Frequency Domain Characterization of Signals and Systems

This chapter investigates the frequency-domain characteristics of a few biomedical signals and the corresponding physiological systems, with particular attention to the phonocardiogram (PCG) and the cardiovascular system. Frequency-domain analysis via power spectral density (PSD) and parameters derived from PSDs can enable one to view the signal from a different perspective than the time domain. Certain signals such as the PCG and electroencephalogram may not lend themselves to easy interpretation in the time domain and, therefore, may benefit from a move to the frequency domain. PSDs and their parameters facilitate investigation of the behavior of physiological systems in terms of rhythms, resonance, and parameters that could be related to the physical characteristics of anatomical entities (for example, the loss of elasticity of the myocardial muscles due to ischemia or infarction, the extent of aortic valvular stenosis, or the extent of calcification and stiffness of bioprosthetic valves). Pathological states may also be derived or simulated by modifying the spectral parameters or representations of the corresponding normal physiological states and signals. The design of biomedical signal analysis techniques requires a thorough understanding of the characteristics and properties of the biomedical systems behind the signals, in addition to detailed knowledge of mathematical principles, computer techniques, and digital signal processing algorithms.