MAIBO: Multi-posture Asymmetry-aware Intelligent Bilateral Observation Dataset for Cardiovascular Monitoring

While continuous cardiovascular monitoring has gained increasing momentum, current digital phenotyping paradigms remain fundamentally limited by unilateral sensing and fixed-state acquisition protocols. These constraints hinder a comprehensive characterization of physiological dynamics, as they fail to capture inter-arm hemodynamic asymmetry and posture-dependent variability that are critical for robust cardiovascular assessment across real-world scenarios. To address these limitations, we present MAIBO, a multi-posture, asymmetry-aware, intelligent bilateral observation dataset for continuous cardiovascular monitoring. MAIBO comprises synchronized photoplethysmography (PPG) signals acquired from both hands using wearable rings under three standardized postural conditions. The dataset includes 1810 participants and 7478 paired blood pressure recordings, with reference annotations of systolic blood pressure (SBP), diastolic blood pressure (DBP), heart rate (HR), and demographic attributes (Age, Gender, and BMI). All data were collected under controlled multi-postural protocols and subjected to stringent quality control procedures, including the removal of incomplete samples and missing signal segments.