pyphysio: A physiological signal processing library for data science approaches in physiology

• Main Authors: Bizzego, Andrea, Battisti, Alessandro, Gabrieli, Giulio, Esposito, Gianluca, Furlanello, Cesare
• Other Authors: School of Social Sciences
• Format: Article
• Language: English
• Published: 2019
• Subjects: Social sciences::Psychology; Physiological Signal Processing; Psychophysiology
• Online Access: https://hdl.handle.net/10356/81261; http://hdl.handle.net/10220/50394
• Institution: Nanyang Technological University

The lack of open-source tools for physiological signal processing hinders the development of standardized pipelines in physiology. Researchers usually must rely on commercial software that, by implementing black-box algorithms, undermines the control on the analysis and prevents the comparison of the results, ultimately affecting the scientific reproducibility. We introduce pyphysio as a step towards a data science approach oriented to compute physiological indicators, in particular of the Autonomic Nervous System activity. pyphysio serves as a basis for machine learning modules and it implements a suite of combinable algorithms for processing of signals from either by wearable or medical-grade quality devices.