In this thesis, numerical and statistical time series analysis techniques have been applied for the purpose of objective quantitative analysis of fetal heart rate (FHR) recording from the antepartum period (period of pregnancy before labour). Currently FHR is routinely analyzed visually and the form of medical interpretations that medics make are simply binary. Medical literature on fetal monitoring, based on FHR, has been extensively searched and the current status of FHR and its components of interest to the medics for screening purposes have been presented coherently. Ambiguities and false expectations have been clarified. All the FHR components of interest used by the medics during the antepartum period have been addressed and analyzed numerically. Prior to any successful analysis of non-stationary FHR, its baseline or trend (an important screening variable by itself) had to be estimated and removed from the original data in a statistically unbiased manner. This has been achieved via a first order bi-directional autoregressive filtering technique. Non-stationarities in both first and second moments had to be taken into account. The detrimental effects of the baseline when analyzing FHR variability through one-dimensional scalar statistic namely the standard deviation have been studied and the need to use detrended data for this purpose has been emphasized. A new non-parametric approach to analyze the development of the FHR variability based on relative frequency histogram analysis has also been proposed. A rule based routine has been devised which makes use of the rate of change of the unbiased estimated baseline and the original antepartum recorded FHR to numerically detect FHR accelerations. The accelerations are a very important screening component of FHR which are usually scanned for visually. This visual scanning technique is inefficient and inaccurate.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:661786 |
| Date | January 1992 |
| Creators | Shariati, M. A. |
| Publisher | University of Edinburgh |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Source | http://hdl.handle.net/1842/14389 |
Page generated in 0.0025 seconds