Mean position of hits in unilateral spatial neglect: An empirical validation

Numerical indices used to detect and quantify neglect in cancellation tests have traditionally relied on omissions count, which may be misleading, as the derived laterality indices (LIs) do not capture the spatial continuum of responses. More recent approaches (e.g., Mean Position of Hits, MPH) quantify neglect based on the spatial position of hits. The MPH approach relies on a theoretically-founded statistical model, which returns a neglect index and a diagnostic label without the need for a control sample and cut-offs. However, such an approach has not yet been empirically validated in patients. Here, we applied the MPH to three common cancellation tests (Letters, Stars, and Lines) in a consecutive sample of right-brain-damaged stroke patients (n = 173) and determined its diagnostic sensitivity and specificity, and its ability to quantify neglect severity against an independent reference standard that combines conventional neuropsychological tests (excluding the three cancellation tests) and a structured clinical observation. Sensitivity and specificity of the MPH were 88.6% and 84.1% for Letters, 87.1% and 95.3% for Stars, 65.2% and 97.7% for Lines. We also compared the MPH diagnostic accuracy with that of two alternative procedures based on a classical cut-off approach (MPH with cut-offs and LIs). Such alternative procedures yielded comparable diagnostic accuracy. Regression analyses showed that the MPH index corresponded more closely to clinical ratings of neglect severity than LIs. Overall, our results show that the MPH is a valid and user-friendly approach for both clinical assessment and research applications, which allows comparisons across time, settings, and cancellation tests.