Physiological fingerprinting of audiovisual warnings in assisted driving conditions: an investigation of fMRI and peripheral physiological indicators.

Physiological responses derived from audiovisual perception during assisted driving are associated with the regulation of the autonomic nervous system (ANS), especially in emergencies. However, the interaction of event-related brain activity and the ANS regulating peripheral physiological indicators (i.e., heart rate variability (HRV) and respiratory rate) is unknown, making it difficult to study the neural mechanism during takeover from the assistance system. In this paper, we established a mapping between the ANS regulation and brain activations of driving events in function magnetic resonance imaging (fMRI)-conditioned audiovisual warnings experiment to add physiological fingerprints for assisted driving. Firstly, we used the general linear model (GLM) to obtain brain activation clusters of driving events and brain activation clusters of peripheral physiological indicators in different frequency bands. Secondly, we redefined the input parameters based on the driving events to calculate the GLM to obtain the brain activation clusters of event-related physiological indicators. Finally, the relationship between the main activation clusters of driving events and the activation of event-related physiological indicators was quantified by the statistical test of the mean-time course of voxels within the region. The results showed that related areas of the brain responsible for movement, visceral autonomic regulation, auditory, and vision actively responded to the audiovisual warnings of automatic driving. The mappings created using them revealed that the correlation between driving event-related activation of brain regions and respiration worked at the onset of audiovisual warnings, especially between the intermediate (IM) and low frequency (LF) bands. For pre-emergency and takeover in audiovisual warnings, the correlations of HRV were dominant, with significant differences among LF, IM and high frequency (HF) bands. At different periods of audiovisual warnings, HRV and respiration play different roles in physiological fingerprints. Compared to respiratory indicators, HRV has higher sensitivity to emergency situations. This study investigates the interaction between driving-related network activity and ANS regulation, revealing the profound connection between driving behavior and neural activity, and contributing to the research of driving assistance systems.