Feature generation and statistical analysis of physiological responses to nerve-agent exposure. Final report, January 1984-January 1985

This report describes research carried out on the problem of automated information extraction from multichannel physiological signals. Analog tapes were obtained of data recorded during an animal study of the ventilatory requirements after nerve-agent exposure. The main objective of the research was to determine if the physiological signals contained information relevant to the state of the subject. It was also of great interest to determine if the feature condition of the subject could be predicted from the data obtained at a point shortly after initial contact with the agent. This study also included development of techniques for compression and processing of large volumes of data. This report covers data digitalization, feature generation, and data analysis. Because of the limited numbers of tapes available with adequate quality signals, only three experiments were analyzed in depth. Feature generation and analysis algorithms were developed and used on data from these experiments. Statistical and physiological feature evaluations of the data revealed that information from several signals, when displayed as biplots, permitted differentiation among the carrying states of the animals. Results of this contract indicate that these techniques may be useful in developing algorithms which predict the consequences for casualties based on data obtained shortlymore » after nerve agent exposure.« less