Abstract
Monitoring arterial blood pressure (ABP) or other cardiovascular signals is routine in intensive care units, for the assessment of deterioration in syndromes such as shock. However, ABP has limited diagnostic value (as macroscopic changes occur late in the disease) and whilst heart rate variability analysis has clinical utility, it ignores most of the information contained within the shape of the waveform and usually requires post hoc processing, a common source of bias.
We have developed a novel non-linear mathematical method, based on Takens’ Embedding Theorem to quantify previously undetectable changes in periodic waveforms. Briefly, the input ABP signal is plotted in phase space by introducing two fixed time delays, generating x, y and z coordinates (Figure 1a) which are plotted three dimensionally as a bounded ‘attractor’ (1b). The 3D image (1c) is then rotated and visualised down a line where x=y=z (v,w plane), simultaneously generating a 2D attractor and minimising the effect of baseline variation in the input signal (1d). Colour is then added to represent density for overlapping points (1e), and scalar measures of the attractor can be extracted for further analysis. The method is uniquely sensitive to changes in waveform shape and has been exemplified with a range of human and rodent waveforms (Figs 1&2). Results from a murine model of septic shock show that the attractor reconstruction method detects changes much earlier than macroscopic examination of ABP (Fig 3).
We hypothesize that attractor reconstruction of ABP provides an early, non-biased diagnostic signal in clinical syndromes e.g. shock.
We have developed a novel non-linear mathematical method, based on Takens’ Embedding Theorem to quantify previously undetectable changes in periodic waveforms. Briefly, the input ABP signal is plotted in phase space by introducing two fixed time delays, generating x, y and z coordinates (Figure 1a) which are plotted three dimensionally as a bounded ‘attractor’ (1b). The 3D image (1c) is then rotated and visualised down a line where x=y=z (v,w plane), simultaneously generating a 2D attractor and minimising the effect of baseline variation in the input signal (1d). Colour is then added to represent density for overlapping points (1e), and scalar measures of the attractor can be extracted for further analysis. The method is uniquely sensitive to changes in waveform shape and has been exemplified with a range of human and rodent waveforms (Figs 1&2). Results from a murine model of septic shock show that the attractor reconstruction method detects changes much earlier than macroscopic examination of ABP (Fig 3).
We hypothesize that attractor reconstruction of ABP provides an early, non-biased diagnostic signal in clinical syndromes e.g. shock.
| Original language | English |
|---|---|
| Title of host publication | Congress of the European Shock Society in conjunction with: |
| Subtitle of host publication | 14th International Conference on Complex Acute Illness |
| Publication status | Published - Sept 2015 |
| Event | International Conference on Complex Acute Illness - Cologne, Germany Duration: 25 Sept 2015 → 27 Sept 2015 |
Conference
| Conference | International Conference on Complex Acute Illness |
|---|---|
| Country/Territory | Germany |
| City | Cologne |
| Period | 25/09/2015 → 27/09/2015 |
