Abstract
Dipper functions are valuable tools for investigating visual sensitivity to various stimulus combinations. Previous studies using luminance-defined and contrast-defined stimuli have measured dipper functions using interleaved static stimuli. In this study, luminance-defined and contrast-defined noise blobs were constructed by adding or multiplying random-dot dynamic noise with a Gaussian (sd = 0.25 deg). Test and pedestal blob stimuli were combined on a single frame and the noise varied dynamically to randomise any luminance clumping cues. Thresholds obtained for these stimuli were compared to those combined neurally, i.e. where test and pedestal stimuli were placed on interleaving frames and presented dichoptically. Foveal detection thresholds for luminance-defined and contrast-defined blobs were measured in the presence of luminance-defined and contrast-defined pedestals. When detecting a luminance-defined blob on a luminance-defined pedestal monocularly, a classical dipper shape of threshold versus pedestal visibility is obtained. Facilitation occurs for pedestals close to the detection threshold and masking, for higher pedestal amplitudes (exponent ∼0.6). Thresholds for detecting a contrast-defined blob combined with a contrast-defined pedestal show masking effects of a similar magnitude. Thresholds for detecting a luminance-defined blob on a contrast-defined pedestal show facilitation (∼30–50%) for pedestals around threshold and above (up to ∼10x threshold). No masking is evident. Thresholds for detecting a contrast-defined blob on a luminance-defined pedestal sometimes show facilitation that reduces with increasing pedestal amplitude, or masking at the highest pedestal visibility. Preliminary results for dichoptic viewing reveal similar dipper functions with facilitation, although the masking effects for same-type stimuli are more pronounced (exponent ∼1). These results support parallel processing streams for detecting luminance and contrast modulations with some evidence of asymmetric crosstalk. It would appear that combining different cues for detection both monocularly and dichoptically, leads to improved target detectability rather than masking.
MIH is supported by a Malaysian Government PhD Scholarship through the University of Kebangsaan Malaysia Support also from the Anglia Ruskin University Research Capacity Fund.