Abstract
<jats:p>The formation of inert gas bubbles during decompression can lead to decompression sickness (DCS), a major operational risk for divers, compressed-gas workers, astronauts, and high-altitude aviators. In diving, DCS risk is typically inferred from post-dive ultrasound detection of venous gas emboli (VGE), precluding modification of decompression schedules based on real-time physiological feedback. Two-dimensional ultrasound imaging could provide additional insight into decompression-related physiological changes; however, its use in hyperbaric environments has been largely precluded by fire risk associated with elevated oxygen partial pressures (ppO₂) in enclosed spaces. Here, we developed a workflow for operating a programmable ultrasound system under hyperbaric conditions and acquiring ultrasound data from the subclavian vein and calf muscle during decompression. A total of 42 dives were conducted by 26 individuals using a previously characterized dive profile to 132 feet seawater (FSW) for 20 min with 9 min of decompression. Three exposure conditions were evaluated: non-exercising, exercising, and a brief pause at 20 FSW during compression. Twelve dives included programmable ultrasound imaging during decompression. Post-dive VGE responses were consistent with prior reports while demonstrating substantial inter-individual variability and sensitivity to modest profile modifications. VGE were detected in the subclavian vein during decompression in two participants and subsequently confirmed by post-dive echocardiography. Calf muscle ultrasound brightness typically increased from pre-dive to decompression measurements, before decreasing below baseline in the 120 min post dive measurement period. These findings demonstrate the feasibility of programmable ultrasound imaging during human decompression and establish a practical framework for ultrasound operation under hyperbaric conditions. This approach may support future physiological studies and development of automated decompression monitoring technologies.</jats:p>