Minimally Invasive Assessment of Mental Stress based on Wearable Wireless Physiological Sensors and Multivariate Biosignal Processing
- Autori: Pernice, Riccardo; Nollo, Giandomenico; Zanetti, Matteo; Cecco, Mariolino De; Busacca, Alessandro; Faes, Luca
- Anno di pubblicazione: 2019
- Tipologia: Contributo in atti di convegno pubblicato in volume
- Parole Chiave: EEG; physiological signals; stress assessment; time series analysis; wearable devices;
- OA Link: http://hdl.handle.net/10447/380104
The development of connected health technologies for the continuous monitoring of the psychophysical state of individuals performing daily life activities requires the aggregation of non-intrusive sensors and the availability of methods and algorithms for extracting the relevant physiological information. The present study proposes an integrated approach for the objective assessment of mental stress which combines wirelessly connected low invasive biosensors with multivariate physiological time series analysis. In a group of 18 healthy subjects monitored in a relaxed resting state and during two experimental conditions inducing mental stress and sustained attention (respectively, mental arithmetic and serious game), we collected simultaneously multichannel EEG, one lead ECG, respiration and blood volume pulse. From these signals, synchronous physiological time series were extracted measuring the _, _, _, and _ EEG amplitudes, the heart period, the sampled respiratory activity and the pulse arrival time. For each condition, five minute windows of each of these seven time series were characterized with measures in the time domain (mean, standard deviation) and in the information domain (self entropy, measuring time series regularity). We show that the dynamical activity of the different physiological systems is affected in a different way by the alteration of the psychophysical state of the subjects induced by stress, and that the measures in the two domains can elicit complementary information about mental stress and sustained attention. These results advocate the feasibility of connected health technology for minimally invasive, automatic classifiers of different levels of mental stress in real life scenarios.