Week 3. What is the mechanism of crowding?
– Attention is needed at multiple stages of processing –
Hunjae Lee (VCC Lab)
There are many theories explaining crowding as much as it is vigorously studied. Although it is under lively competition, now two major explanations seem to be prominent. One is attentional resolution hypothesis (He et al., 1996) that say attention has a limited size of selecting visual stimuli in the region and another is compulsory averaging hypothesis (Parkes et al., 2001) that specifies pooling features of the flankers near targets together makes perception difficult. This look like top-down versus bottom-up, famous visual dichotomy applies them again. But I think they can eventually reach to a compromise; “attention”
Compulsory averaging hypothesis have disadvantages to explain various stages that crowding occurs. Many studies show a slight different region of crowding effect such as ‘at least after V2’ (Arman et al, 2006) , ‘the stage’ that binocular rivalry happens or after it (Kim, 2009), ‘after V5’ that apparent motion involved (Rajimehr et al., 2004) and even holistic crowding at somewhere higher region like FFA (Louie et al, 2007). Successful detection but not integration of features don’t contradict some of those results, but still integrations in relatively earlier cortical areas are not sufficient to explain whole crowding phenomena like holistic crowding.
On the other hand, attentional resolution hypothesis looks more flexible. In order to explain “supercrowding” that breaks Bouma’s law, simply shortage of feature integrators, “attention” can be a good explanation candidate (Vickery, T. J., 2009). This is a point of reconciling two hypothesis.
Without signal from low, physical level object never result in various visual perceptions including crowding effect, whatever. However, almost every stage of visual awareness, attention helps us a lot. In low level of processing it glues features altogether. After integrating features, selecting one object from many distracters and making it conscious need attention in relatively later processing. It seems that both top-down and bottom-up information are working together from very low level of processing indeed. Bilateral advantage (Chakravarthi et al, 2009) and lower visual field advantage (He et al, 1996) in crowding are in the line of this explanation.
Arman, A. C., Chung, S. T. L., & Tjan, B. S. (2006). Neural correlates of letter crowding in the periphery. Jornal of Vision, 6(6), 804.
Chakravarthi, R., Cavanagh, P. (2009). Bilateral field advantage in visual crowding. Vision Research 49, 1638-1646
He et al. (1996). Attentional resolution and the locus of visual awareness. Nature, 383, 334-338.
Kim, S. R. (2009). Interaction between crowding and binocular rivalry. Graduate program in Cognitive Science, Yonsei University, Korea.
Louie, E. G., D. W. Bressler, et al. (2007). “Holistic crowding: selective interference between configural representations of faces in crowded scenes.” Journal of Vision 7(2): 24 1-11.
Parkes et al. (2001). Compulsory averaging of crowded orientation signals in human vision. Nature Neuroscience, 4, 739-744.
Pelli, D. G. and K. A. Tillman (2008). “The uncrowded window of object recognition.” Nature Neuroscience 11(10): 1129-35.
Rajimehr et al. (2004). Adaptation to apparent motion in crowding condition. Vision Research, 44, 925-931.
Vickery, T. J., W. M. Shim, et al. (2009). “Supercrowding: weakly masking a target expands the range of crowding.” Journal of Vision 9(2): 12 1-15.