Which brainstem cells generate the respiration cycles?

We make a major step towards answering the question posed in the title, using as model the mouse foetus in its 17-19 embryonic days. We use (a) 2-photon microscopy to image the brainstem cell activity ([Ca2+]) in the pre-Boetzinger complex, and (b) electrical recordings from the phrenic nerve, which indicate the diaphragm contraction during inspiration. We classify the brainstem regions (individual cells or groups of cells) into `active´ and `inactive´, based on whether they contribute or not to the individual electrical signal peaks. As features, we use the Continuous Wavelet Transform-based Semblance responses, for comparing non-periodic and/or periodic-like signals. We use our novel Generative Mixture Model (GMM) possibilistic clustering to obtain the desired classes robustly. This way, we model the inspiration control as a physiological process, which is a crucial step towards understanding how the living brain controls breathing.