Prism adaptation during target pointing from visible and nonvisible starting locations

The performance of subjects whose starting limb location was visible when pointing to a sagittal target during exposure to prismatic displacement showed immediate target acquisition, but aftereffects of exposure were absent. When starting limb location was not visible, accurate exposure performance was slow to develop, but aftereffects were substantial. Visible starting location evoked a zeroing-in control strategy on the basis of relative-location coding, which rapidly reduced performance error but disabled detection of spatial misalignment between sensorimotor systems. When starting location was not visible, absolute-location coding of the displaced target initiated movement that had to be corrected subsequently by visual feedback. In this case, comparison of the initial erroneous movement code with the limb location that achieved the target enabled misalignment detection and consequent realignment.