Grouping by proximity is the principle of perceptual organization by which the elements of a visual scene which are closer in space tend to be perceived as a coherent ensemble. Research into this topic makes substantial use of the class of stimuli known as dot lattices. The Pure Distance Law (Kubovy et al., 1998) predicts that the probability of grouping by proximity in these stimuli only depends on the relative inter-dot distance between competing organizations. Despite much effort to explain how grouping by proximity is shaped by the basic organization of visual stimuli, its neural mechanisms are still under debate. Moreover, previous studies reported that grouping in dot lattices also occurs according to an orientation bias, by which these stimuli are perceived along a preferred orientation (vertical), regardless of what predicted by the Pure Distance Law. The aim of this thesis is to shed light on the functional and neural mechanisms characterizing grouping by proximity in dot lattices, as well as the trade-off between proximity- and orientation-based grouping. Study 1 investigates the role of high-level visual working memory (VWM) in promoting for the shift between grouping by proximity and orientation bias. Both the quantity (load) and the quality (content) of the information stored in VWM shape online grouping for dot lattices. Study 2 presents a neural network model simulating the dynamics occurring between low- and high-level processing stages during dot lattices perception. The degree of synchrony between the units at low-level module has a key role in accounting for grouping by proximity. Overall, our results show that high-level (Study 1) and low-level (Study 2) operations contribute in parallel to the emergence of grouping by proximity, as well as to its reciprocity with orientation-based grouping.
Perceptual grouping by proximity and orientation bias: experimental and modelling investigations / Esposito, Antonino. - (2020 Dec 11).
|Titolo:||Perceptual grouping by proximity and orientation bias: experimental and modelling investigations|
|Data di discussione:||11-dic-2020|
|Appartiene alla tipologia:||07a Tesi di Dottorato|