The effectiveness of adaptive training for stress inoculation in a simulated astronaut task

Astronauts operate in an environment with multiple hazards that can develop into life-threatening emergency situations. Managing stress emergencies may require cognitive resources and lead to diminishing performance. Stress training aims maintain performance under by methodically increasing stressor levels build inoculation against stress. An adaptive virtual reality (VR) system was developed real-time detection using machine learning on psychophysiological responses. Using a VR simulation of spaceflight fire, classifications were used trigger adaptations the environmental stressors (e.g., smoke, alarms, flashing lights), goal maintaining manageable level during training. Fifty-seven healthy subjects underwent task over eight trials (adaptive, n=19); results compared predetermined gradual increases (graduated, n=18), constant low-level (skill-only, n=20). responses measured through heart rate, rate variability (i.e., root mean squared successive differences (RMSSD), low frequency high (LF/HF) ratio), (distance-from-fire). Heart decreased RMSSD increased pre-post for all experimental conditions. The LF/HF ratio condition, but not other Results suggests conditions had lower stress, condition more successful. Task showed marginal increase across condition. Preliminary suggest prepare individuals responding better than skill-only graduated