Camera PPG waveforms at the forehead

In order to obtain insights into the feasibility of replacing ECG-guided triggering in magnetic resonance imaging (MRI) by a system based on video photoplethysmography (PPG), PPG and ECG data were collected from volunteers in an MRI scanner. PPG waveforms obtained using remote camera PPG directed at the forehead are studied in qualitative and quantitative sense over a number of volunteers. The data analysis considers variations in PPG waveforms across volunteers, modelling of the waveforms in Fourier series, dependencies of waveforms and features on the interbeat interval (IBI) and breath-holding, and models for ECG-blind estimation of R-peak position. The main findings are that the PPG waveform depends on the volunteer and that its shape changes with IBI and does not depend on breath-holding in the given scenario. Low-order harmonic models provide accurate approximations to the PPG waveform, where for higher IBI the waveform shows more temporal details. Accurate predictions (20 ms std) of the delays between markers in ECG and PPG appear feasible from a single PPG feature.