Most people prefer smoothly curved shapes over more angular shapes. We investigated the origin of this effect using abstract shapes and implicit measures of semantic association and preference. In Experiment 1 we used a multidimensional Implicit Association Test (IAT) to verify the strength of the association of curved and angular polygons with danger (safe vs. danger words), valence (positive vs. negative words) and gender (female vs. male names). Results showed that curved polygons were associated with safe and positive concepts and with female names, whereas angular polygons were associated with danger and negative concepts and with male names. Experiment 2 used a different implicit measure, which avoided any need to categorise the stimuli. Using a revised version of the Stimulus Response Compatibility (SRC) task we tested with a stick figure (i.e., the manikin) approach and avoidance reactions to curved and angular polygons. We found that RTs for approaching vs. avoiding angular polygons did not differ, even in the condition where the angles were more pronounced. By contrast participants were faster and more accurate when moving the manikin towards curved shapes. Experiment 2 suggests that preference for curvature cannot derive entirely from an association of angles with threat. We conclude that smoothly curved contours make these abstract shapes more pleasant. Further studies are needed to clarify the nature of such a preference.

Neuroimaging work has shown that visual symmetry activates extrastriate brain areas, most consistently the lateral occipital complex (LOC). LOC activation increases with proportion of symmetrical dots (pSymm) in a degraded display. In the current work, we recorded a posterior ERP called the sustained posterior negativity (SPN), which is relatively negative for symmetrical compared to random patterns. We predicted that SPN would also scale with pSymm, because it is probably generated by the LOC. Twenty-four participants viewed dot patterns with different levels of regularity: 0% regularity (full random configuration) 20%, 40%, 60%, 80%, and 100% (full reflection symmetry). Participants judged if the pattern contained "some regularity" or "no regularity". As expected, the SPN amplitude increased with pSymm, while the latency and duration was the same in all conditions. The SPN was independent of the participant's decision, and it was present on some trials where people reported 'no-regularity'. We conclude that the SPN is generated at an intermediate stage of visual processing, probably in the LOC, where perceptual goodness is represented. This comes after initial visual analysis, but before subsequent decision stages, which apply a threshold to the analog LOC response.

Empirical work on visual aesthetics has found a strong preference for smooth curvature. The use of different tasks and exposures can clarify whether such a preference reflects immediate visual responses or higher cognitive processes, such as semantic associations. In three experiments, we used abstract irregular shapes and manipulated the contour (polygons vs. smoothed versions of polygons) but matched the stimuli for number of protrusions (vertices or curvature extrema) and number of concavities. In Experiment 1, shapes were presented for 120 ms, and observers produced a two-alternative forced-choice response (like or dislike). In Experiments 2 and 3, we used rating scales measuring liking and attractiveness, respectively, and the stimuli were presented until a response was made. Overall, smooth curved contours were preferred over angular contours, especially with immediate responses (Experiment 1). Moreover, shapes were preferred when they contained a balance between number of vertices and concavities (i.e., a lower proportion of concavities for a large number of vertices). However, a preference for shapes with the highest number of vertices and the least number of concavities occurred in the two-alternative forced-choice task (Experiment 1). In contrast, the reverse combination (i.e., star-like shapes) emerged with rating scales (Experiments 2 and 3). We conclude that the curvature effect is stable across tasks, although it is modulated subtly by some parametric variations (vertices and concavities) related to visual complexity.