Reconstruction of Central Aortic Pressure Waveform Using Adaptive Multi-Channel Identification

This paper presents an adaptive multi-channel (AMC) approach for the reconstruction of the central aortic blood pressure (BP) waveform from multiple peripheral BP measurements. In contrast to most of the previously developed single-channel methods for estimating central aortic BP waveform, the key merit of the AMC algorithm is its ability to be individualized without any prior model training or parameter tuning. Preliminary experimental evaluation of the AMC algorithm with respect to a single channel approach in a swine model shows 17.4% improvement in the waveform reconstruction accuracy in terms of RMSE (from 3.3 mmHg for the single-channel approach to 2.8 mmHg for the AMC algorithm) and the superior prediction of several key central aortic BP waveform features (from 0.1512 to 0.9109 for the ejection duration, in terms of r² values). These results suggest that the AMC algorithm can render a very accurate reconstruction of the central aortic BP waveform from two peripheral BP measurements without any prior model training or parameter tuning