A failure to report or respond to stimuli presented in a portion of space is termed hemispatial neglect. Line bisection and line cancellation are two of the tests used most commonly to assess for neglect. Perhaps, because neglect can be induced by a variety of deficits, neither of these tests used alone is as sensitive as both used together. Hence, the primary purpose of this study was to assess the sensitivity, reliability and validity of a new test called the Character-line Bisection Task (CLBT) that combines features of both the bisection and cancellation tests. Since local attention and language are primarily mediated by and activate the left hemisphere, our second goal was to learn if the CLBT and especially the letter version induce a greater rightward bias than the solid-line bisection task. Eighty patients with acute right hemisphere stroke and 81 controls performed the CLBT that consists of two subtests, the Letter-line and Star-line Bisection tasks. All subjects also completed four conventional tests for neglect (Standard solid-line bisection, line cancellation, Star Cancellation, and figure copying). In the bisection tasks both patients and controls bisected to the right with the CLBT than with the solid-line bisection task, suggesting the CLBT induces asymmetrical hemispheric activation. This enhanced rightward deviation with the CLBT was the same for the Letter-line and Star-line Bisection tasks. In regard to sensitivity, we defined the presence of neglect syndrome based on a total score derived from performance of controls on all six tests. This total score detected 55 (68.8%) patients with neglect. Within this group, the Letter-line and Star-line tasks diagnosed neglect in 50 and 48 patients, respectively, resulting in the highest sensitivities (90.9, 87.3%) of the six tests. Thus, the CLBTs demonstrated higher sensitivities than the other commonly used neglect tests and these new tests can be useful for the detection and quantification of unilateral neglect. © 2004 Elsevier Ltd. All rights reserved.