The Mandarin Chinese auditory emotions stimulus database: A validated corpus of monosyllabic Chinese characters

Mengyuan Li ¹, Na Li ², Anqi Zhou ¹

⁰School of Nursing, Peking University Health Science Center, Haidian District, Room 510, 38 Xueyuan Road, Beijing, 100191, China.

Behavior Research Methods +

|

February 3, 2025

View abstract on PubMed

Summary

This study created the Mandarin Chinese Auditory Emotional Monosyllables (MCAE-Monosyllable) dataset, featuring 422 syllables expressing seven emotions. The database aids research in emotional speech decoding and cross-cultural studies.

Area Of Science

Linguistics
Psychology
Computer Science
Speech Processing

Background

Auditory emotional cues are crucial for communication.
Existing datasets lack Mandarin Chinese emotional monosyllables.
Simple syllables can convey complex emotional information.

Purpose Of The Study

To establish and validate the Mandarin Chinese Auditory Emotional Monosyllables (MCAE-Monosyllable) dataset.
To provide a resource for studying emotional speech in Mandarin Chinese.
To support research in neural decoding, cross-cultural communication, and clinical language studies.

Main Methods

Recorded 422 Mandarin Chinese monosyllables by six professional actors expressing seven emotions (neutral, happy, angry, sad, fearful, disgusted, surprised).
Included four tones for neutral expressions.
Validated recordings with 720 participants rating emotional categories and intensity.
Standardized and balanced audio energy.

Main Results

The final dataset comprises 18,089 valid Chinese monosyllabic pronunciations.
Perceptual validation showed high accuracy rates for neutrality (79%), anger (75%), and sadness (75%).
Detailed acoustic and perceptual intensity data are available for each sound.

Conclusions

The MCAE-Monosyllable database is a unique and valuable resource for Mandarin emotional speech research.
It facilitates neural decoding, cross-cultural language studies, and clinical investigations.
The database is publicly accessible via the Open Science Framework.

Keywords:

Acoustic features Affective prosody comprehension Auditory emotion recognition Chinese emotional monosyllable audio dataset Prosody