Brain fMRI Might Not Be As Reliable As Once Believed

When it comes to assessing how the human brain will respond to various stimuli – and using those responses to predict behavior – functional MRI (fMRI) is not as dependable as existing research as suggested.

Although fMRI does a good job of mapping brain activity, it has fallen short of being able to predict how individuals will respond to specific situations or foretell their future mental health (an application called task-based fMRI), despite the hopes of many neurological researchers. A new study from Duke Medicine, published in Psychological Science, shows the image-based blood flow measurements are rarely the same time, revealing that making those predictions falls outside fMRI’s current capabilities.

“Our findings do not question the validity of task-fMRI studies and what they reveal about the average human brain. In fact, task-fMRI does a great job at measuring an average brain’s functioning,” said lead study author Ahmad Hariri, Ph.D., professor of psychology and neurology and head of Duke’s Laboratory of NeuroGenetics. “Rather, our findings more specifically question the application of these same measures to reveal how the brains of different people function differently.”

Hariri’s team used both previous and new analyses of test and retest data to assess the reliability of commonly used task-fMRI brain activity measures. According to their results, 56 test-retest studies were not able to produce the same readings for each individual. This disparity occurred even when the tests were conducted within a short timeframe.

“If a measure gives a different value every time it is administered, it can hardly be used to make predictions about a person,” Hariri explained. “This is important because more and more such studies are being reported, often accompanied by sensationalistic headlines.”

It is also important to note, Hariri said, that many of the large-scale studies designed to measure brain activity in an effort to predict future mental health have adopted unreliable task-fMRI measures.

Based on these results, Hariri said, investigators must use caution when evaluating how to apply task-fMRI. In addition, a more rigorous foundation for promoting and advancing task-fMRI measures as a way to analyze brain-behavior correlations must be created.

“Better research measures, particularly in the age of individualized medicine and precision neuroscience, are needed to achieve clinically useful results,” Hariri concluded.