Imaging’s Groundbreaking Discovery: 30 Years Later

In 2016, Science magazine ranked Randy L. Buckner among the top 10 most influential brain scientists of the modern era. He explains the road to discovering the default network, the pattern of brain activity triggered as we think about the past and the future.

Q: Why don’t we begin with a brief description of what the default network is, how and when it was discovered, and why it’s important.

Randy L. Buckner: In the 1990s, neuroscientists were just starting to do functional imaging studies. For the first time, we had brain scanners that could see the mind at work. We were like kids in a candy store in the sense that we no longer needed a scalpel to see the brain; the new technology allowed us to safely discern information out about what parts of the brain people used when given different tasks and different kinds of visual or auditory stimuli.

I was a graduate student at the time at Washington University and one of my mentors, Marcus Raichle, was at the forefront of positron emission tomography (PET), an imaging technique that measures physiological changes in the brain and shows where blood flow is increasing due to brain activity. This is when many of us first became aware of the Dana Foundation, which was helping fund our work. I was a Dana fellow in those early days, and this was an exciting time in neuroscience.

In early studies, we often asked participants to perform very simple tasks: read and say words, detect colors in pictures, or try to recognize whether a viewed word was on an earlier studied list. The imaging revealed the parts of the brain involved in their responses. But what jumped out at us was something unexpected: When people weren’t asked for a response or given a specific task, much of their brain still remained active.

There were skeptics, but Raichle argued that the brain activity seen in the scanner that was not related to specific tasks or responses was—in and of itself—a groundbreaking phenomenon. At the time we didn’t know what to make of the activity pattern that occurred when people were intended to be at rest. The pattern was anatomically specific, involving the highest-order association regions of the brain and recurred across many separate studies. Nancy Andreasen, the renowned psychiatrist, also early noted the rest pattern of activity and suggested that the brain was at work—spontaneously thinking—when left undirected.

The considerable irony was that when we stopped instructing people to do tasks in the scanner, we saw that the human brain imagines, constructs, and explores mentally.

Look at your own task with this interview, for example. You have the difficult challenge of trying get content from a scientist so that you can communicate it to the public. You’re imagining my perspective, all the ways I’m struggling to explain this, and thinking of a plan to engage me based on what you believe. You also have this sort of imagined audience in the reader and you’re able to listen and hear my words and then mentally construct a plan to deliver it clearly and effectively to a reader. The default network may be helping you do all this, which is extraordinary.

How did the name “default network” come about?

Buckner: The network was first found when people were at rest, so the idea caught on that the network was the default state of the brain. It is the idea that people default to their own processes and make their own mental explorations when not given a test or task in a scanner. However the name is also a bit misleading because we now know the network is active when people intentionally remember and plan. It’s not just used during rest. It’s a network that can be called upon to do many forms of focused mental exploration.

How does the network part fit it?

The network refers to different parts of the brain—the hippocampus, specific regions along the midline of the cortex, for example—acting together, presumably because they are anatomically connected parts of a large, distributed brain system. As study of the default network has progressed, we’ve learned that there are likely multiple separate networks next to one another. There is one network that seems to be involved when people use memory systems to remember especially, if their memory involves constructing a mental scene. Another network physically next to it becomes active when people imagine what other people are thinking.

How has your research evolved?

It’s amazing what’s happened over the last two decades. For the first time, we’re really starting to get a handle on the networks that humans use to mentally explore. When we began using neuroimaging to learn about the brain, we targeted seeing, hearing, and attending to the outside world. What has emerged is the beginning of an understanding of how the brain helps us to be mental explorers. We have an incredible capacity to imagine beyond the present—to think about possible futures, what ifs, and how others who are not ourselves might be perceiving a situation. Our research has evolved to study the networks of the brain responsible for these extraordinary human abilities.

What inspired you to devote your career to studying the default network, or should I say “networks?”

In graduate school, I was more interested in targeted memory functions and using imaging to compare brain functions of people doing tasks. Marcus Raichle was the person going around the lab, saying, ‘You folks, look at this stuff. You should pay attention to it.’ He was referring to the pattern of activity we now know as the default network, but at the time, I personally was slow to appreciate the significance.

One of my earliest papers in graduate school demonstrated this. We had data in 1992, ’93, but it wasn’t until my paper in 1995 that the default network was shown. It wasn’t because of any special insight, but more due to my eagerness to present all the data from an experiment. So just the tendency to describe our methodology and results very thoroughly made for one of the first presentations of the network. In retrospect, it was mother nature saying, ‘Here I am.’ When I’m asked, ‘How was it discovered?’ The answer: Serendipitously. It was just in all the study data.

Imaging has certainly changed neuroscience in so many ways. But how in particular have the default network revelations changed neuroscience?

It’s relevant to psychiatry in particular. Tracking the competition and boundaries between networks may lead to some of the more profound symptoms you see in mental illness. This is a hypothesis. In active psychosis, for example, one can confuse reality with internal constructions. We have these networks and brain systems that allow us to tend to the external world, act, and survive the here and now. But we have other networks—default networks—that may be used when we detach from the outside world and mentally explore. When networks or interactions between them break and seem to be fragile, one might see devastating, cognitive dysfunction.

It’s also the case that we might have stumbled into seeing some of the networks that are highly expanded and evolved in the human lineage, networks that allow us to do these extraordinary human faculties, such as remember, or imagine what we’re going to do tomorrow, and work in social groups, where we can envision what somebody else might be thinking.

Another key insight is that higher-order networks involved in mental thought may be specialized for specific domains of information. It’s easy to appreciate the idea if we begin by thinking about parallels in sensory brain systems. We have different domains and specializations for sensory processing. For example, specialized visual regions deal with where things are, and others with what they are. It’s easy to translate that and tell people that there are parts of the brain’s visual system that become particularly specialized when taking in different bits of sensory and visual information. I suspect that the higher-order networks that we uncovered in our journey to understand the brain’s default network are also specialized, but in more abstracted domains—such as for remembering and, separately, for making social inferences.

Can you change your default mode, or maybe train it?

The short answer is I don’t know. But this clinically relevant question drew my attention when we began studying how the default network is different in older individuals, including those with dementia. An observation is that amyloid, the protein that pathologically builds up in Alzheimer’s disease, deposits preferentially in the default network, which caught our attention in multiple ways.

By better understanding these networks, it might help us to understand why Alzheimer’s disease is so devastating. Since the higher order circuits that people use to remember are so adversely affected, we are seeing neurodegeneration in the pathways used for higher order, internal thought.

When we pointed out that the default network might be vulnerable in Alzheimer’s disease, the immediate question became: ‘Well, then how do you change it? Can meditation, exercise, or diet change it? If we use our networks less for higher order thought, might this delay the accumulation or prevent the spread of amyloid? My hunch is no.

How about memory versus stimulus from the outside world? Does that play into studying default mode?

Absolutely. I think that is a potentially critical bit that we need to understand. It seems that there are predominant constraints on what drives networks, such as sensory stimulation from the retina that drives the activity in the visual system. There are other networks that may be built not off of external stimulation but of things that are generated internally from mnemonic associations, which are the bits and snippets we’ve captured and stored in our brains to help us with memory. This mnemonic information, in part, may be providing the driving information that these other networks elaborate on.

I find that to be potentially critical to this distinction. The brain networks that are expanded and able to utilize internally generated information from mnemonic systems, at their essence, may be different from some of the other networks that we think of as being driven by external events in the world around us, or that we think of in terms of sensory motor function that are primarily driven by taking in sensory stimulation from the outside world and acting upon it. Instead, these networks that utilize internally generated information are about taking information from internal mnemonic systems and elaborating it.

Could further understanding of the default mode help treat autism, Alzheimer’s, or psychiatric problems? What’s the goal?

Many forms of illnesses affect high-order cognition, whether it be neurodegenerative illness or psychiatric illness. Differences in high-order faculties seems to be a common outcome of many routes to atypical brain development and degeneration. Sometimes I wonder—from a clinical perspective—if we focus too much on the default network and problems with high-order mental functions, at least insomuch as we think the origins of the illness result from selective dysfunction in these networks.

It’s true that we notice dysfunction of the default network in illnesses like psychosis associated with schizophrenia. I don’t know if that necessarily means that the illness is due to a specific dysfunction in these networks, as much as it is that you notice the difficulties in the higher order functions first when network coordination or development hasn’t gone in the typical trajectory. That said, being able to measure how someone’s brain works when it’s working well, and knowing how the networks are interacting and competing, might be a window into whether therapies are working.

That’s important, because as we try to mitigate brain dysfunction, we need ways of measuring and seeing the working brain and restoring the working brain. That’s one kind of path I think this work has taken. There are ways of potentially measuring typical interactions between networks and having that knowledge and being able to make those measurements that are part of the translational research effort.

The fact that, in Alzheimer’s disease, we can see the amyloid pathology building up in these networks makes one wonder what it is about these networks—and about their metabolic levels—that is conducive to the pathology and sets up the vulnerability. That’s much more of a direct insight, measuring pathology, which can be now be made with molecular PET scans and eventually, potentially blood biomarkers. In thinking about the mechanisms there, I’ve started to wonder why the metabolic activity sets up risk for Alzheimer’s in your eighth and ninth decades.

But when you ask the question, ‘What should we be prioritizing for clinical translation,’ I think that our ability to see these networks may lead to ways of looking and seeing if therapies, or behavioral interventions, are having positive effects—even if they’re not intimate and mechanistically linked to the origins of the illness.

This article first appeared in the Winter 2022 issue of our Cerebrum magazine. Click the cover for the full e-magazine.