Data for: Accuracy against behavioural or schedule labels does not validate forehead EEG as a drowsiness measure

This deposit contains the HARU-1 forehead EEG recordings of five participants (three channels at AFz, Fp1 and Fp2; 250 Hz; one session each) and the participant- and recording-level results behind every figure and table of the article “Accuracy against behavioural or schedule labels does not validate forehead EEG as a drowsiness measure”. Labels are nominal protocol stages (task, early fixation, late fixation), not verified drowsiness states. The public datasets analysed in the article (Cao et al. 2019; the Cui et al. benchmark; Acı et al. 2019) are not redistributed. The absolute recording clock was replaced by elapsed time, and local computer paths inside the result files were replaced by placeholders; no value was changed. README.md describes the folders, columns and provenance. The recordings were made under approval 2024-518 of the Ethics Review Committee on Research with Human Subjects of Waseda University, with written informed consent.