Abstract
Visual adaptation and spatial frequency content are known to have an effect on face processing. The slope in the log-log plots of the Fourier power spectrum of an image is a relative measure for the energy in low and high spatial frequencies. To investigate whether adaptation to noise with different image statistics as compared to a target face affects face processing differentially, we studied the effect of different Fourier power spectrum slopes in an adaptation paradigm. We used a gray image as a control, and noise images with five different slopes in the Fourier power spectrum. These possessed either the same slope as the face images (matching condition), two different steeper slopes (lower spatial frequencies enhanced) or two different shallower slopes (higher spatial frequencies enhanced). In a block design, participants were first adapted to noise images with one of the five slopes (or a gray image) for three minutes, followed by trials, in which an oval cut-out face appeared on a noise (or gray) image that remained on the screen. We measured event-related potentials (ERPs) for the faces while participants performed an age categorization task. ERP results showed that adaptation to noise images compared to the uniform gray control condition enhances face processing. This was evident by an increase of the N170 and a decrease of the P200 component amplitudes in all noise conditions when compared to the gray condition. Moreover, in the P200 time window, adaptation to shallower slopes increased the signal-to-noise ratio more than to steeper slopes, with the matching condition yielding to intermediate results. In conclusion, our data suggest that adaptation to noise images, especially with enhanced high spatial frequencies, facilitates the neural processing of faces.
Meeting abstract presented at VSS 2015