BIC

We developed data-driven models capable of predicting body core temperature during vigorous physical activity in hot and humid environments. Recent advances in the ability to monitor physiology variables have resulted from the development of new biosensors and information processing capabilities. These capabilities have a direct impact on how closely a person's state can be monitored during civilian activities or during military operations, including the possibility of predicting changes in many vital physiological variables, such as body core temperature, heart and respiratory rates and even such subtleties as level of alertness and performance. The technological breakthroughs in the development of hardware and firmware were also accompanied by an equally profound and significant progress in fields such as data mining and machine learning. The new technology to collect and store relatively large amounts of physiological data in the field allows researchers to explore new opportunities in data-driven methods to forecast physiological variables and status. The models that we are currently using are from the autoregressive integrated moving average (ARIMA) family, which allow us to handle non-stationary signals. An example of application of an ARIMA model to predict future core temperatures is shown in the figure, in which the actual measured core temperature (blue line) and the temperatures that were predicted 20 to 60 minutes beforehand (red, black and blue lines) are displayed.

The power of the purely data-driven approach for near-term predictions comes from the nature of the core temperature signal and the thermal inertia of the human body thermoregulatory process. The low-frequency and smooth nature of the signal lends itself perfectly to AR modeling and predictions, which together with the variability constraints imposed by regularization force the model to produce core-temperature outputs with low variation and excellent predictive capabilities. The relatively large inertia (or time constant) of the body thermoregulatory process is what allows the AR model to make accurate predictions minutes ahead. The thermal inertia, characterized by the specific heat capacity of the human body, regulates and precludes rapid changes in core temperature. This can be explained, for example, by the fact that a significant percentage of the human body (up to 75%) is composed of water and that water has one of the largest specific heat capacity of all substances. This large specific heat capacity allows the human body to absorb a significant amount of energy before its temperature rises, thus permitting accurate short-term predictions.

Our future efforts will leverage our research using ARIMA models and the insight of the practical importance of physiological inertia to refine models to predict the future value of key physiologic parameters, such as body core temperature and blood glucose levels, and to develop algorithms to apply of confidence values to various physiological vital signs measurements.