Objective:
This study aims to investigate the acoustic features of speech and voice in individuals with Major Depressive Disorder (MDD) across two linguistically distinct populations: native German and Cantonese speakers. The objectives are twofold: (1) to identify speech tasks and emotional stimuli that best differentiate individuals with varying degrees of depression, and (2) to examine whether these acoustic markers are consistent across languages, focusing on the impact of linguistic structures on vocal expression in MDD.

Methods:
This cross-sectional, comparative study will recruit 40 participants aged 18–60 years, with 20 diagnosed with MDD (10 mild, 10 moderate) and 20 healthy controls. Participants will be drawn from University Hospital Bonn (Germany) and Queen Mary Hospital (Hong Kong). Each participant will complete a series of emotionally neutral (e.g., vowel prolongation, counting, reading aloud) and emotionally arousing tasks (e.g., picture description, video commentary). Acoustic parameters—including fundamental frequency (f0), jitter, shimmer, harmonic-to-noise ratio (HNR), and speech rate—will be analyzed using Praat and MATLAB. Emotional valence and arousal ratings will be obtained prior to the data collection to standardize the emotional stimuli across tasks. Independent sample t-tests and mixed-effects models will be used to analyze differences between MDD and control groups, as well as cross-linguistic comparisons between German and Cantonese speakers.

Results:
Preliminary findings suggest that specific acoustic features, particularly reduced pitch variability, increased jitter, and slower speech rate, are prevalent in individuals with MDD. Emotionally arousing tasks, such as autobiographical memory recall, appear to elicit more pronounced differences in these voice characteristics compared to neutral speech tasks. Given the tonal nature of Cantonese and the stress-based phonology of German, we anticipate both cross-linguistic consistencies in general depressive markers and language-specific variations, such as differences in pitch modulation and vowel articulation.

Conclusions:
This study will expand the understanding of how linguistic factors influence the manifestation of depression in vocal and paralinguistic parameters. The findings are expected to contribute to the development of non-invasive, language-agnostic diagnostic tools that leverage vocal and speech biomarkers for early detection and monitoring of MDD. Cross-linguistic comparisons may highlight the need for tailored approaches in multilingual populations, paving the way for clinical tools that can be applied across diverse linguistic contexts.