A search for the optimal stimulus.

How do stimulus size and item number relate to the magnitude and direction of error on center estimation and line cancellation tests? How might this relationship inform theories concerning spatial neglect? These questions were addressed by testing twenty patients with right hemisphere lesions, eleven with left hemisphere lesions and eleven normal control subjects on multiple versions of center estimation and line cancellation tests. Patients who made large errors on these tests also demonstrated an optimal or pivotal stimulus value, i.e., a particular size center estimation test or number of lines on cancellation that either minimized error magnitude relative to other size stimuli (optimal) or marked the boundary between normal and abnormal performance (pivotal). Patients with right hemisphere lesions made increasingly greater errors on the center estimation test as stimuli were both larger and smaller than the optimal value, whereas those with left hemisphere lesions made greater errors as stimuli were smaller than a pivotal value. In normal subjects, the direction of errors on center estimation stimuli shifted from the right of true center to the left as stimuli decreased in size (i.e., the crossover effect). Right hemisphere lesions exaggerated this effect, whereas left hemisphere lesions diminished and possibly reversed the direction of crossover. Error direction did not change as a function of stimulus value on cancellation tests. The demonstration of optimal and pivotal stimulus values indicates that performances on center estimation and cancellation tests in neglect are only relative to the stimuli used. In light of other studies, our findings indicate that patients with spatial neglect grossly overestimate the size of small stimuli and underestimate the size of large stimuli, that crossover represents an "apparent" shift in error direction that actually results from normally occurring errors in size perception, and that the left hemisphere is specialized for one aspect of size estimation, whereas the right performs dual roles.