Neural Signatures of Trust During Human-Automation Interactions
Technical Report,01 Jan 2013,31 Dec 2015
George Mason University Fairfax United States
Pagination or Media Count:
The objective of this proposal was to investigate the similarities and differences of the neural systems of human-automation trust HAT and human-human trust HHT in a series of three studies that combined an X-ray luggage-screening task with functional magnetic resonance imaging by manipulating the reliability of advice from a human or automated luggage inspector framed as experts. HAT and HHT were measured as the acceptance rates of advice either giving by the machine or the human agent. Comparing HAT with HHT, those studies provide first neural evidence that reliable study 1 and unreliable false alarm study2 and misses study 3 human-automation interactions evoke unique brain activation patterns linked with the reward network for reinforcement learning e.g., dorsal striatum head, ventromedial prefrontal cortex, the mentalizing network for evaluating personal characteristics and traits e.g., precuneus, temporoparietal junction, and the salience network for interoception e.g.,insula, anterior cingulate cortex. The findings are relevant to the Air Force Office of Scientific Researchs mission aimed at investing in the discovery of the foundational concepts of trust building and trust calibration during complex human-machine interactions.