Exploring Differences in Performance-related Neural Dynamics Between a Well-controlled Laboratory Setting and a Naturalistic Condition

Abstract

Many studies have reported important neural correlates of fluctuations in task performance in sustained-attention tasks. However, nearly all of these studies were conducted in the confines of standard stationary research laboratories. It has been argued that fundamental differences between laboratory-based and naturalistic human behavior may exist. Thus, it remains unclear how well the current knowledge of neural correlates of human performance translates into the highly dynamic, real world. This study systematically explores the link between electroencephalographic (EEG) activities with task performance using an event-related lane-departure paradigm implemented in a virtual-reality (VR) driving simulator on a motion platform. Each subject participated in a ‘motionless’ (inactivated platform that resembles a regular EEG laboratory settings) and a ‘motion’ (activated motion platform that mimics real-world driving with visual/auditory/motion cues) sessions of long-haul driving, during which steering behavior and 30-channel EEG signals were recorded simultaneously. This study also proposes and employs transformative source-level EEG analysis and effective connectivity analysis to examine whether changing the perceptual demand by manipulating amount of sensory inputs would promote a shift of brain activity between different modes of processing, reflected by brain network dynamics on electroencephalographic sources. More importantly, this study demonstrates an integrative and multi-aspect approach to bridge laboratory and real world neuroscience research.