Experimental Analysis of the Attribution of Own Actions to the Intention of Self or Others by the Multiple Forward Models

Human cognitive mechanisms have been studied for designing user-friendly interface. One of the key issues is the attribution of own actions to the intention of self or others. It is known that patients with schizophre- nia who sometimes attribute their own actions to the intentions of others may perceive themselves as caus- ing events which they do not in fact control, when they feel they are in voluntary movement. In this study, we administered similar experiments to normal subjects. We also conducted computational simulations through extending the multiple forward models, which success- fully described the experimental results. Information systems are ubiquitous in the modern in- formation society. People spend a long time in front of the personal computer using the typical input and output de- vices: keyboard, mouse, and display. It is very important to analyze how we adopt to this human-machine interface in order to understand cognitive mechanism of human, be- cause such a mechanism provides us design principle of daily used human-machine interface. In general, people become comfortable using machines when they feel the machines as if they are part of their bodies. Let us imag- ine driving the car. We feel like cars as if they are part of our body when we get used to driving them. Drivers can estimate speeds or angles of the cars when they push the accelerators or rotate the handles. The drivers feel that they are controlling the cars when they can attribute motions of the cars to the intentions of themselves. Like- wise, people feel comfortable using computers when they can attribute the motion of objects in the display, which are caused by their own key or mouse inputs, to their own intentions. Such feeling is called "sense of agency". To date, usability of typical human-machine interface with key board, mouse, and display have not been analyzed in terms of the sense of agency. In this study, we experi- mentally analyze the sense of agency. Studies on sense of agency are overviewed, and experimental results are simulated by the multiple forward models. Our model is validated through computational simulation.