Figure 1.

Experimental task. The stimulus display consisted of 21 dots (13 shown here), with one dot—the target—presented in a unique color. Holding a nonmagnetic robotic arm, participants made a fast out-and-back movement from a yellow starting circle, attempting to intersect the target. Solid lines indicate arm trajectories. A resistive force pulse was provided to the hand as soon as it intersected a red dot on the outward movement. No pulse was delivered when the hand intersected a white dot. This created a 2 × 2 design, where participants either aimed for a red dot and expected a force pulse (left column), or aimed for a white dot and expected to not receive a force pulse (right column). Errors can be signaled by the presence of an unexpected stimulus (intersecting a red dot while aiming for a white target; top right) or the absence of an expected stimulus (intersecting a white dot when aiming for a red target; bottom left). Visual feedback of hand position was present on half of the trials. We increased the error rates by adding a ±7° rotation to hand position (and, when present, visual feedback) on 33% of the trials.