Artifacts in dependency structures of Portapres® signals

Abstract

Portapres® is a unique device that reliably accomplishes a challenging task of continuous and non-invasive recording of
blood pressure (BP) waveforms in moving subjects. The complex procedure of Portapres® signal acquisition includes periodic changes
of cuffed fingers to avoid pain and stress, as well as the blood pressure correction due to the increasing and decreasing elevation of arm
posture. Due to these procedures, the recorded waveforms are corrupted. The aim of this paper is to analyze the influence of inevitable
artifacts on parameters obtained from the blood pressure waveforms. The analyzed waveforms are obtained from healthy volunteers at
Bezanija Kosa Hospital, Belgrade. The parameters include systolic blood pressure (SBP) and pulse interval (PI) extracted by
Beatscope® software. The interrelationship of SBP and PI signals forms a major cardiovascular feedback – baroreflex. It can be
analyzed using the sequence method for spontaneous baroreflex sensitivity, but the tools that reveal more profound dependency
structures include cross-approximate and cross-sample entropy, as well as the copula structures. The influence of artifacts, inevitable in
Portapres® signals, is the main goal of this study. The analyses revealed that automatic artifact correction induced no significant
changes considering the statistical moments and the baroreflex sensitivity; the same applies to the copula density and rank tests. The
entropy analysis, however, turned out to be extremely sensitive so its implementation in Portapres® signal analysis is not recommended.