The Categorization for Single Moving Objects Based on Multiple Motion Characteristics

Object categorization is essential for animals to survive in natural environment. When the form of a moving object is unavailable (e.g., small size, large distance, poorly lighted), object’s categorization is purely relied on object’s motion characteristics. What and how motion characteristics are used for such object categorization remain unclear. To address this question, we trained macaques to categorize two classes of single moving objects (derived from flying bird and drone, respectively) with identical form. We found that after training, macaques were able to categorize moving objects within a short time based on motion characteristics rather than one-by-one memorization. The categorizing time is shorter for moving objects derived from birds than those from drones, which indicated that macaques had animacy preference. We further found that the speed, acceleration, and angular velocity played important roles in categorizing single moving objects, in which speed was most critical. A constructed generalized drift-diffusion model showed that these three motion characteristics may aggregated to momentary evidence. The fast categorization and animacy preference could be explained by collapsing decision boundary and start point bias. These findings offer new insights for understanding the mechanisms of object categorization based on motion characteristics and animacy preference.