Fetal heart rate deceleration detection using a discrete cosine transform implementation of singular spectrum...

Area of Science:

• Biomedical Engineering
• Signal Processing
• Maternal-Fetal Medicine

Background:

• Fetal heart rate (FHR) monitoring is crucial for assessing fetal well-being.
• Accurate detection of deceleration events in FHR is essential for timely intervention.
• Standard change-point detection algorithms may struggle with the unique characteristics of FHR signals, such as successive decelerations.

Purpose of the Study:

• To develop and evaluate a novel singular-spectrum analysis (SSA) based algorithm for improved detection of deceleration events in FHR monitoring.
• To adapt SSA for the specific challenges of FHR analysis, where decelerations can occur in close succession.

Main Methods:

• A modified SSA algorithm, termed "base-hold SSA", was developed using the discrete cosine transform (DCT) for signal decomposition.
• The base-hold SSA approach was adapted to maintain a constant reference model during successive similar events, a common occurrence in FHR decelerations.
• The algorithm was tested on a database of 15 FHR tracings and compared against expert obstetrician annotations.

Main Results:

• The base-hold SSA algorithm demonstrated improved performance in detecting deceleration events compared to standard SSA.
• The number of missed decelerations was reduced from 64 to 49 (a 21.9% improvement).
• The reduction in false-positive decelerations was maintained at the same level as standard SSA.

Conclusions:

• The standard SSA assumption of infrequent changes is not suitable for FHR analysis due to the potential for successive decelerations.
• The proposed base-hold SSA modification effectively addresses the challenge of detecting series of closely occurring deceleration events in FHR.
• This enhanced algorithm offers a promising tool for improving the accuracy and reliability of FHR monitoring.