There are two files in this dataset. The first has data on reaches in the training tasks as well as in the no-cursor reaching tasks, and is called 'reaches.txt'. The second has data on where the participants indicated they though their right hand was by tapping on a touch screen with their left hand. This file is called 'taps.txt'. **Reaches** All reaches are first rotated so that the visual target has a Y coordinate of 0. Then they are splined to a resolution of 10 ms. And then put through a 2.5 Hz lowpass Butterworth filter. After that, some endpoint parameters (X and Y coordinate, angle) are calculated, As well as some parameters of the point of maximum velocity (velocity, X and Y coordinates and angle). Columns with independent variables: 1. subject: subject number (0..20) 2. condition: aligned or rotated (0: aligned, 1: rotated) 3. phase: type of reach (0: training, 1: no-cursor reaches, 2: top-up training) 4. iteration: repetitions of the task (0,1,2,3,4) NB: there are 4 iterations (0..3) in aligned, but 5 in rotated (0..4) 5. trial: trialnumber (0..89; training: 0..xx, no-cursor reaches: 0..21, top-up training: 0..xx) 6. target_angle: target angle in degrees (15,25,35,45,55,65,75) 7. rotation: rotation of the visual feedback in degrees (0..30) Columns with dependent variables: 8. endX: X coordinate of the endpoint 9. endY: Y coordinate of the endpoint 10. end_angle: deviation of the angle of the endpoint of the reach relative to the angle of the visual target in radians 11. max_vel: maximum velocity in the reach (m/s?) 12. max_velX: X coordinate at the point of maximum velocity 13. max_velY: Y coordinate at the point of maximum velocity 14. max_vel_angle: angle at the point of maximum velocity, relative to the angle of the visual target in radians **Taps** This table contains data on where people tapped on the touchscreen in a pre-processed form. Preprocessing consisted of fitting a second degree polynomial to each participants' angular touch screen responses errors in all aligned data. Then separately for each participants' remaining angular errors in delayed and online touch screen responses a linear function was fit. The errors predicted by these functions were subtracted from the screen responses in the aligned as well as the rotated session, in order to remove idiosyncracies in responses. All taps are described as an angle relative to where people indicated the home position was. This has 0 degrees to the right and 90 degrees straight ahead and so on. Columns with in independent variables: 1. subject: subject number (0..20) 2. rotated: tap recorded after rotated training or not (boolean: 0=false, 1=true) 3. active: tap recorded after active movement or not (boolean: 0=false, 1=true) 4. delayed: tap recorded delayed or not (boolean: 0=false - or online response, 1=true) 5. trial: trial number (0..24) Columns with dependent variables: 6. robot_angle: where the robot and hence the hand, actually was when responding on the touchscreen (in degrees angle) 7. tap_angle: where the participant indicate their hand was