Use this URL to cite or link to this record in EThOS:
Title: Phenomenal regression as a potential metric of veridical perception in virtual environments
Author: Elner, Kevin William
ISNI:       0000 0004 6057 6909
Awarding Body: University of Hull
Current Institution: University of Hull
Date of Award: 2015
Availability of Full Text:
Access from EThOS:
Access from Institution:
It is known that limitations of the visual presentation and sense of presence in a virtual environment (VE) can result in deficits of spatial perception such as the documented depth compression phenomena. Investigating size and distance percepts in a VE is an active area of research, where different groups have measured the deficit by employing skill-based tasks such as walking, throwing or simply judging sizes and distances. A psychological trait called phenomenal regression (PR), first identified in the 1930s by Thouless, offers a measure that does not rely on either judgement or skill. PR describes a systematic error made by subjects when asked to match the perspective projections of two stimuli displayed at different distances. Thouless’ work found that this error is not mediated by a subject’s prior knowledge of its existence, nor can it be consciously manipulated, since it measures an individual’s innate reaction to visual stimuli. Furthermore he demonstrated that, in the real world, PR is affected by the depth cues available for viewing a scene. When applied in a VE, PR therefore potentially offers a direct measure of perceptual veracity that is independent of participants’ skill in judging size or distance. Experimental work has been conducted and a statistically significant correlation of individuals’ measured PR values (their ‘Thouless ratio’, or TR) between virtual and physical stimuli was found. A further experiment manipulated focal depth to mitigate the mismatch that occurs between accommodation and vergence cues in a VE. The resulting statistically significant effect on TR demonstrates that it is sensitive to changes in viewing conditions in a VE. Both experiments demonstrate key properties of PR that contribute to establishing it as a robust indicator of VE quality. The first property is that TR exhibits temporal stability during the period of testing and the second is that it differs between individuals. This is advantageous as it yields empirical values that can be investigated using regression analysis. This work contributes to VE domains in which it is desirable to replicate an accurate perception of space, such as training and telepresence, where PR would be a useful tool for comparing subjective experience between a VE and the real world, or between different VEs.
Supervisor: Wright, Helen Sponsor: University of Hull
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available
Keywords: Computer science