Dataset for Validation of AI-Enabled Bioelectronic ECG/EMG Monitoring

This dataset supports the study “AI-Enabled Bioelectronic Interfaces for Continuous ECG/EMG Monitoring.” It provides a structured validation framework for the analysis of electrocardiographic (ECG) and electromyographic (EMG) signals acquired through flexible hydrogel-graphene bioelectronic interfaces. The dataset includes 1,000 structured records designed to reproduce and verify the analytical workflow described in the study, including signal quality assessment, skin-electrode impedance, signal-to-noise ratio, artifact reduction, ECG and EMG feature extraction, classification probabilities, prediction outcomes, Brier scores, inference latency, packet loss, and cybersecurity-related variables. It also includes participant-level summaries for 24 subjects and validation sheets with formula-based calculations for sensitivity, specificity, F1-score, signal quality indicators, usability, energy performance, and system integrity. The dataset is organized to facilitate reproducibility, methodological validation, and secondary analysis of the proposed bioelectronic monitoring framework. It reflects the study’s reported targets, including ECG arrhythmia classification, EMG-based muscle fatigue classification, impedance reduction, SNR improvement, embedded AI processing, energy harvesting, and secure biomedical data handling. The 1,000 records included in the spreadsheet are synthetic and simulated data generated for methodological validation, reproducibility testing, and verification of analytical procedures. They must not be interpreted as original clinical measurements or individual participant records. The workbook is intended to support transparent replication of the computational and statistical workflow and to provide a structured basis for future validation with experimental and clinical data.