Age-Related Changes in Acoustic Cue Weighting for Emotional Prosody Identification by Adult Listeners

OBJECTIVES: Age-related decline in emotional prosody perception has been reported in adults, but the mechanisms underlying this aging effect remain unknown. The primary objective of the present study was to investigate the extent to which individuals aged 19 to 72 years with typical hearing (better-ear hearing thresholds ≤25 dB HL from 0.25 to 2 kHz, with some older individuals showing mild or moderate hearing loss in higher frequencies) weight voice pitch contour, duration, and intensity in an emotional prosody identification task, and how this cue-weighting changes with advancing age. An increasing number of aging adults today have hearing loss, and some of them use cochlear implants to hear. Age-related decline in emotional prosody identification has also been reported in cochlear implant patients, but as with acoustically hearing counterparts, the mechanisms are unknown. A second objective of this study was to investigate changes in cue-weighting patterns when stimuli were spectrally degraded to mimic cochlear implant processing and information transmission. A third objective was to study the relationship between participants' cue-weighting and their emotional prosody identification in a task involving naturally recorded materials. DESIGN: To assess cue-weighting, stimuli were orthogonally manipulated to have combinations of the different cues, from values associated with the "happy" emotion to those associated with the "sad" emotion along each of the three dimensions (pitch, duration, and intensity). Participants heard each item in random order and indicated the emotion (happy or sad) they thought was being expressed. Logistic regression analysis of trial-by-trial data yielded cue-weighting coefficients. Response times obtained on a trial-by-trial basis were analyzed to yield additional information about task load when cues were consistent or conflicting with one another. Participants were 50 adults (all native speakers of US English). All had clinically normal hearing in the low frequencies (0.25, 0.5, 1.0, and 2 kHz). Of the 50, 7 (all middle-aged or older) had mild or moderate hearing loss at higher frequencies. To achieve the third objective, emotional prosody identification was assessed in a subset of the participants who had participated in the cue-weighting study. RESULTS: Results showed a significant age-related decline in voice pitch contour weighting, as well as an age-related reduction in the weighting of duration cues, for emotional prosody identification. The presence of mild/moderate hearing loss in the high frequencies also reduced the utilization of voice pitch contour cues. Spectro-temporal degradation resulted in reduced utilization of voice pitch contour cues, but aging effects remained. Response times obtained in the cue-weighting task showed an age-related increase, as well as elevation when cues conflicted with one another. Finally, mediation analysis showed that weighting of voice pitch contour cues in the cue-weighting experiment partially mediated age-related decline in emotional prosody identification observed with naturally recorded stimuli. CONCLUSIONS: This study confirmed previous findings of age-related decline in emotional prosody identification in individuals with good hearing. The results also showed that aging has a negative impact on listeners' access to dynamic changes in voice pitch cues to emotion, which persists when stimuli are degraded to reduce voice pitch information. Finally, the results showed that age-related decline in emotional prosody identification is partially mediated by age-related decline in the ability to use dynamic changes in voice pitch.