Arcade game testing of generalized cognitive maps in humans

Humans use generalized cognitive maps (GCM) for navigation in dynamic environments. Mechanisms of the generation of GCM in the brain are not well understood. Nevertheless, there is a mathematical model that allows their simulation. Here we propose a methodology for testing the GCM concept in behavioral experiments and comparing results with the model. We perform experiments with human subjects playing a specially design arcade computer game. The resulting trajectories of objects controlled by humans on a PC screen are then compared with those predicted by the GCM model. We show that human trajectories exhibit significant similarity with the model trajectories in dynamical scenes with objects moving at high speed. If the speed of the controlled object decreases, then men and experienced users maintain a similarity with the model predictions, while women and untrained subjects deviate significantly.