Voice as a biomarker of allostatic load: a systematic review of acoustic features and their association with physiological stress

Allostasis refers to the dynamic processes by which an organism maintains physiological stability through changes in response to internal and external stressors. Dysregulation of these mechanisms leads to allostatic load (AL), a state of dysregulated physiology associated with increased risk of chronic disease. Despite its clinical relevance, AL assessment remains challenging due to the absence of a single sensitive biomarker and the reliance on invasive, costly, and time-consuming methods. Recently, non-invasive approaches have gained interest, including acoustic features of the human voice, which may reflect neurophysiological and autonomic dysregulation. This systematic review aims to synthesize current evidence on the associations between voice features and allostatic indicators, in this way evaluating the potential of using voice as a biomarker of AL. A systematic review was conducted according to a Prospero-registered protocol. The research question was structured using the PICO framework. Medline, PubMed, and Scopus were searched for studies published between 2005 and October 2025. Eligible studies included adults aged 18–60 years who reported both acoustic voice analysis and physiological biomarkers related to allostasis. Study selection was performed independently by at least two reviewers in accordance with the PRISMA 2020 guidelines. Risk of bias was assessed using Robins-E. Due to methodological heterogeneity, a meta-analysis was not conducted; findings were synthesized descriptively. Exactly 19 studies ( n = 1045 participants; mean age 26.55 years) met the inclusion criteria. Acoustic parameters, including fundamental frequency (F0), speech rate, and cepstral measures, showed sensitivity to mental, physical, and mixed stressors. Changes in voice were frequently accompanied by physiological indicators of AL, particularly cortisol levels and cardiovascular measures (heart rate, blood pressure). Considerable heterogeneity was observed across study designs, acoustic metrics, and stress-induction protocols. Vocal responses often appeared rapidly, suggesting that voice has the potential to be an early marker of physiological stress. Voice analysis is a promising, non-invasive approach for assessing allostatic load and physiological stress. However, its clinical application is currently limited by methodological inconsistency and lack of standardization. Future research should prioritize the use of multimodal models that integrate both vocal and physiological biomarkers, as well as the use of standard protocols, to improve reliability and translational applicability. PROSPERO CRD420251118646 .