Physiological Signal Dataset on Meditation-Induced Relaxation: GSR, HR, and HRV Measurements in Controlled and Home Environments

This dataset contains physiological signals recorded during repeated guided breathing and meditation sessions performed in controlled and home environments. The study involved 14 participants aged 20–25 years and followed a 30-day measurement protocol. Galvanic skin response (GSR), heart rate (HR), and heart rate variability-related data were collected using a Sensetio wristband and a Polar H10 chest-strap sensor. GSR was sampled at 10 Hz, while HR and RR-interval data were recorded at 1 Hz. Measurements were performed both in the HEALTH.Lab telemetric apartment at VSB – Technical University of Ostrava and in participants’ homes. Each measurement consisted of a pre-meditation phase, a guided breathing and meditation phase, and a post-meditation phase. The beginning and end of the meditation phase were marked during each recording. Laboratory sessions were conducted under supervision, whereas home sessions were performed independently in an uncontrolled environment. The planned protocol comprised 840 sessions. In total, 801 sessions were completed. Because some measurements were repeated following technical interruptions or recording restarts, the dataset contains 897 individual recordings. Of these, 69 recordings were classified as unusable. Among the remaining 828 recordings, 448 contained a visually recognizable meditation phase and 380 contained a meditation phase that could not be reliably distinguished from the surrounding signal. The physiological recordings are accompanied by measurement metadata, pre- and post-measurement questionnaire responses, participant-reported meditation focus, measurement location, and environmental parameters recorded in the telemetric apartment. Coded health and personality questionnaire information is also provided where permitted by the applicable consent and redistribution conditions. The dataset is organized by anonymized participant identifier. Each participant is represented by structured JSON files containing physiological measurements and associated metadata. The associated code repository includes MATLAB scripts that support signal preprocessing, GSR decomposition, feature extraction, statistical analysis, and figure generation. The dataset is intended for research on physiological signal processing, autonomic nervous system activity, meditation and controlled breathing, wearable sensing, affective computing, stress regulation, and comparisons between controlled and naturalistic measurement environments. Important limitations include the small and demographically homogeneous participant group, variability in sensor placement and signal quality, unsupervised home measurements, possible inaccuracies in manually recorded meditation markers, and the absence of expert clinical annotations. The labels describing meditation-phase recognizability represent a signal-quality assessment and should not be interpreted as clinical ground truth or as an objective assessment of meditation success.