An online algorithm for least-square spectral analysis: Applied to time-frequency analysis of heart rate

We propose a novel online algorithm for computing least-square based periodograms, otherwise known as the Lomb-Scargle Periodogram. Our spectral analysis technique has been shown to be superior to traditional discrete Fourier transform (DFT) based methods, and we introduce an algorithm which has O(N) time complexity per input sample. The technique is suitable for real-time embedded implementations and its utility is demonstrated through an application to the high resolution time-frequency domain analysis of heart rate variability (HRV).

[1]  J. Scargle.  Studies in astronomical time series analysis. II - Statistical aspects of spectral analysis of unevenly spaced data , 1982 .

[2]  G. Breithardt,et al.  Heart rate variability: standards of measurement, physiological interpretation and clinical use. Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology. , 1996 .

[3]  Narayanan Vijaykrishnan,et al.  Accelerating the Nonuniform Fast Fourier Transform Using FPGAs , 2010, 2010 18th IEEE Annual International Symposium on Field-Programmable Custom Computing Machines.

[4]  J. Bigger,et al.  Baroreflex sensitivity and heart-rate variability in prediction of total cardiac mortality after myocardial infarction , 1998, The Lancet.

[5]  W. Press,et al.  Fast algorithm for spectral analysis of unevenly sampled data , 1989 .

[6]  J. Miller,et al.  Decreased heart rate variability and its association with increased mortality after acute myocardial infarction. , 1987, The American journal of cardiology.

[7]  Lionel Tarassenko,et al.  Quantifying errors in spectral estimates of HRV due to beat replacement and resampling , 2005, IEEE Transactions on Biomedical Engineering.

[8]  Alexander Holland,et al.  A novel recursive Fourier transform for nonuniform sampled signals: application to heart rate variability spectrum estimation , 2009, Medical & Biological Engineering & Computing.

[9]  Hao He,et al.  Spectral Analysis of Nonuniformly Sampled Data: A New Approach Versus the Periodogram , 2009, IEEE Transactions on Signal Processing.

[10]  N. Lomb.  Least-squares frequency analysis of unequally spaced data , 1976 .

[11]  George B. Moody,et al.  Spectral analysis of heart rate without resampling , 1993, Proceedings of Computers in Cardiology Conference.

[12]  V. Hombach,et al.  THE NORTH AMERICAN SOCIETY OF PACING AND ELECTROPHYSIOLOGY , 1985, Pacing and clinical electrophysiology : PACE.