Provided herein are methods to predict pain sensitivity and pain intensity to prolonged pain in a subject. Electroencephalograms are recorded in a pain-free state or, alternatively, in a pain-free state and after applying a prolonged pain stimulus in a prolonged pain state. Pain-free and prolonged pain peak alpha frequencies or ΔPAF are measured. These values correlate negatively with the likelihood of increased pain sensitivity and increased pain intensity. Also provided is a method for predicting a likelihood of chronic pain in a subject after a medical procedure and designing a plan to treat the chronic pain.