Hemispheric Asymmetry in Visual Perception Arises from Differential Encoding Beyond the Sensory Level

Hemispheric Asymmetry in Visual Perception Arises from Differential Encoding beyond the Sensory Level Janet Hui-wen Hsiao (jhsiao@cs.ucsd.edu) Reza Shahbazi (rshahbaz@ucsd.edu) Garrison W. Cottrell (gary@cs.ucsd.edu) Department of Computer Science & Engineering, University of California San Diego 9500 Gilman Dr. #0404, La Jolla, CA 92093 USA Abstract Hierarchical letter Pattern Perception Hemispheric asymmetries in the perception of local and global features have been consistently reported: there is an advantage for responses to global features in the left visual field/right hemisphere and an advantage for responses to local features in the right visual field/left hemisphere. It has been proposed that this asymmetry originates from differential frequency bias in the two hemispheres (e.g., Ivry & Robertson, 1998). Nevertheless, there is little evidence supporting hemispheric specialization for particular frequency ranges (e.g., Fendrich & Gazzaniga, 1990) or differential frequency tuning in the neurons in the two hemispheres. Here we test the hypothesis that this hemispheric asymmetry in visual perception takes place at the encoding stage beyond the sensory level. We use two autoencoder networks with differential connectivity configurations as the way to develop differential encoding in the two hemispheres, to reflect the anatomical evidence that there is more interconnectivity among the neighboring cortical columns in the right hemisphere than the left hemisphere (e.g. Hutsler & Galuske, 2003). We show that this differential encoding mechanism has a better fit with human data than the model based on differential frequency bias, and thus is a more anatomically realistic and cognitively plausible model in accounting for the hemispheric asymmetry in visual perception. Sergent (1982) used hierarchical letter patterns (Navon, 1977) to examine hemispheric differences in responses to global and local patterns. A hierarchical letter pattern contains two patterns: a global pattern and a local pattern. The global pattern (the large letter in Figure 1(a)) is composed of a number of local patterns (the small letters in Figure 1(a)). She referred to the two levels of the stimulus as having differential spatial frequency contents: low frequency for the global pattern and high frequency for the local pattern. In her experiment, she used four letters to compose the hierarchical letter patterns: H and L were designated as targets, and T and F as distracters. Given that each letter may appear as the local or the global pattern, there are in total 16 patterns, which can be put into six conditions according whether there is a target in the local or global patterns, as shown in Figure 1(a). Stimuli were presented to either the RVF/LH or the LVF/RH for 150 ms, and the participants’ task was to judge whether they saw a target letter or not, regardless of its being in the global pattern or the local pattern. Keywords: Hemispheric asymmetry, visual perception, Double Filtering by Frequency (DFF), autoencoder networks. Introduction The way we analyze and process the global and local forms of visual stimuli has been extensively examined. Navon (1977) proposed the global precedence hypothesis , suggesting that the global form of a visual stimulus is unavoidably recognized before the local forms. This effect was later shown to depend on both the characteristics of the local and global forms and the hemispheric asymmetry in the perception of local and global features (Hoffman, 1980). Follow-up studies further confirm that there is a right visual field (RVF)/left hemisphere (LH) advantage for responses to local features and a left visual field (LVF)/right hemisphere (RH) advantage for responses to global features (Sergent, 1982; Ivry & Robertson, 1998). Nevertheless, studies examining grating detection did not support the existence of hemispheric specialization for particular frequency ranges (e.g. Di Lollo, 1981; Rijsdijk, Kroon, V Peterzell, Harvey, & Hardyck, 1989; Fendrich & Gazzaniga, 1990). It thus remains controversial about why this perceptual asymmetry exists (Peterzell, 1991; Martin, 1979). Figure 1: (a) Stimuli in Sergent’s experiment (1982). “H/L” are targets, “T/F” are distracters. “L+” means the large letter is a target, and “S+” means the small letters are targets. Id. means the local and the global patterns are identical. (b) The RT data for the L+S- and L-S+ stimuli in the LVF and RVF presentation conditions (Sergent, 1982). The stimuli of greatest interest in Sergent's experiment (1982) were the conflict conditions when the target appeared in either the local pattern or the global pattern (i.e., the L+S- and L-S+ cases in Figure 1(a)), since they are the conditions in which interference may arise due to that the