Information theoretic measures of perinatal cardiotocography synchronization

Area of Science:

• Perinatal medicine
• Biomedical signal processing
• Information theory

Background:

• Fetal hypoxia is a critical concern during labor.
• Synchronization between uterine pressure (UP) and fetal heart rate (FHR) can indicate fetal well-being.
• Accurate detection of fetal compromise is essential for timely intervention.

Purpose of the Study:

• To evaluate the efficacy of bivariate mutual information (MI) and transfer entropy (TE) in assessing UP-FHR synchronization.
• To determine if these methods can predict fetal hypoxia and differentiate normal from pathological fetal states.
• To investigate the temporal dynamics of information transfer between UP and FHR signals.

Main Methods:

• Utilized nearest-neighbor based Kraskov entropy estimation for non-Gaussian UP and FHR signals.
• Employed surrogate data testing to ensure robustness of estimates against noise.
• Applied mutual information and transfer entropy to quantify UP-FHR synchronicity and information transfer delays.

Main Results:

• Mutual information (MI) distinguished normal from pathological fetuses 160 minutes before delivery.
• MI further differentiated normal from metabolic acidotic fetuses 110 minutes before delivery, with higher MI in pathological cases.
• Transfer entropy (TE) identified normal from pathological cases 110 minutes before delivery, showing lower TE values and longer delays in pathological cases.

Conclusions:

• Bivariate mutual information and transfer entropy are effective tools for analyzing perinatal UP-FHR dynamics.
• These information-theoretic measures can serve as early biomarkers for fetal hypoxia.
• The findings represent the first report of transfer entropy's utility in detecting fetal compromise via UP-FHR analysis.