基于Hilbert谱的心率变异信号时频分析方法

Time-frequency analysis method of heart rate variability signal based on the Hilbert spectrum

心率变异性(heart rate variability,HRV)信号的频域分析方法是用来表征自主神经调控信息的基本方法。为了评估自主神经系统功能及对心血管活动的影响,提出一种基于Hilbert时频谱的HRV信号的时频特征提取和分析的新方法。对HRV信号进行希尔伯特-黄变换(Hilbert-Huang transform,HHT)以获得HRV信号Hilbert时频谱,依据短时程HRV信号的线性频域分析指标,得到不同生理频带的Hilbert能量棒形图,提取总能量、各生理频带的能量和其归一化能量以及生理频段的能量比值作为评价心率变异性的时频特征。对MIT-BIH数据库的年轻人、老年人样本和健康人、心衰病人样本的HRV信号分析表明,基于Hilbert谱的时频特征的区分性能好,有较清晰的生理意义,能反映人的生理病理变化,为短时程HRV信号分析提供了一种有效方法。

Frequency domain analysis for heart rate variability（HRV） signal is the basic method in testing autonomic nervous activity.In order to assess the function of autonomic nervous system and its influence on cardiovascular activity,a new method of time-frequency feature extraction and analysis of HRV signal is proposed based on the Hilbert spectrum.Hilbert-Huang transform（HHT） is performed on the HRV signal to get Hilbert time-frequency spectrum;according to the characteristic indexes of short-time HRV signal in linear frequency domain,a Hilbert energy bar chart of the HRV signal for different physiological frequency bands is plotted.The total energy,the energies of different physiological frequency bands and their normalized energies,and the energy ratios of physiological frequency bands are extracted and used for evaluating the time-frequency features of the HRV signal.The new approach was applied to the HRV signals of the young,older and heart failure patients in the MIT-BIH standard database.Results show that the time-frequency features based Hilbert time-frequency spectrum have good differentiating performance and clear physiological significance,and can reflect the physiological and pathological changes of the patients,which provides an effective method for short-time HRV signal analysis.