A node-wise analysis of the uterine muscle networks for pregnancy monitoring

TL;DR

This paper introduces a novel network-based analysis of multichannel electrohysterographic signals to monitor pregnancy progression, revealing significant changes in uterine muscle connectivity as labor approaches.

Contribution

The study presents a new node-wise connectivity analysis method using graph theory on EHG signals for pregnancy monitoring, which is a novel application in this context.

Findings

Node strength increases from pregnancy to labor

Electrodes on the median vertical axis are most discriminant

Connectivity measures evolve significantly over pregnancy weeks

Abstract

The recent past years have seen a noticeable increase of interest in the correlation analysis of electrohysterographic (EHG) signals in the perspective of improving the pregnancy monitoring. Here we propose a new approach based on the functional connectivity between multichannel (4x4 matrix) EHG signals recorded from the women abdomen. The proposed pipeline includes i) the computation of the statistical couplings between the multichannel EHG signals, ii) the characterization of the connectivity matrices, computed by using the imaginary part of the coherence, based on the graph-theory analysis and iii) the use of these measures for pregnancy monitoring. The method was evaluated on a dataset of EHGs, in order to track the correlation between EHGs collected by each electrode of the matrix (called node-wise analysis) and follow their evolution along weeks before labor. Results showed that the strength of each node significantly increases from pregnancy to labor. Electrodes located on the median vertical axis of the uterus seemed to be the more discriminant. We speculate that the network-based analysis can be a very promising tool to improve pregnancy monitoring.