Abstract
Mind-wandering, or off-task thought, is believed to be associated with diminished processing of the external sensory environment, while focus is directed at internal dialogue. However, the impact of mind-wandering on driving performance is poorly understood. This study explores the behavioral and EEG oscillatory correlates of mind-wandering while driving in a driving simulator. Drivers followed a lead-vehicle which braked intermittently throughout long drives with random lateral winds. They were asked to report their mind-wandering states both when they caught themselves (self-caught mind-wandering) and when probed by a beep sound (probe-caught mind-wandering) by pressing a button labeled 'yes' on the steering wheel. An additional 'no' button indicated they had been attentive when probed (attentive). Driving dynamics and 32-channel EEG data were compared in the ten second period prior to auditory probes using moving window averaging. Subjects reported mind-wandering about 37% of the time when probed. Results show that mindless drivers drove at higher speed and followed the lead-vehicle at a further distance prior to probes compared to when they were attentive. Evoked ERP activity elicited by probes occurring during mind-wandering showed attenuation of both the N2 and P3 components compared to those presented during attentive periods, revealing diminished processing of external stimuli. The power of delta, alpha, and theta oscillations was also larger during pre-probe mind-wandering than when attentive. Using a multivariate Gaussian mixture model on this time-frequency EEG data over frequency, time, and channels, within-subject classifiers were able to classify untrained pre-probe data as mind-wandering vs. attentive with over 90% sensitivity and specificity, revealing important markers of mind wandering and dangerous driving. This study reveals behavioral and electrophysiological indices that distinguish mindless from mindful driving and provides a foundation for monitoring of drivers whose minds wander behind the wheel.
Meeting abstract presented at VSS 2014