Abstract
It has been suggested that adaptation to texture density only ever reduces, i.e. never increases, perceived density, implying that density adaptation is 'uni-directional' and that texture density is coded as a scalar attribute (Durgin & Huk, 1997). However we have recently shown that simultaneous density contrast, which describes the effect of a surround texture on the perceived density of a centre region, is 'bi-directional' - that is, not only do denser surrounds reduce perceived density of the center but sparser surrounds enhance it (Sun, Baker, & Kingdom, 2016). Therefore we decided to re-examine the directionality of density adaptation. To do this we measured the density aftereffect in random dot patterns using a 2AFC matching procedure that established a PSE (point-of-subjective-equality) between an adapted test patch and an unadapted match patch. The adaptors and test were presented at the same position, either at top left or bottom right of the fixation. The match was presented at bottom left or top right correspondingly. These positions were fixed within a block and switched between blocks. In the first experiment we established that bi-directionality could indeed be obtained, provided the test and match were presented sequentially not simultaneously. Then, using sequential presentation, we measured the density aftereffect for a wide range of adaptor and test densities. We found bi-directionality for all combinations of adaptor and test densities, with the exception of one of the test conditions in one of the four observers. In line with our previous results with simultaneous density contrast, this evidence supports the idea that there are multiple channels selective to texture density in human vision.
Meeting abstract presented at VSS 2017