We have liftoff: A discovery study to use artificial intelligence to identify adaptative profiles for future space missions

Long-duration space missions will challenge astronauts' adaptive capacities. Interoception and heart rate variability (HRV), reflecting parasympathetic activity, are increasingly recognized as predictors of adaptation and health. This study investigated whether artificial intelligence may predict adaptation profiles from interoceptive and HRV responses accross different space analogs. Data were collected from 84 participants in four environments: parabolic flight, nuclear submarine, sea survival simulation, and chemical/biological hazard exercises. Interoceptive sensitivity and HRV were measured to identify adaptation profiles using clustering. Baseline data were then used to train a support vector machine (SVM) to predict these profiles. Three adaptation profiles emerged, differing in interoceptive awareness, body-mind integration, and neuroception. The SVM model predicted these profiles with 79% accuracy. These findings demonstrate the feasibility of using machine learning to anticipate adaptation outcomes based on physiological and interoceptive markers. They emphasize the embodied nature of adaptation and the relevance of interoceptive pathways in HRV dynamics. This work provides new directions for optimizing astronaut training by tailoring preparation to individual physiological profiles. Tomorrow is here, we are ready for take-off.