Quantitative analysis of cancellation tasks in neglect.

Patients with spatial neglect do not explore contralateral space effectively. Although cancellation tasks are used widely to assess this visual search deficit, their methods of analysis are not well established. We introduce logistic regression analyses for cancellation tasks in 7 patients with left neglect. We investigated the influences of spatial location, stimuli number, and target discriminability on the probability of canceling a target. As a group, neglect patients showed left and near neglect. They also explored and canceled targets further into contralateral space on arrays with fewer visual stimuli. Individual analyses revealed exceptions to these group patterns, such as two patients with far rather than near neglect. Only patients with relatively mild neglect canceled more targets when they were more easily discriminated from distracters. Logistic regression models accounted for 0.68 the variance in cancellation performances of the entire group. Shifting the unit of analysis from the proportion of targets canceled to the probability of detecting individual targets offers a powerful parametric method to analyze group and individual performances on cancellation tasks and can reveal functional dissociations in neglect.