Probing virtual boundaries and the perception of delayed stiffness

During interaction with robotic manipulanda, the human brain constructs internal representations of the environment imposed by the robotic device. These representations (i) provide cognitive interpretation of the interfaced environment and (ii) generate motor commands for future interaction with the imposed environment. Interestingly, cognitive and motor representations are not always mutually consistent. We consider a simple environment consisting of a spring-like surface, where either the delay between force and position or the location of the boundary is experimentally altered. We explored the cognitive representation of rigidity by asking subjects which of two surfaces is stiffer. We also considered the motor representation by investigating adaptation to the same virtual environments. We asked subject to reach a target inside virtual surface, and observed adaptation and its after effects in catch trials. In the cognitive study, we constructed psychometric curves based on the verbal reports of the subjects. In the motor study, we constructed analogous curves, which we name 'motormetric curves', describing the implicit motor expectation of rigidity, as expressed not verbally but by the errors in catch trials, where the delay was unexpectedly removed. We simulated motormetric curves from a simplified mechanical model of the arm and neural controller. We found that the cognitive reports reflected our measure of the motor behavior in the case of delayed stiffness, but not in the case of shifted boundary.