De-identified session-level dataset for temporal interpretability of post-lesson EEG and HRV signals in AI-based versus human tutoring

This record contains a de-identified session-level minimum dataset from a within-learner crossover repeated-measures study comparing AI-based tutoring (AI-CATS) and human 1:1 tutoring in middle-school mathematics. Physiological indicators were measured immediately after each lesson and include EEG Theta/Beta Ratio (TBR; eyes-closed segment) and heart rate variability (HRV; RMSSD). Temporal spacing between sessions is represented using calendar-day proxies (Δdays), with same-day sessions coded as Δdays = 0 due to unlogged within-day washout duration. Direct identifiers and exact calendar dates were removed. Participant identifiers were recoded (P01–P11), and study days were encoded relative to each participant’s first session. This dataset is provided to support transparency and reproducibility of analyses focusing on the temporal interpretability of post-lesson physiological signals, rather than to assert instructional effectiveness.