Date of Award

January 2020

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Psychology

First Advisor

Dmitri Poltavski

Abstract

Automated systems (e.g., self-driving cars, autopilot) can reduce an operator’s (i.e., driver, pilot, baggage screener) task engagement, which can result in mind wandering, distraction, and loss of concentration. Consequently, unfavorable performance outcomes, such as missed critical signals and slow responses to emergency events, can occur. Because automation reverts the operator to a “visual monitoring” role, the oculomotor accommodative-vergence responses (the oculomotor responses that maintain a single focused image on the retina) may play a vital role in human-automation interactions. Prior research has shown that individuals with deficits in the accommodative-vergence responses can exhibit inattentive symptoms (e.g., poor concentration) characteristic of attention-deficit/hyperactivity disorder (ADHD) while performing prolonged close work (e.g., reading). Given the behavioral symptoms present in those experiencing accommodative-vergence stress, automated systems may exacerbate these negative effects. The current study examined the effects of accommodative-vergence stress in combination with automation on aspects of operator task engagement. Participants (N = 95) under accommodative-vergence stress wearing -2.0 diopter lenses or normal viewing conditions completed a 40 min flight simulation task either with or without automation. Physiological dependent measures included electroencephalographic (EEG) parietal-occipital alpha power spectral density (PSD), an EEG multivariate metric of engagement, and pupil diameter. Self-report measures of task engagement, cognitive fatigue, and visual fatigue symptoms were also collected along with oculomotor measurements (accommodation and convergence) and flight simulation task performance. Multivariate analyses indicated that the application of -2.0 diopter lenses did not significantly alter oculomotor measurements or subjective reports of visual fatigue. Oculomotor stress modestly affected task performance and tended to result in increased EEG measures of engagement, while subsequently increasing feelings of fatigue, potentially indicating a compensatory effort response. Participants performing the simulation with automation exhibited significantly lower task engagement, as indicated by greater parietal-occipital alpha PSD, less multivariate EEG engagement, smaller pupil diameter, and lower self-reported engagement. Overall, oculomotor stress and automation did not interact synergistically to affect task engagement and associated performance outcomes. Automation and time on task were the main determinants of task engagement. These results underscore the negative effects automation can have on underlying operator cognitive states and the associated need to carefully design automation to combat reduced task engagement. Applications for system design and the use of EEG in augmented cognition systems involving automation are discussed.

Share

COinS