+44 1223 790975
This study investigated how age-related-hearing loss (ARHL) contributes to memory deficits and whether decreasing listening effort by enhancing the auditory-verbal message can facilitate memory performance. Recall of complex medical prescription instructions presented in degraded (65% time-compressed speech in babble in sound field) and enhanced (120% expanded speech in quiet with insertion earphones) listening conditions was compared for older adults with various configurations of hearing loss to younger adults without hearing loss. In addition, a third group of older musicians (‘expert listeners’) was included. Results demonstrated that enhancements of the auditoryverbal message during encoding facilitated memory at retrieval for all groups, but more so for the hearing-impaired older adult individuals. The older adult musicians showed additional enhancement in listening such that their memory performance was more similar to the younger non-musician than to a group of older adults matched for age and hearing ability. These findings support the effortful listening hypothesis. We propose that ARHL increases the processing load required to efficiently decode the message for communication at the perceptual, lexical and cognitive levels. These processing loads result in fewer attentional and cognitive-linguistic resources available for elaborate encoding for later recall. Enhancements to the auditory-verbal message in an ecologically valid task demonstrated that memory performance can be improved in older adults with hearing loss. These findings lend support to ARHL as a potential underlying causal mechanism contributing to declining memory performance in the aging adult population.
Keywords: Memory; Learning; Aging; Speech perception; Comprehension; Age-related hearing loss; Musicianship; Auditory processing; Aging cognition
There is no doubt that older adults perform worse than younger adults on a variety of different memory tasks . Cognitive explanations for this age-related memory decline have been focused on older adults’ reduced working memory capacity , slowed speed of processing  and decreased inhibitory control [4,5]. However, some types of memory appear to be either less prone to decline as a function of aging or are differentially affected by aging. For example, episodic memory (e.g., the recall of studied lists of words), shows a steep trajectory of decline, compared to semantic memory, (e.g., facts, vocabulary, general knowledge) which shows a gradual increase with age with a shallow trajectory of decline in the oldest adults . Further, even within the same type of memory task (free-recall) there are large within and between-individual variations in performance .
Salthouse summarized the age differences in memory as follows: older adults have been shown to perform more poorly than younger adults on episodic, short- and long-term memory, and memory tasks that require explicit and controlled processes for a free-recall type response (e.g., the stimuli and task that will be used in the present study; recalling a novel set of medical instructions). Conversely, older adults perform similarly to younger adults when the memory is testing semantic information (such as facts or vocabulary), remote memories (from young adulthood), and those tasks that engage automatic and implicit processes for a recognition-type response (e.g. recognizing a synonym for a given word).
Thus, although there is agreement that some aspects of memory decline as a function of age, not all older adults will be similarly affected and not all types of memory abilities will decline to the same degree or with the same pattern . Understanding the source of this variability advances understanding of the causal mechanisms that underlie the differences in performance between and within age groups. A growing body of evidence points to the possibility that reduced cognitive functioning is related to and could be substantially affected by, reductions in lower-level perceptual processing [9-14]. Therefore, the present study considers the hearing-listening changes that occur as a function of aging as a potential source for this variability .
Sensory and cognitive abilities are highly correlated. Understanding the nature of the relationship between cognition and hearing is particularly relevant since age-related hearing loss (ARHL) is the 3rd most prevalent chronic disorder among older adults . Baltes and Lindenberger found that 94% of age-related variance in intellectual functioning was accounted for by perceptual functioning (vision and hearing). These authors concluded that the findings of a strong connection between sensory-perceptual and cognitive function in the aging adult requires investigations into the sources, factors and the mechanisms that are common to both domains .
In order to examine this relationship between ARHL and memory performance, we considered the complex ways in which hearinglistening abilities change as a function of aging. Therefore, we used a definition of ARHL that includes those aspects of hearing-listening that influence signal detection and processing for speech discrimination and language comprehension for communication purposes. ARHL is defined broadly as a combination of those auditory perceptual and processing deficits that occur as a function of age, with these changes beginning early in midlife [17-22].
There are several ways in which aspects of ARHL might interfere with the sub-lexical and lexical processes for communication success and listening ease [23-31]. The increased cognitive load arises from the need to recruit the additional top-down processes such as working memory, inhibition, monitoring, and attention for comprehension of the message for communication [32-40].
Importantly, ARHL is correlated with cognitive decline in the older adult population, specifically on tests of memory and executive dysfunction. Several recent studies demonstrated that the more significant the hearing loss, the greater the risk of developing dementia [41,42]. In addition, greater hearing loss was associated with a faster rate of incident cognitive impairment . However, as  indicate, these findings demonstrate associative relationships through correlational analysis and therefore, do not imply causation or the direction of causation. The authors do suggest further research is needed into the potential causal mechanisms that may account for these associative relationships. The present study considers one suggested causal mechanism, that is, it is the listening effort arising from ARHL which interferes with the ability to encode sufficiently for recall memory performance .
Listening effort has been described previously as those attention and cognitive resources required for perceptual processing that supports speech perception for communication. This listening effort is greater for older adults compared to younger adults [45-47].
Therefore, two related hypotheses were considered for the present study. According to the effortfulness hypothesis [11,47-50] when individuals listen to a degraded signal (speaker, listener or environmental issues), successful speech discrimination comes at the cost of the limited capacity attentional resources for decoding the message for communication success . In a similar manner, according to the Ease of Language Understanding model (ELU) , if the match between the stimuli and the long-term representation of the target in memory is rapid, automatic and implicit, then fewer explicit resources will be needed for comprehension of the message. If we can modify the speech message with enhancements (e.g., time-expansion) that precisely increases the fidelity of the speech message (e.g., increased vowel space, pause lengths, durations of consonant-vowel transitions) that promotes listening ease or a rapid match, then the explicit cognitive-linguistic resources should become more available for perceptual learning, comprehension, and elaborate encoding for later recall. Both of these hypotheses suggest that easier sub-lexical and lexical decoding of the auditory-verbal message for communication should result in more efficient learning and memory encoding for later recall. Also the suggestion is that resources for listening, learning, and remembering processes are limited and must be shared or re-allocated as needed. Therefore, these hypotheses suggest that those with ease of auditory processing (sub-lexical), exceptional listening abilities (precisely tuned neural encoding of pitch, timing and timbre), or greater cognitive-linguistic resources should experience less effort or cognitive load and have more resources to allocate to encoding the information for later recall .
The main objective of this research is to determine whether the cognitive load (the accumulative processing demands) arising from the listening effort is a possible mechanistic pathway through which ARHL influences memory performance decline in the aging adult. Additionally, this study was designed to determine if the effortful listening arising from ARHL is a modifiable risk factor that could be potentially managed by realistic hearing-listening enhancements and/or listening training/expertise. If so, will these factors narrow the gap between the younger and older adults’ memory performances and perhaps provide a mechanism to slow the cognitive decline in older adults?
The present study investigated how age-related-hearing changes might contribute to memory deficits and whether an enhanced message can facilitate memory employing more ecologically valid methods than has been done previously . Auditory enhancements should reduce the listening effort or reduce the cognitive load and free up those resources required for memory encoding resulting in better free-recall memory performance. The research question is, if older adults can listen like younger adults will they remember more similarly to the younger adult, and if younger adults listen like older adults will they remember more similarly to older adults? Ultimately the goal is to investigate whether the interaction of age-related acuity deficits and age-related spectral-temporal processing changes (the timecompressed listening condition) contributes to listening effort, and whether auditory-verbal enhancements and/or listening training/ expertise mitigate these deficits. If the hypotheses are supported, there should be a main effect of listening condition: Relative to the degraded listening condition, the enhanced listening condition will result in better immediate and delayed memory performance. Further, considering the growing empirical support that musicianship enhances the sub-cortical or sub-lexical processes for speech perception, we hypothesized that these more preserved temporal-spectral processing abilities would contribute to listening ease and/or a more rapid match of the stimuli to the target in memory . Therefore, consistent with the ELU  and effortfulness hypotheses , the older musicians should perform more similarly to the younger adults, and demonstrate significantly better memory performance compared to the older nonmusicians. If listening ‘expertise’ further reduces the listening effort then there should be a main effect of group with the younger and older musicians performing more similarly to each other and better in immediate and delayed memory than the older non-musicians.
If experience-dependent perceptual learning or adaptation  is influenced by the enhanced listening condition then listening order and listening condition will interact. This finding would suggest that the previous experience with a higher fidelity speech message effectively decreased the degradation effect so that the within-subject difference for the memory performances in the two listening condition is smaller when enhanced listening condition is heard first compared to degraded heard first .
In addition, if listening effort interacts with cognitive abilities then those individuals with relative strengths in cognitive-linguistic abilities should demonstrate better memory performance than those with relative weaknesses in those areas. Finally, strengths in hearinglistening abilities and/or cognitive-linguistic processes should be associated with better delayed memory performance for the degraded than the enhanced listening conditions.
Previous work in our lab investigated whether an enhanced message (i.e., spoken with a “clear speech” technique) decreased listening effort by promoting speech intelligibility, and improved both comprehension and recall of medical instructions in two groups of older adults, one group who heard the passages in quiet and a second group who listened in noise . The results showed that when older adults with normal to moderate hearing loss heard complex medical instructions in a relatively enhanced listening condition (clear speech) their learning and memory of these passages was better compared to their performance while listening to conversational speech.
Further, explicit cognitive-linguistic abilities (working memory, executive control and lexical abilities) were positively associated with memory performance with a greater magnitude in the suboptimal listening condition (conversational speech). We concluded that, consistent with the ELU  and the effortfulness hypotheses , the relatively degraded listening (i.e., typical conversational speech) required the allocation of cognitive-linguistic resources to decipher the message for comprehension, where the relative ease of listening of the clear speech passages freed up these same limited capacity resources for encoding into memory for later recall. However, since the effects were small, the motivation for the present study is to use more controlled stimulus manipulations for enhancements (to mimic younger adult listening in optimal listening) and degradation (to mimic older adult listening in typically adverse listening) to further explore how aspects of age-related hearing changes interact with cognitive processes and influence learning and memory performance.
Ethics clearance was obtained from Memorial University’s Interdisciplinary Committee on Ethics in Human Research (ICEHR) in accordance with the Tri-Council Policy Statement on Ethical Conduct involving Humans (TCPS-2). This study was carried out in accordance with the recommendations of TCPS-2 guidelines, and ICEHR with written informed consent from all participants prior to engaging in this study. Sixty-one adults participated, divided into three groups: older musicians (21), younger non-musicians (20), and older non-musicians (20). All participants received $10 an hour for their participation. Inclusion criteria: healthy community-dwelling adults (younger 18-30 years; older 55+ years). Exclusion criteria: a known medical event that may affect cognition; failed cognitive screening metric , inadequate vision for completing the experiment; and hearing loss that exceeded the capacity of the speakers (90dB).
Musicians were defined as those individuals who considered themselves to be musicians, had initiated formal musical training by 10 years of age or younger, had a minimum of 12 years musical experience and had been actively engaged in music, currently performing, teaching and/or practicing on average 6 times a week for 1 hour or more daily (Appendix A). They were 55-84 years old (12 females). A musicianship score was calculated based on the responses regarding life-long musical experience/training that were included on the demographic questionnaire (Appendix B). The musicianship classification score was an ordinal scale in which a higher value reflected more experience/training with music. These criteria were established to be similar to studies that have investigated musical training and its impact on hearing and listening performance [56,57] and its relationship with auditory perceptual and processing abilities in behavioral and electrophysiological studies [58-61]. One older musician wore bilateral hearing aids and wore the hearing aids for the entire study except in the enhanced listening in which he wore the 3A E.A.R.toneTM insertion earphones that all the participants used in the experiment for that condition. Forty non-musicians were selected as a comparison group. The two non-musician groups had very minimal to no exposure to music. The younger non-musicians were 20 Memorial University of Newfoundland students, 19-26 years old (12 females). The older non-musicians were 20 community-dwelling adults, 56-84 years old (10 females) (Table 1 for demographic means and standard deviations; Figure 1 for audiogram data).
|Sensation Level dB SLa
|MCL in dB HLb
Table 1: Intensity level of stimuli presentation;Means and (standard deviations in parenthesis) for the intensity level of the stimuli in the degraded and enhanced listening conditions. aSensation Level (SL) is the difference in decibels (dB) of the participants’ speech reception threshold (SRT) and the presentation level in dB Hearing Level (HL) of the stimuli. This dB SL level reflects the intensity level of the stimuli perceived by the participant. bMost Comfortable Loudness listening level (MCL) is the presentation level in which the stimuli were delivered to the speaker in sound field or to the insert earphones. Bolded mean value indicates this group differed from other groups on this variable(**p < .01).
Preliminary measures: No participant was excluded from the study due to vision, hearing or cognitive screening (e.g., passing score on MMSE >23. The hearing-listening and cognitive-linguistic measures obtained for all participants were the same as a previous study,  for a more detailed description of each test.
Hearing-listening measures: Audiometric tests were conducted in a single-walled sound attenuated chamber using a Grason Stadler Instruments Audiometer (GSI-61), Telephonics TDH50P headphones, E.a.r.ToneTM 3A insert earphones and free-field speakers calibrated to specification (American National Standards Institute ANSI S3.62004, 2004). Standardized procedures were used to obtain pure-tone hearing thresholds for right (R) and left (L) ears (Katz, 1978). Pure tone average-4 (PTA4), the average threshold of the four speech frequencies, (0.5 kHz, 1 kHz, 2 kHz, and 4 kHz) was the metric used to indicate degree of auditory acuity deficit consistent with the WHO definition (PTA4>25 dB HL) (World Health Organization (WHO) Prevention of Blindness and Deafness Program, 2014). The Speech Reception Threshold SRT and the phonetically balanced (PB) maxmost comfortable loudness level (PB max-MCL) were used to calculate the sensation level in which participants experienced the stimuli.
The Quick Speech-In-Noise test (QuickSIN): Etymotic Research, Elk Grove, IL;  was the metric used to assess listening-in-noise ability. The Hearing Handicap Inventory for Adults HHIA  was the standardized and normed self-assessment used to determine the individual’s self-perception of the degree to which they experience a handicap due to hearing loss .
Cognitive-linguistic measures: All participants completed the following cognitive-linguistic metrics: Listening span (L-span) a working memory (WM) task that is similar to the reading span measure except sentences are presented auditorily [65,66]. Backward Digit Span (BDS), in which participants hear lists of digits and recreate them in reverse order . Boston Naming Test (BNT) is a standardized and normed confrontation picture-naming task . Verbal Fluency Measure (FAS) correlates with other metrics that measure executive function .
The auditory-verbal stimuli
The fictionalized medical prescription vignettes created and used in this study were the same ones used in  (Appendix C for the two vignettes and the training passage). The vignettes were matched for linguistic and non-linguistic aspects of speech to equate them for complexity, while at the same time maintaining their ecological validity . Each vignette comprised 10 sentences, with 37 critical units (CU) to report. The 37 CU were the content words within each phrase that carried the most important salient meaning in order to use these fictional medications. Critical units may be a single word, compound word or multiple words (e.g., out of reach). The two vignettes were spoken at their original-conversational rate, 192.5 syllables per minute (spm). Avid Pro-tools 8.0.5 computer software was used to manipulate the original sound files for the training passage and experimental vignettes to ensure that the recordings were equated for loudness across the stimuli and throughout the passages via Root Mean Squared (RMS) for amplitude and to create the two listening conditions.
Degraded speech listening condition: Using an algorithm that uses a pseudo-sampling technique to alter the wave file, the original speech sound file was compressed to 65% of the original length, while maintaining normal speech contours so that it sounded naturally fast. At a specified rate throughout the sound file, small acoustic bits were deleted equally in the voiced and voiceless segments of the wave file, the remaining sound file was abutted in time, so that the sound file was compressed relative to its original length. This method deletes segments from both words and pauses at a specified rate throughout; the resultant stimulus retains the temporal patterning of the original preserving the pitch and prosody .
Enhanced speech listening condition: Using the original sound file the speech was expanded to 120% of the original length, while maintaining normal speech contours so that it sounded naturally slow. At a specified rate throughout the sound file, small acoustic bits were reiteratively resampled equally in the voiced and voiceless segments of the wave file, the entire sound file was then abutted in time, so that the sound file is expanded relative to its original length. In this way the duration of the speech elements such as vowel duration and silent intervals were lengthened equally throughout; the resultant expanded speech retains again the temporal patterning of the original speech and preserves the pitch and prosody .
Figure 2 depicts the waveforms of the sentence ‘wash your hands’ from the vignette ‘medipatch’ in its original format, conversational speech technique, 196 (spm), with the clear speech technique (152 spm) , and for this study enhanced, 120% time-expanded (165 spm), and degraded, 65% time-compressed (304 spm) (Praat; Boersma & Weenick, 2014). In this experiment these two listening conditions were relatively degraded/enhanced with respect to each other such that degraded mimics ‘older-listening’ and enhanced mimics ‘youngerlistening’.
Figure 2: The Praat waveforms: Four listening conditions. The waveforms depict the phrase “wash your hands” from the medipatch vignette. The two listening conditions for clear vs. conversational speech : A) 0.97 seconds, original format, conversational speech technique (196 spm); B) 1.24 seconds, spoken with clear speech technique (152 spm). The two listening conditions for this study degraded vs. enhanced: C) 1.19 seconds, 120% timeexpanded (165 spm); and D) 0.62 seconds, 65% time-compressed (304 spm).
Each participant listened to two passages (medipatch and puffer), in two listening conditions (degraded and enhanced), and all preliminary measures and filler/interference tasks (set A and set B). This resulted in eight different combinations of order conditions (e.g., 1st Enhancedmedipatch Set A/2nd Degraded-puffer Set B). The order in which participants performed the listening conditions, passages, or tasks (set A and B) was counterbalanced and participants were randomly assigned to one of the order conditions. Figure 3 which illustrates the procedures for the experiment .
Filler/interference tasks were conducted to: 1) provide a delay between listening and delayed recall; and 2) assess participants on various cognitive and linguistic measures. All the tasks within each set were conducted in the same order. Set A included the (FAS), the BDS, the Philadelphia naming test (PNT) items 1-87 , and a demographic questionnaire. Set B included PNT test items 88-175, the BNT, the MMSE, and the HHIA.
The three dependent measures for the listening conditions were operationally defined and calculated as follows: Learning efficiency was the total sum of the number of CU reported at each of the trials of learning divided by the number of trials required to reach criteria. Criteria were established a priori as either 100% reporting of the 37 critical units or if the participant demonstrated no increase in reporting of the critical units over 3 consecutive trials. In this way there was a single value for learning efficiency during the degraded listening, and a single value for learning efficiency during the enhanced listening condition. Immediate memory was the sum total of the CU that had been reported during any of the learning trials for that listening condition, to the maximum of a possible total of 37 units. Delayed memory was the total number of reported CU after the filler tasks for that listening condition, to the maximum of 37 CU. However, due to space limitations, we focus the analysis and highlight the results for the two memory measures (immediate, delayed) as the area of interest for this study.
Participants were informed of the tasks with a written script that was read aloud to them, while they read along. They were instructed that they would have multiple trials to learn each vignette and to repeat all that they had heard and remembered after each trial of listening. Participants were instructed that gist reporting was acceptable but were encouraged to use as close to verbatim as possible. The participants were not under any time constraint. Responses were spoken aloud and were audio-recorded for later transcription and off-line scoring by a research assistant blinded to listening condition and expertise group. A training vignette was used to ensure the participant’s ability to perform the experimental task, and to confirm that the intensity level identified as PB max-MCL was comfortably loud.
Intra-rater reliabilities of coded participant responses
To determine the consistency and accuracy of the coding of the participant sound files, a research assistant, blinded to the listening condition, coded all the participant files and then re-coded 20% of the total of the files randomly selected from the experiment. Intra-rater reliabilities for coding of blinded scoring were assessed using intraclass correlation coefficient (ICC) with a two-way mixed effects model and absolute agreement type . The ICC for single measures for the reported-recalled CU for each trial was 0.99. An ICC value between 0.75-1.00 is considered excellent . The high ICC intra-rater reliabilities suggest that minimal amount of measurement .
Presentation of the auditory condition
The stimuli were routed from a MacBook Pro computer via Apogee One, a studio quality USB music interface, to the auxiliary channels of the GSI-61 to the transducers (insert earphones or free-field speaker). The intensity level was set at each individual participant’s PB max- MCL obtained during the audiometric testing. This individualized audibility level is consistent with an intensity level that reflected their best performance for recognizing a list of open-set words in quiet in a sound attenuated chamber.
Despite the advantages of using MCL in dB HL , the actual sensation levels or hearing levels for the presentation of the stimuli may have varied by group. Therefore, the sensation level that the participants experienced the stimuli was calculated for all participants (e.g., MCL – SRT=sensation level in dB SL). There were no significant differences among the three groups for sensation level by ANOVA with Bonferroni correction. However, there were the expected differences in MCL in dB HL among the groups, F (2,60)=19.48, p<.001. This difference was only between the younger and older groups (musicians/ non-musicians) and not between the two older groups (Table 1).
Degraded speech listening condition – ‘older listening’: The compressed speech was presented binaurally via a free-field speaker calibrated to a 1 kHz tone. The free-field presentation was used for this listening condition to mimic listening in natural listening environments. All participants were seated and positioned 1 meter distance and 0 degree azimuth to the speaker. The degraded speech vignette and competing speech babble noise at +5 dB SNR were routed to the speaker. The single older-musician participant who wore hearing aids did so for this listening condition only.
Enhanced speech listening condition – ‘younger listening’: The expanded speech stimuli were presented binaurally via disposable 3A E.A.R.tone TM insert earphones in quiet. This was intended to simulate enhancements for listening by optimizing SNR benefit easily captured in the natural environment (i.e., heard with either a personal FM system, head phones, or through a looped hearing aid).
There was one between-subjects variable, listening expertise (older musicians vs. younger and older non-musicians) and two withinsubjects variables, listening condition (degraded vs. enhanced) and time delay (immediate vs. delayed). This study used a modification of the learn-relearn paradigm . Participants listened to, immediately repeated what they had heard (immediate memory), and learned the vignettes as precisely as they could over a series of trials (learning efficiency). They then recalled the vignettes after the completion of 20 minutes of interference/filler tasks (delayed memory). The participants completed the study in two sessions. In the first session they completed the vision screening, audiometric tests and the listening span (L-span). In the second session they completed the experiment as well as the other measures of hearing-listening and cognitive-linguistic abilities (included in the interference/filler task sets A and B).
Comparing groups on demographic, hearing and cognitive measures
There were the expected differences on demographic, hearing and cognitive measures among the three groups by ANOVA or Kruskal- Wallis tests and post-hoc tests with a Bonferroni correction for multiple comparisons (where appropriate) as follows.
There were the expected difference in age and auditory acuity between the younger and the two older adult groups, but no difference between the older non-musician and the older musicians. The older musicians reported more experience with musical training than either of the two non-musician groups. Further, there were no differences among the groups on the QuickSIN, HHIA, BDS, or the BNT, all p=ns (Table 2).
|Younger Non- Musician
|Older Non- Musician
|RPTA4 (dB HL)
|LPTA4 (dB HL)
Table 2: Demographics, Hearing, and Cognitive Characteristics (aEducation: self-reported category: 1=some High school, 2=High School, 3=some University/College, 4=University/college degree, 5=Graduate/professional degree. bHealth: self-reported category: 1=very poor, 2=poor, 3=good, 4=very good, 5=excellent. cQuickSIN=Quick Speech-in-Noise, dHHIA-Hearing Handicap Inventory for Adults, eMusicianship: ordinal scale 0-10 points (higher number reflects greater musicianship experience), fFAS- verbal fluency-executive function task, gBNT-Boston Naming Test, hBackDigit Span-backwards digit span, iL-Span-Listening span. Bolded mean value indicates this group differed from other groups on this variable (*p<.05, **p<.01).
However, the groups demonstrated the expected finding that the younger non-musicians and the older musicians had better auditoryworking memory compared to the older-non-musicians (L-span). In addition, the older musicians demonstrated better executive function (FAS) when compared to the two non-musician groups, whose FAS scores did not differ from each other. These findings of superior auditory-working memory (L-span), executive control (FAS) in addition to other cognitive-linguistic abilities (auditory attention) of musicians compared to non-musician groups are consistent with previous research (58-61,77,78).
Order of experiment effects
There were 8 different orders in which the participants completed the experiment. To determine whether the order of the conditions affected performance, a series of mixed design ANOVAs were conducted. Learning efficiency, immediate memory and delayed memory scores were analyzed, with a 2 (listening condition: degraded vs . enhanced) × 2 (listen order: degraded first vs. enhanced first) × 2(passage order: medipatch first vs. puffer first) × 2 (interference/filler task set order: Set A first vs . Set B first) mixed factors ANOVA, with listening condition as a within-subjects factor, and the three order variables as between-subjects factors. This was conducted for each of the dependent variables separately. There was no effect of order or interactions for passage or interference/filler task set on Learning efficiency, Immediate or Delayed memory performance.
Listening condition order and Listening condition interactions: There was an interaction between listening condition order (degradedenhanced vs. enhanced-degraded) and listening condition on learning efficiency, F(1, 53)=10.66, p=.002, and on delayed memory, F(1, 53)=7.19, p=.01, however not on immediate memory, (p=ns). The interactions of listening order and listening condition on learning and delayed memory are as follows: Performance was always better for the subgroups who experienced either of the listening condition as their second task compared to the subgroups who experienced that same listening condition as their first task.
Learning efficiency was better for second versus first listening condition in both the degraded, Mdegraded 1st=20.20, SD=5.65; Mdegraded 2nd=22.82, SD=5.21; and the enhanced listening, Menhanced 1st=23.47, SD=4.47; Menhanced 2nd=25.08, SD=4.08.
Delayed recall was better during the second versus first listening condition in both the degraded listening, Mdegraded 1st=23.97, SD=6.07; Mdegraded 2nd=25.30, SD=5.27; and the enhanced listening, Menhanced 1st=27.03, SD=4.92; Menhanced 2nd=28.71, SD=5.31.
This reflects general learning-practice effects which interacted with the listening condition such that the difference between learning efficiency and delayed memory performance in the two listening conditions was greater for the subgroups who had degraded-1stenhanced- 2nd order compared to the subgroups who had enhanced-1st-degraded-2nd order. As a result of these interactions, listening order was entered as a covariate for further statistical analyses.
Listening condition, listening expertise and interaction effects on immediate and delayed memory performance
Immediate and delayed memory scores were analyzed with a 3 (listening expertise groups: younger non-musicians, older musicians, older non-musicians) × 2 (listening condition: degraded, enhanced speech) × 2 (delay: immediate, delayed) mixed design ANOVA in which listening condition and delay were entered as the repeated measure within-subjects variables and listening expertise group was a between-subjects variable.
Main effects of delay (immediate and delayed memory), listening condition, and listening expertise groups: There were main effects of delay, F(1,57)=36.73, p<.001; listening condition, F(1,57)=19.38, p<. 001; and group, F (2,57)=10.97, p<.001 on memory performance. Participants reported back immediately a greater number of CU (+4.97) compared to their delayed recall performance. The enhanced listening condition increased recall on average by approximately 2.45 CU (Table 3). The younger adults recalled more than the older nonmusicians (+5.34 CU) p<.001, and older musician groups (+3.02), p=. 03; and overall memory performance between the two older groups, (musicians, non-musicians) (+2.32 CU) was better in the older musicians (p=.04).
Table 3: Expertise group for Learning Efficiency, Immediate and Delayed Memory performance in degraded and enhanced listening conditions. Means and Standard Deviations (CU) (M=Means and SD=(standard deviations in parenthesis).
Interaction effects of listening condition, delay, and expertise group: There were significant interactions of delay × group, F(2,57)=5.11, p=. 009; delay × listening condition, F(2,57)=4.79, p=.03; and for listening condition × expertise group, F(2, 57)=3.41, p=.04. There was no threeway interaction.
The interaction of delay (immediate, delayed) and listening condition (degraded, enhanced) is such that (for all three groups combined) in the immediate memory performance there is less difference between the two listening conditions (1.88 CU); whereas delayed memory performance results in a larger difference (3.03 CU). Planned follow up paired samples t-test confirmed this interaction as significant, t(60)=2.075, p=.04.
The interaction of delay by group is that there were the smallest numerical differences between immediately repeating the passages (degraded and enhanced combined) and then later recalling these passages in the younger non-musicians (7.85 CU) compared to older musicians (10.14 CU) and compared to older non-musicians (11.8 CU).
The interaction of listening condition by group is that the groups were differentially affected by the two listening conditions for their overall memory performance (immediate and delayed combined). There were the smallest numerical differences in overall memory performance between the two listening conditions in the younger nonmusician group (2.05 CU); followed by the older non-musicians (5.00 CU); followed by the older musicians (7.67 CU) (Figure 3).
Narrowing the gap: Does ‘younger listening’ result in functional memory performance for recall of medical instructions that is more similar to the younger adult’s memory performance?
Figure 3 depicts the immediate and delayed memory performance in the degraded and the enhanced listening condition by group and shows the predicted trend based on the effortfulness hypothesis. Overall the younger group demonstrated the highest recall scores followed by the older musicians followed by the older non-musicians. The interaction of listening condition by group was that this relationship of memory performance and group changes depending on the listening condition. In order to answer the research question above, in relation to the hypotheses proposed, and to further examine the interaction of listening condition x group, we performed planned follow up comparisons with an ANOVA (Bonferroni correction for multiple comparisons) and examined the groups’ delayed recall by listening conditions. This was conducted only for the delayed memory performance since it was the ecologically valid metric that captures the functional memory performance for the experimental task (e.g., episodic free-recall of novel medical instructions after a delay).
However, first to address the possibility that the significantly better auditory working memory abilities (L-span), the executive function (FAS) and the greater amount of education of the older musicians may have fully accounted for the group differences for delayed memory performance, an ANCOVA with education, L-span and FAS as covariates for delayed memory (with Bonferroni correction) was conducted. The significant differences among the groups for the overall delayed memory performance when controlling for education, auditory working memory and executive function, F(2, 55)=6.62, p=. 003, η2 =.194; and without the covariates entered, F(2,58)=11.72, p<. 001, η2=.288 remained the same, however the effect size decreased with the covariates in the analysis. Figure 4 shows the interaction of listening condition × group for delayed memory performance with the auditory working memory (L-span) and listening order entered as covariates. Since removing the covariates did not change the results, the following analysis was performed without the covariates.
Figure 4: Delayed memory performance: Group x listening condition interaction; Degraded (65% time-compressed in noise) and Enhanced (120% time-expanded in quiet) listening conditions for younger non-musicians and older musicians and older nonmusician groups (estimated marginal means). Covariates appearing in the model are evaluated at the following values: L-span=22.02; Listen-order=1.49. Error bars are standard error of the mean.
Delayed memory performance in degraded “older listening” for typically adverse listening conditions: In the degraded listening condition there was a difference among the groups, F(2, 58)=9.22, p<. 001, η2=.241. The younger group demonstrated better delayed recall compared to the older non-musicians (p<.001) and the older musicians (p=.01). There was no difference between the two older groups in the degraded listening condition (p=ns). In the degraded listening condition, the younger group demonstrated the highest number of critical units recalled in delayed memory performance compared to both older adult groups, musicianship did not significantly improve delayed memory performance when the listening was degraded for older adults.
Delayed memory performance in enhanced “younger listening” for optimized listening conditions: In the enhanced listening condition there was a difference among the groups, F(2,58)=8.61, p=.001, η2=. 229. The younger non-musicians demonstrated better delayed recall compared to the older non-musicians (p<.001), but not compared to the older musicians (ns). Further, the older musicians demonstrated a higher number of critical units recalled compared to the older nonmusicians (p=.05). In this condition, the younger non-musicians and the older musicians performed similarly in delayed memory performances and both of these groups performed better than the older non-musicians.
Figure 3 reveals another interesting trend. The older musicians’ delayed memory performance in the degraded listening appeared to match the older non-musicians’ memory performance in the enhanced listening condition. A post hoc independent t-test was used to confirm this visual trend. Indeed, results indicated that when older musicians listened in the degraded listening condition their delayed memory performance was not significantly different from the older nonmusicians when they listened in the enhanced listening condition, t(39)=0.479, p=.64.
A degraded listening condition: 65% time-compressed speech in noise (speech babble at +5 dB SNR), presented via loud speakers, in number of critical units reported (CU). benhanced listening condition: 120% time-expanded speech in quiet presented via insertion earphones, in number of critical units reported. Bolded mean value indicates this group differed from other groups on this variable (mean difference is significant at the 0.05 level with Bonferroni correction).
Delayed memory performance and the relationship with hearing-listening and cognitive-linguistic abilities
Correlational analyses were conducted to further explore the contribution of the individual’s hearing-listening (LPTA4 and RPTA4, QuickSIN, and the HHIA) and cognitive-linguistic abilities (L-span, FAS, BDS, and the BNT) on delayed memory performance in the degraded and enhanced listening conditions for the entire sample. Further, since the amount of education was significantly higher for the older musicians compared to the younger/older non-musicians, the education variable was also entered into the correlation analysis to determine if this variable predicts memory performance. Education did not significantly predict delayed memory performance in enhanced r=03, p=.79; or degraded listening r=.23, p=.07. The variables of interest and/or significance appear in Table 4. The significant correlations described below would all be considered to reflect small to medium effect sizes [77-79]. However, generally the magnitude of the relationships between hearing-listening and cognitive-linguistic factors with the delayed memory performance became smaller and/or non-significant when the listening condition was more favorable.
Table 4: Correlation analysis between delayed memory performance in the degraded and enhanced listening conditions and hearing and cognitive abilities – entire sample. (*p < .05, **p < .01)
Hearing-listening abilities and cognitive-linguistic abilities: There were no significant correlations between auditory acuity (LPTA4 and RPTA4) and delayed memory performance in the degraded and enhanced listening conditions. However, there were significant negative correlations for listening-in-noise (QuickSIN) and selfperception of hearing handicap (HHIA) with delayed memory in the degraded but not in the enhanced listening condition.
There were significant positive correlations for both the L-span and the FAS scores with delayed memory performance in the degraded listening condition; however, in the enhanced listening, only L-Span remained significant, but the magnitude of this effect size became smaller. There were significant positive correlations for the BNT and delayed memory performance in the degraded and in the enhanced listening conditions. The magnitude of this effect was slightly numerically larger in the enhanced listening condition but this did not change the significance.
The purpose of this study was to examine how auditory perception and processing of a relatively enhanced vs . degraded speech message affected immediate and delayed memory performance in three groups. This was examined by using auditory-verbal stimuli that were precisely manipulated to mimic conversational speech as it would be typically perceived by older adults with ARHL and to mimic how younger adults perceive speech. We tested three groups of adults who differed by age, auditory acuity, and musical training/expertise. The method was based on real-life listening scenarios that were degraded in a manner that would be considered to be a typically adverse listening scenario . As well, enhancements were constructed in a manner that could be reasonably obtained in vivo by using either a hearing aid or personal listening device. The materials chosen (medical prescription information) were intentionally relevant, somewhat familiar and important to all adults.
The results supported the predictions as there were main effects of listening condition, group and delay. Also there were interactions of delay with listening condition, delay with group, and group with listening condition. Enhanced relative to degraded listening resulted in better memory performance. The groups’ overall performance was significantly different from each other, with the younger non-musician group performing better than the older musicians, and the older musicians performing better than older non-musicians, a musicianship benefit. As well the groups were differentially affected by the two listening conditions, the younger group had the smallest difference between the two listening conditions, the two older groups had larger differences, with the older musicians demonstrating the largest difference. The interaction of listening condition x group was evident in how the two older groups performed in the two listening conditions in relation to each other and to the younger group. As predicted, the older musicians performed more similarly to younger non-musicians when the fidelity of the message was enhanced. When the message was degraded, the two older groups performed less well than the younger group and more similarly to each other (Figures 3 and 4).
Ultimately one of the research questions was whether listening training/expertise commensurate with musicianship mitigates the sources of adverse or effortful listening (speaker, listener or environmental issues). The positive finding of a musicianship benefit supports both the ELU  and/or the effortfulness hypothesis . The results suggest that the presumed more preserved neural encoding of sound that has been found in other studies in older musicians [53,81-88] may decrease the distorting aspects of ARHL, particularly those aspects that disrupt the temporal-spectral processing that supports categorical speech perception (since this was the experimental manipulation used and the two older adult groups were matched for age and for auditory acuity). This potentially more precise and stable temporal-spectral processing in older musicians compared to older non-musicians translates to an enhanced ability to rapidly capture the acoustic information from word-onsets (syllabic stress) that signals word meaning [34,89]. In this manner, the musicianship benefit (listening training/expertise) mitigates the distortions from ARHL by requiring less acoustic information (they are able to make better use of the redundancies) and fewer cognitive resources (inhibition) such that attention to processing the rapid auditory-verbal message for meaning happens more automatically and implicitly (ELU hypothesis) [90-92]. These cognitive resources become available and are re-allocated for elaborate encoding promoting better learning of and memory for these medical instructions (effortfulness hypothesis).
However, since older musician had more education than the nonmusicians, and better auditory working memory (L-span) and executive function (FAS) than the older non-musicians, these variables needed to be considered regarding the influence they may have had on the differences for delayed memory performances among the groups. The education variable did not predict delayed memory performance in either the degraded or enhanced listening. L-span and FAS did predict delayed memory performance with the magnitude of that relationship becoming smaller in the enhanced listening condition. However, when L-span and FAS were controlled for (entered as covariates), the effect size of the expertise group as a predictor for delayed memory performance did decrease, further supporting the hypothesis that those with strengths in these cognitive abilities would have more resources for encoding for later recall. Overall though the results were highly similar, the main effects did not change (e.g., delay, listening condition, and group) and the relationships among the groups remained the same for the interactions of listening condition by group (Figure 4).
Further, the older musicians’ ‘trained/expert listening’ seemed to further enhance their memory performance relative to the older nonmusician group, providing a musicianship benefit demonstrated by the following: 1) Older musicians benefitted more so from the enhancements. 2) Older musicians with ARHL performed more similarly to the younger adults in the enhanced listening condition and significantly better than the older non-musicians’ even though the older groups were matched for age and hearing loss. 3) The older nonmusicians’ delayed memory performance in the enhanced listening was not significantly different from the older musicians’ in the degraded listening condition. These findings suggest that the ‘trainedexpert listening’ of the older musician preserves neural encoding in a way that may be similar to the experimental enhancements used in the effortless listening condition. However, since the temporal processing ability of the older musicians in this study was not directly assessed, the suggestion that it is an enhanced temporal processing ability of the musician that enhanced memory performance in this study is purely speculative. Musicianship may have provided a temporal, a spectral, or a temporal-spectral benefit.
Summary and implications for hearing-listening and cognitive-linguistic abilities and delayed memory performance
Despite finding no relationship between PTA4 and delayed memory performance, there were the expected negative correlations in the metrics that examine other aspects of listening abilities. Listening-innoise ability (QuickSIN) and self-perception of hearing handicap (HHIA) were significantly correlated with delayed memory performance. The magnitude of this effect was larger in the degraded listening condition and became non-significant in the enhanced listening condition.
These findings suggest that the acuity deficits (LPTA4, RPTA4) were not associated with the listening effort, likely since the stimuli were sufficiently audible for both listening conditions (individualized MCL) and the groups were equated on the audibility of the stimuli (i.e., the sensation levels did not differ by group). Further, these findings are consistent with studies that demonstrate that poorer speech recognition scores and reports of effortful listening are not consistently predicted based on the audiometric profiles for older adults .
In relation to cognitive-linguistic characteristics, there were positive correlations between working memory, executive function and lexical abilities and delayed memory performance. The delayed memory scores were positively correlated with strengths in these variables, more so when listening was more difficult as in the degraded listening condition but became less significant in the enhanced listening.
As a within-subject design, differential effect sizes for the relationships of these hearing-listening and cognitive-linguistic measures for delayed memory performance in the two listening conditions suggests that these factors play a greater role when the listening is more adverse or effortful and a lesser role and sometimes non-significant role when the listening is effortless .
Further the enhanced speech in this study promoted better learning and memory performance in a similar manner that the clear speech style did in . The slower rate, increased pause lengths particularly at the salient linguistic boundaries in which they were spoken, and an increase in the acoustic information for the vowels (increased vowel space) enhanced the temporal-spectral aspects of the stimuli such that it was more similar to how the younger adult perceives speech. Relative to younger adults with normal hearing, older adults with normal audiograms have relatively less stable and less precise temporal processing of specific speech cues such as timing, frequency and harmonics . These auditory temporal-spectral processes are necessary for sub-lexical and lexical processing of the message and the discrimination of the regularities in the speaker’s voice and speech pattern  including the ability to detect prosody for syllabic stress for word recognition [88,93]. This instability of the acoustic information, found in older non-musicians, would therefore, interfere with the ability to detect the regularities of the speaker’s speech and voice patterns and therefore, further compromise identification of the targeted speaker from the background noise thereby increasing the cognitive load .
Musicianship: Cognitive enhancement and listening enhancement
Several studies that have identified those with more optimized auditory perceptual-processing ability (younger adults, musicians, and bilinguals) using neurophysiological measures have demonstrated that these groups have both superior listening-in-noise ability and better executive function compared to older adults, non-musicians, and monolinguals [57,60,61,78,94-96]. However, it has yet to be determined whether this positive associative relationship between executive function and those with superior auditory-neural encoding is a causal one and if so the direction of the causation. This study was not designed to adjudicate cause or direction of musicianship benefit with superior executive function (e.g., the significantly higher FAS scores for the older musicians compared to the two non-musician groups in this study). Musicians may have greater cognitive-linguistic resources, better auditory working memory (L-span) and/or executive control, which may serve to increase implicit, rapid-automatic processing of sound (verbal and music alike). Potentially they experience greater ease in understanding the message (ELU hypothesis) and/or this ease in sound processing further promotes an affinity toward music (nature or expertise). However, it is still likely that the musician’s superior neural-encoding of sound arises from their training (nurture or fine tuning of the neural networks), perhaps by allowing for more efficient adaptation to the speaker’s speech and voice, which results in decreased cognitive-linguistic resources required for processing of the sound so that re-allocation of these same processes can be employed for memory encoding (effortfulness hypothesis).
Learning-practice effects: Order of listening condition influences learning efficiency and delayed memory performance
Learning effect benefit can be defined as better performance with greater experience of a task or test compared to one’s performance with no previous experience . Learning effects were expected due to the nature of the repeated measure design of this experiment. Since learning effects are often evident in clinical tests and experimental studies, and can arise from decreased anxiety and strategies employed to perform in subsequent trials, listening condition order was counterbalanced, examined, and controlled for in the analysis [98,99].
The results here further support the idea that experience-dependent perceptual learning or adaptation to the speaker’s speech and voice pattern is influenced by the fidelity of the message . When comparing the within-subject differences between the two listening conditions, there was a much smaller and non-significant difference when enhanced listening was heard first, compared to the larger and significant difference when degraded listening was heard first (for both learning efficiency and delayed memory). Enhanced-1st listening condition narrowed the gap between the two listening conditions, whereas degraded-1st widened the gap between the two listening conditions. In this way, the ability to adapt to the speaker’s speech rate, prosody and intonation positively influenced performance such that perceptual experience-dependent learning decreased the degradation effect to non-significance or marginally significant when enhanced listening was heard first. However, for the subgroup that heard the degraded message first, there remained a very significant difference between their performances for the two listening conditions.
This study demonstrated that when younger and older groups experienced effortful listening they recalled less of the information, and when they experienced effortless listening they recalled more of the information.
Aspects of ARHL, such as listening-in-noise ability and perception of hearing handicap were negatively related to delayed memory performance with a greater magnitude of that effect when the listening was effortful compared to when it was relatively more effortless. Cognitive-linguistic abilities were positively related to delayed memory performance with greater magnitude of that effect when the listening was effortful compared to when it was effortless. These findings demonstrate that when the individual experiences effortful listening, those cognitive-linguistic abilities are employed for the comprehension of the message for communication purposes, perhaps in a compensatory manner . Those individuals with a greater capacity or efficiency in employing explicit use of these resources will have residual resources for subsequent memory encoding.
Finally, the two older groups’ delayed memory performance in the degraded listening condition not being significantly different from each other suggests an interaction between the ARHL with the temporally degraded stimuli and the noise. This last finding suggests a confounding effect of the degraded listening condition on memory performance. It appears that these degrading factors reached a threshold in which an instability or inefficiency to employ compensatory processes to offset the impact of ARHL resulted. This indeed may be the scenario the older adult experiences in listening in the noisy real-world environment. Finally, the results of the older musicians’ immediate and delayed memory performance not being significantly different from the younger adults’ in the enhanced listening condition suggests effortful listening arising from distortion due to ARHL may be a causal underlying mechanism contributing to memory decline. This suggests that the accumulative processing loads required to efficiently decode the message for communication at the perceptual, lexical and cognitive levels (e.g., cognitive load) may be a possible mechanistic pathway in which ARHL influence declining memory performance in the aging adult.
Limitations and future directions
Although this study did measure musicianship with an ordinal scale and identified the groups based on this metric, we did not have an independent measurement of temporal-spectral auditory processing abilities for any of the participants. The younger adults with normal hearing and the older musician group were presumed to exhibit the better temporal-spectral processing abilities previously reported in several other studies when compared to a group of older nonmusicians (with and without hearing loss) [85-87]. The investigation of the influence of the temporal-spectral manipulation of the stimuli on recall performance is consistent with an age-related deficits for temporal-spectral processing (degraded) and musicianship benefit of the more preserved temporal-spectral processing (enhanced). Ideally, obtaining precise and individualized neurophysiological measures of the participants’ auditory-neural encoding of sound or behavioral measures of gap detection [101,102] would have allowed for a more direct analysis of how temporal-spectral processing ability may have influenced memory performance in the two listening conditions in relation to the other hearing-listening and cognitive-linguistic variables. The effortfulness hypothesis would predict that stable and precise temporal-spectral processing would positively correlate with learning and memory performance perhaps more so in the degraded listening.
The results clearly demonstrate a musicianship benefit, however, what is less clear is how musicianship contributed to these findings. Musicianship may have contributed to better learning and memory performance either from domain specific bottom-up auditory perceptual/processing strengths (better temporal-spectral processing), and/or from top-down domain general cognitive abilities (e.g., experience-dependent perceptual learning, attention, inhibition) arising from musical training (nurture) or expertise (nature). Further, consideration of other latent variables not measured that are commensurate with musicianship would also need to be examined in future. For example, musicianship may demand active social engagement that potentially enhances hearing-listening, cognitive or emotional factors .
Future research that investigates the auditory temporal and spectral processing abilities in younger, middle-aged, and older adults, with and without specialized auditory training (musicians and multi-linguals) [104-109], could then examine how these variables influence the domain general cognitive functions recruited for comprehension of the message and subsequent memory encoding [110-112]. These investigations would elucidate how sensory-perceptual and cognitive factors interact throughout the lifespan, as well potentially identify interventions and/or prevention of cognitive decline in the aging population.
This work was funded by Newfoundland and Labrador Centre for Applied Health Research (Healthy Aging Research Program) and Canadian Institutes of Health Research.