PPG Waveform Morphology: Features and Analysis
PPG morphology analysis examines the shape characteristics of individual pulse waveforms to extract cardiovascular information beyond heart rate. Key morphological features include systolic peak amplitude and timing, dicrotic notch position, diastolic peak, pulse width at half-maximum, and features from the first and second time derivatives.
A single PPG beat consists of several anatomically interpretable components. The systolic upstroke reflects rapid ventricular ejection and arterial filling. The systolic peak marks peak arterial volume. The dicrotic notch corresponds to aortic valve closure and the transition from systole to diastole. The diastolic wave reflects the reflected pressure wave from peripheral sites. Beat-to-beat variation in these features encodes information about cardiac contractility, preload, afterload, and vascular tone.
Feature extraction from PPG morphology has been applied to numerous clinical prediction tasks. Systolic peak width (SPW) and pulse area below the systolic peak are associated with cardiac output. The ratio of the diastolic-to-systolic amplitude correlates with peripheral resistance. Pulse wave analysis combining 15+ morphological features extracted by machine learning can predict diabetes (AUC 0.75–0.85), hypertension (AUC 0.80–0.88), and atrial fibrillation (AUC 0.90–0.97) in population studies.
Deep learning approaches have dramatically advanced PPG morphology analysis. Convolutional neural networks trained end-to-end on raw waveforms outperform hand-crafted feature extraction for AF detection, age estimation, and blood pressure prediction. Transformer models with attention mechanisms over multi-beat sequences capture temporal dependencies that single-beat features miss, achieving state-of-the-art performance in cardiovascular risk stratification from continuous wearable PPG.
Frequently Asked Questions
What is the dicrotic notch and why does it matter?
The dicrotic notch is a brief downward deflection in the PPG waveform corresponding to aortic valve closure. Its absence or blunting indicates increased arterial stiffness or high peripheral resistance.
Can PPG morphology predict atrial fibrillation?
Yes. CNN models analyzing PPG beat irregularity and morphological variability achieve 94–97% accuracy for AF detection, comparable to single-lead ECG algorithms in some validation studies.
What is the second derivative of PPG (SDPPG)?
The SDPPG (or APG, acceleration plethysmogram) reveals five characteristic waves (a–e) that are difficult to see in the raw PPG. These waves quantify early and late systolic components and the diastolic reflection wave, enabling vascular age and stiffness assessment.