This dataset contains raw data from a behavioral neuroscience experiment conducted at Northeastern University, Boston, MA, USA. Ten participants rhythmically manipulated a virtual cup containing a ball via a robotic manipulandum which controlled the cup position and provided haptic feedback of the force applied by the ball. The cup movement was constrained to 1 dimension. Participants were free to choose the frequency of oscillation, while the amplitude was imposed by visual constraints. The cup and ball system was represented mathematically by a cart and pendulum system. The following dimensions were used for the experiment: pendulum length = 0.45m, pendulum mass = 0.6kg, cart mass = 2.4kg. Each participant performed 5 blocks of 10 trials. For each data file, the first 2 letters identify (anonymously) a participant, the first number identifies the block and the second number identifies the trial within the block. The data are in Matlab data format. "RackPosition" (resp. Velocity, Acceleration) corresponds to the cart position (resp. velocity, acceleration). Y is the axis of the movement. "BallTheta" and "BallOmega" are the pendulum angular position and velocity. Please note that in this dataset angles are positive in clockwise direction (contrary to usual mechanical conventions). "BallForce" is the force applied by the pendulum on the cart. This experimental dataset was confronted to simulation results obtained with two different models. One ("Uncoupled Model") only simulates the dynamics of a cart and pendulum system (Matlab script "cup_task_uncoupled_model_inverde_dynamics.m" and "simulation_cup_task_uncoupled_model_inverse_dynamics.m"). Simulations of this Uncoupled Model are run using inverse dynamics, assuming a sinusoidal trajectory of the cart. The second model ("Coupled Model") includes a simplified model of hand dynamics, represented by an ideal force generator in parallel with a spring and a damper to simulate hand impedance (Matlab and Simulink scripts "cup_task_coupled_model.slx" and "simulation_coupled_model_forward_dynamics.m"). Simulations of this Coupled Model are run using forward dynamics

Acknowledgments: Dagmar Sternad was supported by the National Institutes of Health R01-HD087089, R01-HD081346, and R21-DC013095 and the National Science Foundation NSF-NRI 1637854 and NSF-EAGER-1548514. Neville Hogan was supported by NIH R01-HD087089, NSF-NRI 1637814, NSF-EAGER-1548501 and by the Eric P. and Evelyn E. Newman fund.