Using fNIRS to Identify Transparency- and Reliability-Sensitive Markers of Trust Across Multiple Timescales in Collaborative Human-Human-Agent Triads
DOI: 10.3389/fnrgo.2022.838625
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 monitoring and calibrating trust in human-agent teams, specifically within collaborative triads consisting of two humans and one artificial agent. As intelligent agents evolve from tools to teammates, maintaining appropriate trust levels is critical to prevent over-reliance or under-utilization. The authors identify a gap in existing research, which often relies on intrusive surveys or limited dyadic interactions, and aims to identify neurophysiological markers of trust using functional near-infrared spectroscopy (fNIRS). The research investigates how agent transparency (providing explanations and certainty levels) and reliability (accuracy of recommendations) influence trust, mental demand, and team cohesion across multiple timescales. The experimental design utilized a geospatial resource allocation task where teams placed crime-prevention resources based on historical data. Thirty-eight participants formed 19 teams, each collaborating with a simulated agent. The researchers manipulated the agent’s transparency (high vs. low) and reliability (high vs. low) across eight rounds. In high-transparency conditions, the agent provided certainty levels (“high” or “medium”), whereas low-transparency conditions labeled certainty as “unknown.” Reliability was manipulated by varying the accuracy of the agent’s internal predictions. Data collection included fNIRS measurements of brain activity, galvanic skin response, eye-tracking, behavioral logs, and post-round surveys assessing trust, mental demand, and team processes. The results demonstrated that both transparency and reliability significantly affected trust in the agent. High transparency facilitated calibrated trust without increasing mental demand, whereas low transparency required greater cognitive effort for credibility assessment. Reducing agent communication disrupted interpersonal trust and team cohesion, mirroring dynamics in human-human teams. Neuroimaging analysis revealed that dorsal medial prefrontal cortex (DMPFC) activation was specific to assessing the agent’s transparency explanations. Increases in mental demand were characterized by activation in the dorsal lateral prefrontal cortex (DLPFC) and frontopolar area (FPA). Furthermore, short-scale event-level analysis showed that fNIRS data from 15 seconds prior to a decision could predict whether an individual would trust the agent, highlighting the predictive power of these neural markers. The significance of this work lies in its demonstration that fNIRS can effectively identify neural correlates of trust in ecologically valid, multi-person human-agent teams. By linking specific brain regions to transparency assessment and mental demand, the study provides a foundation for developing real-time, intelligent trust-modulation systems. These systems could dynamically adjust agent behavior to optimize collaboration, ensuring that human operators maintain appropriate trust levels without excessive cognitive load. The findings support the integration of neuroergonomics into human-computer interaction, offering a pathway to continuous, unobtrusive monitoring of team states in complex collaborative environments.
Key finding
Transparency and reliability manipulations significantly affected trust in the agent, with specific neural markers in the prefrontal cortex enabling the prediction of trust decisions from fNIRS data collected shortly before the decision.
Methodology
lab_experiment
Sample size: 38
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. Discovered via author_sweep_intake on 2026-05-28.
| Stage | Outcome | Tool | Model | Prompt | Attempts | Completed |
|---|---|---|---|---|---|---|
| discover | success | author_sweep | — | — | 2 | 2026-05-28 |
| archive | success | canonical_url | — | — | 1 | 2026-06-04 |
| extract | success | cached | — | — | 3 | 2026-06-10 |
| clean | success | clean | — | — | 1 | 2026-06-04 |
| chunk | success | chunk | — | — | 1 | 2026-06-04 |
| embed | success | embed | Qwen/Qwen3-Embedding-8B | — | 1 | 2026-06-04 |
| enrich | success | — | — | — | 1 | 2026-05-28 |
| promote | success | — | — | — | 1 | 2026-06-04 |
| summarize | success | llm | qwen3.6-27b-prismaquant | summ-v5 | 2 | 2026-06-10 |
| tag | success | vector_similarity | — | — | 15 | 2026-06-11 |
| verify | success | — | — | — | 2 | 2026-06-10 |
Summary generated by qwen3.6-27b-prismaquant on 2026-06-10; verification: verified.
Topics
Ranked by relevance to this paper. Hover a topic for its definition.
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: self report data