Investigating Differences in Behavior and Brain in Human-Human and Human-Autonomous Vehicle Interactions in Time-Critical Situations

Unni, Anirudh; Trende, Alexander; Pauley, Claire; Weber, Lars; Biebl, Bianca; Kacianka, Severin; Lüdtke, Andreas; Bengler, Klaus; Pretschner, Alexander; Fränzle, Martin; Rieger, Jochem W. · 2022 · Crossref

DOI: 10.3389/fnrgo.2022.836518

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 human drivers behave and process information differently when interacting with autonomous vehicles (AVs) versus human-driven vehicles (HVs) in time-critical traffic situations. Motivated by evidence that humans may exploit the risk-averse programming of AVs, the researchers aimed to determine if drivers value interactions with AVs differently than with HVs and whether these differences manifest in behavioral metrics and brain activity. The study specifically examined gap acceptance behavior at unsignalized intersections, a common site for traffic accidents, to assess if drivers take greater risks or exhibit different decision-making certainty when facing AVs. The experiment involved 13 participants driving in a full-scale fixed-base simulator while their whole-head brain activity was recorded using functional near-infrared spectroscopy (fNIRS). Participants were instructed that they were driving under time pressure and that AVs (visually distinct yellow cars) were programmed to be defensive and had faster braking reaction times than HVs, although both vehicle types actually followed identical driving behaviors. The task required participants to stop at intersections and accept gaps in oncoming traffic to turn left. To incentivize speed, participants received monetary bonuses for completing blocks within a time limit. The study analyzed behavioral data, including gap acceptance probabilities and safety margins, and neurophysiological data using multivariate logistic ridge regression to decode whether participants were interacting with an AV or HV based on brain activation patterns during the decision-making phase. Behavioral results showed no significant difference in the safety margins used when turning in front of AVs compared to HVs. However, participants demonstrated greater certainty in their decision-making when interacting with AVs, evidenced by a smaller variance in the gap sizes they accepted. Neurophysiologically, the multivariate model successfully predicted whether a participant was deciding to turn in front of an AV or an HV with a mean accuracy of 67.2%. Channel-wise analysis revealed increased brain activation differences for AV interactions in prefrontal areas associated with the valuation of actions, including the dorsolateral and ventromedial prefrontal cortices and the anterior cingulate cortex. These findings suggest that while overt safety behaviors may not change, the cognitive valuation of actions differs significantly depending on the type of traffic agent. The significance of this research lies in its demonstration that human-autonomous vehicle interactions involve distinct neural correlates for decision-making compared to human-human interactions. The ability to decode interaction types from brain activity provides objective measures for understanding mental models and trust in automation. These insights are crucial for developing control systems and human-machine interfaces that can better assess and manage interactions in future mixed traffic environments, potentially mitigating risks associated with the misuse or overreliance on autonomous systems.

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
• automation surprise
• anticipation
• driver vru interaction
• gap acceptance
• automation

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: behavioral performance data
• Theoretical Contribution: computational model, theory or model