Abstract
We present a bio-hybrid sensing system that uses a living plant (Tradescantia pallida) as an ambient biosensor for human stress states as a single-participant proof-of-concept study. An AD8232 biosignal amplifier captures plant bioelectric activity, while a Happimeter smartwatch simultaneously measures the wearer's mood via machine learning on wrist-worn sensor data. Over 129 paired observations across eleven days in a naturalistic desk-work setting, a within-day fixed-effects analysis reveals robust stress-plant coupling: seven correlations survive Benjamini-Hochberg false discovery rate correction (q = 0.05), with two also surviving Bonferroni correction. The strongest results are stress_rolling vs. plant mean (r = +0.36, p = 3.3 × 10(-5)) and RMS (r = +0.34, p = 7.8 × 10(-5)). An incidental electrode reattachment mid-experiment created a natural control: mean/RMS correlation signs flipped with electrode polarity, while the coefficient of variation remained consistently negative across both configurations (r = -0.32, p = 2.6 × 10(-4)). This electrode-invariant finding-higher stress associated with lower relative signal variability-provides the strongest evidence for genuine bio-hybrid sensing. The results position living plants as bio-inspired ambient sensing elements for workplace wellbeing monitoring.