Assessing Velocity and Directionality of Uterine Electrical Activity for Preterm Birth Prediction Using EHG Surface...

Area of Science:

• Biomedical Engineering
• Obstetrics and Gynecology
• Signal Processing

Background:

• Distinguishing between preterm and term labor is crucial for neonatal outcomes.
• Electrohysterography (EHG) measures uterine electrical activity, offering potential for labor assessment.
• Current EHG analysis methods may not fully capture subtle differences between preterm and term pregnancies.

Purpose of the Study:

• To evaluate the utility of electrohysterogram (EHG) wave conduction velocity and direction for differentiating preterm from term EHG recordings.
• To identify novel EHG signal features that improve classification accuracy for preterm versus term labor.

Main Methods:

• Utilized the Term-Preterm EHG DataSet with Tocogram (TPEHGT DS).
• Applied short-time cross-correlation to bipolar EHG signals to estimate conduction velocities and directions.
• Analyzed EHG data in frequency bands below and above 1.0 Hz, corresponding to uterine contractions and maternal heart rate influences.

Main Results:

• No consistent direction of EHG wave propagation was observed during contraction or non-contraction intervals.
• Preterm non-contraction (dummy) intervals above 1.0 Hz showed a significantly lower percentage of vertical velocity vectors and a higher percentage of horizontal velocity vectors.
• A combination of velocity vector directions and sample entropy in the vertical direction achieved 86.8% classification accuracy (AUC=90.3%) for distinguishing preterm from term EHG records.

Conclusions:

• Conduction velocity direction and signal entropy in specific frequency bands are promising biomarkers for differentiating preterm from term EHG signals.
• These novel features, particularly from non-contraction intervals above 1.0 Hz, enhance the diagnostic capability of EHG for preterm labor prediction.