Transient Alerting Effects of Karolinska Sleepiness Scale Ratings on Physiological and Driving Performance Indicators

The objective of this study was to determine whether subjectively rating sleepiness on the Karolinska Sleepiness Scale (KSS) gives rise to an alerting effect and, if so, how long it persists. Data from five driving simulator studies involving 83 participants were pooled. Time-locked trajectories of blink duration, blink closing velocity, heart rate, electroencephalographic (EEG) parietal α and frontal θ power, and lateral lane position variability were analysed around each KSS rating using linear mixed-effects models with condition-specific KSS baselines, a transient post-rating component, a late post-rating offset, and time-on-task terms. Immediately after the rating, several indicators shifted in an alerting direction: blink duration and standard deviation of lateral position (SDLP) decreased, blink closing velocity and heart rate increased, and parietal α power generally decreased. Frontal θ showed a more condition-dependent response. The estimated duration of the transient component was indicator dependent, ranging from 40 s for heart rate to the upper bound of the 120-s analysis window for SDLP. These findings have practical implications for driver sleepiness detection system development, validation protocols, and type-approval processes that rely on KSS as a reference measure. If ratings themselves are associated with short-lived physiological or behavioural changes, the timing and modality of KSS administration become important for ensuring valid ground truth labels. Future studies with control conditions, randomized or jittered rating times, and real-road data are needed to determine whether the rating-associated changes are causal and generalisable.