Comparison of Two Algorithms Analysing the Intracranial Pressure Curve in Terms of the Accuracy of Their Start-Point Detection and Resistance to Artefacts

Acta Neurochir Suppl. 2021:131:243-248. doi: 10.1007/978-3-030-59436-7_46.

Abstract

Objectives: For further insight into the possibly predictive quality of the intracranial pressure (ICP) waveform morphology a definite and reliable identification of its components is a prerequisite but presents the problem of artefacts in physiological signals.

Methods: ICP and electrocardiogram (ECG) data were recorded to depict not only their numerical value but also their respective waveforms and were analysed by two algorithms, which were then compared for their artefact resistance.The algorithms in question identify the start point of every ICP wave, one (AR[SA]) by scale analysis, the other (AR[ECG]) by analysing the ICP wave linked to the ECG.

Results: Start-point identification accuracy in rhythmic patients showed sensitivity of 95.14% for AR[SA] and 99.99% for AR[ECG], with a positive predictive value (ppv) of 98.30% for AR[SA] and 99.76% for AR[ECG].In arrhythmic patients sensitivity was 98.05% for AR[SA] and 99.73% for AR[ECG], with a ppv of 100% for AR[SA] and 99.78% for AR[ECG].

Conclusions: AR[ECG] has proven to be more resistant to artefacts than AR[SA], even in cases such as cardiac arrhythmia. It facilitates reliable, three-dimensional visualisation of long-term changes in ICP-wave morphology and is thus suited for analysis in cases of more complex or irregular vital parameters.

Keywords: Algorithm design; Artefact resistance; Intracranial pressure waveform analysis; Peak detection; Start-point detection; Wavelet transform.

MeSH terms

  • Algorithms
  • Artifacts*
  • Electrocardiography
  • Humans
  • Intracranial Pressure*
  • Signal Processing, Computer-Assisted