PPG Waveform Decomposition: Extracting Physiological Components from the Pulse Wave
**The photoplethysmography waveform is not a simple cardiac pulse but a composite signal encoding information about cardiac function, vascular complia...
- PPG & Signal Processing
The photoplethysmography waveform is not a simple cardiac pulse but a composite signal encoding information about cardiac function, vascular compliance, respiratory mechanics, autonomic nervous system activity, and peripheral hemodynamics. Decomposing this complex waveform into its constituent components unlocks clinical and physiological insights far beyond basic heart rate. This article examines the physiological origins of each PPG waveform component, the mathematical methods used to extract them, and the diagnostic potential of component-level analysis.
For background on how the PPG signal is generated and its basic characteristics, see our introduction to PPG technology.
Anatomy of the PPG Pulse Wave
A single PPG pulse cycle contains several morphological features, each carrying distinct physiological information. Understanding these features at the waveform level is the prerequisite for meaningful decomposition.
The Systolic Phase
The systolic phase begins at the foot of the pulse (the diastolic minimum) and rises to the systolic peak. This ascending limb represents the arrival of the arterial pressure wave at the measurement site, driven by left ventricular ejection. The slope of the systolic upstroke is determined by the rate of pressure rise in the arterial system, which depends on cardiac contractility, stroke volume, and proximal arterial compliance.
The maximum slope (maximum first derivative) of the systolic upstroke occurs approximately one-third of the way up the rising edge and correlates with the rate of pressure change (dP/dt) in the ascending aorta.
The Dicrotic Notch
The dicrotic notch is a small but physiologically significant dip that appears on the descending limb of the PPG pulse, separating the systolic and diastolic phases. It corresponds temporally to aortic valve closure and the subsequent brief retrograde flow in the aorta. In the peripheral PPG signal, the notch also encodes information about the reflected pressure wave returning from the lower body vasculature.
The visibility and timing of the dicrotic notch depend strongly on arterial compliance, peripheral vascular resistance, and measurement site. At the fingertip, where the signal amplitude is highest, the dicrotic notch is typically well-defined in healthy young adults but diminishes with age, hypertension, and atherosclerosis. Dawber et al. (1973, doi:10.1177/000331977302400407) related the characteristics of the dicrotic notch of the finger pulse wave to coronary heart disease. At the wrist, the dicrotic notch is often attenuated or absent even in young subjects due to the lower signal quality in reflectance mode.
The Diastolic Phase
After the dicrotic notch, the diastolic phase represents the gradual runoff of blood from the arterial system through peripheral resistance vessels during cardiac relaxation. The diastolic decay rate follows an approximately exponential curve, with the time constant determined by the product of arterial compliance and peripheral resistance (the Windkessel time constant, ).
The diastolic time constant can be extracted from the PPG waveform by fitting an exponential function to the diastolic decay segment.
Multi-Component Signal Model
The total PPG signal can be conceptualized as the sum of several physiological components operating at different time scales. This multi-component model provides the framework for decomposition.
DC Component (Baseline)
The largest portion of the PPG signal (typically 95-99% of total light absorption) is the DC or non-pulsatile component. This represents steady-state light absorption by skin pigmentation (melanin), tissue matrix, non-pulsatile venous blood, and steady-state arterial blood volume. The DC component is not clinically uninteresting: changes in venous blood volume, tissue edema, and skin perfusion all modulate the DC level. However, because the DC component is dominated by non-vascular absorption and is sensitive to sensor placement and contact pressure, extracting reliable physiological information from it requires careful calibration.
AC Cardiac Component (0.5-4 Hz)
The pulsatile cardiac component, oscillating at the heart rate frequency and its harmonics, represents the primary signal of interest in most PPG applications. This component has a typical amplitude of 1-2% of the DC level at the fingertip and 0.1-0.5% at the wrist. The cardiac AC component contains the morphological features described above: systolic upstroke, peak, dicrotic notch, and diastolic decay.
The harmonic content of the cardiac component is itself informative. A fundamental-only pulse (pure sinusoid) would indicate a highly damped vascular system, while a pulse rich in harmonics indicates sharp waveform features associated with wave reflections and preserved vascular compliance. Fourier analysis of individual pulses typically reveals 4-8 significant harmonics, with the relative harmonic amplitudes changing with arterial stiffness and age.
Respiratory Component (0.15-0.5 Hz)
Respiration modulates the PPG signal through three distinct mechanisms, each contributing to a respiratory component in the 0.15-0.5 Hz band.
First, respiratory-induced intrathoracic pressure changes modulate venous return, which in turn affects stroke volume through the Frank-Starling mechanism. This produces beat-to-beat amplitude modulation of the PPG pulse. During inspiration, reduced intrathoracic pressure increases venous return, which (after a 1-2 beat delay) increases stroke volume and PPG pulse amplitude.
Second, respiration modulates heart rate through respiratory sinus arrhythmia (RSA), producing baseline frequency modulation of the pulse intervals. This component can be extracted from the IBI time series as described in our guide to PPG inter-beat interval measurement.
Third, changes in intrathoracic pressure directly affect peripheral arterial pressure, causing baseline wander in the PPG signal. This mechanism is most apparent during deep breathing or mechanical ventilation, where the baseline modulation can approach 5-10% of the AC amplitude.
Sympathetic/Vasomotor Component (0.04-0.15 Hz)
Low-frequency oscillations in vascular smooth muscle tone, known as Mayer waves or vasomotor waves, produce oscillations in the PPG baseline and amplitude at approximately 0.1 Hz (roughly a 10-second cycle). These oscillations reflect baroreflex-mediated sympathetic nervous system modulation of peripheral vascular resistance and are of interest in autonomic function assessment.
These low-frequency PPG oscillations are generally attributed to vasomotor activity under sympathetic control, and their amplitude tends to change with sympathetic stimuli such as the cold pressor test or orthostatic stress.
Thermoregulatory and Very Low Frequency Components (< 0.04 Hz)
Very slow oscillations in the PPG baseline reflect thermoregulatory vasoconstriction-vasodilation cycles, which operate on time scales of minutes. These components are relevant for long-duration PPG recordings but are typically filtered out in standard cardiac monitoring applications.
Mathematical Decomposition Methods
Several mathematical frameworks have been applied to separate PPG waveform components, each with distinct strengths and physiological interpretability.
Gaussian Decomposition
The PPG pulse can be modeled as a sum of Gaussian functions, where each Gaussian represents a physiological sub-wave. The forward-traveling systolic wave is modeled as a primary Gaussian centered near the systolic peak. Reflected waves from major arterial bifurcation sites (renal, iliac, aortic) are modeled as additional Gaussians arriving at later times.
Rubins (2008; doi:10.1007/s11517-008-0406-z) separated the systolic and diastolic waves of finger and ear PPG pulses from 40 healthy subjects and fitted each with the sum of two Gaussian functions, deriving the timing of the direct wave and three reflected waves. The timing and amplitude parameters of each Gaussian carry physiological meaning: the delay between the primary and reflected wave Gaussians relates to pulse wave velocity and arterial path length, while the amplitude ratio encodes the reflection coefficient of the peripheral vasculature.
Empirical Mode Decomposition (EMD)
EMD adaptively decomposes the PPG signal into a set of intrinsic mode functions (IMFs) without requiring predefined basis functions. Each IMF represents an oscillatory component at a characteristic time scale, extracted through an iterative sifting process.
For PPG signals, EMD typically produces the following hierarchy of IMFs. IMF1 captures high-frequency noise and sensor artifacts. IMF2-3 contain the cardiac pulse waveform details. IMF4-5 capture the respiratory modulation. IMF6-7 represent the sympathetic/vasomotor oscillations. Higher-order IMFs and the residual represent very slow trends and the DC component.
For respiratory extraction, the advantage of EMD over fixed-frequency filtering is its ability to track non-stationary respiratory patterns where the breathing rate varies over time.
The limitation of EMD is mode mixing, where components from different physiological sources contaminate the same IMF. Ensemble EMD (EEMD), which averages multiple EMD decompositions with added white noise, partially addresses this problem (Wu & Huang, 2009; doi:10.1142/S1793536909000047).
Wavelet Decomposition
The discrete wavelet transform (DWT) decomposes the PPG signal into approximation and detail coefficients at multiple resolution levels. Each level corresponds to a specific frequency band, allowing separation of cardiac, respiratory, and vasomotor components based on their characteristic frequencies.
For a PPG signal sampled at 125 Hz, the typical wavelet decomposition levels align as follows. Level 1-2 details (31-62 Hz) capture high-frequency noise. Level 3-4 details (8-31 Hz) contain cardiac harmonic content. Level 5 details (2-4 Hz) capture the fundamental cardiac component. Level 6 details (1-2 Hz) overlap cardiac and respiratory bands. Level 7-8 details (0.24-1 Hz) contain respiratory modulation. Level 9-10 details (0.06-0.24 Hz) capture sympathetic oscillations.
Addison (2005; doi:10.1088/0967-3334/26/5/R01) provided a comprehensive review of wavelet methods for physiological signal analysis, including specific applications to PPG decomposition. The choice of mother wavelet affects the decomposition quality, with the Daubechies-4 (db4) and symlet-4 (sym4) wavelets commonly used for PPG due to their morphological similarity to the pulse waveform.
Second Derivative Analysis (SDPPG)
The second derivative of the PPG waveform, known as the acceleration plethysmogram (APG) or SDPPG, provides an alternative decomposition into characteristic wave components labeled a, b, c, d, and e. The a-wave represents initial systolic positive acceleration. The b-wave represents systolic negative acceleration (early systolic deceleration). The c-wave is a mid-systolic re-acceleration related to the reflected wave. The d-wave is a late systolic deceleration. The e-wave represents early diastolic acceleration after the dicrotic notch.
Takazawa et al. (1998; doi:10.1161/01.HYP.32.2.365) demonstrated in 600 subjects (100 per decade from the 20s to the 70s) that b/a increases and c/a, d/a and e/a decrease with age. The composite aging index (b-c-d-e)/a increased with age (r = 0.80) and was higher in subjects with diabetes, hypertension, hypercholesterolemia or ischemic heart disease, and it has been proposed as a non-invasive vascular aging biomarker.
For more on how waveform analysis algorithms are implemented in PPG processing pipelines, see our algorithms documentation.
Forward and Reflected Wave Separation
A particularly informative decomposition separates the PPG pulse into its forward-traveling (incident) wave and backward-traveling (reflected) wave components. The forward wave originates from cardiac ejection, while the reflected wave returns from peripheral impedance mismatches, primarily at major arterial bifurcations.
Wave Separation Analysis
Wave separation analysis (WSA), developed by Westerhof and colleagues for arterial pressure waveforms, has been adapted for PPG. The method assumes that the measured PPG waveform is the superposition of a forward wave Pf and a reflected wave Pb:
The forward and reflected components can be estimated using the characteristic impedance of the arterial segment and the flow waveform. For PPG, where direct flow measurement is not available, the forward-reflected separation is approximated using waveform modeling approaches that fit the measured pulse as a sum of sub-waves with constrained temporal relationships.
Augmentation Index
The augmentation index (AIx) quantifies the contribution of the reflected wave to the systolic peak. It is defined as the difference between the second and first systolic peaks (or the systolic peak and the inflection point) divided by the pulse pressure. Higher AIx values indicate greater wave reflection and are associated with arterial stiffness, increased cardiovascular risk, and aging.
Millasseau et al. (2006; doi:10.1097/01.hjh.0000239277.05068.87) reviewed digital volume pulse contour analysis and noted that, as with the pressure pulse, its contour is sensitive to vasoactive drugs, ageing and large artery stiffness. The PPG-derived reflection index (RI), calculated as the diastolic peak amplitude divided by the systolic peak amplitude, provides a simpler alternative that is less sensitive to waveform calibration.
Clinical Applications of Waveform Decomposition
PPG waveform decomposition has enabled several clinical applications beyond basic heart rate and SpO2 monitoring. These applications represent the cutting edge of non-invasive cardiovascular assessment.
Respiratory Rate Estimation
Extracting the respiratory component from decomposed PPG signals enables continuous, contactless respiratory rate monitoring. This is particularly valuable in clinical settings where dedicated respiratory monitoring equipment is not available. Charlton et al. (2016; doi:10.1088/0967-3334/37/4/610) systematically assessed 314 respiratory rate algorithms in healthy participants; the best algorithm operating on PPG had a bias of 1.0 breaths per minute with 95% limits of agreement of -5.1 to 7.2 breaths per minute.
Arterial Stiffness Assessment
The timing and amplitude of the reflected wave component provide non-invasive indices of arterial stiffness. The stiffness index (SI), calculated as body height divided by the time delay between systolic and diastolic peaks, reflects pulse wave velocity in the aorta and large arteries. Millasseau et al. (2002, doi:10.1042/cs1030371) validated SI against carotid-femoral PWV with correlation r = 0.65 in 87 subjects. While not replacing clinical PWV measurement, PPG-derived stiffness indices enable longitudinal tracking and screening.
Cardiac Output Estimation
The amplitude and area of the systolic component of the decomposed PPG waveform correlate with stroke volume.
Conclusion
PPG waveform decomposition transforms a seemingly simple pulse signal into a rich source of cardiovascular and autonomic information. The mathematical methods available, from Gaussian fitting to wavelet analysis to SDPPG, offer complementary perspectives on the underlying physiology. The key challenge remains ensuring that decomposition results are robust to the signal quality limitations of real-world PPG measurement, including motion artifacts, sensor contact variability, and inter-individual morphological differences. As computational methods improve and multi-wavelength PPG sensors become standard, the clinical utility of waveform-level analysis will continue to expand.
For deeper exploration of PPG signal processing and analysis methods, visit our algorithms reference and PPG learning center.
References
- Rubins U. Finger and ear photoplethysmogram waveform analysis by fitting with Gaussians. Med Biol Eng Comput. 2008;46(12):1271-1276. https://doi.org/10.1007/s11517-008-0406-z
- Takazawa K, Tanaka N, Fujita M, et al. Assessment of vasoactive agents and vascular aging by the second derivative of photoplethysmogram waveform. Hypertension. 1998;32(2):365-370. https://doi.org/10.1161/01.HYP.32.2.365
- Millasseau SC, Ritter JM, Takazawa K, Chowienczyk PJ. Contour analysis of the photoplethysmographic pulse measured at the finger. J Hypertens. 2006;24(8):1449-1456. https://doi.org/10.1097/01.hjh.0000239277.05068.87
- Charlton PH, Bonnici T, Tarassenko L, et al. An assessment of algorithms to estimate respiratory rate from the electrocardiogram and photoplethysmogram. Physiol Meas. 2016;37(4):610-626. https://doi.org/10.1088/0967-3334/37/4/610
- Dawber TR, Thomas HE, McNamara PM. Characteristics of the dicrotic notch of the arterial pulse wave in coronary heart disease. Angiology. 1973;24(4):244-255. https://doi.org/10.1177/000331977302400407
- Millasseau SC, Kelly RP, Ritter JM, Chowienczyk PJ. Determination of age-related increases in large artery stiffness by digital pulse contour analysis. Clin Sci. 2002;103(4):371-377. https://doi.org/10.1042/cs1030371
Core guide: PPG Waveform Morphology: Features and Analysis
Related articles
PPG Vascular Age Assessment: Arterial Stiffness Measurement from Pulse Wave Analysis
**Arterial stiffness is one of the earliest detectable markers of cardiovascular disease, and the PPG waveform encodes surprisingly rich information a...
Pulse Wave Analysis Using PPG: Waveform Features, Vascular Health, and Clinical Applications
Complete guide to pulse wave analysis from PPG signals. Covers augmentation index, stiffness index, derivatives, fiducial points, and vascular assessment.
PPG Stress Detection Methods: Physiological Stress Measurement from PPG Features
**Stress detection is one of the most commercially visible applications of PPG technology, yet it remains one of the most scientifically challenging.*...
PPG Respiratory Rate Estimation: Extracting Breathing Rate from Photoplethysmography
**Respiratory rate is the vital sign most sensitive to clinical deterioration, yet it is the least frequently and least accurately recorded in routine...
Pulse Transit Time from PPG for Cuffless BP
How pulse transit time (PTT) from PPG tracks blood pressure without a cuff: PTT vs PAT, calibration, sensor sites, and where validation stands today.
Pulsus Bisferiens: Double-Peaked Pulse Waveform
What pulsus bisferiens is, how to identify the double-peaked arterial pulse, its causes (aortic regurgitation, HOCM), and detection via PPG waveforms.