Abstract
We performed a meta-analysis of suprathreshold discrimination experiments with overlaid and surround "masks" performed our lab over the past decade. We compared our results to those found in crowding and in masking experiments to determine which phenomenon they more closely resembled. Overlay masking, lateral interactions, surround suppression, and crowding all reduce performance in many psychophysical tasks, but also show distinctive differences suggesting that they are separate processes with different neural substrates (Levi, 2008; Petrov, Popple & McKee, 2007). We describe a series of previously reported and new fine spatial frequency and orientation discrimination results that show reduced performance in the presence of "masks", but taken together, do not fit neatly into either of these categories. In all experiments, observers were highly trained, experienced undergraduates naïve to the purpose of the studies. In most studies, observers discriminated between 2 sinusoidal patterns. On any one trial, a single stimulus was presented. Highly experienced, naïve observers rated their certainty, on a six-point-scale, that pattern A or B had been presented. Performance was measured in d'. In different studies, test and "mask" stimuli were overlaid, were center-surround patterns that abutted or not, were 1[sup]st[/sup] order luminance-modulated gratings or 2[sup]nd[/sup] order contrast-modulated gratings or a mixture, differed in apparent depth or were in the same plane, were in or out of phase, differed in mean luminance, etc. The collective results provide a rich set of fairly homogeneous data, collected under identical laboratory conditions in, to describe characteristics of the mechanisms underlying fine spatial discriminations. Our results resembled both phenomena in some, but not all respects. The comparison is limited in that our experiments were foveal, where it is not clear that crowding occurs at all. In a companion poster, Mingliang Gong (VSS 2013, under review) presents new orientation discrimination data collected under peripheral viewing conditions.
Meeting abstract presented at VSS 2013