Ultra-short objective alertness assessment: An adaptive duration version of the 3 minute PVT (PVT-BA) accurately tracks changes in psychomotor vigilance induced by sleep restriction

Abstract

Study Objectives: The Psychomotor Vigilance Test (PVT) is a widely used and sensitive assay of the cognitive deficits associated with sleep loss and circadian misalignment. As even shorter versions of the PVT are often considered too long, I developed and validated an adaptive duration version of the 3 min PVT (PVT-BA). Methods: The PVT-BA algorithm was trained on data from 31 subjects participating in a total sleep deprivation protocol and validated in 43 subjects undergoing 5 days of partial sleep restriction under controlled laboratory conditions. With each subject response, the algorithm updated the odds of the test being high, medium or low performance based on lapses plus false starts on the full 3 min PVT-B. Results: With a decision threshold of 99.619%, PVT-BA classified 95.1% of training data tests correctly without incorrect classifications across two performance categories (i.e. high as low or low as high) and resulted in an average test duration of 1 min 43 s with a minimum duration of 16.4 s. Agreement corrected for chance between PVT-B and PVT-BA was "almost perfect"for both the training (kappa = 0.92) and validation data (kappa = 0.85). Across the three performance categories and data sets, sensitivity averaged 92.2% (range 74.9-100%) and specificity averaged 96.0% (range 88.3-99.2%). Conclusions: PVT-BA is an accurate adaptive version of PVT-B and, to my knowledge, the shortest version to date that maintains key properties of the standard 10 min duration PVT. PVT-BA will facilitate the use of the PVT in settings in which it was previously considered impractical.