Microsleep episodes and vigilance failure in driving: A systematic review of definitions, detection, and transport safety relevance

Drowsy driving remains a major contributor to severe transportation accidents, yet operational systems still lack sensitive markers of rapid vigilance failure. Microsleep episodes (MSEs), brief sleep-like intrusions during intended wakefulness, may provide an event-level marker of imminent performance breakdown. This systematic review synthesized 45 peer-reviewed studies published from 2011 to 1 February 2026 using a PRISMA-aligned workflow. Evidence was organized by construct validity, sensing modality, validation design, and deployment context. EEG/EOG-based criteria provided the strongest reference standard, especially for ultra-short events. Selected automated EEG studies reported accuracies from 84.7% to 98.2%, but the studies could not be ranked or statistically pooled because thresholds, datasets, class balance, temporal windows, and validation protocols differed substantially. Vision-based and behavioral systems offered greater deployability but often measured drowsiness proxies rather than EEG-confirmed MSEs. Wearable and multimodal systems provided a compromise between neurophysiological specificity and operational feasibility. The review introduces a three-level construct-validity hierarchy, distinguishes direct MSE detection from calibrated and uncalibrated proxies, and maps sensing choices to clinical, simulator, and fleet deployment contexts. Major barriers were inconsistent definitions, limited participant-independent and cross-dataset validation, underreported false-alarm rates, and insufficient evidence on acceptance, privacy, latency, and energy use. Future work should standardize duration bands and reference labels, calibrate surrogate measures against synchronized EEG/EOG, and evaluate models in diverse real-world driving conditions.