Skip to Main Content
ASME Press Select Proceedings

International Conference on Computer and Computer Intelligence (ICCCI 2011)

Yi Xie
Yi Xie
Search for other works by this author on:
No. of Pages:
ASME Press
Publication date:

Advancements in clinical, portable, and wearable equipment for real-time collection of physiological data provide new opportunities for computerized diagnostics of developed pathologies, early detection of emerging abnormalities, and prediction of acute and critical events. However, many conceptual and algorithmic challenges for robust quantitative modeling in such applications remain unresolved. Variability analysis of physiological time series provides a generic framework for quantification of normal and abnormal states and their discrimination. Unfortunately, in many clinically significant cases it is hard to achieve robust 1œnormal-abnormal1 classification using this framework or other established diagnostic modalities. Recently, we have demonstrated that many problems in heart rate variability (HRV) analysis could be overcome when several complementary nonlinear dynamics (NLD) indicators (complexity measures) are combined using boosting-like algorithms. Such generic meta-indicators are capable to detect both single abnormalities irrespective of their specific type, and conditions specified by complex combination of different pathologies. Here we argue that aggregated probability-like output of these multi-component models could be effective for more detailed quantification of psycho-physiological states. These robust state representations could be used as early signals of emerging abnormalities and other negative physiological changes as well as for real-time prediction of acute and critical events. In addition, we propose some extensions to HRV analysis and possible ways of its application to other physiological channels.

This content is only available via PDF.
Close Modal
This Feature Is Available To Subscribers Only

Sign In or Create an Account

Close Modal
Close Modal