Abstract
Visual awareness is thought to have a limited capacity. Therefore, selection of information for awareness can massively impact how we experience the world. To investigate the relative priority for different images to enter awareness, many scientists have turned to an interocular masking paradigm known as Continuous Flash Suppression (CFS). The benefit of CFS is that it allows a certain level of control over which eye's image is perceived first. Namely, while one eye's image is a dynamic mask, the other eye's static image is suppressed from awareness. Due to the dynamic nature of the mask, the similarity between the two eye's images (Interocular Image Similarity; IIS) is not necessarily the same between trials or images. Such variations in IIS may affect the degree of interocular suppression and interfere with the interpretation of CFS suppression durations. In fact, previous studies using schematic images have indeed shown that suppression is more effective when IIS is higher. However, IIS can be computed in many ways and the predicted influence on CFS suppression durations is not necessarily equivalent over different computations. We first asked if CFS suppression durations for natural images are related to IIS. Results show significant relations between CFS suppression durations across different computations of IIS. To be able to suggest a preferred approach for future CFS studies, we next compared reliability, quantified using a cross validation procedure, between the different indices of IIS. Our results suggest that local-luminance similarity indices are more reliable compared to Fourier amplitude spectrum based similarity indices. Finally, we go on to show that removing the influence of IIS substantially influences the outcome of statistical comparisons of CFS suppression durations for different classes of images. We suggest that removing IIS-based influences on CFS suppression durations should be a standard pre-processing step when analyzing access to awareness results.
Meeting abstract presented at VSS 2018