Abstract
The discrimination of two stimuli can be measured as a contrast threshold at a constant stimulus difference (i.e., +/− 45 deg orientation) or as an orientation threshold at a given contrast. Both can be understood together within an elaborated perceptual template model (ePTM) that incorporates the effects of judgment precision, or non-orthogonal stimuli for discrimination (Jeon, Lu, & Dosher, 2008). Perceptual learning improvements in contrast thresholds have been studied in different external noise conditions (Dosher & Lu, 1998, 1999), while stimulus difference thresholds have not. Here we investigate orientation psychometric functions and thresholds and perceptual learning. (1) Orientation thresholds in zero noise and in high noise improve with practice in a peripheral orientation discrimination experiment in which the angles of discrimination are altered by an adaptive staircase; threshold angle difference narrowed gradually with practice. (2) Orientation psychometric functions were measured at three contrast levels and eight external noise levels to measure orientation threshold TvC functions at fovea; consistent with prior reports (Lu & Dosher, 2004) foveal performance showed little or no effect of learning comparing the first to the last half of sessions. (3) A similar experiment measured orientation discrimination at 5 deg in the periphery; peripheral discrimination performance showed perceptual learning between the first and last half of sessions, with larger learning effects in high external noise. In both (2) and (3) the ePTM provided a very good account of the orientation psychometric functions in different levels of external noise. These results widely validate the elaborated perceptual template model in accounting for accuracy as a joint function of orientation difference, contrast, and external noise, with perceptual learning resulting in reduced impact of external and additive internal noises.
Supported by the National Eye Institute.