Many doctors rely on patients to indicate levels of pain and discomfort, and while these self-reports are certainly subjective, they can still provide physicians with valuable clinical information. But what happens when patients are unable to talk with their doctors, like in cases of severe cognitive or communicative impairment?
For years, researchers have searched for a way to measure pain by physiological assessment alone. Now, advances in neuroimaging have allowed researchers to identify a promising way to do just that, allowing them to detect pain in patients without requiring them to communicate whatsoever.
A team of researchers led by Sean Mackey, chief of Stanford School of Medicine's Pain Management Division, developed the technique by training a computer to recognize patterns of human brain activity that emerge when the person is in pain.
Scientists used functional magnetic resonance imaging (fMRI) to identify how numerous regions of the brain interact in response to both painful and non-painful heat-based stimuli, then trained a sophisticated computer algorithm to recognize these brain interactions on its own.
Incredibly, the computer was able to predict whether a test subject was experiencing pain 81% of the time. But what makes the research so groundbreaking is that the people used to train the computer algorithm how to classify pain were not the same people that the computer model was ultimately tested on.
Here's why that's significant: In 2010, a similar study, led by neuroscientist Andre Marquand, trained a computer algorithm using fMRI data for individual test participants. The trained computer models were then used to accurately predict levels of self-reported pain for the same person, which, while certainly noteworthy, is a little like taking an exam that you've just written out the answer key for.
In the study led by Mackey, however, computer algorithms were first trained using fMRI data from one group of test subjects, and then asked to classify pain in a completely different set of people. These results suggest that a similar system could one day be used to help doctors objectively measure things like the severity of chronic pain — even in patients who can't communicate their discomfort, or in patients the doctor has never met before.
While the scientists classify their results as a "major development," they emphasize that physiology-based pain assessment has a long way to go before it can be used in a clinical setting. The researchers suggest that future studies focus on the ability of fMRI-trained computer algorithms to distinguish between varying degrees of pain, or recognize the cognitive effects of persistent pain (the current study shows that it is feasible to classify transient pain experiences, but the researchers say that this does not easily translate to assessments of chronic pain).
"A key thing to remember is that this approach objectively measured thermal pain in a controlled lab setting," Mackey said. "We should take care not to extrapolate these findings to say we can measure and detect pain in all circumstances."
Be that as it may, the results of the experiment are certainly encouraging, and represent a significant step in the development of an objective solution to one of the most subjective problems in pain medicine today.
The researcher's findings are published in today's issue of PLoS ONE, and are accessible free of charge
Thanks to Dr. Sean Mackey for the advance copy of the article
Top image via sportgraphic/Shutterstock