Cognition and Emotion ISSN: 0269-9931 (Print) 1464-0600 (Online) Journal homepage: http://www.tandfonline.com/loi/pcem20 Age differences in vocal emotion perception: on the role of speaker age and listener sex Antarika Sen, Derek Isaacowitz & Annett Schirmer To cite this article: Antarika Sen, Derek Isaacowitz & Annett Schirmer (2017): Age differences in vocal emotion perception: on the role of speaker age and listener sex, Cognition and Emotion, DOI: 10.1080/02699931.2017.1393399 To link to this article: http://dx.doi.org/10.1080/02699931.2017.1393399 View supplementary material Published online: 24 Oct 2017. Submit your article to this journal Article views: 3 View related articles View Crossmark data Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=pcem20 Download by: [University of Florida] Date: 26 October 2017, At: 08:19 COGNITION AND EMOTION, 2017 https://doi.org/10.1080/02699931.2017.1393399 Age differences in vocal emotion perception: on the role of speaker age and listener sex Antarika Sena, Derek Isaacowitzb and Annett Schirmerc,d,e Neurobiology and Aging Programme, National University of Singapore, Singapore, Singapore; bDepartment of Psychology, Northeastern University, Boston, USA; cDepartment of Psychology, The Chinese University of Hong Kong, Hong Kong, Hong Kong; d The Mind and Brain Institute, The Chinese University of Hong Kong, Hong Kong, Hong Kong; eMax Planck Institute for Human Cognitive and Brain Sciences, Leipzig, Germany Downloaded by [University of Florida] at 08:19 26 October 2017 a ABSTRACT ARTICLE HISTORY Older adults have greater difficulty than younger adults perceiving vocal emotions. To better characterise this effect, we explored its relation to age differences in sensory, cognitive and emotional functioning. Additionally, we examined the role of speaker age and listener sex. Participants (N = 163) aged 19–34 years and 60–85 years categorised neutral sentences spoken by ten younger and ten older speakers with a happy, neutral, sad, or angry voice. Acoustic analyses indicated that expressions from younger and older speakers denoted the intended emotion with similar accuracy. As expected, younger participants outperformed older participants and this effect was statistically mediated by an age-related decline in both optimism and working-memory. Additionally, age differences in emotion perception were larger for younger as compared to older speakers and a better perception of younger as compared to older speakers was greater in younger as compared to older participants. Last, a female perception benefit was less pervasive in the older than the younger group. Together, these findings suggest that the role of age for emotion perception is multi-faceted. It is linked to emotional and cognitive change, to processing biases that benefit young and own-age expressions, and to the different aptitudes of women and men. Received 3 May 2017 Revised 28 September 2017 Accepted 6 October 2017 KEYWORDS Prosody; affective; ageing; acoustic; motivation; working memory; age bias Introduction Vocal emotion perception and ageing It has long been established that healthy ageing is associated with a decline in mental functioning (e.g. Colsher & Wallace, 1991; Riddle, 2007). Among other processes, emotional perception has been implicated (Ruffman, Henry, Livingstone, & Phillips, 2008). For example, older adults make more mistakes than younger adults when interpreting the emotion of nonverbal signals such as vocal expressions. Here, we sought to explore and to better describe this phenomenon. Specifically, we aimed to replicate previous reports of sensory, cognitive, and emotional ageing statistically mediating age differences in vocal emotion perception. More importantly, we explored whether and how these differences are characterised by speaker age effects and whether they vary between women and men. The last few decades produced a growing interest in the emotional processes of older adults (Schirmer, 2014). Many studies appeared comparing individuals aged 60 and above with middle-aged and/or younger adults showing age-related performance decrements (for a review see Isaacowitz et al., 2007; Ruffman et al., 2008). In the context of voice perception, participants were presented with nonverbal exclamations such as “Ah” (Fecteau, Armony, Joanette, & Belin, 2005; Ryan, Murray, & Ruffman, 2009) or verbal material that was semantically neutral (Mitchell, Kingston, & Barbosa Bouças, 2011; Ruffman, Halberstadt, & Murray, 2009), non-sensical (Demenescu, Mathiak, & Mathiak, 2014; Ruffman, Halberstadt, et al., 2009), or low-pass filtered, removing segmental information and leaving basic aspects of prosody intact (Mitchell CONTACT Annett Schirmer firstname.lastname@example.org Supplemental data for this article can be accessed https://doi.org/10.1080/02699931.2017.1393399 © 2017 Informa UK Limited, trading as Taylor & Francis Group Downloaded by [University of Florida] at 08:19 26 October 2017 2 A. SEN ET AL. et al., 2011). A typical experimental task entailed indicating whether the vocal tone of two successive stimuli was identical (Orbelo, Grim, Talbott, & Ross, 2005) or which of several expression options best characterised the tone (Demenescu et al., 2014). Together, available results suggest that the recognition of sadness, anger, fear, and happiness are difficult for older adults (Demenescu et al., 2014; Lambrecht, Kreifelts, & Wildgruber, 2012; Lima, Alves, Scott, & Castro, 2014; Paulmann, Pell, & Kotz, 2008; Ryan et al., 2009; but see Hunter, Phillips, & MacPherson, 2010; Ruffman, Sullivan, & Dittrich, 2009), whereas the recognition of surprise seems relatively preserved (Hunter et al., 2010; Paulmann et al., 2008; Ruffman, Halberstadt, et al., 2009; Ryan et al., 2009). Attempts have been made to uncover the mechanisms by which ageing affects performance in emotion recognition tasks. The dominant approach has been a correlation or mediation analysis linking age differences in control measures with those in emotion judgments. Results suggest that hearing loss (Lambrecht et al., 2012; Mitchell & Kingston, 2014) and cognitive decline (Krendl & Ambady, 2010; Lambrecht et al., 2012; Orbelo et al., 2005; Pichora-Fuller, 2003) could be relevant. Additionally, current theories of emotional ageing hold that a reduction in feeling emotions (Cacioppo, Berntson, Bechara, Tranel, & Hawkley, 2011) and changes in the importance of certain emotions (Carstensen, Isaacowitz, & Charles, 1999; Isaacowitz, Wadlinger, Goren, & Wilson, 2006; Levenson, Carstensen, Friesen, & Ekman, 1991; Tsai, Levenson, & Carstensen, 2000) affect how older adults engage with emotional expressions and how they respond to them. Nevertheless, evidence for these different mechanisms is equivocal and the mechanisms themselves are still debated (Ruffman et al., 2008). Own-age vs other-age emotion perception Apart from uncertainty as to the roles of sensory, cognitive, and emotional ageing, little is known about how situational factors affect vocal emotion recognition in young and old. Of particular interest here is the expresser’s age. This interest derives from face perception research pointing to processing differences between signals from young as compared to old individuals and own-age as compared to other-age individuals (but see Ebner & Johnson, 2009). Compared to signals from older individuals, signals from younger individuals are perceived more accurately by both younger and older decoders (Riediger, Voelkle, Ebner, & Lindenberger, 2011). Additionally, compared to other-age faces, own-age faces produce fewer recognition errors (Malatesta, Izard, Culver, & Nicolich, 1987) and reduce associated age differences (although not entirely consistently, Riediger et al., 2011). Related to this, own-age faces attract more gazing (Ebner, He, & Johnson, 2011; He, Ebner, & Johnson, 2011) and activate emotion areas in the brain more readily (Ebner et al., 2013). These effects have been attributed to a range of causes (Fölster, Hess, Hühnel, & Werheid, 2015) of which three will be relevant here. First, it has been argued that age affects the accuracy with which emotions are encoded. Structural changes of the face may impact emotion signalling making it harder to identify the expressions of older as compared to younger individuals (Malatesta et al., 1987). Second, the culture associated with a particular peer group may engender generational differences in emotion signalling. Moreover, the familiarity individuals have with a given group may then influence their ability to recognise emotions. Younger individuals may be more familiar with younger as compared to older age groups, whereas older individuals, due to their life experience, may be equally familiar with both (Wiese, Schweinberger, & Hansen, 2008). Last, motivation may be relevant in that individuals may be more interested in interacting with young (Gordon & Arvey, 2004) and own-age individuals (for a review see Rhodes & Anastasi, 2012) as compared with old and other-age individuals because they expect these interactions to be more relevant or rewarding. To date, these possibilities have been rarely tested. Moreover, available evidence has come from face perception only (e.g. Fölster et al., 2015; Rhodes & Anastasi, 2012). Insights are lacking from other modalities, like the voice, which may differ from the face in terms of age-related effects on emotion expression and perception. In terms of expression, vocal ageing produces a loss in collagenic and elastic fibers, connective tissue degeneration (e.g. articular cartilage), and muscle atrophy (Colton, Casper, Leonard, Thibeault, & Kelly, 2011; Gracco & Kahane, 1989; Kahane, 1987). Thus, it impairs the vibratory properties of the vocal folds (e.g. Ohno & Hirano, 2014) making older voices sound less stable than younger voices. Moreover, compared to younger voices, older voices have a higher harmonic-to-noise ratio and show a pitch decrease or increase for women and men, respectively, thus marking the speaker’s sex less clearly (Decoster & Debruyne, Downloaded by [University of Florida] at 08:19 26 October 2017 COGNITION AND EMOTION 1997; Dehqan, Scherer, Dashti, Ansari-Moghaddam, & Fanaie, 2012). Analogously, facial ageing produces wrinkles and tissue sagging that may impact expressive clarity. Yet, specific effects may differentiate from vocal ageing both qualitatively and quantitatively. In terms of perception, age effects have been documented for auditory and visual systems. Auditory deficits, referred to as presbycusis, emerge gradually, most likely through the accumulative impact of noise, toxins (e.g. smoking), and genetic mechanisms on the hair cells in the inner ear (Moser, Predoehl, & Starr, 2013; Yamasoba et al., 2013). As a consequence, older adults have a poorer frequency resolution and an increased hearing threshold especially at higher frequencies. Visual deficits result from macular degeneration, increasing lense opacity and myopia among others (Andersen, 2012). Compared to agerelated hearing loss, however, these deficits are typically more readily noticed and corrected (Oyler, 2012). Given these modality differences, it seems critical to explore the role of encoder age for nonverbal channels other than the face. Moreover, new insights from the voice could help identify and differentiate modality specific and unspecific mechanisms. The female emotion perception benefit in young and old Apart from the expresser’s age, the receiver’s sex may be relevant for age effects in emotion perception. Support for this idea comes from a range of implicit and explicit paradigms showing greater emotion sensitivity in women than in men. Implicit paradigms have been used with both behavioural and neuronal measures. Behavioural responses have been recorded in priming tasks where participants classify a target that is preceded by a task-irrelevant prime. In such settings, women are more readily influenced by the affective primetarget relationship (Donges, Kersting, & Suslow, 2012; Schirmer, Seow, & Penney, 2013). Sex differences in neuronal responses have been documented fairly consistently using event-related potentials (ERPs). Among other findings, women show greater ERP differences than men between emotional and neutral vocalisations that are task irrelevant (Hung & Cheng, 2014; Schirmer, Striano, & Friederici, 2005) or for which verbal but not vocal content has to be remembered (Schirmer, Chen, Ching, Tan, & Hong, 2013). Past explicit paradigms involved tasks akin to those used to study ageing, and women were found to be 3 more accurate than men (Bonebright, Thompson, & Leger, 1996; Hall, 1978; Mill, Allik, Realo, & Valk, 2009). Notably, however, this difference shows inconsistently across published work (e.g. Lima et al., 2014; Paulmann et al., 2008) leading some to speculate that it may be concealed by ceiling effects. In line with this, a study comparing emotion perception on subtle and highly-expressive displays found that women outperformed men for the former but not the latter (Hoffmann, Kessler, Eppel, Rukavina, & Traue, 2010). To date, few studies explored the developmental course of sex differences in emotion perception and results are mixed. Some studies collapsing over young and old cohorts reported a female advantage in emotion perception accuracy (Lambrecht, Kreifelts, & Wildgruber, 2014; Mill et al., 2009; Williams et al., 2009). However, of the few studies that investigated the interaction of age and sex, two found a main effect of age only (Lima et al., 2014; Paulmann et al., 2008), two reported main effects of both age and sex (Ruffman, Murray, Halberstadt, & Taumoepeau, 2010; Sullivan, Campbell, Hutton, & Ruffman, 2017), and two found a sex effect that increased with age. In other words, the age-related decline in task accuracy was smaller in women than in men suggesting that women were relatively better protected (Campbell, Ruffman, Murray, & Glue, 2014; Demenescu et al., 2014). An improved understanding of the developmental trajectory of sex differences would be relevant in two ways. First, it would help specify the roles of genetic (for a review see Hines, 2010) and environmental factors (e.g. Koenig & Eagly, 2005) in shaping male and female emotion sensitivity. Although their respective contributions are now well recognised, the mechanisms and significance of nature and nurture for behavioural outcomes are still incompletely understood. Considering age would help with this by, for example, elucidating the relation between developmental changes in the levels of sex hormones and emotion recognition abilities. Second, insights into the intersection of age and sex would be instructive as concerns the mechanisms underpinning the effect of ageing on emotion recognition. Specifically, factors that promote and protect against performance decline may be gleaned from sexspecific rates of decline (Campbell et al., 2014). The present study In sum, past research revealed age-related impairments in vocal emotion perception. Yet, the role of Downloaded by [University of Florida] at 08:19 26 October 2017 4 A. SEN ET AL. perceptual, cognitive, and emotional processes is still debated. Moreover, how age effects are shaped by biases in the perception of young and own-age stimuli as well as listener sex remains unclear. To address these gaps, we recorded emotional expressions from a set of younger and older individuals and subjected recordings to an acoustic analysis aimed at elucidating potential age effects on emotion expression. Previous work examining the relation between voice acoustics and emotions has shown that different parameters like pitch, intensity or duration take on different values for different emotions, thus enabling emotion identification (Banse & Scherer, 1996). Subsequently, we used the recordings in an explicit emotion perception task with naive, age-matched men and women whose hearing, cognitive, and emotional state were carefully assessed. Specifically, we applied an extensive test battery including pure tone thresholds, basic measures of intelligence, working memory, and a range of emotion measures targeted at positive and negative affect, anxiety, depression, optimism, and emotion regulation. In line with the evidence reviewed above, we expected emotion perception to be less accurate for older as compared to younger listeners and for this difference to be statistically related to sensory, cognitive, and/or emotional age effects. More importantly, we anticipated an effect of speaker age on listener accuracy. If age and own-age biases were additive, younger listeners should do better when listening to young as compared to old speakers because those speakers are both young and of the same age. Older listeners on the other hand might perform comparably when listening to young and old speakers. Moreover, performance differences between younger and older listeners should be more pronounced for young as compared to old speakers. Acoustic analysis of vocal expressions should reveal whether age differences in expression accuracy and/or expressive style account for these effects. Specifically, reduced and qualitatively different emotion differentiation as a function of age would speak for the former and the latter, respectively. Last, we hypothesised that female listeners outperform male listeners especially for more difficult expressions. If this sex difference were to buffer ageing effects in women as suggested previously (Campbell et al., 2014; Demenescu et al., 2014), it should be greater in older relative to younger individuals. Method Participants This study recruited 169 participants. Of those, four had to be excluded because they dropped out from the study with incomplete data. Data from an additional two participants was discarded because their accuracy was below three standard deviations of the mean. The remaining participants comprised 40 older men and 43 older women aged 68.6 years (6.2 SD, range 60–85) as well as 40 younger men and 40 younger women aged 23.1 years (3.6 SD, range 19–34). All participants were Englishspeaking Singaporeans. Older participants were recruited through online advertisements and by reaching out to various elder-care centres, senior volunteer groups, and senior activity centres. Younger participants were recruited from the National University of Singapore and the community via online university portals and online advertisements, respectively. The final sample is comparable in terms of age, sex, and socio-economic background to previous studies (Ruffman et al., 2008). However, two novel aspects include its large size and East-Asian origin. Participants provided written informed consent and received financial compensation of S$15 per hour. Stimulus material Twenty Chinese Singaporean speakers who were layactors and had taken part in non-professional theatre performances contributed to stimulus construction. Half the speakers (6 women, 4 men) had a mean age of 66.4 years (8.6 SD, range 50–81) and produced age-matched stimuli for our older participants. The other speakers (6 women, 4 men) had a mean age of 21.9 years (1.6 SD, range 20–24 years) and produced age-matched stimuli for our younger participants. Because the number of speakers was fairly small (N = 20) and the ratio of men and women unbalanced, we considered it best to ignore speaker sex and to focus on speaker age effects instead. The sentence material for this study comprised of six semantically neutral, subject-verb-object sentences (e.g. “The man walked to the office”). Although one may argue that an emotional voice would render these sentences emotionally incongruous, we considered this issue negligible as much of everyday language is neutral and emotions are often added nonverbally. Moreover, our approach has been Downloaded by [University of Florida] at 08:19 26 October 2017 COGNITION AND EMOTION established previously (Mitchell et al., 2011; Ruffman, Halberstadt, et al., 2009) and circumvents other, more serious issues, arising from other approaches such filtering speech or using pseudowords. Speakers were invited individually. After a vocal warm-up, they were given the list of sentences and were informed about the emotion categories they should portray. For each category, they were given a short scenario (e.g. you have received very good news) and asked to act out the associated emotion while producing the sentences on the list. Thus, each speaker portrayed each of the given sentences with angry, happy, neutral, and sad vocal expressions resulting in 480 sentences (120 per emotion, 240 per speaker age). Recordings were done in a soundproof chamber and digitised at a 16 bit/44.1 KHz sampling rate. The amplitude of the recorded sentences was normalised at the root-mean-square value using Adobe Audition 2.0. The quality of the recordings and their emotional expressiveness were verified subjectively by the authors of this study. Additionally, as detailed in the results section, stimulus quality was assessed objectively via acoustic analysis. Praat (Boersma & Weenink, 2013) was used to extract stimulus duration, harmonics-to-noise ratio (HNR; the periodicity of the signal), fundamental frequency (F0) mean (the lowest frequency band of an utterance perceived as pitch or speech melody), F0 standard deviation (F0 SD), F0 range (the pitch difference between the lowest and the highest value in an utterance), formant 1 frequency mean (F1; the next highest frequency band in the utterance following F0), F1 bandwidth (F1B), formant 2 frequency mean (F2; the next highest frequency band in the utterance following F1), jitter (cycle-to-cycle variation in F0; measures short-term variability/perturbations in pitch), and shimmer (cycle-to-cycle variation in amplitude; measures short-term variability/perturbations in voice intensity). Procedure Participants completed a screening and an experimental session. The sessions were separated by a minimum of 15 min and a maximum of 30 days constrained largely by the availability of older adults and the centres through which they were recruited. The average delay between sessions did not differ significantly between older (mean 4.3, SD 6.5 days) and younger (mean 2.9, SD 4.9 days) participants as demonstrated with a Welch t-test (t(152) = 1.49, p > .1). 5 Screening session The screening session comprised the following tests which were selected based on prior research in this area (e.g. Stanley & Isaacowitz, 2015). First, participants completed a set of questionnaires and paper-pencil tests including a demographics questionnaire, the State-Trait Anxiety Inventory (STAI1 and STAI2; Spielberger, Gorsuch, Lushene, Vagg, & Jacobs, 1983), the Center for Epidemiologic Studies Depression Scale (CES-D; Radloff, 1977), the Emotion Regulation Questionnaire (ERQ; Gross & John, 2003), the Positive and Negative Affect Schedule (PANAS; Watson, Clark, & Tellegan, 1988), the Revised Life Orientation Test (LOT-R; Scheier, Carver, & Bridges, 1994), the Digit Symbol Substitution Test of the Wechsler Adult Intelligence Scale (DSST), the Digit Span Tests (Digit Forward, DF; Digit Backward, DB) of the Wechsler Adult Intelligence Scale, and the Mini–Mental State Examination (MMSE; Folstein, Folstein, & McHugh 1975). Subsequently, we assessed participants’ hearing. To this end, pure tones were presented via headphones into one ear and participants indicated whether they had heard a tone with their left or right ear by raising their left or right arm. Tones were played at 1000, 2000, 4000, and 500 Hz, in that order for all participants. For the first 83 participants (20 older men, 23 older women, 20 younger men, 20 younger women), each frequency was first presented at 25 dB, whereas for later participants, each frequency was first presented at 0 dB. Measurements were chosen in accordance with the audiometric ISO values as mentioned by the World Health Organization (WHO, http://www.who. int/pbd/deafness/hearing_impairment_grades/en/) and included frequencies and intensities crucial for understanding speech (French & Steinberg, 1947). If participants heard a tone presented to them, the current dB level was recorded as the threshold for that particular frequency and ear. Upon failing to hear that tone, intensity was increased in steps of 5 dB until loud enough for the participant to hear. A single pure-tone average was derived by calculating mean hearing thresholds across both ears, and across all frequencies. Screening measures of study participants were subjected to a series of Welch t-tests comparing old and young (Table 1). With the exception of the use of suppression in emotion regulation, both groups differed on all measures. Experimental session During the experimental session, participants completed an explicit emotion perception task. Each trial 6 A. SEN ET AL. Table 1. Listener demographic and screening measures. Young Downloaded by [University of Florida] at 08:19 26 October 2017 Characteristic Old Mean (SD) t(df) p-value Education Years 14.16(1.8) 9.26(3.3) 11.92(128) <.001 1st Recruitment Drive Hearing Threshold (dB) 25 (0) 31.30(7.0) 6.17(42) <.001 2nd Recruitment Drive Hearing Threshold (dB) 13.38(4.4) 37.32 (9.0) 12.89(153) <.001 State-Trait Anxiety Inventory I 34.96(8.0) 29.95(10.1) 3.5(155) .001 State-Trait Anxiety Inventory II 42.78(9.6) 33.48(9.3) 6.29(160) <.001 CES-D 12.51(7) 9.32(8.1) 2.69(159) .008 ERQ Reappraisal 31.69(6.0) 33.71(6.1) 2.03(161) .034 ERQ Suppression 16.55 (4.4) 17.72(6.6) 1.34(143) .18 PANAS-P 30.76(6.8) 34.92(7.6) 3.65(160) <.001 PANAS-N 18.24(5.9) 15.17(6.0) 3.3(161) .001 LOT-R 9.83(4.3) 7.05(3.9) 4.31(157) <.001 Digit Symbol Substitution Test 76.49(9.6) 41.10(15.3) 17.75(139) <.001 Digit Span Forward 11.25(3.4) 9.06(2.7) 4.63(151) <.001 Digit Span Backward 9.05(3.0) 5.84(2.6) 7.33(157) <.001 MMSE 29.30(.86) 27.69(1.62) 7.97(126) <.001 a Professional Background (Count Data) Retired/Housewife 0/1 49/15 Student 67 0 Unemployed 4 3 Full-time employed 5 6 Part-time employed 3 10 a Our participants had a broad professional background including untrained positions such as drivers and highly trained positions such as teachers, air traffic controllers, and managers. Both young and old individuals were recruited from institutions catering to low and high-income families. began with a fixation cross. After 200 ms, a sentence played over headphones while the fixation cross remained on the screen. The fixation cross disappeared at sentence offset and was followed by five choices including “1-angry”, “2-happy”, “3-neutral”, “4-sad” and “5-other”. Participants made their choice by pressing the appropriate number on a keyboard. If participants selected “other”, they were prompted to type in the emotion that they felt described the speaker’s state. Participants submitted their choice using the enter key after which they were asked to rate emotion intensity and arousal on 5-point scales ranging from 1 (very weak) to 5 (very strong). Then the screen turned blank and the next trial began after 1000 ms. Prior to completing the experimental task, participants were instructed about the general procedure and performed eight practice trials that were not part of the stimulus set of the experiment proper. The actual experiment comprised 480 trials. Spoken sentences were presented only once and in a randomised order. Results Emotional expression Mean values for each acoustic parameter, speaker group, and emotion condition are presented in Supplementary Figure S1. Parameter values were subjected to separate two-way ANOVAs with Emotion (angry, happy, sad, neutral) as a repeated measures factor and Speaker Age (old, young) as a between items factor. Emotion main effects were pursued Table 2. Summary of results from the ANOVA analyses for emotion expression. Parameter Duration HNR F0 mean F0 SD F0 range F1 mean F1 bandwidth F2 mean Jitter Shimmer Emotion Effect Statistic F(3, 54) = 41.2, p < .001, η 2 = .53 F(3, 54) = 26.9, p < .001, η 2 = .33 F(3, 54) = 35.85, p < .001, η 2 = .38 F(3, 54) = 4.82, p = .005, η 2 = .13 p > .1 F(3, 54) = 1.31, p < .001, η 2 = .32 p > .1 p > .1 F(3, 54) = 4.85, p = .004, η 2 = .08 F(3, 54) = 2.27, p = .09, η 2 = .02 Post-hoc comparisons sad > angry = happy = neutral sad > happy = neutral > angry angry = happy > sad > neutral happy = angry > neutral; angry, neutral = sad < happy n.a. angry > happy > neutral = sad n.a. n.a. neutral > angry = happy; all other ps > .1 angry > happy; all other ps > .1 Downloaded by [University of Florida] at 08:19 26 October 2017 COGNITION AND EMOTION with six paired t-tests, the results of which were corrected for multiple comparisons using the BenjaminiHochberg procedure. With the exception of F0 range, F1 bandwidth, and F2 mean, expressions differed by Emotion (Table 2). Importantly, none of the parameters produced significant main or interaction effects involving Speaker Age (ps > .16). We conducted two discriminant analyses, one for each speaker age group, to further probe potential speaker age effects and to ascertain reliable acoustic discrimination between the emotion conditions. Acoustic parameters served as the independent variables and emotion served as the dependent variable. The analysis yielded three discriminant functions for both older and younger speakers. Figure 1 illustrates the first and second discriminant function of each group. The first function accounted for 76.7% variance in older speakers and 77.9% variance in younger speakers. Correlations between the acoustic parameters and this function established F0 mean and HNR as the two best discriminators in older speakers (rs = .63, .44), and duration and HNR as the two best discriminators in younger speakers (rs = .84, .30). The second function accounted for 19.5% and 18.7% variance in older and younger speakers, respectively. It correlated most with F0 mean and duration in both older (rs = .72, .53) and younger 7 speakers (rs = .71, .45). The third function accounted for 3.8% and 3.4% variance in older and younger speakers, respectively. It correlated most with F0 SD and F0 range in both older (rs = .63, .42) and younger speakers (rs = .45, .48). The older speaker model accurately categorised 70.4% of the stimuli (angry, 63.3%; happy, 61.7%; neutral, 81.7%; sad, 75%), while the younger speaker model accurately categorised 78.3% of the stimuli (angry, 80%; happy, 48%; neutral, 88.3%; sad, 96.7%). A chi-square analysis indicated that results did not differ as a combination of speaker age and emotion (χ 2 = 5.02, df = 3, p = .170). Thus, we conclude that acoustic parameters enabled classification of emotional expressions well above chance (> 25%) for all age groups and expressions. Emotion perception Emotion perception performance is illustrated in Figures 2 and 3. Our interest focused on the accuracy with which participants perceived emotions. Arousal and intensity ratings were collected for exploratory purposes only and are documented in the Supplementary Materials. Incidentally, their patterns compare to those of accuracy. Response times were unsuitable for statistical analysis because participants Figure 1. Discriminant analysis results for older (left) and younger (right) speakers, respectively. Emotions (angry, happy, sad, neutral) were predicted on the basis of the acoustic parameters described in the results section. Each vocal expression is plotted according to its discrimination scores for discriminant functions 1 and 2. The different emotion categories are represented by different geometrical symbols and their centroids are marked as 1 for the angry, 2 for the happy, 3 for the neutral, and 4 for the sad condition. Of interest is the difference between centroids, how closely individual expressions cluster around their condition centroid, and how much they overlap with the expressions of another condition. Larger centroid distances, tighter expression clusters, and less overlap between expressions from different conditions indicate better discrimination. Downloaded by [University of Florida] at 08:19 26 October 2017 8 A. SEN ET AL. first chose or typed an emotion before pressing the enter key. Emotion categorisation responses were converted into unbiased hit rates (Hu; Wagner, 1993) and arcsine transformed before being subjected to an ANOVA with Emotion (angry, happy, neutral, sad) and Speaker Age (young, old) as repeated measures factors and Listener Age (young, old) and Listener Sex (female, male) as between subjects factors. The analysis revealed main effects of Emotion (F(3, 477) = 248.99, p < .001, η 2 = .31) indicating that sad voices had the highest accuracy followed by angry, neutral, and happy voices (all pairs significant, p < .001). Additionally there were main effects of Listener Age (F(1, 159) = 49.44, p < .001, η 2 = .16) and Speaker Age (F(1, 159) = 733.84, p < .001, η 2 = .16) as well as interactions involving Emotion and Listener Age (F(3, 477) = 12.68, p < .001, η 2 = .02), Emotion and Speaker Age (F(3, 477) = 49.79, p < .001, η 2 = .02), Listener Age and Speaker Age (F(1, 159) = 106.99, p < .001, η 2 = .03), Emotion, Listener Age and Listener Sex (F(3, 477) = 2.74, p < .05, η 2 = .00), and Emotion, Listener Age and Speaker Age (F(3, 477) = 35.07, p < .001, η 2 = .01). All other effects were non-significant (p > .1). We pursued the interaction of Emotion, Listener Age, and Listener Sex for each level of Emotion. The interaction of Listener Age and Listener Sex was significant for neutral voices (F(1, 159) = 4.33, p = .04, η 2 = .03) indicating that younger (F(1, 78) = 6.19, p = .01, η 2 = .07) but not older women (p > .25) outperformed their male peers. For happy voices, only the main effect of Listener Sex reached significance indicating that both younger and older women outperformed younger and older men, respectively (F(1, 159) = 6.17, p = .01, η 2 = .04). Main and interaction effects of Sex were non-significant for sad and angry voices (ps > .25). The Emotion, Listener Age, and Speaker Age interaction was explored by analyzing each level of Emotion. This revealed a significant interaction of Listener Age and Speaker Age in the angry (F(1, 161) = 5.94, p = 0.02, η 2 = .00), happy (F(1, 161) = 188.23, p < 0.001, η 2 = .13), neutral (F(1, 161) = 57.69, p < 0.001, η 2 = .03), and sad (F(1, 161) = 7.01, p = .01, η 2 = .01) conditions. Based on our interest in listener age effects, we pursued these two-way interactions for each speaker age. This revealed that younger listeners outperformed older listeners in the case of younger speakers for angry (F(1, 161) = 21.17, p < 0.001, η 2 = .12), happy (F(1, 161) = 161.08, p < 0.001, η 2 = .50), neutral (F(1, 161) = 56.04, p < 0.001, η 2 = .26), and sad (F(1, 161) = 12.65, p < 0.001, η 2 = .07) voices. Similarly, in the case of older speakers, younger listeners were significantly better than older listeners for angry (F(1, 161) = 18.82, p < 0.001, η 2 = .10), happy (F(1, 161) = 16.87, p < 0.001, η 2 = .09), neutral (F(1, 161) = 24.42, p < .001, η 2 = .13), and sad (F(1, 161) = 4.38, p = .04, η 2 = .03) voices. Importantly, however, differences between younger and older listeners were larger for younger than older speakers. Due to our interest in speaker age effects, we explored each two way interaction a second time by listener age. This revealed that in the case of older listeners, younger speakers were better recognised than older speakers for angry (F(1, 82) = 92.23, p < .001, η 2 = .06), neutral (F(1, 82) = 29.91, p < .001, η 2 = .02) and sad (F(1, 82) = 187.82, p < .001, η 2 = .19) voices. However, performances were comparable for both speaker age groups in the case of happy voices (p > .25). In the case of younger listeners, younger speakers were better recognised than older speakers for angry (F(1, 79) = 116.84, p < .001, η 2 = .16), happy (F(1, 79) = 347.39, p < .001, η 2 = .43), neutral (F(1, 79) = 287.66, p < .001, η 2 = .29), and sad (F(1, 79) = 290.66, p < .001, η 2 = .42) voices. Notably, differences between younger and older speakers were larger for younger than older listeners. Statistical mediation of age differences in emotion perception We conducted a multiple mediation analysis to assess whether age-related differences in screening measures statistically explain age-related differences in emotion perception. The model estimated total, direct, and indirect (individual as well as combined) effects of multiple simultaneous mediators. Method and terminology are described in much detail by Preacher and Hayes (Preacher & Hayes, 2008) and have been applied to the study of ageing and emotion previously (Lambrecht et al., 2012, 2014; Lima et al., 2014). In short, the estimated total effect refers to the statistical relationship between a primary dependent and independent variable, in our case emotion recognition and age. The direct effect refers to that same relationship when the effect of potential mediators is controlled. The indirect effect refers to the relationship of dependent and independent variable that is explained by one or more mediators. Again, in our case, an indirect effect may be that working memory statistically explains age differences in emotion recognition. Downloaded by [University of Florida] at 08:19 26 October 2017 COGNITION AND EMOTION 9 Figure 2. Emotion perception plots illustrating the interaction of listener age and listener sex for each emotion. Error bars represent the betweensubject standard error. Age was defined as the independent variable and Hu averaged across all variables besides listener age was defined as the dependent variable. Mediators were selected as follows. We first identified screening measures that correlated with age and, among those, identified measures with conceptual overlap (e.g. STAI-1 and STAI-2). In case of such overlap, we chose the measure showing the greater correlation with age. Last, if mediators were strongly inter-correlated (r ≥ .6), the one with the smaller age effect was dropped. Together, these steps helped reduce problems arising from collinearity (Preacher & Hayes, 2008). The mediators that entered the model included trait scores of the State-Trait Anxiety Inventory (STAI2), reappraisal scores of the Emotion Regulation Questionnaire (ERQ-R), Positive Affect (PANAS-P), Revised Life Orientation Test (LOT-R), the Digit Symbol Substitution Test (DSST), and the Digit Forward Span Test (DF). The mediation analysis revealed that the total effect coefficient ± SE (−.0034 ± −.0005) of the regression of emotion perception performance on age was significant (t(161) = −7.38, p < .001) and that the direct effect coefficient ± SE (−.0016 ± .0009), when controlling for all mediators, was only marginally significant (t(155) = −1.77, p = .08). Thus, there was an indirect effect, which was estimated via bootstrapping (5,000 resamples; Hayes, 2009). Results in the form of 95% confidence intervals of the estimated effects revealed a significant effect across all mediators, 95% CI [−.0039, −.0003], and for DSST [−.0036, −.0004] and LOT-R scores [−.0009, −.0001], individually. Lower scores were associated with an increased emotion perception deficit for older relative Downloaded by [University of Florida] at 08:19 26 October 2017 10 A. SEN ET AL. Figure 3. Emotion perception plots illustrating the interaction of speaker age and listener age for each emotion. Error bars represent the withinsubject standard error. to younger adults. None of the other individual mediators crossed the significance threshold (STAI2 [−.0003, .0007], ERQ-R [−.0001, .0003], PANAS-P [−.0001, .0006], DF [−.0007, .0001]). We conducted a similar analysis with Sex as a co-variate to explore a possible role of this factor in the relationship between screening measures and emotion perception performance. Results were unchanged, suggesting that effects of sex were negligible. Discussion Although it is well established that older adults experience increasing difficulties with emotion perception, past research has described the phenomenon incompletely. Specifically, previous studies paid little attention to possible speaker age and listener sex effects. Additionally, there has been conflicting evidence on the role of sensory, emotional, and cognitive changes. The present study addressed this situation. Replicating previous work, we found that older adults perform worse than younger adults when categorising vocal expressions and that cognitive and emotional age effects are statistically related to this. Extending previous work, we demonstrate a role of the vocaliser’s age and the receiver’s sex in modulating performance differences between young and old. The following paragraphs deal with these findings in more detail. Own-age versus other-age emotion perception In line with our predictions, we found that listener age effects were more pronounced for younger as Downloaded by [University of Florida] at 08:19 26 October 2017 COGNITION AND EMOTION compared with older speakers. Moreover, emotion perception was better for younger than for older speakers and this effect was more prominent in younger as compared to older listeners. These results agree with the simple additive model of age and an own-age biases proposed in the introduction. They differ only in that both younger and older listeners performed better for younger as compared to older speakers suggesting that the age bias affects listener performance more strongly than the own-age bias. The present findings contradict one previous vocal perception study (Dupuis & Pichora-Fuller, 2015) that found no effects of speaker age in two separate experiments. However, this former study used only two female speakers and compared only 28 or 56 younger adults with 28 older adults and was likely less sensitive than the present protocol. On the other hand, the findings obtained here map onto visual work showing that expressions from young individuals produce overall better emotion perception (Riediger et al., 2011) and that own-age faces dampen emotion perception deficits in older adults (Riediger et al., 2011; but see Ebner & Johnson, 2009). Hence, they extend these visual phenomena into the auditory modality. Additionally, they speak to the mechanisms underpinning speaker age effects proposed previously. The mechanisms that were of interest here concerned age differences in (1) expression accuracy, (2) expression style and the familiarity with that style, as well as (3) the motivation to engage with young and old individuals. The present results offer no support for the first mechanism. A discriminant analysis on acoustic parameters categorised young and old expressions with comparable accuracy and an ANOVA conducted for each acoustic parameter failed to identify speaker age effects. However, because their absence may be due to the relatively small speaker sample or the fact that our acoustic analysis, although detailed, was far from exhaustive, more work is needed to determine whether indeed the ability to encode emotions vocally is preserved in older age. The second mechanism received partial support from both the expression and perception results. The discriminant analysis, but not the ANOVA on voice acoustics, showed that the different vocal parameters differed in the way they contributed to the emotional expressions of younger and older speakers. Whereas younger speakers relied most on expression 11 duration, older speakers relied most on pitch when emotionally modulating speech. This was complemented by an interaction of speaker age and listener age in emotion recognition. As mentioned above, the age gap in listener performance was smaller in the context of older as compared to younger speakers in line with the notion that vocal expressions change with age. Familiarity with these expressions (Wiese et al., 2008) may then differentiate recognition processes in an age-dependent fashion and buffer the performance of older adults when listening to peers. The present results also support the third mechanism, which is based on motivation. This mechanism does not presuppose age differences in expression accuracy or style, but merely assumes that agerelated changes in general voice acoustics allow listeners to infer speaker age and that this in turn influences their motivation to engage with the speaker. Young and own-age speakers are preferred over old and other-age speakers (Gordon & Arvey, 2004). The interaction of speaker age and listener age evident in the perceptual data is in line with this. However, as this interaction agrees also with the familiarity mechanism mentioned above, future research is needed to clearly differentiate the two. This may be achieved, for example, by establishing expressive style differences more clearly using a larger sample of speakers or by manipulating motivational processes through age priming (e.g. Dijksterhuis, Aarts, Bargh, & van Knippenberg, 2000; Kelley, Tang, & Schmeichel, 2014; Packer & Chasteen, 2006). The female emotion perception benefit in young and old Past research has shown greater emotion sensitivity in women than in men at both young and old ages. However, whereas reported sex differences are relatively robust when explored with implicit paradigms (e.g. Schirmer, Chen, et al., 2013), they are more fickle when explored with explicit paradigms, tending to show only when expressions are subtle (Hoffmann et al., 2010). In line with previous studies, we found that women performed more accurately than men. Moreover, as reported by others (Lambrecht et al., 2014), this effect was present for happy and neutral expressions only, which were overall less well-perceived than sad and angry expressions. Although the superior performance of women is well established in the literature, its developmental course has been rarely investigated and results are Downloaded by [University of Florida] at 08:19 26 October 2017 12 A. SEN ET AL. equivocal. Some studies find no sex effects (Lima et al., 2014; Paulmann et al., 2008), whereas others find sex effects independently of age (Ruffman et al., 2010; Sullivan et al., 2017) or show that sex effects increase as individuals get older (Campbell et al., 2014; Demenescu et al., 2014). The present results add to this controversy by suggesting that sex effects decline with age. Whereas for happy expressions, women outperformed men irrespective of age, for neutral expressions the effect was confined to younger women. These differences may be due to methodological choices (e.g. faces vs voices). Additionally, sample size and sampling biases may make it difficult to estimate true sex effects in cross-sectional designs. Moreover, this issue may be especially relevant for studies with small group sizes of less than 20 individuals (Campbell et al., 2014; Demenescu et al., 2014; Lima et al., 2014; Paulmann et al., 2008). Although pending replication, the present study – with 40–43 individuals per group – disagrees with the idea that women are buffered against an agerelated performance decrement. On the contrary, it suggests that emotion recognition deficits become an equal concern for the two sexes. This could be due to sex-specific environmental influences declining with age. For example, stereotype threat concerning the nonverbal abilities of men and women (Koenig & Eagly, 2005) may be greater for young than old individuals because of differences in life experience. Additionally, the biological factors underpinning sex differences may change with age. Declining amounts of circulating sex hormones may reduce the female benefit in emotion recognition. This possibility accords with existing work linking sex hormones with sex differences in nonverbal tasks (Ebner, Kamin, Diaz, Cohen, & MacDonald, 2015; Schirmer et al., 2008). Possible explanations of age effects in emotion perception Although speaker age and listener sex modulated accuracy differences between the two age groups examined here, they did not explain these differences in full. Moreover, the presence of a robust listener age main effect pointed to more fundamental processes compromising the task performance of older participants. Several suggestions have been made as to what these processes might be and the present results contribute to this debate. A first, obvious candidate is an emerging hearing deficit termed presbycusis. Some findings suggest it impairs auditory emotion perception (Lambrecht et al., 2012; Mitchell & Kingston, 2014). However, other studies show it to be irrelevant (Dupuis & Pichora-Fuller, 2015; Orbelo et al., 2005) or to explain age differences in emotion perception incompletely (Lambrecht et al., 2012). The present results align with this latter perspective. Although older participants had poorer hearing than younger participants, this difference was statistically unrelated to the difference in emotion perception. A second possible causal mechanism concerns agerelated changes in the experience of emotions. At present, the literature is divided as to whether all emotions or only positive emotions change. In support of the former position, research has shown that age weakens subjective and bodily responses to both negative (Curtis, Aunger, & Rabie, 2004) and positive events (Levenson et al., 1991; Tsai et al., 2000). Additionally, there is evidence that the amygdala, a brain structure thought to represent stimulus relevance and to be instrumental in both positive and negative affect (Sander, 2012), shows an age-related decline (Allen, Bruss, Brown, & Damasio, 2005; Cacioppo et al., 2011; Malykhin, Bouchard, Camicioli, & Coupland, 2008; Pressman et al., 2016) albeit the relevance of this decline for emotion processing is still debated (Mather, 2012). Taking the latter position, socio-emotional selectivity theory (SST), holds that ageing reduces an individual’s temporal horizon, making short-term outcomes more relevant than long-term outcomes (Carstensen, 2006; Carstensen et al., 1999). Thus, older adults see less value in tolerating immediate distress for a delayed reward and, when given the choice, opt for short-term, hedonic pay-offs. In line with this, there is some evidence that their amygdala reactivity to negative but not positive events is reduced relative to younger adults (Mather et al., 2004). The present study documents emotion changes for a number of measures including positive affect (PANAS-P), negative affect (PANAS-N), optimism (LOT-R) and the use of reappraisal in emotion regulation (ERQ Reappraisal). Compared to younger participants, older participants had a more positive and a less negative affective state, had a less optimistic outlook on their future, and used reappraisal more frequently. Although these effects partially concur with SST, together they failed to explain age differences in emotion perception. The mediation analysis indicated that age differences in optimism were positively related to age differences in emotion perception and that lower optimism partially explained poorer Downloaded by [University of Florida] at 08:19 26 October 2017 COGNITION AND EMOTION emotion perception in older relative to younger adults. Moreover, other emotion measures were irrelevant. Last, it has been argued that declining cognitive processes impair emotion perception. In particular, the abilities to attend to multiple information units and to manipulate these units in working memory (typically conceived of as fluid intelligence; Kyllonen & Christal, 1990) are necessary when listening to sounds and deciding which of several response options is appropriate. As these abilities peak in the early twenties and then slowly wane (Filley & Cullum, 1994; Gazzaley, Sheridan, Cooney, & D’Esposito, 2007; Wechsler, 1955), one can expect them to contribute to age differences in explicit tests of emotion perception. Accordingly, such a contribution has been demonstrated (Krendl & Ambady, 2010; Lambrecht et al., 2012; Lima et al., 2014; Orbelo et al., 2005; Pichora-Fuller, 2003) and was replicated here. Specifically, we observed that working memory (DSST) declined with age and that this decline statistically mediated the relationship between age and emotion perception. Outlook and conclusions Before concluding, we wish to discuss a few questions arising from the present work. One question concerns an apparent difference between emotion effects in the visual and the auditory domain. Whereas happy expressions elicit the highest levels of performance in face perception studies, they elicited the lowest levels of performance here. Moreover, whereas the happy face advantage shows consistently (Kirita & Endo, 1995; Lipp, Craig, & Dat, 2015), results for happy voices are mixed with some converging (e.g. Dupuis & Pichora-Fuller, 2015; Paulmann et al., 2008) and others diverging from the present results (Demenescu et al., 2014). A modality difference in the recognition of happy expressions might be due to differences in the frequency with which people encounter them in the face and in the voice. Perhaps smiling is more prevalent and more readily posed than vocal happiness. Alternatively, the expressive features of facial happiness may be quite salient and readily discerned from those of other emotions. By contrast, the expressive features of vocal happiness may be less discriminating. These and other possibilities are theoretically interesting and should be explored in future studies. A second question arising from this work is whether the statistical relation between age effects 13 on emotion recognition and both working memory and optimism could be replicated with a longitudinal design. Although popular in the literature, the crosssectional approach is confounded by economic progress and life-style changes that may produce generations with different emotion sensitivity (see Lindenberger, von Oertzen, Ghisletta, & Hertzog, 2011 for a discussion of problems with the cross-sectional approach). Future research should tackle this possibility and explore emotion perception with longitudinal data. Moreover, it should aim for a larger age range and the testing of more and other expressions including less common social signals like those of confidence or trust (Jiang & Pell, 2015; Lima et al., 2014). Although not without shortcomings, many methodological features make the present study a valuable addition to existing work on ageing and emotion. Among other features, this includes a large sample size, extensive participant screening, as well as the use of age-matched stimuli that were subjected to extensive acoustic analyses. Its results are hence informative and contribute to our understanding of the relationship between age and vocal emotion perception. First, the results suggest a relevance of both optimism and working-memory in explaining performance differences between young and old. Second, they show that age deficits are greater for expressions from younger as compared to older speakers in line with an additive bias towards young and own-age individuals. Last, our findings indicate that the female benefit in emotion perception persists in older adults. Yet, this benefit becomes smaller, making emotion perception deficits a relevant concern for both older men and women. Disclosure statement No potential conflict of interest was reported by the authors. References Allen, J. S., Bruss, J., Brown, C. K., & Damasio, H. (2005). Normal neuroanatomical variation due to age: The major lobes and a parcellation of the temporal region. Neurobiology of Aging, 26, 1245–1260. Andersen, G. J. (2012). Aging and vision: Changes in function and performance from optics to perception. Wiley Interdisciplinary Reviews: Cognitive Science, 3, 403–410. Banse, R., & Scherer, K. R. (1996). Acoustic profiles in vocal emotion expression. Journal of Personality and Social Psychology, 70, 614–636. Downloaded by [University of Florida] at 08:19 26 October 2017 14 A. SEN ET AL. Boersma, P., & Weenink, D. (2013). Praat: Doing phonetics by computer [Computer program]. Version 5.3.51. Retrieved from http://www.praat.org/ Bonebright, T. L., Thompson, J. L., & Leger, D. W. (1996). Gender stereotypes in the expression and perception of vocal affect. Sex Roles, 34, 429–445. Cacioppo, J. T., Berntson, G. G., Bechara, A., Tranel, D., & Hawkley, L. C. (2011). Could an aging brain contribute to subjective well-being? The value added by a social neuroscience perspective. In A. Todorov, S. Fiske, & D. Prentice (Eds.), Social neuroscience: Toward understanding the underpinnings of the social mind (pp. 249–277). New York: Oxford University Press. Campbell, A., Ruffman, T., Murray, J. E., & Glue, P. (2014). Oxytocin improves emotion recognition for older males. Neurobiology of Aging, 35, 2246–2248. Carstensen, L. L., Isaacowitz, D. M., & Charles, S. T. (1999). Taking time seriously. A theory of socioemotional selectivity. American Psychologist, 54, 165–181. Carstensen, L. L. (2006). The influence of a sense of time on human development. Science, 312, 1913–1915. Colsher, P. L., & Wallace, R. B. (1991). Longitudinal application of cognitive function measures in a defined population of community-dwelling elders. Annals of Epidemiology, 1, 215–230. Colton, R. H., Casper, J. K., Leonard, R., Thibeault, S., & Kelly, R. (2011). Understanding voice problems: A physiological perspective for diagnosis and treatment (4th ed.). Philadelphia: LWW. Curtis, V., Aunger, R., & Rabie, T. (2004). Evidence that disgust evolved to protect from risk of disease. Proceedings of the Royal Society B: Biological Sciences, 271, S131–S133. Decoster, W., & Debruyne, F. (1997). The ageing voice: Changes in fundamental frequency, waveform stability and spectrum. Acta Oto-Rhino-Laryngologica Belgica, 51, 105–112. Dehqan, A., Scherer, R. C., Dashti, G., Ansari-Moghaddam, A., & Fanaie, S. (2012). The effects of aging on acoustic parameters of voice. Folia Phoniatrica et Logopaedica: Official Organization of the International Association of Logopedics and Phoniatrics (IALP), 64, 265–270. Demenescu, L. R., Mathiak, K. A., & Mathiak, K. (2014). Age- and gender-related variations of emotion recognition in pseudowords and faces. Experimental Aging Research, 40, 187–207. Dijksterhuis, A., Aarts, H., Bargh, J. A., & van Knippenberg, A. (2000). On the relation between associative strength and automatic behavior. Journal of Experimental Social Psychology, 36, 531–544. Donges, U.-S., Kersting, A., & Suslow, T. (2012). Women’s greater ability to perceive happy facial emotion automatically: Gender differences in affective priming. PloS One, 7, e41745. Dupuis, K., & Pichora-Fuller, M. K. (2015). Aging affects identification of vocal emotions in semantically neutral sentences. Journal of Speech, Language, and Hearing Research: JSLHR, 58, 1061–1076. doi:10.1044/2015_JSLHR-H-14-0256 Ebner, N. C., He, Y., & Johnson, M. K. (2011). Age and emotion affect how we look at a face: Visual scan patterns differ for own-age versus other-age emotional faces. Cognition & Emotion, 25, 983–997. Ebner, N. C., & Johnson, M. K. (2009). Young and older emotional faces: Are there age group differences in expression identification and memory? Emotion, 9, 329–339. Ebner, N. C., Johnson, M. R., Rieckmann, A., Durbin, K. A., Johnson, M. K., & Fischer, H. (2013). Processing own-age vs. other-age faces: Neuro-behavioral correlates and effects of emotion. NeuroImage, 78, 363–371. Ebner, N. C., Kamin, H., Diaz, V., Cohen, R. A., & MacDonald, K. (2015). Hormones as “difference makers” in cognitive and socioemotional aging processes. Frontiers in Psychology, 5, 505. doi:10.3389/fpsyg.2014.01595 Fecteau, S., Armony, J. L., Joanette, Y., & Belin, P. (2005). Judgment of emotional nonlinguistic vocalizations: Agerelated differences. Applied Neuropsychology, 12, 40–48. Filley, C. M., & Cullum, C. M. (1994). Attention and vigilance functions in normal aging. Applied Neuropsychology, 1, 29– 32. Folstein, M. F., Folstein, S. E., & McHugh, P. R. (1975). “Mini-mental state”. A practical method for grading the cognitive state of patients for the clinician. Journal of Psychiatric Research, 12 (3), 189–198. Fölster, M., Hess, U., Hühnel, I., & Werheid, K. (2015). Age-related response bias in the decoding of sad facial expressions. Behavioral Sciences, 5, 443–460. French, N. R., & Steinberg, J. C. (1947). Factors governing the intelligibility of speech sounds. The Journal of the Acoustical Society of America, 19, 90–119. Gazzaley, A., Sheridan, M. A., Cooney, J. W., & D’Esposito, M. (2007). Age-related deficits in component processes of working memory. Neuropsychology, 21, 532–539. Gordon, R. A., & Arvey, R. D. (2004). Age bias in laboratory and field settings: A meta-analytic investigation1. Journal of Applied Social Psychology, 34, 468–492. Gracco, C., & Kahane, J. C. (1989). Age-related changes in the vestibular folds of the human larynx: A histomorphometric study. Journal of Voice, 3, 204–212. Gross, J. J., & John, O. P. (2003). Individual differences in two emotion regulation processes: Implications for affectrelationships, and well-being. Journal of Personality and Social Psychology, 85, 348–362. Hall, J. A. (1978). Gender effects in decoding nonverbal cues. Psychological Bulletin, 85, 845–857. Hayes, A. F. (2009). Beyond Baron and Kenny: Statistical mediation analysis in the new millennium. Communication Monographs, 76, 408–420. He, Y., Ebner, N. C., & Johnson, M. K. (2011). What predicts the own-age bias in face recognition memory? Social Cognition, 29, 97–109. Hines, M. (2010). Sex-related variation in human behavior and the brain. Trends in Cognitive Sciences, 14, 448–456. Hoffmann, H., Kessler, H., Eppel, T., Rukavina, S., & Traue, H. C. (2010). Expression intensity, gender and facial emotion recognition: Women recognize only subtle facial emotions better than men. Acta Psychologica, 135, 278–283. Hung, A.-Y., & Cheng, Y. (2014). Sex differences in preattentive perception of emotional voices and acoustic attributes. Neuroreport, 25, 464–469. Hunter, E. M., Phillips, L. H., & MacPherson, S. E. (2010). Effects of age on cross-modal emotion perception. Psychology and Aging, 25, 779–787. Isaacowitz, D. M., Löckenhoff, C. E., Lane, R. D., Wright, R., Sechrest, L., Riedel, R., & Costa, P. T. (2007). Age differences in recognition of emotion in lexical stimuli and facial expressions. Psychology and Aging, 22, 147–159. Isaacowitz, D. M., Wadlinger, H. A., Goren, D., & Wilson, H. R. (2006). Is there an age-related positivity effect in visual Downloaded by [University of Florida] at 08:19 26 October 2017 COGNITION AND EMOTION attention? A comparison of two methodologies. Emotion, 6, 511–516. Jiang, X., & Pell, M. D. (2015). On how the brain decodes vocal cues about speaker confidence. Cortex; a journal devoted to the study of the nervous system and behavior, 66, 9–34. Kahane, J. C. (1987). Connective tissue changes in the larynx and their effects on voice. Journal of Voice, 1, 27–30. Kelley, N. J., Tang, D., & Schmeichel, B. J. (2014). Mortality salience biases attention to positive versus negative images among individuals higher in trait self-control. Cognition and Emotion, 28, 550–559. Kirita, T., & Endo, M. (1995). Happy face advantage in recognizing facial expressions. Acta Psychologica, 89, 149–163. Koenig, A. M., & Eagly, A. H. (2005). Stereotype threat in men on a test of social sensitivity. Sex Roles, 52, 489–496. Krendl, A. C., & Ambady, N. (2010). Older adults’ decoding of emotions: Role of dynamic versus static cues and agerelated cognitive decline. Psychology and Aging, 25, 788–793. Kyllonen, P. C., & Christal, R. E. (1990). Reasoning ability is (little more than) working-memory capacity?! Intelligence, 14, 389– 433. Lambrecht, L., Kreifelts, B., & Wildgruber, D. (2012). Age-related decrease in recognition of emotional facial and prosodic expressions. Emotion, 12, 529–539. doi:10.1037/a0026827 Lambrecht, L., Kreifelts, B., & Wildgruber, D. (2014). Gender differences in emotion recognition: Impact of sensory modality and emotional category. Cognition and Emotion, 28, 452–469. Levenson, R. W., Carstensen, L. L., Friesen, W. V., & Ekman, P. (1991). Emotion, physiology, and expression in old age. Psychology and Aging, 6, 28–35. Lima, C. F., Alves, T., Scott, S. K., & Castro, S. L. (2014). In the ear of the beholder: How age shapes emotion processing in nonverbal vocalizations. Emotion, 14, 145–160. Lindenberger, U., von Oertzen, T., Ghisletta, P., & Hertzog, C. (2011). Cross-sectional age variance extraction: What’s change got to do with it? Psychology and Aging, 26, 34–47. Lipp, O. V., Craig, B. M., & Dat, M. C. (2015). A happy face advantage with male caucasian faces: It depends on the company you keep. Social Psychological and Personality Science, 6, 109–115. Malatesta, C. Z., Izard, C. E., Culver, C., & Nicolich, M. (1987). Emotion communication skills in young, middle-aged, and older women. Psychology and Aging, 2, 193–203. Malykhin, N. V., Bouchard, T. P., Camicioli, R., & Coupland, N. J. (2008). Aging hippocampus and amygdala. Neuroreport, 19, 543–547. Mather, M., Canli, T., English, T., Whitfield, S., Wais, P., Ochsner, K., … Carstensen, L. L. (2004). Amygdala responses to emotionally valenced stimuli in older and younger adults. Psychological Science, 15, 259–263. Mather, M. (2012). The emotion paradox in the aging brain. Annals of the New York Academy of Sciences, 1251, 33–49. Mill, A., Allik, J., Realo, A., & Valk, R. (2009). Age-related differences in emotion recognition ability: A cross-sectional study. Emotion, 9, 619–630. Mitchell, R. L. C., Kingston, R. A., & Barbosa Bouças, S. L. (2011). The specificity of age-related decline in interpretation of emotion cues from prosody. Psychology and Aging, 26, 406–414. Mitchell, R. L. C., & Kingston, R. A. (2014). Age-related decline in emotional prosody discrimination. Experimental Psychology, 61, 215–223. 15 Moser, T., Predoehl, F., & Starr, A. (2013). Review of hair cell synapse defects in sensorineural hearing impairment. Otology & Neurotology, 34, 995–1004. Ohno, T., & Hirano, S. (2014). Treatment of aging vocal folds: Novel approaches. Current Opinion in Otolaryngology & Head and Neck Surgery, 22, 472–476. Orbelo, D. M., Grim, M. A., Talbott, R. E., & Ross, E. D. (2005). Impaired comprehension of affective prosody in elderly subjects is not predicted by age-related hearing loss or agerelated cognitive decline. Journal of Geriatric Psychiatry and Neurology, 18, 25–32. Oyler, A. L. (2012). Untreated hearing loss in adults—a growing national epidemic. Retrieved from http://www.asha.org/ Articles/Untreated-Hearing-Loss-in-Adults/ Packer, D. J., & Chasteen, A. L. (2006). Looking to the future: How possible aged selves influence prejudice toward older adults. Social Cognition, 24, 218–247. Paulmann, S., Pell, M. D., & Kotz, S. A. (2008). How aging affects the recognition of emotional speech. Brain and Language, 104, 262–269. Pichora-Fuller, M. K. (2003). Cognitive aging and auditory information processing. International Journal of Audiology, 42 (Suppl. 2), 2S26–2S32. Preacher, K. J., & Hayes, A. F. (2008). Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behavior Research Methods, 40, 879– 891. Pressman, P. S., Noniyeva, Y., Bott, N., Dutt, S., Sturm, V., Miller, B. L., & Kramer, J. H. (2016). Comparing volume loss in neuroanatomical regions of emotion versus regions of cognition in healthy aging. PloS One, 11, e0158187. Radloff, L. S. (1977). The CES-D Scale: A self-report depression scale for research in the general population. Applied Psychological Measurement, 1, 385–401. Rhodes, M. G., & Anastasi, J. S. (2012). The own-age bias in face recognition: A meta-analytic and theoretical review. Psychological Bulletin, 138, 146–174. Riddle, D. R. (Ed.). (2007). Brain aging: Models, methods, and mechanisms. Boca Raton, FL: CRC Press. Retrieved from http://www. ncbi.nlm.nih.gov/books/NBK1834/ Riediger, M., Voelkle, M. C., Ebner, N. C., & Lindenberger, U. (2011). Beyond “happy, angry, or sad?”: Age-of-poser and age-of-rater effects on multi-dimensional emotion perception. Cognition and Emotion, 25, 968–982. Ruffman, T., Halberstadt, J., & Murray, J. (2009). Recognition of facial, auditory, and bodily emotions in older adults. The Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 64B, 696–703. Ruffman, T., Henry, J. D., Livingstone, V., & Phillips, L. H. (2008). A meta-analytic review of emotion recognition and aging: Implications for neuropsychological models of aging. Neuroscience & Biobehavioral Reviews, 32, 863–881. Ruffman, T., Murray, J., Halberstadt, J., & Taumoepeau, M. (2010). Verbosity and emotion recognition in older adults. Psychology and Aging, 25, 492–497. Ruffman, T., Sullivan, S., & Dittrich, W. (2009). Older adults’ recognition of bodily and auditory expressions of emotion. Psychology and Aging, 24, 614–622. Ryan, M., Murray, J., & Ruffman, T. (2009). Aging and the perception of emotion: Processing vocal expressions alone and with faces. Experimental Aging Research, 36, 1–22. Downloaded by [University of Florida] at 08:19 26 October 2017 16 A. SEN ET AL. Sander, D. (2012). The role of the amygdala in the appraising brain. Behavioral and Brain Sciences, 35, 161. Scheier, M. F., Carver, C. S., & Bridges, M. W. (1994). Distinguishing optimism from neuroticism (and trait anxiety, selfmastery and self-esteem): A re-evaluation of the Life Orientation Test. Journal of Personality and Social Psychology, 67, 1063–1078. Schirmer, A. (2014). Emotion. (1st ed.). Thousand Oaks, CA: SAGE Publications, Inc. Schirmer, A., Chen, C.-B., Ching, A., Tan, L., & Hong, R. Y. (2013). Vocal emotions influence verbal memory: Neural correlates and interindividual differences. Cognitive, Affective, & Behavioral Neuroscience, 13, 80–93. Schirmer, A., Escoffier, N., Li, Q. Y., Li, H., Strafford-Wilson, J., & Li, W.-I. (2008). What grabs his attention but not hers? Estrogen correlates with neurophysiological measures of vocal change detection. Psychoneuroendocrinology, 33, 718–727. Schirmer, A., Seow, C. S., & Penney, T. B. (2013). Humans process dog and human facial affect in similar ways. PLoS ONE, 8, e74591. Schirmer, A., Striano, T., & Friederici, A. D. (2005). Sex differences in the preattentive processing of vocal emotional expressions. Neuroreport, 16, 635–639. Spielberger, C. D., Gorsuch, R. L., Lushene, R., Vagg, P. R., & Jacobs, G. A. (1983). Manual for the State-Trait Anxiety Inventory. Palo Alto, CA: Consulting Psychologists Press. Stanley, J. T., & Isaacowitz, D. M. (2015). Caring more and knowing more reduces age-related differences in emotion perception. Psychology and Aging, 30, 383–395. Sullivan, S., Campbell, A., Hutton, S. B., & Ruffman, T. (2017). What’s good for the goose is not good for the gander: Age and gender differences in scanning emotion faces. The Journals of Gerontology: Series B, 72, 441–447. Tsai, J. L., Levenson, R. W., & Carstensen, L. L. (2000). Autonomic, subjective, and expressive responses to emotional films in older and younger Chinese Americans and european Americans. Psychology and Aging, 15, 684–693. Wagner, H. L. (1993). On measuring performance in category judgment studies of nonverbal behavior. Journal of Nonverbal Behavior, 17, 3–28. Watson, D., Clark, L. A., & Tellegan, A. (1988). Development and validation of brief measures of positive and negative affect: The PANAS scales. Journal of Personality and Social Psychology, 54(6), 1063–1070. Wechsler, D. (1955). Manual for the Wechsler adult intelligence scale (Vol. VI). Oxford: Psychological Corp. Wiese, H., Schweinberger, S. R., & Hansen, K. (2008). The age of the beholder: ERP evidence of an own-age bias in face memory. Neuropsychologia, 46, 2973–2985. Williams, L. M., Mathersul, D., Palmer, D. M., Gur, R. C., Gur, R. E., & Gordon, E. (2009). Explicit identification and implicit recognition of facial emotions: I. Age effects in males and females across 10 decades. Journal of Clinical and Experimental Neuropsychology, 31, 257–277. Yamasoba, T., Lin, F. R., Someya, S., Kashio, A., Sakamoto, T., & Kondo, K. (2013). Current concepts in age-related hearing loss: Epidemiology and mechanistic pathways. Hearing Research, 303, 30–38.