Traditional signal spectral analysis examines both temporal and spectral content distinctly. This method of analysis is suitable for signals where the spectral content is stationary and time-invariant. However, hemodynamic signals such as heartrate variability and pulse-mean blood velocity variability are often non-stationary and time-variant. Here, time-dependent Fourier analysis may be used to transform the data into a quasi-stationary signal using a windowing function and thereby performinga short-time Fourier transform. However, even with time-dependent Fourier analysis, the signal spectral content may not be adequately determined due to inherent limitations of this technique (Cohen, 1995). The primary objective of time-frequency analysisusing non-Fourier based techniques is to create a distribution function that will accurately describe the energy density of the signal in both time and frequency, concurrently. In this abstract we use non-Fourier based time-frequency distributions, knownas reduced-interference-distributions (Jeong and Williams, 1992), to explore whether functional hemodynamic variability measures can be suitable indicators of health and injury in the chicken embryo during primary cardiovascular development.
展开▼
