Distinguishing the roles of edge, color, and other surface information in basic and superordinate scene representation

The human brain possesses a remarkable ability to recognize scenes depicted in line drawings, despite these drawings contain only edge information. It remains unclear how the brain uses this information alongside surface information in scene recognition. Here, we combined electroencephalogram (EEG) and multivariate pattern analysis (MVPA) to distinguish the role of edge and surface information in scene representation at the basic and superordinate levels over time. First, we found that surface information is involved in initial neural scene representation at the basic level, but not at the superordinate level. Second, edge information is sufficient and more effective for scene recognition at the superordinate level. Third, time-generalization analysis revealed that edge information in line drawings is crucial for representation at every level of abstraction. These findings provide novel neural evidence that edge and surface information play different roles in dynamic neural scene representations at different levels of abstraction, and help to enhance our understanding of how the human brain represents scenes.