View larger image to better understand the brain

Summary: Zooming out to image larger areas of the brain using fMRI technology allows researchers to obtain additional relevant data, offering a better understanding of neural interactions.

A source: Yale

Scientists have learned a lot about the human brain using functional magnetic resonance imaging (fMRI), a technique that helps us understand how the brain works. However, typical fMRI techniques may be missing key information and provide only part of the picture, according to the Yale researchers.

In the new study, they evaluated the different methods and found that zooming out and taking a wider field of view captures additional relevant information that is left behind by a narrow focus, leading to a better understanding of neural interactions.

Furthermore, these broader findings may help address the reproducibility problem in neuroimaging, where some findings in studies cannot be reproduced by other researchers.

The results were announced in August. 4 inches Proceedings of the National Academy of Sciences.

Studies using fMRI typically focus on small areas of the brain. As an example of this approach, researchers look for small areas of the brain that are most strongly activated, looking for areas of the brain that become more “active” when a particular activity is performed. But growing evidence suggests that brain processes, and particularly complex processes, are not confined to small parts of the brain.

“The brain is a network. It’s complicated, said Dustin Sheinos, associate professor of radiology and biomedical imaging and senior author of the study. Oversimplification, he said, leads to false conclusions.

“For complex cognitive processes, it is unlikely that many areas of the brain are fully involved,” added Stephanie Noble, a postdoctoral fellow in Sheinos’ lab at Yale School of Medicine and lead author of the study.

Focusing on small areas leaves out other areas that may be involved in the behavior or process being studied, which also affects the direction of future research.

“You’re misrepresenting what’s going on in the brain,” he said.

For the study, the researchers evaluated how well fMRI analyzes at a range of scales were able to detect effects, or changes in fMRI signals, that revealed which parts of the brain were activated as participants performed different activities.

They used data from the Human Connectome Project, which scans the brains of people performing various tasks related to complex processes such as emotion, language and social interaction.

The research team looked for effects in very small parts of a brain network, such as connections between just two regions, as well as in clusters of connections, in widespread networks, and throughout the entire brain.

They found that the larger the scale, the better they could detect the effects. The ability to detect these effects is called “power”.

“We get better power with these large-scale approaches,” Noble said.

At the smallest scale, the researchers were only able to detect about 10% of the effects. However, he was able to identify more than 80% of them at the network level.

The trade-off for the added power was that larger views did not provide the same spatially accurate information as smaller-scale analyses. For example, at the smallest scale, researchers could say with confidence that the effects they observed were occurring in a small area.

At the network level, they were only able to say that the effects were occurring across most of the network, without pinpointing exactly where the effects were.

The goal, Noble says, is to balance the benefits and harms of different approaches.

“Do you want to be too confident about a small piece of relevant information?” he said.

“Or maybe you want to have a really big picture of the whole iceberg that’s a little fuzzy, but gives you a sense of the complexity and the vast spatial scale of what’s going on in the brain?

The method is easy for other researchers to implement, and Noble said he’s excited to see how other scientists use it.

Furthermore, these broader findings may help address the reproducibility problem in neuroimaging, where some findings in studies cannot be reproduced by other researchers. Image is in the public domain

He notes that the fields of psychology and neurobiology, including neuroimaging, are facing a renaissance problem. And low power in fMRI analyzes contributes to it: low-powered studies reveal only small parts of the story, which can be viewed as contradictory rather than parts of the whole.

Increasing the power of fMRI, as he and his colleagues did here by scaling up their analyses, may be one way to address the reproducibility problem by showing that seemingly contradictory results can actually be harmonious.

“Moving up the food chain, so to speak, from very low levels to more complex industries, buys you more power,” Scheinos said. “It’s one of the tools you can use to solve the problem of duplication.”

See also

It shows a man playing the banjo

And scientists shouldn’t throw out the baby with the bathwater, Noble said. A lot of good work is being done to improve methods and increase rigor, and fMRI is still a valuable tool, he said: “I think that assessing power, rigor and reproducibility is useful for the whole field. It is especially concerned with the complexity of living beings and mental processes.’

Noble is now developing a “power calculator” for fMRI to help others achieve desired power levels in their studies.

This is about neuroimaging research news

Author: Mallory Locklear
A source: Yale
The connection: Mallory Locklear – Yale
Photo: Image is in the public domain

Original research: Open access.
Stephanie Noble et al. PNAS


Improving power in functional magnetic resonance imaging beyond cluster-level inference

Neuroimaging findings usually appear in focal brain regions or at the circuit level. However, increasingly robust studies paint a much richer picture of widespread brain effects, and many focal reports suggest that the underlying effects may represent only the tip of the iceberg.

How focal and broader perspectives affect our conclusions has not been comprehensively evaluated using real data.

Here, we compare sensitivity and specificity across procedures representing multiple levels of inference using an empirical benchmarking procedure that samples task-based connectomes (~1000 subjects, 7 tasks, 3 resampling group sizes, 7 inference procedures) from the Human Connectome Project dataset.

Only broad-scale (network and whole-brain) procedures received the traditional 80% statistical power to detect a medium effect, reflecting >20% greater statistical power than focal (peripheral and cluster) procedures. Compared to family error control procedures, power for false discovery rate was also increased.

The downsides are fairly limited; Loss of specificity for large-scale and FDR procedures was relatively modest compared to gains in power. Furthermore, the large-scale methods we introduce are simple, fast, and easy to use, providing a straightforward starting point for researchers.

It also shows the promise of more sophisticated large-scale approaches not only for functional connectivity but also for related networks, including task-based activation.

Overall, this work demonstrates that inference rescaling and FDR control selection are straightforward, and helps to overcome statistical power problems that plague typical studies in this field.

Leave a Comment

Your email address will not be published.