Abstract
Adaptation is expected to improve visual performance for prevailing stimuli. Previous studies yielded conflicting results: no change (Barlow, MacLeod & van Meeteren, 1976), gain (Clifford & Wenderoth, 1999), or loss (DeValois, 1977) of sensitivity at the adapting conditions, or change of sensitivity to stimuli vastly different from the adapting conditions (DeValois, 1977; Krekelberg, van Wezel & Albright, 2006). Many of the inconsistencies are resolved, however, when effects of adaptation are studied comprehensively, across the entire range of visible spatiotemporal stimuli: the domain of the spatiotemporal contrast sensitivity function (CSF; Kelly, 1979). A characteristic pattern of local gains and losses of sensitivity is consistent with a global shift of CSF, predicted by a theory of efficient allocation of receptive fields (Gepshtein, Lesmes & Albright, 2013). In the latter study, adaptation was instantiated by varying the distribution of stimuli across trials, so that different distributions of stimulus speeds on different days created a change of stimulus statistics: high-speed and low-speed stimulus contexts ("context adaptation"). This is in contrast to the single adapter typically used in previous studies ("point adaptation"). Does the systematic global pattern of sensitivity changes observed in context adaptation generalize to point adaptation? We measured a "slice" of CSF at six spatial frequencies (between 0.2 and 8 c/deg) at the same temporal frequency (0.5 Hz) using a direction discrimination task. On every trial, the stimulus was preceded by a static or dynamic high-contrast "point adapter" at the same spatial frequency within an experiment. In all conditions and observers, the contrast sensitivity was suppressed by adaptation. But the peak of CSF shifted away from the adapting spatial frequency for both static and dynamic point adapters, consistent with the notion that adaptation generally causes global shifts of CSF.
Meeting abstract presented at VSS 2014