Photoplethysmography (PPG) is a non-invasive optical technique used widely for monitoring vital signs via detecting dynamic changes of optical signals modulated by blood flow. However, there is still a knowledge gap regarding the variations of PPG signal under different spectra on different anatomical sites, especially on hands. This study customized a multispectral imaging platform to collect PPG signals at the distal and middle phalanges of fingers from ten healthy participants. The pulsatile component of the PPG signals was computed and evaluated. The experimental findings revealed significant variations in the pulsatile component across anatomic regions and spectral ranges. Drawing from a mathematical model on the intensity variation of reflected lights, we attributed these variations to factors such as light penetration depth, vascular network density, and epidermal thickness. The experiments provide evidence for further investigation into the characteristics of PPG signals and contribute to enhancing its accuracy and applicability in medical monitoring.