How Do Drivers Perceive Risks During Automated Driving Scenarios? An fNIRS Neuroimaging Study

Perello-March, Jaume; Burns, Christopher G.; Woodman, Roger; Birrell, Stewart; Elliott, Mark T. · 2023 · Crossref

DOI: 10.1177/00187208231185705

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 investigates how drivers perceive risk during highly automated driving (HAD) scenarios, addressing a critical gap in understanding the cognitive states of drivers who are disengaged from active control. While risk perception is well-documented in manual driving, its neural correlates during automation remain unclear. The authors posit that functional near-infrared spectroscopy (fNIRS) can serve as a novel, objective measure of perceived risk by monitoring prefrontal cortical haemoglobin oxygenation, which reflects central executive cognitive processes involved in hazard evaluation. The research aims to determine if drivers maintain situation awareness and perceive risk when traffic complexity increases or when hazardous events occur, despite delegating control to the automated system. The experiment involved 20 participants (after excluding three for motion sickness) in a high-fidelity driving simulator equipped with automated driving capabilities. Participants experienced two primary conditions: a suburban/urban scenario with gradually increasing traffic complexity and potential unmaterialised hazards, followed by a materialised hazardous event involving a sudden collision with a truck. Neurophysiological data were collected using a wearable fNIRS device measuring oxygenated (HbO) and deoxygenated (HbR) haemoglobin in the prefrontal cortex. Subjective risk perception and trust in automation were assessed via self-report questionnaires. Data preprocessing included motion artefact correction and band-pass filtering, with analysis focusing on block-averaged haemodynamic response functions standardized against baseline resting states. Results indicated that prefrontal cortical haemoglobin oxygenation levels significantly increased in response to both self-reported perceived risk and traffic complexity. Specifically, HbO levels rose during the urban driving condition with moderate traffic complexity, suggesting that drivers actively engaged in risk assessment even when not manually driving. This activation intensified significantly during the hazardous scenario involving the sudden truck appearance. The findings support the hypothesis that drivers do not remain entirely "out-of-the-loop"; rather, they perceive moderate risk as traffic complexity builds and exhibit heightened neural activity during acute hazards. These objective neural measures aligned with subjective self-reports, confirming that fNIRS can effectively capture variations in risk perception. The study concludes that fNIRS is a valuable tool for assessing driver mental states during automated driving, offering insights beyond traditional peripheral physiology metrics like heart rate. The findings imply that drivers maintain a degree of situation awareness and risk perception during HAD, which has important implications for the design of human-automation interaction systems. Understanding these cognitive processes can inform the development of better monitoring systems and interfaces that support safe take-over requests, ultimately promoting the safe adoption of automated driving technology.

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.

• situational awareness
• risk taking
• automation
• automation surprise
• decision making risk perception
• hazard perception

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
• Theoretical Contribution: computational model, theory or model