Exploring cognition in visual search and vigilance tasks with eye tracking and pupillometry

Recent findings in experimental psychology suggest that pupillometry, the measurement of pupil size, can provide insight into cognitive processes associated with effort and target detection in visual search tasks and monitoring performance in vigilance tasks. With the increasing availability, affordability and flexibility of video-based eye tracking hardware, these experimental findings point to lucrative practical applications such as real-time biobehavioural monitoring systems to assist with socially important tasks in operational settings. The aim of the current thesis was to explore this potential with further experimental work paying close attention to methodological issues which complicate cognitive interpretations of pupillary responses, such as physical stimulus confounds and eye movement-related measurement error in video-based systems. Six original experiments were designed to specifically explore the relationship between pupil size, cognition and behavioural performance in classic visual search and vigilance paradigms. Experiments 1-2 examined the pupillometric effects of effort and target detection in visual search with briefly presented stimuli. Pupil responses showed small variability with respect to manipulations of set size and target presence but were influenced substantially by the requirement for a motor response. Experiments 3-4 explored the cognitive pupil dynamics of free-viewing visual search with data-driven correction for eye movement artefacts. Group-level averages revealed small transient pupil dilations following fixations on targets but not distractors, an effect which was not contingent on a motor response or correction for gaze position artefacts. Experiments 5-6 looked at the relationship between pupil size and detection performance measures in two types of vigilance task. Changes in baseline and stimulus-evoked pupil responses loosely mirrored changes in performance, but the relationships were neither linear nor consistent. Overall, the thesis affirms the practical potential for using cognitive pupillometry in research and applied settings, but emphasises the constraints arising from methodological and theoretical limitations