Computing perceived gaze direction

In this thesis, I will briefly introduce the perception of others’ gaze direction as an integral part of social cognition. I will then discuss experimental findings that underpin our understanding of how the perception of gaze direction is achieved in the human brain. The main body of this thesis presents a number of experiments that were conducted to further address how a variety of sensory cues are weighted in the computation of perceived gaze direction. Chapter One examines a bias-minimising measure of the effect of head orientation on perceived gaze direction. This measure produced pronounced differences from previous measures, which is explored further in Chapter Two. The findings of Chapter Two suggest that these differences are not the result of some perceptual artefact concerning differences in stimulus presentation, but rather in the way observers are using cues to gaze direction depending on whether observers are taking an allocentric or egocentric perspective. This could be the result of integrating nondirectional cues to direct gaze when making judgements of whether gaze is direct. The use of one such cue, the circularity of the pupil/iris, is examined in Chapter Three. Although it is found that the circularity cue is not driving the task dependent differences in the effect of head orientation on perceived gaze direction, the results suggest a further, computational, difference in the integration of information for allocentric (directional) and egocentric (non-directional) gaze judgements. This difference is explored in Chapter Four, where a double judgements procedure is used to elucidate the computational differences between detecting direct gaze and identifying gaze direction, and the implications of this difference on perceived gaze direction. Finally, Chapter Five accounts for perceived gaze direction, which can differ from veridically presented gaze direction, in objective measures of the range of gaze deviations the observer is willing to accept as directed at them. These experiments suggest that observers are able to flexibly integrate cues to gaze direction depending on the judgement to be made, and that consideration of decision-making factors is important in understanding the perception of gaze direction from observers’ responses in an experimental setting