Speech Rate as a Potential Biomarker of Respiratory Health: Evidence from Mandarin Repetitive Articulation Task

This study investigates the relationship between speech rate and respiratory function using a repetitive articulation task in Mandarin Chinese, validating the cross-linguistic applicability of speech-based respiratory assessments. Seventy-four native Mandarin speakers performed a 30-second articulation of the phrase “pīng pāng qiú,” (table tennis or ping pong in Mandarin) during which speech and breathing metrics were collected and analyzed alongside pulmonary function test results. ANOVA results revealed significant effects of breathing group and articulation run on speech rate, while regression analyses demonstrated that Peak Expiratory Flow (PEF) showed the strongest correlation with speech rate among all respiratory parameters. Unlike previous studies, no significant sex differences were observed in speech rate or breathing patterns, which may be due to the analysis being limited to the first two breathing groups. In addition, machine learning regression models were developed for predicting respiratory indices. Random Forest and AdaBoost, outperformed linear statistical methods in this task, highlighting the potential of machine learning in non-invasive respiratory health monitoring. These findings underscore the utility of speech rate as a proxy for expiratory strength and support the integration of AI and language-specific tasks in speech-based respiratory diagnostics.