Autonoumous Nervous System biosignal processing via EDA and HRV from a wearable device

• Autor(a) principal: Lima, Rodrigo Olival
• Data de Publicação: 2018
• Tipo de documento: Dissertação
• Idioma: eng
• Título da fonte: Repositórios Científicos de Acesso Aberto de Portugal (RCAAP)
• Texto Completo: http://hdl.handle.net/10362/59608
• topic: Heart Rate Variability; Electrodermal Activity; Photoplethysmography; Biosignals; Autonomous Nervous System; Wearable Device; Domínio/Área Científica::Engenharia e Tecnologia::Outras Engenharias e Tecnologias

The assessment of changes in the autonomous nervous system (ANS) with certain diseases and pathologies conditions, has been demonstrated to have important prognostic and diagnostic value, so delineating the role of autonomous activity is important to prevent health diseases. There are many approaches to directly measure the sympathetic and parasympathetic nervous system, although, most of themare invasive and unable to provide continuous monitoring, leading to inaccurate assessment of the autonomous nervous system. Heart rate variability (HRV) and Electrodermal activity (EDA) are presented has noninvasive methods to assess the ANS, by computing the spectral analysis of both HRV and EDA biosignals. The combination of these signals is necessary to correctly measure the activity of the sympathetic and parasympathetic system, due to the fact that frequency analysis of HRV only provides the level of unbalance between these two systems, while EDA reflects only activity from the sympathetic system. ANS biosignal processing via HRV and EDA from a wearable device was studied in this thesis, in order to provide continuous monitoring. A wearable device is the ideal solution, as HRV can be calculated with photoplethysmography signals from the wrist and EDA from the fingers, providing wireless and continuous monitoring of the subjects. The extraction of the HRV and EDA features, that describe the activity of the sympathetic and parasympathetic system, were obtained by submitting the subjects to a mental arithmetic stress test, and then compared to the baseline values, in order to verify changes in the autonomous nervous system between the two situations. The distinct response to stress for the subjects was then predicted usingmachine-learning classification mechanisms, with the ability to predict how the subject will respond when submitted to a situation of stress, using only time-domain features, instead of frequency-domain features, which reduces the time needed to performthe classification.