Disorders of visuomotor function are common following traumatic brain injury (TBI), but spatially directed visual attention has received little study in this population. 'Cancellation' testing is a common bedside method for assessing directed attention, which can provide information on how task properties influence visual scanning and search following severe TBI. Groups of 20 individuals after severe TBI and 21 healthy control subjects were matched for age and education. Participants performed finger tapping tests to assess motor speed as well as four cancellation tests employing letter and geometric figure stimuli in random and structured arrays. Control and TBI groups differed significantly on measures of accuracy, task completion time, and search quality. There was no significant effect of stimulus or array type on accuracy or time. Figure targets in a higher search quality, suggesting a right hemispheric dominance effect on these tasks. The findings support a deficit in visuomotor scanning performance in TBI beyond a purely motor effect. Interactions between stimulus and array types suggest that hemispheric cooperation is required for the optimal performance of these tasks, and that interhemispheric communication may be preferentially compromised by TBI.