Functional Magnetic Resonance Imaging and Computers Help Researchers Determine When A Patient Feels Pain.

Pain is a subjective experience—people feel pain differently and have different abilities to tolerate pain. Historically, doctors have relied on reports from patients to determine whether they are in pain and how severe the pain is, because there have been no other means of assessing a patient’s pain. This method may be adequate for most people, but it cannot be used for patients who have a limited ability to communicate as a result of dementia or because illness or injury has robbed them of the ability to speak, move, or make a facial expression indicating they are in pain.

To help resolve this problem, researchers in Stanford’s Pain Medicien Division have been working to find a way to reliably assess pain without depending on patient self-reports.

Modern brain scanning—“neuroimaging”—techniques such as magnetic resonance imaging (MRI) are increasingly providing a window on how the brain responds to pain. Stanford’s pain researchers designed a study combining functional magnetic resonance imaging (fMRI) with a support vector machine (SVM) to see if they can use brain activity to determine when a person is experiencing pain and to see which areas of the brain are involved.

Functional MRIs track the brain’s response to various stimuli in real time. The SVM is a computerized device that uses artificial intelligence called a machine-learning algorithm to decipher the data provided by the fMRI. A machine-learning algorithm is a computer program that enables a computer to read, interpret, and classify data without the continual help of a computer programmer, analyst, or researcher.

The study looked at the brain’s response to pain in 24 people. Each person received several painful heat stimuli on the inside of one forearm and hot but not painful stimuli on the other forearm. The fMRI recorded the brain’s response to each, and the SVM interpreted the data. Eight of the 24 people were used to train the algorithm.  The algorithm was then tested on the other 16 subjects.

The study found that the combination of the fMRI and SVM accurately distinguished between the painful and nonpainful stimuli more than 80% of the time. It also showed which areas of the brain were most active when exposed to painful stimuli.

Although this is a small study, researchers are hopeful that it may lead to an objective means of assessing a patient’s pain even when he or she is unable to communicate. This will enable doctors to better treat these patients. In addition, by looking at the whole brain’s response to pain, the combination of neuroimaging and computerized machine-learning techniques will help researchers understand how the brain perceives and responds to pain and to various therapies used to treat pain. This will help lead to more effective treatments for the 100 million Americans suffering from pain.

Brown JE, Chatterjee N, Younger J, Mackey S. Towards a physiology-based measure of pain: patterns of human brain activity distinguish painful from nonpainful thermal stimulation. PLoS One 2011: 6(9): e24124.