Diagnostic Sensitivity of Cancellation-Based Pen-and-Paper Tests in Visuospatial Neglect: Effects of Severity and Subtype

Visuospatial neglect is a frequent and disabling consequence of right-hemisphere stroke and is commonly assessed using pen-and-paper tests, particularly cancellation tasks. Although cancellation is generally regarded as the most sensitive of these measures, sensitivity differs markedly between versions of the task, and the factors determining this are not fully understood. Sarri et al. (2009) showed that top-down task demands can modulate the extent of neglect revealed by cancellation while holding the visual display constant, isolating task-related from stimulus-related influences. If task-demand modulation is to be proposed as an improvement to diagnostic sensitivity, knowing which patients it actually helps is a prerequisite for its clinical application. The present thesis addresses this question through a secondary analysis of the individual patient data underlying Sarri et al. (2009), obtained from the first author. Baseline neglect severity, defined as the percentage of targets detected at clinical screening, was correlated with the magnitude of the task-demand effect in each of the three experiments (n = 17, 13, and 9). Two outcomes were examined: the change in a column-weighted laterality index, capturing the spatial distribution of search, and the change in detection accuracy. Correlations were positive in the predicted direction for detection accuracy in all three experiments and reached significance in Experiment 2, indicating larger task-demand effects in less severely impaired patients. No relationship with severity was found for the laterality index. Individual data converged on the same pattern: patients whose search was already strongly rightward showed minimal further change, including the two patients for whom Sarri et al. reported no significant task effect. Patients with severe neglect may already omit a substantial portion of contralesional stimuli under simpler task demands, leaving limited room for what a harder task can reveal. Task-demand manipulation therefore appears most useful in milder cases, where standard cancellation tests are least informative.