Clinical application and research based on the analysis of short-term physiological series with base scale entropy
Yang Xi,Ning Xin-Bao,He Ai-Jun,Zhuang Jian-Jun,Bian Chun-Hua
DOI: https://doi.org/10.3321/j.issn:0469-5097.2008.04.005
2008-01-01
Abstract:The heart rate variability(HRV) signals reflect the reactive nerve's regulation on the cardiac rhythm,and HRV parameters and measurements of the time series of the beat-to-beat-intervals have been used to predict the mortality risk in patients with structural heart diseases.Complexity in HRV time series is such a measurement that can reflect cardiac physiological function and health.Varieties of methods for complexity in a complex dynamic system have been developed for time series analysis,such as correlation dimension,entropy measurement,multi-fractal,and Lyapunov exponent,and so on.However,these methods are less efficient when processing short term time series.This paper introduces a recently developed method,base scale entropy method,to detect evolving characteristic parameters for the complexity in HRV signals,and then apply these parameters to application in the clinical diagnoses field.The basic scale entropy method evolves from symbol dynamics,which have been shown to be an efficient approach to analyze dynamic aspects of HRV,yet the basic scale entropy method is much more effective than symbol dynamics methods in our lab's experiment and research,for the former can better eliminate noises interspersing in the raw time series and better analyze short-term time series.To begin with,we need reconstruct the HRV time series into a vector space of dimension m×(N-m+1),where N is the length and m is the embedded dimension of the HRV time series.Then it involves a transformation of the time series into symbol sequences with symbols form a given alphabet.The basic scale entropy method chooses the alphabet {0,1,2,3},and transforms referring to three given levels where BS denotes the root mean standard deviation of neighbor data values' differences and a is a special parameter that we choose by test.Finally this method provides an effective parameter,basic scale entropy,as a measurement for the complexity in time series.Larger values of basic scale entropy refer to higher complexity in the corresponding tachograms and lower values to lower ones.Besides this parameter,we introduce another two,one is forbidden words number,the other is m-wordsdistribution histogram,where m-words refers to the combination formats of symbols from the alphabet in the vectors of dimension m having been transferred from the reconstructive vector space.Forbidden words number unveils otherwise the complexity in time series and shows adverse tendency that from basic scale entropy,larger values of basic scale entropy company lower values of forbidden words number and values company larger ones.The other effective parameter,the m-words distribution histogram,refers to the wave modes' details of the HRV time series.Based on the aforementioned theory and method,we compile software which can put forth an analysis report of the HRV time series having been analyzed.The report shows the histogram,and lists both values of basic scale entropy and forbidden words number under the picture of histogram.We use five groups of data,got from 50 healthy young volunteers and 49 hypertension,15 angina,64 coronary heart disease and 37 heart failure patients to test these parameters' availabilities,with a hope to give a reference to the clinical diagnoses.The five groups of data are collected in the PLA 414 hospital with cooperation of our lab.All these data are collected under the standard measuring criterions.In the data's pretreatment,those data not locating in the range of 500-1500ms are dealt simply with deletion.The results of five groups of data show an indication accordant with the basic theories' conclusion.Larger the values of forbidden words number refer to higher certainty of the combination formats distribution,and correspondingly the lower values of the basic scale entropy.The mean value of entropy is lowest and of forbidden words number is highest in the group of young healthy men,from which the other four groups all deviate to some extents.We also find out that,the histogram can show clear and intuitionistic references for the clinical diagnoses.Only when the four highest distributions locate on the four fixed points on the abscissa: 95,125,215,245,which denote the corresponding m-words combination formats(1,1,3,3)(1,3,3,1)(3,1,1,3)(3,3,1,1),the result can be primarily considered as of health condition.As a result,this application can be served to provide significant references for the clinical diagnoses and further heart diseases researches.