Integrated machine-learning-assisted wearable platform for stress state assessment based on sweat cortisol and multimodal physiological biosensing
Huang et al.
The finding, in our words
A wearable platform that combines sweat cortisol measurement with continuous physiological monitoring was evaluated in ten participants. The machine-learning model fused these signals to distinguish stress from rest states with 94.75% accuracy, and the immunosensor's cortisol readings agreed well with laboratory ELISA.
A paraphrase to the Library’s standard, never the abstract. The source is one link away and is always the authority.
Wearable biochemical monitoring requires a complete measurement chain that integrates quality-controlled biofluid sampling, sensor fusion and metadata capture to convert raw sensor outputs into clinically valid digital biomarkers.
The wearable microfluidic sensor measured sweat cortisol across a range of 0.1 fM to 100 μM with a 0.043 fM detection limit. In human volunteers it tracked diurnal rhythms and stress-induced elevations from cold-pressor and exercise tests, showing agreement with ELISA, and combined with electromyography it profiled exercise fatigue by linking muscle activity to endocrine response.
Chen et al., Biosensors & bioelectronics (paywalled) · source ↗
Salivary cortisol is suitable as an adjunctive measure in decentralised monitoring, with lateral-flow assays providing rapid low-cost discrete measurements and wearable platforms enabling higher-frequency collection but facing biofouling and calibration challenges. Suitability requires validation in authentic saliva against reference methods, not merely low detection limits.
The NanoTracker measured sweat metoprolol and heart rate in real time in people given metoprolol, and sweat levels correlated with plasma metoprolol while heart rate fell, showing the device can track both pharmacokinetic and pharmacodynamic effects for personalised dosing.
The study demonstrated that longitudinal remote monitoring using wearable sleep electroencephalography, actigraphy, and self-collected saliva samples is feasible in older adults with mild cognitive impairment or dementia, with high retention and data completeness rates over eight weeks.