The following is a summary of “Endogenous attention biases transformational apparent motion based on high-level shape representations” published in the November 2022 issue of Vision by Saleki, et al.

Viewers do not see the bar appearing all at once when a bar suddenly appears between two previously separated, pre-existing squares, connecting them. However, they observe an illusory evolution of the initial squares over time. Changing the transformational apparent motion’s (TAM) direction is possible by using endogenous attention before the connecting bar appears. For a study, researchers sought to determine whether endogenous attention impacted TAM through operations over lower-level form representations characterized by particular visual characteristics instead of operations over higher-level, feature-independent shape representations. 

To achieve this, they examined the impact of endogenous attention on TAM in first- and second-order displays, which had similar forms but various shape-defining characteristics (luminance and texture contrast, respectively). They discovered that endogenous attention had a comparable effect on first- and second-order items regarding the strength of directional bias and time. The findings suggested that endogenous attention operated on high-level shape representations independent of the low-level visual elements that characterize them to skew the perceived direction of TAM. 

The findings were consistent with a four-stage TAM model, where a feature encoding stage transferred a layout specific to certain features to a parsing stage, which creates discrete, high-level meta-featural shapes. The shapes were then matched and visually interpolated over time.