Detection of Pain Caused by a Thermal Stimulus Using EEG and Machine Learning.

Pain is a subjective feeling difficult to explain since it depends on the history, sexuality, cultural and social context of each person, and worsens its communication in cases of people with certain motor or language diseases. In this research the pain response of 6 healthy participants is studied using an electroencephalographic device. To have a first approach to the problem and a clearer understanding of the pain response, a non-traumatic thermal stimuli is used, which simulates phasic pain that is known for being gradual. The use of machine learning is important to find a discrimination model among 3 conditions proposed (pre-stimulus, beginning of the stimuli and intolerance to the stimuli) and the EEG data recorded. Different machine learning models found a solid distinction between the 3 different conditions in the case of the 4 males. Females could not be classified since the stimulus does not generate any kind of pain to them.