Observer-Based Deconvolution of Deterministic Input in Coprime Multichannel Systems With Its Application to Noninvasive Central Blood Pressure Monitoring

Aortic pressure waveform reconstruction using a multi-channel Newton blind system identification algorithm

BACKGROUND

Central aortic pressure (CAP) as the major load on the left heart is of great importance in the diagnosis of cardiovascular disease. Studies have pointed out that CAP has a higher predictive value for cardiovascular disease than peripheral artery pressure (PAP) measured by means of traditional sphygmomanometry. However, direct measurement of the CAP waveform is invasive and expensive, so there remains a need for a reliable and well validated non-invasive approach.

METHODS

In this study, a multi-channel Newton (MCN) blind system identification algorithm was employed to noninvasively reconstruct the CAP waveform from two PAP waveforms. In simulation experiments, CAP waveforms were recorded in a previous study, on 25 patients and the PAP waveforms (radial and femoral artery pressure) were generated by FIR models. To analyse the noise-tolerance of the MCN method, variable amounts of noise were added to the peripheral signals, to give a range of signal-to-noise ratios. In animal experiments, central aortic, brachial and femoral pressure waveforms were simultaneously recorded from 2 Sprague-Dawley rats. The performance of the proposed MCN algorithm was compared with the previously reported cross-relation and canonical correlation analysis methods.

RESULTS

The results showed that the root mean square error of the measured and reconstructed CAP waveforms and less noise-sensitive using the MCN algorithm was smaller than those of the cross-relation and canonical correlation analysis approaches.

CONCLUSION

The MCN method can be exploited to reconstruct the CAP waveform. Reliable estimation of the CAP waveform from non-invasive measurements may aid in early diagnosis of cardiovascular disease.