Tapping responses were recorded in a group of patients who sustained lesions from cerebrovascular accidents and from an age-matched group of healthy subjects. In each experiment the stimuli were an auditory signal, a flash, or a light touch on the arm, delivered either at random intervals or at fixed time intervals of 1, 2, 3, or 4 s. The tapping response was done with each hand separately (first experiment) or with each hand alternately (second experiment). Healthy subjects, ''predicted'' the stimuli (response time less than 150 ms), when the ''preferred'' fixed time interval was used. One group of patients (n = 3), did not predict the stimuli at the preferred fixed intervals when tapping was done with the hand related to the involved hemisphere. The second group of patients (n = 3) did not predict the stimuli when tapping was done with either hand.