In the United States 12.5% of babies are born prematurely. The premature birth rate has increased 27% over the past 20 years. The three principal components of the preterm-birth syndrome are: premature rupture of membranes, myometrial activity causing preterm labor, and premature cervical ripening. We investigated myometrial activities which are noninvasively collected multi-channel magnetomyographic (MMG) data from patients' uterus. On the temporal data mining side, we have developed an advanced signal processing technique to further investigate this data beyond Fourier and Wavelet analysis. On the spatial data mining side, we have developed a method to further explore the interactions between different locations of the uterine muscle. The key part of the latter method is combining synchronization theory (applicable to multi-channel data) with an advanced signal processing technique for nonlinear and non-stationary data analysis. Since it is adaptive it has wide application area on non-stationary and nonlinear signals with multi-channel recordings.