Objective pain assessment using wearable sensors

Pain is an unpleasant sensation and emotional experience linked to tissue damage. It's purpose is to allow the body to react and prevent further tissue damage. We feel pain when a signal is sent through nerve fibers to the brain for interpretation. The experience of pain is different for everyone, and there are different ways of feeling and describing pain. This can make it difficult to define and treat. In this study, a computational solution for objective assessment of pain using vital signs was offered. This thesis evaluates the performances of computational methods that take the sequence of vital signs acquired until pain onset as input and report the predicted existence of pain. In this study, 129,267 number of data related to pain were obtained by using MIMIC-III database. 76,310 of the data are marked as pain. Existance of pain estimation was made using this dataset. The highest AUROC score obtained was 0.711 using the Random Forest algorithm. This thesis argue that the use of computational intelligence methods can promote computer-aided monitoring of pain in hospitalized environment to a certain degree.