Abstract
Introduction. In psychophysical experiments stimuli are sometimes scaled (magnified) at each eccentricity in an attempt to compensate for eccentricity-dependent resolution losses. Applying a preselected magnification factor in some cases leads to non-monotonic changes in performance as a function of eccentricity. We argue that such non-monotonic changes arise when performance at each eccentricity is limited by more than one type of resolution loss.
Methods. Poirier and Gurnsey (2002, Vision Research) proposed a method for characterizing the scalings needed to compensate for multiple, eccentricity-dependent resolution losses. This method generalizes several earlier scaling procedures (e.g., Watson, 1987, JOSA). In several experiments we have found that different sources of sensitivity loss may scale differently with eccentricity. Therefore, a stimulus magnified at each eccentricity by a predetermined scaling factor may be unresolved foveally by a one mechanism and unresolved peripherally by another mechanism. Performance may therefore vary non-monotonically over a range of eccentricities.
Results. Using the analytical framework of Poirier and Gurnsey (2002, Vision Research) we show how predetermined scaling factors lead to non-monotonic changes in performance in an orientation discrimination task (Poirier & Gurnsey, 1998, Spatial Vision). We also show how the analysis can explain non-monotonic changes in performance arising from unscaled stimuli; specifically, the central performance drop reported by Kehrer (1987, Spatial Vision) and a case of ‘reverse scaling’ reported by Tyler (1999, Visual Neuroscience).
Discussion. We conclude that most eccentricity research can be explained by multiple scaling theory as extended here, where all underlying mechanisms in a task increase in size linearly with eccentricity, but not necessarily at the same rate.