women, seem to be able to keep their advantage as they age. This result dovetails with other evidence showing that extensive experience with using visual speech earlier in life can help to mitigate deterioration of speechreading in midlife. Middle-aged adults with sensorineural hearing loss that started earlier in life (between 7 and 25 years prior to when they participated in the study) were better speechreaders than middle-aged adults with normal hearing (Pelson and Prather, 1974). Among middle-aged adults with sensorineural hearing loss, the onset of the hearing loss, and not its severity, predicted their speechreading ability (Tillberg et al., 1996). A group of middle-aged adults with earlier onset hearing loss (between 4 years of age and 34 years of age, average 15 years) performed worse than a group with later onset hearing loss (27–55 years of age, average 45 years) when presented with auditory-only speech but outperformed them on visual-only presentations. However, both groups achieved similar recognition rates when tested with audiovisual presentations. Taken together, these results suggest that individuals who presumably have more experience with speechreading at an early age (due to hearing loss) can preserve some of their ability to effectively process information from visual speech as they age.

Few studies directly assessed audiovisual speech perception in middle-aged adults. To be able to compare audiovisual processing across age groups, it is crucial to distinguish the relative contributions of differences in the processing of auditory and visual speech from differences in the integration of information across the two modalities. A large-scale cross-sectional study (Tye-Murray et al., 2016) of adults between the ages of 22 and 92 showed that when controlling for age-related declines in visual and auditory speech processing, the audiovisual benefit remained stable across the adult lifespan. Likewise, despite the presence of age-related decline in speechreading, older and middle-aged adults with bilateral moderate-to-severe hearing loss can achieve equivalently high recognition rates for audiovisually-presented speech materials, when matched on their auditory-only performance (Walden et al., 1993). Similar results were found for older and middle-aged adults with cochlear implants (Hay-McCutcheon et al., 2009). For at least simple tasks involving the recognition of a small set of consonant vowel syllables, middle-aged adults can achieve the same audiovisual benefit as younger adults (Alm and Behne, 2015).

In summary, middle-aged adults retain the ability to benefit from seeing the talker’s face. Age-related differences found in processing audiovisual speech seem entirely due to differences in auditory and visual processing rather than in audiovisual integration (see also Sommers et al., 2005).

4. Listening effort in middle-aged adults

Even if middle-aged adults can understand speech adequately in challenging listening situations, they may need to expend more effort to do so. One framework for examining how effort relates to speech understanding is the Ease of Language Understanding (ELU) model (Rönnberg, 2003; Rönnberg et al., 2013). The ELU model has been used to account for the influence of cognition (particularly working memory) on speech understanding. According to the ELU, speech perception is rapid and automatic under ideal circumstances, necessitating few cognitive resources. When there is a mismatch between the stored representation of speech and auditory input (due to hearing loss or the masking effects of noise), speech understanding is more effortful and requires explicit use of cognitive resources, especially working memory (Rönnberg, 2003; Rönnberg, Holmer, and Rudner, 2019; Stenfeld and Rönnberg, 2009). The ELU (like many other models of working memory) posits that there is a limited capacity system; when more cognitive capacity is required for listening, less remains for encoding information into memory or performing another task at the same time.

Also relevant is the FUEL (Framework for Understanding Effortful Listening; Pichora-Fuller et al., 2016) model, which takes into account the importance of motivation and rewards when considering listening effort. Individuals likely differ not only in how much effort is required for them to successfully complete a task, but also in the amount of effort they are willing to expend on a task, depending on the cost-benefit ratio (Eckert, Teubner-Rhodes, and Vaden, 2016; Matthen, 2016; Richter, 2016). If a conversation is inordinately difficult to follow, a listener may not be willing to devote sufficient effort to understand what is being said, particularly if he/she perceives that the cost of investing the necessary effort is not worth the potential benefit that would be derived from participating in that conversation. Hence, effort has been characterized by an inverted u-shaped curve, with maximum effort expended in the midrange of performance (e.g., Wendt et al., 2018; Wu et al., 2016).

Effort or cognitive load can be estimated as an increase in dual-task costs: the extent to which performance on a task completed in isolation is negatively affected when individuals execute another task simultaneously (e.g., Broadbent 1958). For example, participants may be asked to complete a speech understanding test (the primary task) while also pressing a button in response to another stimulus (the secondary task), and are instructed to optimize their performance of the primary task. Listening effort is quantified by the difference in performance on the secondary task (e.g., increased button-press reaction time) when completing both tasks together as compared to when just completing the secondary task. A decline in performance on the secondary task during dual-task conditions is taken to indicate an increase in effort needed to maintain performance on the primary task. These dual-task costs are assumed to be caused by competition for resources: when the primary task is more difficult or effortful, fewer resources remain for performing the secondary task (e.g., Kahneman, 1973). Results of research using dual-task paradigms support the idea that understanding speech in noise is more effortful for older than for younger adults (e.g., Desjardins and Dougherty, 2013; Gosselin and Gagne, 2011; Tun, McCoy, and Wingfield, 2009). This increase in listening effort with aging, as estimated from dual-task costs, begins in middle age (Cramer and Donai, 2019; Degeest, Keppler, and Corthal, 2015; Xia et al., 2015). For example, Degeest and colleagues (2015) measured the impact of performing a visual memory task (the secondary task) on recognizing five-digit utterances in the presence of steady-state noise (the primary task) in adults 20–77 years. Listening effort, as estimated by dual-task costs, began to increase between 40 and 45 years of age.

One of our labs has recently focused on using more realistic dual-task studies to index listening effort in younger and middle-aged adults. The first of these investigations (Helfer et al., 2020) used a balancing-while-listening task. In this study, participants performed a speech understanding task with sentences presented in steady-state noise or in a two-talker competing speech background at two different signal-to-noise ratios (SNRs). During this study, participants were instructed to stand as still as possible while either in normal stance, with their feet approximately shoulder-width apart (the easy postural condition) or in tandem stance, with one foot behind the other (the harder postural condition). Results were consistent with the idea that participants were able to maintain their level of performance on the speech perception task during dual-task conditions, but at a cost to postural control. Although performance on both the speech perception task and the balance task were similar between groups when each task was administered in isolation, dual-task costs for postural control were larger in middle-aged (vs. younger) adults, especially when the masker was competing speech. Moreover, postural control was poorer for both groups of participants when the speech perception task was conducted at the more adverse SNR, vs. when the SNR was more advantageous. Speech perception did not differ significantly between single- and dual-task conditions. Overall, the results suggested that understanding sentences in the presence of competing speech was especially effortful for our middle-aged participants, as demonstrated by the magnitude of dual-task costs.

As discussed earlier, analysis of self-report data that include questions about how well people believe they perform different listening situations shows that middle-aged people tend to rate their functional hearing ability as poorer than what would be expected based on their pure-tone thresholds (Kamil et al., 2014; Bainbridge and Wallhagen, 2015). Another dimension that can be examined with self-report data is how much effort people report they need to expend in adverse listening situations. Two of the questions on the abbreviated version of the SSQ that is used in one of our labs can give insight into the cognitive load or effort that individuals feel they expend when listening in the real world: “Do you have to put in a lot of effort to hear what is being said in conversations with others?” and “Do you have to concentrate very much when listening to someone or something?”. We have collected responses to these questions in nine studies that included middle-aged and/or older adults (n = 69 participants ≥ 65 years of age and 169 participants 40–64 years of age). Although, as anticipated, high-frequency pure-tone thresholds were significantly higher for the older group (F(1,282 = 30.24, p < .001), responses to these two SSQ questions were essentially equivalent across age groups. ANOVA on these data showed that the difference between groups was not statistically significant for either question (Effort question: F(1, 236) = 3.18, p = .076); Concentration question: F(1,236) = 2.4, p = .120). In fact, self-assessed effort and concentration changed little between ages 45 and 75 years (Figure 2). Consistent with the overall SSQ results reported earlier in this paper, middle-aged participants in our studies perceived that they expended about as much effort while listening as our older participants who had considerably more hearing loss, and that they needed to concentrate as much as did the older adults.

Figure 2.

Responses to the SSQ effort question (blue dotted line), the SSQ concentration question (red dashed line), and better-ear pure-tone average (yellow solid line) across 238 adults between 40 years of age and 75 years of age. Error bars represent the standard error.

5. Cognition and hearing in middle age

A portion of the inconsistency between objectively-measured hearing loss and middle-aged adults’ subjective problems understanding speech could be related to changes in cognition associated with aging. The situations that are most problematic for middle-aged adults tend to entail listening to one talker embedded in a background of other intelligible speech messages. These environments likely require more cognitive mediation than when the competition is less complex noise. Indeed, complaints about cognitive changes (especially short-term memory and concentration) are common among middle-aged adults (e.g., Begum et al., 2014; Vercammen et al., 2017), and it is feasible that changes in these functions that occur early in the aging process can influence individuals’ ability to cope in complex listening situations. It should be noted, however, that there is abundant inter-subject variability when memory and executive functioning are assessed in middle-aged adults (e.g., Gunstad et al., 2006).

One question that has been addressed in a number of studies is the extent to which hearing loss in middle-aged adults mediates these cognitive changes. Recent evidence is mixed, with some work suggesting that only weak links exist between hearing and cognition when measured in adults who are middle-aged (Merten et al., 2019; Schubert et al., 2019), especially once other variables such as age, education, and depression are controlled (Bucks et al., 2016). However, a recent analysis of a large sample of individuals (n = 6451, 50–76 years of age, mean 59 years) concluded that the presence of even slightly elevated hearing thresholds (less than what would be considered clinical hearing impairment) was associated with poorer performance for a range of cognitive measures including psychomotor speed, attention, verbal fluency, verbal memory, and global cognitive function. This relationship persisted even after accounting for potential confounding variables such as age, educational level, and the presence of cardiovascular risk factors (Golub et al., 2019).

6. Summary and Future directions

Age-related problems with speech understanding that begin in middle age cannot be explained entirely by the amount of audiometric hearing loss that individuals experience. Underlying changes in cognitive and higher-level auditory processing likely contribute to these challenges. Moreover, middle-aged individuals may need to expend more effort to reach the same level of performance as younger individuals, which also likely helps to explain why middle-aged people report that understanding speech in adverse acoustic environments is difficult. Seeing the face of a speaker can provide valuable information to at least partially mitigate some of these challenges, yet it remains unclear whether access to audiovisual speech reduces or increases listening effort in middle-aged adults (for other age groups, see e.g. Gosselin and Gagné, 2011; Mishra et al., 2013; Sommers and Phelps, 2016).

The types of listening situations that appear to be most problematic and/or effortful for middle-aged adults (understanding speech when there are intelligible competing messages) typically are not evaluated during routine clinical assessment. Current clinical protocols also do not assess how individuals use visual speech cues and so are not accurate indicators of how people function in face-to-face communication situations. Clinically, this points to the perils of over-relying on objective test results and confirms the importance of making sure that we get a clear picture of how middle-aged individuals perceive they function in the outside world.

How should the hearing challenges of middle-aged adults be remediated? The importance of not ignoring these problems is bolstered by the frequently-cited conclusion by Livingston et al. (2017) that hearing loss in midlife is a modifiable risk factor for dementia. Currently, the primary way to treat hearing loss is through hearing aids. Hearing aids can help reduce listening effort (e.g., Johnson et al., 2016; Picou and Ricketts, 2017; Rovetti et al, 2019; Shehorn et al., 2017) and emerging data suggests that treating hearing loss may lead to a reduction in cognitive decline (Amieva et al., 2015; Dawes et al., 2015; Deal et al., 2015; Maharani et al., 2018; Sarant et al., 2020). More research needs to be conducted to determine if these findings hold up under the gold-standard of longitudinal placebo-controlled clinical trials. If that is the case, it would confirm the importance of urging middle-aged adults with hearing loss to seek treatment.

Overall, there is a paucity of research that focuses specifically on speech understanding in middle-age. We challenge researchers to devote more attention to middle-aged adults, not only to address clinical questions but also to further document the nature and time course of age-related changes that lead to problems understanding speech. Doing so will allow for a more complete picture of how, when, and why the ability to process speech deteriorates as people age, providing the basis for developing new methods and protocols for remediation.

Highlights.

• Changes to subjective and objective hearing are common among middle-aged adults.

• Elevated pure-tone thresholds can explain only a portion of objective problems.

• Audiovisual speech processing remains intact with aging.

• Middle-aged adults exert more effort in speech understanding than younger adults.

• Research on middle-aged adults is important to discover pathways to healthy aging.