Fig. 2

Comparison of photoplethysmography (PPG) signals and their derivatives from real and in-silico datasets. (a–c) Multicycle real PPG signal and its first and second derivatives, showing natural variability and noise representative of physiological conditions. (d–f) In-silico PPG signal and its derivatives exhibiting smoother and idealized waveform characteristics. The first derivatives emphasize the rate of change in signal amplitude, while the second derivatives highlight inflection points and subtle waveform features. While in-silico signals are valuable for controlled modeling and initial algorithm development, the real PPG data are essential for evaluating performance under realistic conditions, capturing inter-cycle variability, and biological noise important for robust and clinically relevant analysis.