Topological Nonlinear Analysis of Dynamical Systems for Patient-Ventilator Asynchrony Events Recognition in Mechanical Ventilation

Mechanical ventilation is critical to patients suffering from respiratory disorders. However, patient-ventilator asynchrony(PVA) events seriously impact patients’ comfort in rehabilitation or treatment progress and, sometimes, threaten clinical safety. Thus, detecting and altering abnormal mechanical ventilation events are vital in biomedical state monitoring. This paper proposes a topological nonlinear analysis approach to identifying PVA events, addressing the recognition challenge during mechanical ventilation. The proposed approach distinguished itself in the experimental validations performed with clinical data collected from 22 subjects for recognizing common PVA events, including double triggering, ineffective efforts, premature cycling, and delayed cycling. The proposed topological feature shows advantages compared to other related nonlinear dynamics descriptors, which is supposed to be a novel solution for PVA analysis.