(PP1606) Auditory Emotion Recognition: Test Parameters and Stimuli Optimization

Abstract: The purpose of this study is to determine optimal parameters and stimuli to use for auditory emotion recognition testing. This study specifically looks at how listeners hear, process, and understand information about the person speaking (i.e., how it is being said via emotion recognition). Including emotion recognition testing as part of the clinical audiometric test battery may assess a listener’s social and pragmatic communication beyond word recognition; it offers audiologists and clinicians an additional tool for counseling their patients with hearing loss on realistic expectations for recognizing emotional speechand its impact on conversation.

Summary: Rationale: We commonly are expected to make correct social judgments in conversational situations where communication is difficult (e.g., in noisy backgrounds). Previous research suggests that individuals with normal hearing are able toaccurately perceive socialinformation about a talker (e.g., age, sex, emotional state) even from the talker’s voice in complex listening environments. Many studies have investigated stimulus effects for word recognition, but less well researched is the impact that stimuli selection has on auditory emotion recognition tasks. For example,varying the number of stimuli, the number of talkers, and the number of options in a forced-response paradigm all influence word recognition, and may have similar effects for emotion recognition. Identifying normal auditory processing of talker informationsuch as their emotional statewill require an understanding of how stimuli used in these assessments affect performance. These emotion recognition assessments may yield an improvedunderstanding of the social communication process and how listeners makejudgments (correct or incorrect) about talkers and messages shared in complex listening environments. This study aims to assess the impact of test parameters on emotion recognition performance in order to optimize stimuli to utilize in auditory emotion recognition testing.

Design: Younger adult participants(age 18-35 years) with normal hearingwho are native speakers of American Englishparticipated in this study.Emotion-based questionnaires were administeredused to assess the participant’s emotional state prior to testing.Listeners completedan auditory emotion recognition task by listening to lists of speech stimuli and selecting the emotion theythoughtthey heard in each stimulus from a list of options. We manipulated the emotion recognition tasks using fourprimary conditions, all completed in a single session: 1) stimulus length (words vs sentences), 2) number of stimuli per emotion category (5-20), 3) number of emotion categories (2-8), and 4) number of talkers (2-12).

Results: Emotion recognition performance is higher with sentence stimuli compared to word stimuli. In line with simulations, variability in emotion recognition scores will be lower for an increased number of stimuli per emotion category. Performance will decrease as the number of emotion categories increases. Emotion recognition is reduced with increased variability as the number of talkers increases.

Conclusions: Though there are likely interaction effects between these stimuli parameters, it seems likely that a combination of emotion categories, number of talkers, stimulus length, and number of stimuli will yield different emotion recognition results. We encourage those using auditory emotion recognition testing in research or clinic to carefully consider these variables and anticipated effects when designing their studies. Based on the results of this study, individuals can select appropriate test parameters and optimal stimuli to utilize in auditory emotion recognition testing within the clinical practice for audiologists.