Quantitative analysis of externally-induced patterns and natural oscillations in the human cardiovascular response: Implications for development of a biofeedback system. (July 2017)

• Record Type: Journal Article
• Authors: Sarabadani Tafreshi, Amirehsan; Riener, Robert; Klamroth-Marganska, Verena
• Is Part Of: Biomedical signal processing and control. Volume 36(2017)
• Page Start: 76
• Page End: 83
• Publication Date: 2017-07
• Subjects: Rehabilitation -- Robotic tilt table -- Closed-loop control -- Signal difference to noise ratio (SDNR) -- Heart rate (HR) and blood pressure (BP) variability; Signal processing -- Periodicals; Biomedical engineering -- Periodicals; Signal Processing, Computer-Assisted -- Periodicals; Image Processing, Computer-Assisted -- Periodicals; Biomedical Engineering -- Periodicals; 610.28
• DOI: 10.1016/j.bspc.2017.03.021
• Languages: English
• ISSNs: 1746-8094

Abstract

Highlights: We used signal-difference-to-noise ratio to assess feasibility of developing a closed-loop system to control the human cardiovascular system. The inputs provided with a tilt table were body inclination, passive leg exercise, and functional electrical stimulation; outputs were the cardiovascular variables heart rate, and systolic and diastolic blood pressures. Only body inclination induced clinically relevant changes in the cardiovascular variables. It is feasible to develop a biofeedback system based on body inclination. However, such a system has limited clinical value. Abstract: Immobilization and prolonged bed rest can result in secondary complications that threaten survival chances of patients. Early mobilization can be achieved with a robotic tilt table that allows modulation of the external stimuli body inclination, passive robotic leg exercise, and functional electrical stimulation. However, due to the prevalence of cardiovascular instability in the patients, it may not be safe. A closed-loop biofeedback system with the external stimuli as inputs to control three relevant cardiovascular variables (i.e., heart rate, and systolic and diastolic blood pressures as outputs) would allow safe and early mobilization. To evaluate the feasibility of developing such a biofeedback system, we conducted a series of open-loop experiments with ten healthy subjects. Steady-state changes in the cardiovascular variables that were induced by the inputs (i.e., induced biosignal differences) were defined as gains of the system. Natural oscillations during these steady-state changes were considered as noise. We analyzed these features to test (1) whether the externally-induced changes are differentiable from natural oscillations and (2) whether the gains of the system are clinically relevant (i.e., >2.5 beats/min for heart rate, and >5 mmHg for systolic and diastolic blood pressures). To this end, we compared signal differences (i.e., induced changes in the biosignals; the gains of the system) with noise (i.e., natural oscillations) from a statistical perspective. Results showed that (1) only the changes induced by body inclination were differentiable from natural oscillations, and (2) only change of the body inclination produced reliable clinically relevant changes, and this only held for heart rate and diastolic blood pressure but not systolic blood pressure. We conclude that it is feasible to utilize body inclination for the development of a biofeedback system to control either heart rate or diastolic blood pressure. However, such a system where only two of the three relevant cardiovascular variables can be controlled, and these not simultaneously, would have limited clinical value.