Card playing enhances speech perception among aging adults: comparison with aging musicians

Abstract

Speech perception and auditory processing have been shown to be enhanced among aging musicians as compared to non-musicians. In the present study, the aim was to test whether these functions are also enhanced among those who are engaged in a non-musical mentally challenging leisure activity (card playing). Three groups of 23 aging adults, aged 60–80 years, were recruited for the study: Musicians, Card players, and Controls. Participants were matched for age, gender, Wechsler Adult Intelligence Scale-III Matrix Reasoning, and Digit Span scores. Their performance was measured using auditory spectral and spatial temporal order judgment tests, and four tasks of speech perception in conditions of: no background noise, background noise of speech frequencies, background noise of white noise, and 60% compressed speech. Musicians were better in auditory and speech perception than the other two groups. Card players were similar to Controls in auditory perception tasks, but were better in the speech perception tasks. Non-musician aging adults may be able to improve their speech perception ability by engaging in leisure activity requiring cognitive effort.

Introduction

A common complaint among the elderly is difficulty in understanding speech, especially when it is rapid, or accompanied by background noise or multi-talkers (Ben-David et al. 2011; Fostick et al. 2013; Heinrich et al. 2016; Taitelbaum-Swead and Fostick 2016). This difficulty can be explained by age-related changes in several mechanisms, such as a decline in hearing sensitivity (Fitzgibbons and Gordon-Salant 2010; Souza et al. 2007), poor temporal resolution (Babkoff and Fostick, in press; Ben-David et al. 2012; Fitzgibbons and Gordon-Salant 2010; Fostick and Babkoff 2013; Fostick et al. 2014), and cognitive slowness (Heinrich and Schneider 2011; Heinrich et al. 2015; Pichora-Fuller and Souza 2003).

Studies show that engagement in musical activity enhances cognitive performance in general, and speech perception in specific (Parbery-Clark et al. 2009a, b, 2011). This enhancement is hypothesized to be related to mechanisms of auditory perception through mechanisms of encoding (Alain and McDonald 2007; Snyder and Alain 2005), attention (Alain and Bernstein 2008; Carlyon 2004), stream segregation (Ben-David et al. 2012; Ben-David et al. 2016; Ezzatian et al. 2015; Getzmann and Näätänen 2015), and information processing (Herholz and Zatorre 2012; Humes 2007). Along this line, musicians were found to perform better on tasks related to auditory perception (e.g., temporal discrimination and speech-in-noise) and cognitive processing (e.g., working memory) than non-musicians (Parbery-Clark et al. 2011; Rammsayer and Altenmüller 2006). This is also reflected in subcortical responses to speech and music that were found to be earlier and larger than in non-musicians (Musacchia et al. 2007; Strait et al. 2009).

The existing evidence suggests that being a musician helps preserve speech perception in both simple and complex settings (i.e., with background noise). However, for those who have not practiced music or acquired a musical education, can other leisure activities similarly preserve speech perception? The literature, although extensive in studying aging musicians, does not provide a clear answer. Table 1 presents studies that compared groups of aging adults (not included are studies comparing young to aging adults). As shown in Table 1, these studies used non-musicians as control group, and only characterized participants by their musical education. However, whether the control participants were engaged in other leisure activities (and which types) was not reported. Therefore, we are limited to knowing only about their involvement in musical education.

Table 1.

Studies of music engagement and speech perception among aging adults

Study	Groups
Amer et al. (2013)	18 musicians (mean age = 59.2) 24 non-musicians (mean age = 60.8)
Bidelman and Alain (2015)	10 musicians (mean age = 70.1 years) 10 non-musicians (mean age = 69.6 years)
O’Brien et al. (2015)	8 musicians (mean age = 62.5 years) 8 non-musicians (mean age = 61.4 years)
Parbery-Clark et al. (2011)	18 musicians (mean age = 55 years) 19 non-musicians (mean age = 54 years)
Parbery-Clark et al. (2012)	20 musicians (mean age = 55 years) 17 non-musicians (mean age = 57 years)
White-Schwoch et al. (2013)	27 musicians with different amounts of musical training (mean age = 65.3 years) 15 non-musicians (mean age = 63.6 years)
Zendel and Alain (2013)	74 musicians (aged 19–91 years) 89 non-musicians (aged 18–86 years)

Beyond studying aging musicians, the literature on the effects of leisure activity on aging adults is focused on general cognitive ability or risk of dementia and Alzheimer’s disease. Mixed results were reported regarding the effect of engagement in mentally stimulating activities on cognitive performance among aging adults. Some studies show that higher levels of engagement in mentally stimulating activities such as reading or playing board games are associated with reduced rates of cognitive decline (e.g., Bosma et al. 2002; Dartigues et al. 2013; Hultsch et al. 1999; Schooler and Mulatu 2001; Wang et al. 2006; Wilson et al. 2002, 2003, 2007), but some fail to show such a relationship (Aartsen et al. 2002; Hultsch et al. 1999). A large body of literature is focused on the relationship between leisure activity and the development of dementia (Coyle 2003; Dartigues et al. 2013; Gow et al. 2014; Hertzog et al. 2008; Park et al. 2014; Stern and Munn 2010; Verghese et al. 2003; Wang et al. 2006). Although the causal relationship between engagement in leisure activity and the development of dementia and Alzheimer’s disease is not clear, studies consistently show that those who are engaged in leisure activities tend to develop less dementia and Alzheimer’s disease, with more leisure activity being associated with reduced risk of these conditions. The present study, however, is focused on the effects of leisure activities on auditory and speech perception among the healthy aging.

The aim of present study was to compare the effects of musical and non-musical leisure activities on perceptual abilities, specifically auditory temporal processing and speech perception, of aging adults. To facilitate the comparison, three groups of 60–80 years old were recruited for the study: (1) non-professional musicians who met once a week to play in a symphonic orchestra, (2) card players who met once a week to play cards, and (3) a control group of participants that did not engage in any regular leisure activity. We chose to compare musicians with card players in order to test whether aging adults who are engaged in non-musical leisure activity that is mentally demanding and involves meeting with other people once a week would differ in auditory processing and speech perception from non-professional musicians. Although card playing may be considered a less demanding activity than music, because it involves strategic planning and rule-mastering, and is a highly common leisure activity among aging adults that involves social interaction, it was considered worthwhile to test relative to musical activity. The control group was utilized in order to estimate the effect of the leisure activities on auditory temporal processing and speech perception. The hypothesis was that Musicians would replicate previous findings and show enhanced auditory temporal processing and speech perception compared to Controls. Card players were hypothesized to show lower auditory perception due to lack of any specific auditory exposure. Yet, they were predicted to show better speech perception than Controls due to their engagement in regular mentally demanding activity and social interaction. Overall, the expectation was for incrementally improved effects in auditory processing and speech perception across the study groups: least among the Controls, more among the Card players, and highest among the Musicians.

Methods

Participants

Sixty-nine individuals, aged 60–80 years old (mean = 65.57, SD = 5.75, 47.8% females) participated in the study in one of the three groups: Musicians, Card players, or Controls. Participants were native Hebrew speakers or started speaking Hebrew before the age of 12 and used it as their main language for speaking, reading, and writing. All participants were screened for hearing ability using Danplex Diagnostic Audiometer DA64. They had hearing thresholds ≤ 35 dB HL for 0.5, 1, 2, and 4 kHz, and less than a 5-dB threshold difference between ears. They reported being healthy and independent in their functioning with no history of diseases related to the central nervous system, and had scores of 29–30 in the Mini Mental State Examination (MMSE, Folstein et al. 1975), indicating a high level of mental ability. The groups were matched for gender, age, and Wechsler Adult Intelligence Scale-III Matrix Reasoning and Digit Span scores (Wechsler 1997). Table 2 presents means and standard deviations for each group in these variables. No difference was found between the groups. The matching process was done in order to prevent alternative explanations to group differences and strengthen the study results and conclusions. It was done by recruiting a larger pool of participants, consisting of 87 Musicians, Card players, and Controls that met the above criteria for hearing level, health, and mental ability, and who performed the study tests. For each participant in the Musician group, a Card player participant and a Control participant of the same age and gender, and with similar WAIS-III subtests scores, were chosen.

Table 2.

Mean (SD) and comparisons between study groups for variables used for matching

	Musicians	Card players	Controls	F (2,66)
Age	65.57 (5.84)	65.39 (5.57)	65.61 (6.36)	.009
Matrix reasoning	20.09 (2.04)	20.00 (2.15)	19.96 (2.12)	.023
Digit span	17.57 (2.27)	17.70 (2.48)	17.78 (2.52)	.047

Musicians, in our study, were non-professional instrument players engaged in playing music at least 7 h a week, with 3 h of them dedicated to playing in an orchestra. None of the musicians played cards regularly. Card players were non-professional who met regularly once a week to play cards (usually “Bridge”) for approximately 3 h. The Bridge game involves adherence to game rules and strategic planning, and therefore is considered demanding mental activity. None of the card players played any musical instrument, or were engaged in any musical activity (such as singing or teaching music), aside from occasional passive listening to music. The Control participants were neither musicians nor card players: They neither played any musical instrument nor had any formal music education in the past, and were also not engaged in any musical activity (such as singing or teaching music), aside from occasional passive listening to music. The Control group also did not play cards regularly.

Tasks and stimuli

Spectral temporal order judgement (TOJ)

This task measured participants’ auditory temporal processing by reporting the order of two brief high and low tones. Participants were required to repeat the order of 1 (“low”) and 1.8 (“high”) kHz, 5-ms tones, with 1-ms cosine-squared, rise/fall envelopes, that were presented binaurally one after the other. Participants pressed the relevant keyboard keys in the order corresponding to what they heard (either high-low or low–high) that were delivered to both ears simultaneously. This task was performed in the same manner as in Fostick and Babkoff (2013). The tones were presented with an inter-stimulus interval (ISI, the time separating the offset of the first stimulus and the onset of the second one) of either of 2, 5, 10, 15, 30, 60, 90, 120, and 240 ms, repeating 16 times for each high–low/low–high combination. This resulted in a total of 288 trials that were presented randomly. TOJ threshold was calculated as the ISI needed for 75% correct responses. When the accuracy rate was higher than 75% for the shortest ISI value, the Spectral TOJ threshold was defined as 0 ms.

Although all participants were screened for a good hearing level, in order to avoid a possible effect of differences in hearing thresholds, the tones were delivered at the same level (40 dB SL) for all the participants. Therefore, before participants performed the TOJ task, an absolute threshold task was performed using the same equipment and experimental conditions as in the TOJ task. The hearing threshold was measured for 1 and 2.8 kHz, 15-ms tones using a two-alternative forced-choice two-down-one-up adaptive staircase procedure. Hearing thresholds were calculated as the average of the last 8 out of 10 reversals.

Spatial TOJ

Similarly to Spectral TOJ, this task also measured participants’ auditory temporal processing, by reporting the order of two brief tones. In this task, the tones differ by the ear to which they are delivered. Both tones were 1 kHz 15 ms, and each presented to a different ear (participants’ responses were either “right–left” or “left–right”). This task was performed in the same manner as in Fostick et al. (2014) and Fostick and Babkoff (2013). The tone pairs were presented with an ISI of either 5, 10, 15, 30, 60, 90, 120 or 240 ms. Each value repeated 16 times in “left–right” order and 16 times in “right–left” order, resulting in 256 trials that were presented in a different and random order for each participant. The TOJ threshold was calculated as the ISI needed for 75% correct responses.

Speech perception

The Hebrew version of the AB words test (Boothroyd 1968; Fostick et al. 2014) was used. The test includes lists of ten one-syllable consonant–vowel–consonant meaningful words that are phonemically balanced (i.e., in each list every consonant appears once and every vowel appears twice). The words were presented binaurally, with a 3-second gap between them, in four conditions: (1) no background noise; (2) with background speech-noise frequencies (SpN) consisting of steady-state narrowband noise comprised of frequencies within the speech-range (0.5–4 kHz); (3) with background white noise (WN) consisting of wide-band noise evenly distributed over the entire range of frequencies (0.25–8 kHz); and (4) compressed speech at a 60% compression rate. A compression rate of 60% was chosen since it was found previously to be sensitive to age-related difficulty in speech perception (e.g., Babkoff and Fostick 2017; Fostick et al. 2013, 2014). The words were compressed to be 60% of their original length and presented with no background noise. Taking into account a baseline speech rate of 140–160 words per minute (wpm), a 60% compression resulted in a rate of about 200 wpm. Two lists (20 words) were presented in each condition. Participants were required to verbally repeat each word immediately after hearing it.

Procedure

The study was approved by the Institutional Review Board and was conducted in accordance with Good Clinical Practice (GCP) guidelines. All participants received a full explanation about the study and provided signed informed consent. Participants were first screened for hearing level and for mental ability. Following screening, participants completed the study tests. The order of the tests was counterbalanced between participants.

Statistical analysis

Group effects were tested separately for auditory temporal processing (Spectral TOJ and Spatial TOJ) and for speech perception (speech perception with no background noise, speech perception with SpN, speech perception with WN, 60% compressed speech) using multiple analysis of variance (MANOVA) for each group of measures. Levene’s test of equality of error variances was used to test for possible group differences in variance. Univariate ANOVA was used to measure group effects in tests with no group differences in variance (Spatial TOJ, speech perception with SpN, and speech perception with WN) and Kruskal–Wallis tests were performed for tests in which there were group differences in variance (Spectral TOJ, and in speech perception with no background noise and with compressed speech). Post-hoc analyses were done using the Mann–Whitney test.

Results

MANOVA revealed significant group effects both for auditory temporal processing tests (F₍₄₁₃₂₎ = 9.469, p < .001, η²_p = .223) and for speech perception tests (F₍₈₁₂₈₎ = 7.292, p < .001, η²_p = .313). Group differences in variance were found in Spectral TOJ, in speech perception with no background noise, and in 60% compressed speech (Table 3). Univariate ANOVA and Kruskal–Wallis tests showed significant group effects in all study measures (Table 3).

Table 3.

Results of Levene’s test of equality of error variances and group differences for all study measures

	Levene’s test of equality of error variances (df = 2,66)	Group differences
Spectral TOJa	390.587***	44.550***d
Spatial TOJa	.776	5.699**e
Speech perception no noise	17.676***	23.366***d
Speech perception with SpNb	1.748	8.255**e
Speech perception with WNc	.764	14.684***e
Compressed speech perception	9.144***	36.837***d

**p < .01

***p < .001

^aTemporal order judgment

^bSpeech noise

^cWhite noise

^dKruskal–Wallis test (df = 2)

^eANOVA (df = 2,66)

Figure 1 presents mean TOJ thresholds for all three study groups. Musicians were found to have lower Spectral and Spatial TOJ thresholds than both Card players (U = 10, p < . 001, η²_p = 1.528, and U = 135.5, p = .005, η²_p = .365, respectively) and Controls (U = 5, p < .001, η²_p = 1.583, and U = 125.5, p = .002, η²_p = .424, respectively). Most of the Musicians (78%) had an accuracy higher than 75% in the shortest ISI value of Spectral TOJ (2 ms); therefore, their threshold was defined as 0 ms. The remaining five participants had Spectral TOJ thresholds of 3 to 8 ms. No difference was found between Card players and Controls in both Spectral and Spatial TOJ thresholds (U = 255, p = .829, η²_p = .002, and U = 244, p = .642, η²_p = .009, respectively).

Fig. 1.

Mean and standard error of temporal order judgment (TOJ) thresholds for all study groups. *p < .01; **p < .001

Figure 2 presents speech perception accuracy for all study groups. Musicians had a higher accuracy level than both Card players and Controls in all speech perception conditions (no noise: U = 138, p = .002, η²_p = .422, and U = 63, p < .001, η²_p = .977; SpN: U = 176, p = .04, η²_p = .213, and U = 126, p = .002, η²_p = .423; WN: U = 161, p = .022, η²_p = .238, and U = 62, p < .001, η²_p = .907; compressed speech: U = 80.5, p < .001, η²_p = .773 and U = 7, p < .001, η²_p = 1.488). Interestingly, Card players also had higher speech perception accuracy than Controls in all speech perception conditions (no noise: U = 162, p < .05, η²_p = .243; SpN: U = 139, p = .02, η²_p = .431; WN: U = 155, p < .015, η²_p = .270; compressed speech: U = 157, p < .05, η²_p = .260).

Fig. 2.

Mean and standard error of speech perception accuracy for study groups

Discussion

The current study was undertaken in light of the literature showing the superiority of aging musicians’ speech perception relative to that of non-musicians. The goal of the study was to uncover whether aging adults who engage in regular non-musical cognitively demanding leisure activity also benefit from enhanced speech perception compared to musicians, and to those who are not engaged in any such leisure activity. Our findings replicate previous studies showing that aging musicians indeed have superior auditory processing and speech perception. However, the more remarkable finding of the study was that although not to the same degree as musicians, card players showed better speech perception than aging participants not engaged in any regular leisure activity.

Enhancement of auditory processing has been shown to be related to better speech perception in children (François et al. 2013; Kraus et al. 2014), adults (Parbery-Clark et al. 2009a, b), and more recently in aging adults (Amer et al. 2013; Bidelman and Krishnan 2010; O’Brien et al. 2015; Parbery-Clark et al. 2011, 2012; White-Schwoch et al. 2013; Zendel and Alain 2013). Therefore, it is not surprising that in the current study we found both high auditory processing and speech perception scores among aging musicians, compared to non-musicians.

It was also not surprising to find that auditory perception was not enhanced among Card players, given that their leisure activity does not include components of auditory processing. Their performance was similar to the Controls who were not engaging in any regular, defined leisure activity. Still, it was interesting to discover that card players showed better speech perception than controls. It is possible that the Card players were able to generalize their effortful leisure activity to enhanced speech perception. Most studies show little generalization from trained to non-trained tasks (Ball et al. 2002; Willis et al. 2006), but there are some studies that do show such generalization (Ball et al. 2007; Willis et al. 2006). Inferring generalization from card playing to speech perception, or of speech perception training while playing cards, should be done cautiously, since the present study was not design to test causality; it is possible that those with better speech perception self-selectively chose to play cards, those with lower speech perception self-selectively chose not to play cards, or that a third factor affected both variables.

Our findings evoke the question: Why would aging card players’ speech perception be better retained? Possible answers can be divided into two explanations: functional and social. Functionally speaking, as mentioned previously, this increased retention might be explained by the self-selection of those with good speech perception to engage in an activity that requires spoken discourse. Alternatively, engaging in card playing may strengthen speech perception either directly (i.e., exposure to speech during card playing serves as speech perception “training”) or indirectly (i.e., the cognitive demands of playing cards strengthen various perceptual and cognitive abilities, included among them also speech perception). Mechanisms underlying the relationship between engagement in leisure activity and cognitive ability have been investigated previously in the context of dementia. According to that line of research, having more intellectual resources buffers the initial damage of dementia, delaying the outcomes of its development (the “cognitive reserve” hypothesis, Coyle 2003; Verghese et al. 2003; Vinkers et al. 2003). Moreover, the cognitive functioning that is involved in leisure activity can strengthen existing synaptic connections, generate new ones, and stimulate neurogenesis (“mental workout” and “use it or lose it” hypotheses, Coyle 2003; Gow et al. 2014; Hertzog et al. 2008). These mechanisms might also play a role in healthy aging and explain the retention of speech perception among aging adults that engage in cognitively challenging leisure activity.

One should bear in mind that, although compelling, the hypothesis of leisure activity protecting one from dementia should be regarded with caution since the causal relationship is not yet clear (Stern and Munn 2010). It might be that leisure activity and cognitive ability protect from dementia; however, it also might be that experiencing some cognitive decline in the pre-dementia period results in a reduction in leisure activities. The findings of prospective long-term follow-up studies on participants with no initial dementia tip the scales toward the protective role of leisure activity (Verghese et al. 2003; Vinkers et al. 2003), but more studies are needed.

Previous studies have also examined the social aspect of leisure activity as a potential factor to preserve cognitive and perceptual abilities (Barnes et al. 2004; Bassuk et al. 1999; Zunzunegui et al. 2003). Two particular types of social engagements have been the object of these studies: activities with social interactions (e.g., membership in groups) and maintaining social networks (regular contact with family and/or friends). Studies on social interactions generally show that engagement in social activities is associated with less cognitive and perceptual decline (Barnes et al. 2004; Bassuk et al. 1999; Zunzunegui et al. 2003), while studies on social networking provide mixed results (Barnes et al. 2004; Zunzunegui et al. 2003). The connection between social activity and cognitive/perceptual functioning is hypothesized to be related to a strengthening of neural connections (Barnes et al. 2004; Bassuk et al. 1999) and protection against the negative effects of stress (Cohen 2004). It might be that the results of the present study reflect the effect of combining perceptual and cognitive activities with social interaction. Therefore, future studies may wish to add a comparison group that involves solitary leisure activity, in order to control for the potential impact of social engagement.

It should be noted that the present study refers to perceptual abilities, including auditory temporal processing and speech perception. In addition, study participants were matched on working memory and non-verbal intelligence in order to avoid confounding factors, but this decreased the study’s external validity. It also should be noted that in the present study “leisure activity” referred to card playing. This is a very narrow scope of leisure activities. If a leisure activity is any activity done during leisure time, then it includes all kinds of activities, both cognitively challenging and not. Future studies should focus on other types of leisure activities, with different degrees of social engagement and cognitive effort, and assess their effects on other perceptual and cognitive abilities among the general aging population. Future studies also should take into account participants’ other activities (i.e., activities during employment) that might also affect auditory and speech perception.

It also should be noted that although the Musicians and Card players in our study had the same amount of weekly group practice (either orchestra rehearsal or card games), the musicians usually also engaged in a few hours a week of individual practice. Thus, overall the musicians engaged in more activity and training than the Card players. This difference could serve as an alternative explanation of the superior performance of the Musicians. Future studies should control also for total practice time. Another possible confounding factor in study might be that some of our participants immigrated to Israel during their childhood and only then started speaking Hebrew. Although this was their main language for most of their life, it can serve as an additional explanation for the difference between the study groups.

The critical question at hand is whether individuals can influence their future speech perception by engaging in various leisure activities. Until now, studies have shown that, among healthy aging adults, the ability to perceive speech is more highly preserved among musicians; engagement in other leisure activities has been studied mainly in the context of the development of dementia. The results of the present study communicate a possible message to non-musicians, if confirmed in future studies: Even if one did not receive musical education and is not involved in making or teaching music, engaging in card playing regularly with others (and perhaps other activities that involve cognitive effort and social interactions, to be determined in future studies) may help to preserve one’s speech perception at older ages, a critical ability for daily life that tends to degrade with aging.

Funding

This research did not receive any specific Grant from funding agencies in the public, commercial, or not-for-profit sectors.