Abstract
Despite our constantly changing environment, the visual system remains efficient through the process of visual adaptation. While this rapid form of neural plasticity has been widely studied, it remains unknown whether the rate at which it operates can change through experience. Here we addressed this question by repeatedly adapting subjects to naturalistic visual input with stereotypical orientation statistics, which we term a context. In 4 hours session on 3 consecutive days, subjects (n=8) viewed their surroundings via a head-mounted display with an attached camera, whose input was filtered so as to produce novel visual statistics. In each session, filtering alternated every 30 min between two visual contexts: One where contrasts of orientations around vertical were enhanced and contrast around horizontal was reduced, and one with the opposite pattern of filtering. Viewing inputs under both these conditions produced contrast adaptation, which we measured using the tilt-aftereffect (TAE), i.e. shifted perception of orientations near the adapted axes. To measure the TAE, subject completed 2.5 min blocks of trials where they viewed a plaid pattern (100 ms presentation with 2 sec ISI) consisting of +/- 45 deg gratings, and used the mouse to adjust the gratings' orientations until the plaid appeared to contain square checks. These measures were taken after 2 and 30 min of adaptation within each context, and were compared across days. Our paradigm induced robust TAE in both conditions (0.6 deg and 1.5 deg within the first 30 sec of the block, both p < 10-4), which decayed over the course of the block. However, comparison across days showed no significant changes in TAE magnitude or decay rates. These results suggest that the degree of familiarity with the statistical structure of visual environments does not alter the rate of visual adaptation, at least for the timescales and contexts tested here.
Meeting abstract presented at VSS 2016