Purpose: To assess the importance of 'disengagement failure' and 'attentional gradient' in unilateral spatial neglect (USN) and in recovery from neglect. Method: Eight right-hemisphere-damaged stroke patients performed the standardized Behavioural-Inattention-Test battery for visual neglect, line-bisection tests, and two computerized reaction-time (RT) tasks: a variant of Posner's 'Spatial-Cueing' paradigm (with special emphasis on the magnitude of leftward disengagement time) and a signal-detection task (marking the spatial gradient of attention by the distribution of RTs to target stimuli in different spatial locations). The correlation between the different measures was assessed at two points in time, before and after a period of rehabilitation treatment. Results: A recovery pattern could be identified in both RT paradigms. However, the correlation between standard measures of neglect and performance on both, spatial-cueing and signal-detection tasks, was weak. Conclusion: Neither difficulty disengaging attention from an ipsilesional stimulus nor changes in the attentional gradient can fully explain the processes underlying USN and its recovery. A large interpersonal variance exists among USN patients in the expression of disengagement and other spatial-attention deficits. Hence, individual patients should be tested by measuring different factors known to play a role in USN. This information is crucial for assigning the appropriate treatment for each patient in accord with the specific deficit revealed.