Vocal Instability as a Sensitive Biomarker for Driving Stress: Decoupling Cognitive Load and Environmental Friction in a Real-World Dual-Task Protocol

Doğan, Dağhan · 2026 · Crossref

DOI: 10.17798/bitlisfen.1833468

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This study addresses the challenge of distinguishing between internal cognitive load (CL) and external environmental friction (EF) in driver stress detection, a critical gap in existing literature that often lacks ecological validity. While vocal acoustic analysis offers a non-invasive monitoring alternative to physiological sensors, previous research has struggled to decouple stress sources in real-world settings. The authors propose that vocal instability, specifically pitch standard deviation, serves as a sensitive biomarker capable of differentiating these stressors. The research aims to validate this hypothesis through a real-world dual-task protocol, demonstrating how continuous conversation interacts with varying traffic conditions to influence vocal characteristics. The methodology employed a single-subject (N=1) case study design to minimize inter-driver variability. A highly experienced male driver completed six drives across two distinct routes: a 2.5 km urban congestion segment and a 6.5 km hybrid urban-intercity segment. Each route was driven three times at different times of day (11:00, 14:00, and 17:00) to account for temporal traffic variations. The driver maintained continuous verbal communication with a remote interviewer via headset, ensuring a constant cognitive load. Data acquisition included synchronized vehicle kinematics via a mobile GPS application, video recordings of the driver and road, and high-fidelity audio of the driver’s speech. Acoustic features, including pitch, jitter, shimmer, and harmonics-to-noise ratio, were extracted using a custom MATLAB algorithm. A weighted acoustic stress index was calculated, with pitch standard deviation assigned significant weight due to its established correlation with autonomic nervous system arousal. The results indicated that the constant dual-task demand established a stable, moderate stress baseline (approximately 34–36%) across all drives, regardless of traffic congestion levels. This consistency demonstrated that the measured vocal stress was decoupled from routine environmental fluctuations. However, pitch standard deviation proved more sensitive to environmental context than the overall stress index. The long route, which included lower-demand highway segments, exhibited significantly lower pitch standard deviation compared to the pure urban segment. This finding confirms that low-demand driving environments create a stabilizing effect on the voice, even when overall cognitive load remains moderate. Additionally, the 17:00 drive on the short route showed the lowest vocal instability, suggesting potential diurnal variations in driver alertness despite higher traffic impedance. The study concludes that vocal instability is a validated, sensitive biomarker for distinguishing between distraction-related cognitive load and environmental stress. By effectively decoupling these factors, the findings support the development of context-aware in-vehicle systems capable of accurately assessing driver state. This approach enhances the ecological validity of stress detection methods, moving beyond laboratory settings to provide robust, real-time monitoring solutions that can improve transportation safety by identifying the specific sources of driver impairment.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-nvidia on 2026-08-10; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• stress driving
• mental demand
• workload measurement
• music audio mood
• auditory

Information type

What kind of knowledge this paper contributes, grouped by family — independent of topic (what it is about) and method (how it was studied).

• Empirical Findings: physiological data
• Methodological Resource: validation psychometrics
• Theoretical Contribution: theory or model