Cross-over, completion and confabulation in unilateral spatial neglect.

Patients with left-sided neglect generally mis-bisect horizontal lines to the right of mid-position. However, with short lines they frequently cross over and place their marks to the left; a phenomenon not easily explained by current theories of neglect. It is difficult to ascertain whether patients neglect the ipsilateral segment of short lines, perceptually distort these lines, or extend these lines leftward at a representational level. Reading horizontally arrayed letter strings is an alternative task potentially capable of sorting between these hypotheses, providing this task could be demonstrated to be comparable to line bisections. In this study, we exploit the observation that line bisection performance may be described by a power function: psi = K phi beta, in which psi represents the subjective awareness of lines and the objective measurements of these lines. The constant and exponent are empirically derived. We have investigated whether reading task performance also subscribes to power functions. Five patients with left-sided neglect were tested on line bisections and reading tasks with stimuli of varying horizontal lengths. Performance on both tasks could be described by power functions and patients were similarly ordered by severity across both tasks. Since power function analyses established that these tasks were comparable, reading performance on short words provided insight into how patients bisected short lines. Patients reported additional letters to the left end of short words suggesting that cross-over bisections are analogously a consequence of completion or confabulation of short lines beyond their objective left end-point. Although most neglect patients have difficulties forming veridical left-sided representations, this cross-over phenomenon is interpreted as resulting from a lack of inhibition of leftward confabulatory mental representations. Since both bisections and reading tasks are described by a continuous function, it is proposed that similar lack of inhibition must also be present, albeit hidden, when patients confront stimuli with longer lateral extension. Confabulatory phenomena at the left edge of these patients' rightwardly constricted attentional window, may be far more common than hitherto appreciated.