Building an open-source community to enhance autonomic nervous system signal analysis: DBDP-autonomic

Figures

• 

  Figure 1. This figure demonstrates the need for additional context when analyzing ambulatory physiological signals. We used a stress-prediction model on heart rate variability data to predict a probability of physiological stress/arousal for a person over 24 h (a). From a naïve interpretation, it seems there are three major stressful (or high arousal) periods (b). However, when asked to self-report stress levels, the user rated a mix of low to high stress for those periods, contradictory to the purely physiological interpretation (c). Considering additional contextual information, we realize that only one high-arousal episode was stressful since the user was undergoing an exam. The other periods were when the user was in a class and exercising later during the day, which showed similar physiological arousal but were not stressful (d).

• 

  Figure 2. DBDP Autonomic extends the functionalities of DBDP for biobehavioral research. (A) As signals from various modalities enter the analysis pipeline, DBDP Autonomic provides additional features on top of the existing modules in DBDP. These features extract and add contextual information, provide domain knowledge (for parameter tuning), and support multimodal data fusion. (B) Researchers can then integrate the processed features such as HR, RR, and BP to understand the autonomic constructs in the context of major domains of basic human neurobehavioral functioning. ECG, electrocardiogram; PPG, photoplethysmography; RIP, respiratory inductance plethysmography; BP, blood pressure; EDA, exploratory data analysis; HR, heart rate; HRV, heart rate variability; SBP, systolic BP; DBP, diastolic BP; SCR, skin conductance response; SCL, skin conductance level.