Directional Sensors for Recording Uterine EMG During Pregnancy

Abstract

Multichannel uterine electromyography (uEMG) during pregnancy is traditionally performed with electrocardiography (ECG) sensors. Similar signals are often observed in two or more channels, suggesting the ECG sensors report activities originating from the same location on the uterus. To improve signal source localization, we designed a directional sensor or "Area Sensor". Here we compare Area Sensors with ECG sensors for source localization. Subjects were ≥ 38 wks experiencing regular contractions. 6 Area Sensors (n = 8) or 6 to 7 ECG sensors (n = 7) were used to record multichannel uEMG for 60 min. For each sensor type, the similarity of signals observed in pairs of channels during contractions was assessed by quantifying channel crosstalk. Since crosstalk depends on the separation between sensors, analyses were performed within distance groups: A 9-12 cm; B 13-16 cm; C 17-20 cm; D 21-24 cm; E ≥ 25 cm. For ECG sensors, crosstalk was 67.9 ± 14.4% in group A, decreasing to 27.8 ± 17.5% in group E. For Area Sensors, crosstalk was 24.6 ± 18.6% in Group A, decreasing to 12.5 ± 13.8% in group E. Area Sensors showed less crosstalk than ECG sensors in distance groups A, B, C and D, with all p < 0.002. Compared with ECG sensors, Area Sensors are more directional and report uterine activity from a smaller area of the uterine wall. Using 6 Area Sensors separated by at least 17 cm provides acceptably independent multichannel recording. This introduces the possibility of non-invasively evaluating uterine synchronization and the strength of individual uterine contractions in real time.

Keywords: Directional sensors; Local uterine contractions; Pregnancy; Uterine electromyography.