Scientists map the hidden architecture of the brain’s default mode network

A new study published in Nature Neuroscience sheds light on the structural foundations of the brain’s default mode network, a system of regions long associated with internally focused thought, memory, and self-reflection. Using postmortem brain tissue and advanced neuroimaging, researchers found that this network is composed of distinct anatomical types of brain tissue, each with different roles in processing information. The findings help explain why the default mode network is involved in such a wide variety of mental states—from introspection to decision-making—and suggest that its structure enables a unique balance of communication across the brain.

The default mode network, or DMN, is one of the most widely studied yet poorly understood brain systems in neuroscience. It was originally discovered through brain scans showing that certain regions become less active when a person focuses on an external task, like solving a math problem. But over time, researchers noticed these same regions were also active during a wide range of cognitive activities, such as daydreaming, remembering the past, imagining the future, and even making complex decisions. This unexpected versatility raised fundamental questions about what the DMN does and how it manages to participate in such seemingly contradictory mental functions.

One key to solving this puzzle, the researchers hypothesized, lies in the network’s anatomy. Much of the past research on the DMN has used functional MRI to track patterns of activity, but less attention has been paid to the underlying microstructure that might shape those functions. The authors of the new study believed that a deeper understanding of the DMN’s cellular and anatomical features could clarify how it supports such a diverse array of mental processes.

“The default mode network has a fascinating history in neuroscience. It was first identified as a group of brain regions that become less active when people engage in a specific task,” explained study author Casey Paquola, the head of the Multiscale Neurodevelopment Lab at the Institute of Neuroscience and Medicine (INM-7) at the Helmholtz Associations’ Research Center Jülich.

“But over time, researchers noticed that these same regions were actually activated during a wide variety of tasks — from recognizing faces to making decisions. This led to a lot of diverse theories about what the DMN is and what its role is in cognition. So, despite being discussed in tens of thousands of studies since 2001, it remained enigmatic.”

“Notably, the vast majority of those studies used fMRI to study the DMN. As I work at the intersection of neuroanatomy and neuroimaging, I thought it would be a novel angle to investigate this functional entity through the lens of neuroanatomy. We wanted to test whether the hypotheses based on functional MRI were supported or rejected based on the architecture of the DMN.”

To do this, the research team combined two powerful tools: postmortem brain histology and in vivo neuroimaging. The histological data came from a high-resolution 3D reconstruction of a human brain donated after death, allowing for precise mapping of cell densities and tissue types across the cortex. The team analyzed nearly 7,400 thinly sliced brain sections, stained to reveal the shapes and layers of cells, and reconstructed them into a 3D model known as “BigBrain.” These anatomical maps were then compared with functional brain networks defined by resting-state MRI scans in living participants.

The researchers focused on how different parts of the DMN varied in their cytoarchitecture—the arrangement and characteristics of cells across different layers of the cortex. They identified several types of cortical microstructure within the DMN, ranging from areas specialized for processing sensory information to regions associated with memory and internal thought. This showed that the DMN is not uniform but rather includes a mix of cell types, each potentially suited for different tasks.

Using data-driven modeling, the team found a spectrum—or axis—of cytoarchitectural variation across the DMN. Some regions had highly layered structures with dense mid-level cell populations, while others had flatter, less differentiated profiles. These differences aligned with known anatomical types, such as eulaminate regions that process external information and agranular areas often linked to self-generated thought and emotion.

The researchers then examined how these anatomical differences relate to the DMN’s connectivity. By analyzing diffusion MRI data from healthy adults, they assessed how efficiently information could travel along structural pathways linking different brain regions. They found that parts of the DMN with highly layered cytoarchitecture were more strongly connected to other brain regions, particularly those involved in sensory processing. These regions appeared to serve as “receivers,” efficiently gathering input from across the brain.

In contrast, areas with flatter cytoarchitecture were relatively isolated from sensory input, suggesting that they form an “insulated core” within the DMN. These more internally focused regions, like the anterior cingulate cortex, may support self-referential processing or maintain mental representations that are less dependent on immediate sensory information.

To further understand the flow of information within the DMN, the researchers used a modeling technique called regression dynamic causal modeling. This method estimates how activity in one brain region influences another, providing a measure of functional input and output. The results confirmed that “receiver” regions in the DMN received strong input from a variety of other brain systems, while the more insulated areas were relatively unaffected by outside signals.

Perhaps most strikingly, the researchers found that the DMN differs from other brain networks in how it sends information back out. While many networks tend to favor certain types of connections, the DMN distributed its output evenly across all levels of the brain’s processing hierarchy—from low-level sensory areas to high-level association regions. This suggests that the DMN plays a unique integrative role, capable of influencing thought and behavior at multiple levels.

To verify that these patterns weren’t just artifacts of group-level analyses, the team conducted a replication study using ultra-high field 7-Tesla MRI in individual participants. This allowed them to map microstructural variation and connectivity within each person’s brain. The results were consistent with the earlier findings, supporting the idea that the DMN’s unique structure and connectivity patterns are present at the level of individual brains.

“Given the DMN is largely expanded in humans relative to other mammals, more so than any other functional network, an interesting take-away from this study is that this unique network of brain regions is capable of enriching our interpretation of the world, by coloring our world view with other types of information, from autobiographical memory or social perspectives,” Paquola told PsyPost. “In other words, this network operates in a very different way to standard sensory processing, and that may help humans to have a richer understanding of the world around us.”

But as with all research, there are limitations. The detailed anatomical mapping was based on a single postmortem brain, and while replication was attempted with high-resolution imaging in living subjects, further validation across more diverse samples is needed. The study also focused on healthy adults, leaving open questions about how the DMN might develop during childhood or change in mental illness.

Future research will aim to understand how the DMN’s anatomy evolves over time and how it interacts with cognitive development and psychiatric symptoms. For example, if certain subregions mature more slowly or become disrupted in disorders such as depression or schizophrenia, this could help explain why these conditions involve changes in self-perception or thought patterns.

“I’m very interested in how the brain develops, that is how it changes from infancy to adulthood,” Paquola said. “Moving forward, we’d like to understand more about how the maturation of the DMN coincides with cognitive maturation and changes in mental health.”

The study, “The architecture of the human default mode network explored through cytoarchitecture, wiring and signal flow,” was authored by Casey Paquola, Margaret Garber, Stefan Frässle, Jessica Royer, Yigu Zhou, Shahin Tavakol, Raul Rodriguez-Cruces, Donna Gift Cabalo, Sofie Valk, Simon B. Eickhoff, Daniel S. Margulies, Alan Evans, Katrin Amunts, Elizabeth Jefferies, Jonathan Smallwood, and Boris C. Bernhardt.