| Abstrakt: |
BACKGROUND: The applications for AI regarding aerospace medicine are broad and have the potential to dramatically improve safety by recognizing hazards and preconditions that may predispose aviators to mishaps. How may aeromedical researchers leverage this emerging technology to combat the most common human pre-conditions that contribute to aviation mishaps? OVERVIEW: Upon review of the cited human pre-conditions involved in naval aviation mishaps between 2012 to 2022, some of the most common contributing pre-conditions involved were fatigue, spatial disorientation, visual illusions, hypoxia and hyperventilation. Recently, both fatigue and spatial disorientation have been evaluated in simulated aviation environments with semi-dry and dry EEG systems. Lee et al. (2023), utilized a semi-dry 30 channel head-set, during simulated flight and were able to identify abnormal mental states, such as fatigue, high workload and distraction, finding 9 EEG indices that were significantly different with varied flight tasks. Likewise, Geva, et al. (2023), utilized a 32 channel dry-eeg, during Barani chair induced spatial disorientation and noted a 52% reduction in theta power complemented by nystagmus in 72% of the trials. What was lacking in both studies, as demonstrated by relatively low accuracy was a method of cleaning their EEG data and reducing the variance within specific bandwidths. Realizing the need for real-time sensors on cognitive performance in military aviation, Rice (2019) evaluated dryelectroencephalogram (EEG) technology ability to detect hypoxia. Their research suggested that a reduction in overall dry-EEG power could identify hypoxia in lieu of aviators not recognizing their own meaningful decreases in oxygen saturation and cognitive performance. Linnville (2021) and Snider (2022) respectively advanced this work further by reducing the variance of the data sets through principal component analysis (PCA) and then applying three common AI algorithms. By doing so, these researchers increased the sensitivity and specificity of dry-eeg technology to detect hypoxia to greater than 97%. DISCUSSION: This presentation will provide a framework for future researchers to investigate and mitigate the most commonly associated preconditions for aviation mishaps with AI. The ultimate goal being to provide the aviator a useful real-time, helmet embedded sensor to prevent mishaps in our next generation aircraft. Learning Objectives 1. The participant will be able to identify the common pre-conditions associated with aviation mishaps that may potentially be identifiable through the acquisition of data obtained through multi-dimensional biosensors such as electroencephalogram (EEG) and electrocardiogram (EEG) and subsequently analyzed utilizing artificial intelligence algorithms. 2. The participant will understand broadly common computer programs and artificial intelligence algorithms to clean and produce models that accurately predict and identify common human preconditions that are associated with aircraft mishaps. [ABSTRACT FROM AUTHOR] |
