Data collection, analysis, and statistics

Data collection was performed using commercially available hard- and software (hardware: MicroLabBox, dSPACE GmbH, Paderborn, Germany; software: ControlDesk 7.1, dSPACE GmbH, Paderborn, Germany; MATLAB Version R2022b, MathWorks, Natick, USA). All data from the ICU-ventilator was collected using custom-built software developed at the Chair for Medical Information Technology (MedIT, Faculty of Electrical Engineering and Information Technology, RWTH Aachen University, Germany) as published in 2020³⁰.

Data analysis was performed using commercially available software packages (Microsoft Excel, Microsoft 365 MSO, Version 2112, Redmond, USA; MATLAB Version R2022b, MathWorks, Natick, USA). From the second animal onwards, tidal volumes were quantified by area-under-curve calculation using flow-curve data in MATLAB and aggregated as mean per hour.

Statistics were performed using GraphPad Prism 10 (Version 10.0.3, GraphPad Software Inc., San Diego, USA) and SPSS version 28 (IBM Corp., Armonk, NY). Paired t-test was used for parametric values. Simple linear regression was performed to calculate the slope of tidal volume change over experimental time. When appropriate, data is presented as mean ± SD, median (range), absolute numbers or for the slope, as value with 95% confidence interval [95%-CI]. All statistical tests are two-tailed. Significance was defined as p < 0.05. Since this was an exploratory study and thus the effect size was unknown, no a priori power analysis could be performed. However, based on previous investigations of phrenic nerve stimulation in pigs^31,32, it was expected that feasibility could be shown reliably with a respective group size of n = 6 animals.

Figure 1 was created using BioRender.com. Figures 2, ​,3,3, ​,44 and Supplementary Figs. ^S1 and ^S2 were compiled and annotated using Photoshop CS6 (Version: 13.0). Shotcut Video Editor (Shotcut, Version 23.11.29, Meltytech, Oceanside, USA) was used to compile the supplementary videos. Sonograms and B-mode video-loops were captured using the portable ultrasound device with a linear probe (Sonosite Edge II, HFL50x linear probe, 15–6 MHz, FUJIFILM SonoSite Inc., Bothell, USA).

Comparison of mechanical ventilated and phrenic nerve stimulated breaths. (a) Flow, airway pressure, stimulation voltage and left-side diaphragmatic sonography are plotted over a 10 s period (Pig6, BW: 49 kg; target tidal volume set: 294 ml [6 ml/kg BW]). Phrenic nerve stimulated breaths (PNS, white background) were solely induced by bilateral electric stimulation of the phrenic nerves (here 9.5 V, see voltage curve). This resulted in the contraction of the diaphragm with a 2.2-fold inspiratory-thickening [M-mode diaphragm: 11 mm (end-inspiratory diameter, d_I); 5 mm (end-expiratory diameter, d_E)], which led to a decrease in airway pressure (Paw) and consecutive inspiratory flow. Mechanical ventilated breaths (MV; orange background) were applied using pressure-controlled ventilator mode. The exact tidal volume is indicated for each PNS (Tv_s) and MV breath (Tv_m). (b) The ratio between end-inspiratory and end-expiratory diaphragm thickness (d_I/d_E-ratio) was quantified bilaterally at two time points by two investigators independently. The baseline (dashed grey line) represents the respective end-expiratory diaphragmatic thickness (100%). Data from each measurement is plotted individually, mean ± SD is indicated (n = 4⁺). (c) Over six hours, the mean values for all MV and PNS breaths per hour (normalized to the body weight of the respective pig) were plotted separately (n = 5). The horizontal continuous lines illustrate the respective trend over time. The dashed grey lines indicate the target corridor for lung-protective ventilation (4-8 ml/kg BW). (d) The averaged intratracheal pressure (Ptrach)-volume curves were calculated for PNS and MV (100 consecutive breaths per ventilation mode) separately for each pig (n = 5). ⁺systematic sonographic quantification not performed for the second pig (Pig2), *p < 0.001. Ex expiration, In inspiration, Paw airway pressure, Ptrach intratracheal pressure, Tv_s phrenic nerve stimulated tidal volume, Tv_m mechanical ventilated tidal volume.

EIT-based evaluation of caudal ventilation in PNS and MV. (a) Shows a qualitative representation of inspiration for MV (top, orange background) and PNS (bottom, white background) in relation to the chest cross-section. The stronger the intensity of the EIT signal, the brighter the blue color. Regions of interest (ROI) analyzed in (b) are shown as numbered rectangles, where region 1 is ventral and region 4 is caudal. (b) Depicts an exemplary EIT quantification for six breaths. The EIT impedance intensity (y-axis) is displayed over time (x-axis). In addition, to the global EIT-signal depicted in the top graph, a differentiation and quantification of the regional EIT-signal in four layers (ROIs) from ventral to dorsal was performed. A breath-by-breath comparison illustrates differences in regional EIT-signal between MV (orange background) and PNS (white background) breaths. (c) Shows the comparison of the Center of Ventilation (CoV) for MV and PNS, where higher y-values indicate a more dorsal location of the CoV. The CoV was calculated separately for 300 PNS and 300 MV breaths, which alternated breath-by-breath. Ventral is at the top in all figures/ROIs. *p < 0.001. CoV center of ventilation, EIT electrical impedance tomography, MV mechanical ventilation, PNS phrenic nerve stimulation, ROI Region of interest.