Interaction between central and peripheral vision: Influence of distance and spatial frequencies

Abstract

Visual scene perception is based on reciprocal interactions between central and peripheral information. Such interactions are commonly investigated through the semantic congruence effect, which usually reveals a congruence effect of central vision on peripheral vision as strong as the reverse. The aim of the present study was to further investigate the mechanisms underlying central-peripheral visual interactions using a central-peripheral congruence paradigm through three behavioral experiments. We presented simultaneously a central and a peripheral stimulus, that could be either semantically congruent or incongruent. To assess the congruence effect of central vision on peripheral vision, participants had to categorize the peripheral target stimulus while ignoring the central distractor stimulus. To assess the congruence effect of the peripheral vision on central vision, they had to categorize the central target stimulus while ignoring the peripheral distractor stimulus. Experiment 1 revealed that the physical distance between central and peripheral stimuli influences central-peripheral visual interactions: Congruence effect of central vision is stronger when the distance between the target and the distractor is the shortest. Experiments 2 and 3 revealed that the spatial frequency content of distractors also influence central-peripheral interactions: Congruence effect of central vision is observed only when the distractor contained high spatial frequencies while congruence effect of peripheral vision is observed only when the distractor contained low spatial frequencies. These results raise the question of how these influences are exerted (bottom-up vs. top-down) and are discussed based on the retinocortical properties of the visual system and the predictive brain hypothesis.

Figure 1.

(A) Illustration of the eccentricities (in degree of visual angle) of the stimuli's inner and outer edges. (B) Example of stimuli displayed in Experiment 1. The central stimulus (central disk or central ring) and the peripheral stimulus were presented simultaneously. Stimuli could be either congruent (a central outdoor scene and the same peripheral outdoor scene or a central indoor scene and the same peripheral outdoor scene) or incongruent (a central outdoor scene and a peripheral indoor scene or a central indoor scene and a peripheral outdoor scene).

Figure 2.

Box plots of (A) error rates in percentage, ERs and (B) correct response times in milliseconds (ms), RTs for categorizing the central or the peripheral target simultaneously displayed with a peripheral or a central distractor, respectively. The distractor is either semantically congruent (in green) or incongruent (in blue). Black dots indicate the mean and error bars indicate standard errors. Boxes represent medians and quartiles and whiskers represent the minimum and maximum sample without the outliers. Congruence effect (incongruent condition minus congruent condition) for (C) error rates in percentage and (D) correct response times in milliseconds (ms). Black dots indicate the mean congruence effect and error bars indicate standard errors. Red dots correspond to the congruence effect for each participant. Values above 0 indicate the presence of a congruent effect. Stimuli are illustrated on the x-axis: the target is in red and the distractor is in gray. The asterisk (*) means that the difference is significant (p < 0.05) for the comparisons tested.

Figure 3.

Example of congruent stimuli displayed in Experiment 2. A central disk and a peripheral ring extracted from the same scene were presented simultaneously. The distractor (either in central or peripheral position) was filtered either in HSF or LSF. It should be noted that the perception of spatial frequencies could be affected by the reduction of stimuli for an illustrative purpose.

Figure 4.

Box plots of (A) error rates in percentage, ERs and (B) correct response times in milliseconds (ms), RTs for categorizing the central or the peripheral target simultaneously displayed with a peripheral or a central distractor, respectively. The distractor is either semantically congruent (in green) or incongruent (in blue) and it is either filtered in LSF or HSF. Black dots indicate the mean and error bars indicate standard errors. Boxes represent medians and quartiles and whiskers represent the minimum and maximum sample without the outliers. Congruence effect (incongruent condition minus congruent condition) for (C) error rates in percentage and (D) correct response times in milliseconds (ms). Black dots indicate the mean congruence effect and error bars indicate standard errors. Red dots correspond to the congruence effect for each participant. Stimuli are illustrated on the x-axis: the target is in red and the distractor is in gray. The asterisk means that the difference tested is significant (p < 0.05) for the comparisons tested.

Figure 5.

Example of stimuli displayed in Experiment 3. The central disk and the peripheral ring were presented simultaneously. When participants had to categorize the central disk (upper row), the peripheral ring distractor was filtered in LSF. It could be either a noise pattern, a semantically congruent scene or a semantically incongruent scene. When participants had to categorize the peripheral ring (lower row), the central disk distractor was filtered in HSF. It could be either a noise pattern, a semantically congruent scene or a semantically incongruent scene. It should be noted that the perception of spatial frequencies could be affected by the reduction of stimuli for an illustrative purpose.

Figure 6.

Box plots of (A) error rates in percentage, ERs and (B) correct response times in milliseconds (ms), RTs for categorizing the central or the peripheral target simultaneously displayed with a peripheral or a central distractor, respectively. The distractor is either semantically congruent (in green) or incongruent (in blue) or a noise pattern (in gray). The distractor was filtered in LSF when peripheral or HSF when central. Black dots indicate the mean and error bars indicate standard errors. Boxes represent medians and quartiles and whiskers represent the minimum and maximum sample without the outliers. Congruence effect (incongruent condition minus congruent condition) for (C) error rates in percentage and (D) correct response times in milliseconds (ms). Black dots indicate the mean congruence effect and error bars indicate standard errors. Red dots correspond to the congruence effect for each participant. Stimuli are illustrated on the x-axis: the target is in red and the distractor is in gray. The asterisk (*) means that the difference tested is significant (p < 0.05) for the comparisons tested.