Since the development of fMRI in the 1990s, reliance on neuroimaging has increased exponentially as researchers investigate how fMRI data from the resting brain, and brain anatomical structure itself, can be used to predict individual traits, such as depression, cognitive decline, and brain disorders.
Brain imaging has the potential to reveal the neural underpinnings of many traits, from disorders such as depression and chronic diffuse pain to why one person has a better memory than another, and why some people’s memories are malleable as they age. But how reliable brain imaging is for detecting traits has been the subject of wide debate.
Previous research in brain-wide association studies (called “BWAS”) has shown that the links between brain function, structure, and traits are so tenuous that thousands of participants would be needed to detect reproducible effects. Research on this scale requires an investment of millions of dollars in each study, which limits the traits and brain disorders that can be studied.
However, according to a new commentary published in natureAnd Stronger links between measures of brain and traits can be obtained when using modern pattern recognition (or ‘machine learning’) algorithms, which can obtain high-powered results from moderate sample sizes.
In their article, researchers from Dartmouth and University of Essen Medicine provide a response to an earlier analysis of brain-level association studies led by Scott Marek at Washington University School of Medicine in St. Louis, Brendan Truffaut-Clemence at Massachusetts General Hospital/Harvard Medical. school and colleagues. The previous study found very weak associations across a range of traits in several large brain imaging studies, and concluded that thousands of participants would be needed to detect these associations.
The new article explains that the very weak effects found in the previous paper do not apply to all brain images and all traits, but are limited to specific cases. Shows how fMRI data from hundreds of participants, versus thousands, can be better leveraged to provide important diagnostic information about individuals.
One of the keys to stronger links between images and brain traits such as memory and intelligence is the use of modern pattern recognition algorithms. “Because there is almost no mental function performed by a single region of the entire brain, we recommend using pattern recognition to develop models of how multiple brain regions contribute to predicting traits, rather than testing brain regions individually,” says Tor Wager senior author Diana L. Taylor is Distinguished Professor of Psychology and Brain and Director of the Center for Brain Imaging at Dartmouth.
“If models of multiple brain regions working together rather than in isolation are applied, this provides a more powerful approach in neuroimaging studies, resulting in predictive effects four times greater than when testing brain regions in isolation,” says lead author Tamas Spisak. . , Head of the Laboratory of Predictive Neuroimaging at the Institute for Diagnostic, Interventional and Neuroradiology at the University of Medicine Essen.
However, not all pattern recognition algorithms are created equal, and finding algorithms that work best for certain types of brain imaging data is an active research area. Marek’s previous paper, Tervo-Clemmens et al. They also tested whether pattern recognition could be used to predict traits from brain images, but Spisak and his colleagues found that the algorithm they used was suboptimal.
When the researchers applied a more robust algorithm, the effects became larger and reliable associations could be detected in much smaller samples. “When you power the calculations on the number of participants required to detect replicable effects, the number drops below 500,” says Spisak.
says co-author Ulrike Bingel of the University of Medicine in Essen, who is head of the University Center for Pain Medicine. “Identification of markers, including those involving the central nervous system, is urgently needed because they are essential for improved diagnosis and individually tailored treatment approaches. We need to move towards a personalized medical approach based on neuroscience. And the potential for multivariate BWAS to move us towards this goal It should not be underestimated.”
The team explains that the weak associations found in previous analysis, particularly through brain images, were collected while subjects were simply resting in the scanner, rather than performing tasks. But fMRI can also capture brain activity associated with specific, moment-by-moment thoughts and experiences.
Wager believes that linking brain patterns to these experiences may be key to understanding and predicting differences between individuals. “One of the challenges associated with using brain imaging to predict traits is that many traits are not stable or reliable. If we use brain imaging to focus on studying mental states and experiences, such as pain, empathy, and craving for drugs, the effects can be much larger and more reliable,” she says. Wager says. “The key is finding the right mission to take over the state.”
“For example, showing people with substance use disorders images of drugs can trigger cravings for drugs, according to a previous study revealing a neural predictor of cravings,” Wager says.
“Identifying the approaches by which understanding the brain and mind is most likely to succeed is important, as this affects how stakeholders view and ultimately fund translational research in neuroimaging,” says Bingel. “Discovering limitations and working together to overcome them is key to developing new ways of diagnosing and caring for patients with brain disorders and mental health disorders.”