Abstract
It has been argued that natural scene statistics influence perception of structure in naturalistic environments. Specifically, the distribution of oriented content in the environment inversely matches the perception of broadband, orientated structure (e.g., Essock, et al., Vis. Res., 2003; Essock, Haun, & Kim, JOV, 2009). We have previously shown that adapting observers to an environment with atypical content can reliably change perception of structure after only two hours of experience (Schweinhart, Shafto, & Essock, in press). Specifically, observers' visual perception changed in ways that are predicted by a Bayesian model that assumes adaptation to the statistical content of the new environment (Girshick, Landy, & Simoncelli, Nat. Neuro., 2011). Here, we investigated the time course of adaptation to a horizontally-decremented environment. We modified the distribution of amplitudes of visual orientations that subjects encountered via FFT filtering of their environment in near-real-time. The duration during which the subjects experienced the filtered environment was varied from 5 to 240 minutes. Analysis revealed a plateau in the strength of adaptation effects as adaptation duration increased consistent with previous research showing that adaptation effects are best fit by a power function (Dong, Engel, & Bao, Perception, 2014; Hakk, et.al., Curr, Bio ., 2014). By weighting the contribution to the prior of the typical environmental distribution and that of the newly experienced distribution, we were able to effectively model the change in observers' orientation perception across different durations of adaptation in an atypical environment.
Meeting abstract presented at VSS 2017