Retinal versus physical stimulus size as determinants of visual perception in simultanagnosia

Abstract

Patients with simultanagnosia following bilateral parieto-temporo-occipital brain damage show a characteristic impairment of global gestalt perception, while their perception of individual objects or elements remains intact. For instance, when shown 'hierarchical' stimuli comprising a larger global object (e.g. a large letter) made up from smaller components (e.g. multiple small letters), they typically report seeing one of the smaller components but not the global figure. Recent work on simultanagnosia revealed that global perception can be improved if local element spacing is reduced. However, it is still unclear whether the retinal separation or the physical (post-size-constancy) spatial separation is critical. Here, we presented various hierarchical global/local letter stimuli at different viewing distances and sizes to separate the impacts of retinal versus physical size. Our findings indicate a key role for visual angle in determining simultanagnosic perception. We observed that not only retinal spacing (in terms of visual angle) between local elements had a major impact on global perception in simultanagnosia, but also the physical size of the separation between local elements, provided that binocular cues to viewing distance were available. The results indicate both pre-size-constancy retinal influences and binocular-post-constancy influences upon conscious perception in simultanagnosia.

Fig. 1

18-Fluorodeoxyglucose positron emission tomography (FDG-PET) scans overlayed with the anatomical magnetic resonance imaging (MRI) scans for patient HW reveal reduced metabolism in the parieto-temporo-occipital cortex bilaterally; see main text.

Fig. 2

Navon-like hierarchical letter stimuli, comprising a larger ‘global’ letter constructed from multiple repeated smaller ‘local’ letters, were used in all experiments. Five different letter identities could appear at the global or repeated local level (A, B, E, H, N), with the global letter identity always differing from the local (a). The spacing between letters at the local scale varied parametrically across five conditions (schematically labelled as ‘Numbers 1 to 5’) by adding local letters while keeping the size of the global letter unchanged (b). The stimuli were presented at two different viewing distances (‘close’ or ‘far’) between the patient's eyes and the presentation monitor. We either kept the physical displays themselves constant (Experiment 1, ‘Constant Physical Size’) while varying viewing distance; or else adjusted their physical size for the far viewing distance such that the retinal size of the stimuli then remained identical across difference viewing distances (Experiments 2 and 3, ‘Constant Retinal Size’).

Fig. 3

Average percent (with standard errors shown) of correct global-letter report for patient HW in the sub-conditions of (a) Experiment 1, (b) Experiments 2, and (c) Experiment 3. The solid horizontal line in each graph indicates chance-level performance; the dashed lines indicate the binomial 95%-confidence interval for above-chance performance.