Distinctive and Complementary Roles of Default Mode Network and Frontoparietal Control Network in Fluid Intelligence

The Unity and Diversity of Executive Functions

Executive function is an umbrella term for the cognitive processes necessary to manage diverse challenges. From recalling items in short-term memory to solving tasks with changing rules, executive tasks tap into a range of abilities that are crucial for everyday problem-solving and adaptive behavior.

Theoretical models suggest that executive functions rely on both domain-general and domain-specific processes. On one hand, executive tasks tend to positively correlate, pointing to a shared “common executive function.” On the other hand, distinctions have been made between specific components like updating, set shifting, and inhibition.

This unity and diversity is reflected in the brain’s functional organization. Neuroimaging and lesion studies have revealed that executive activations converge within broad multiple-demand (MD) regions distributed across the frontal, parietal, and temporal cortices. These MD regions are thought to support domain-general cognitive control. However, executive activations also show task-specific topographies, suggesting an interplay between domain-general and domain-specific processes.

The lack of precise anatomical mappings has long impeded our understanding of how executive functions are implemented in the brain.

Mapping Executive Functions with High-Resolution Neuroimaging

To address this challenge, a recent study (Assem et al., 2024) used the advanced multimodal neuroimaging approach of the Human Connectome Project (HCP) to investigate executive functions at an unprecedented level of spatial detail.

The researchers scanned participants performing three canonical executive tasks: n-back, rule switching, and stop signal. Their results revealed that:

1. Overlapping Activations in Multiple-Demand Regions: At the individual subject level, different executive activations converged within nine domain-general MD territories distributed across frontal, parietal, and temporal cortices.

2. Unique Topographical Shifts within MD Regions: Each task exhibited a unique activation topography, characterized by fine-grained gradients within the MD territories. These shifts were systematically oriented toward adjacent resting-state networks (RSNs):

3. n-back activations shifted toward the Default Mode Network (DMN)
4. Rule switching shifted toward the Dorsal Attention Network (DAN)

5. Stop signal shifted toward the Cingulo-Opercular Network (CON)

6. Peak Activations at Network Borders: The strongest activations often arose at the intersection of task-specific networks and the core MD regions, suggesting a key role for these border zones in information integration.

These findings provide a novel mechanistic insight into how executive functions emerge from the interplay between domain-general MD regions and adjacent domain-specific networks.

The Default Mode Network: From Internal Thought to External Control

The DMN has long been associated with internally-directed cognition, such as mind-wandering, autobiographical memory, and future planning. However, recent studies have challenged this narrow view, implicating the DMN in externally-focused cognitive control as well.

A study by Crittenden et al. (2015) found that DMN regions, particularly the core and medial temporal lobe (MTL) subnetworks, showed increased activity during demanding task switches that involved a change in stimulus domain. This suggested a role for the DMN in implementing and controlling externally-directed behavior, not just internal thought.

The current study (Smith et al., 2023) further explored the DMN’s involvement in executive control. By incorporating brief rest periods into a task-switching paradigm, the researchers were able to dissociate DMN activity related to:

1. Relaxation of Previous Task Set: DMN regions showed increased activity when participants switched from a task to a rest period.

2. Establishment of a New Task Set: DMN regions were also strongly engaged when participants switched back from rest to a new task.

These findings challenge theories that strictly link the DMN to internal, self-directed cognition. Instead, the researchers propose that the DMN encodes representations of the broader cognitive context, which can include both internally constructed and externally perceived information.

The DMN’s contextual representations may play a direct role in implementing and controlling current behavior, not just in internally-focused thought.

Complementary Roles of Default Mode and Multiple-Demand Networks

The study also revealed some intriguing parallels and differences between the DMN and the frontoparietal MD network in supporting executive control:

• Shared Involvement in Cognitive Transitions: Both the DMN and MD network showed increased activity during major cognitive transitions, such as switches between tasks or from rest to task.

• DMN: Coarse Context Representation: The DMN exhibited task-specific activity patterns, but these were relatively coarse, distinguishing only between broad task domains (e.g., faces vs. buildings).

• MD Network: Fine-Grained Task Representations: In contrast, the MD network was able to discriminate between all six individual task rules, suggesting more fine-grained representations.

These results suggest a complementary division of labor:

• The DMN may provide a broad contextual framework to guide and constrain behavior.
• The MD network then implements the specific task rules and operations within this broader context.

At cognitive transitions, the DMN may reactivate to allow rereference to the current context, enabling the MD network to establish a new, appropriate task set.

The interplay between the DMN’s contextual representations and the MD network’s task-specific control mechanisms may be a key neural substrate for flexible, goal-directed cognition.

Implications and Future Directions

This study’s high-resolution approach has provided unprecedented insights into the functional architecture supporting executive functions. By revealing the distinct yet complementary roles of the DMN and MD network, it challenges simplistic views of these systems and points to a more nuanced understanding of how the brain orchestrates complex, adaptive behavior.

The findings also raise intriguing questions for future research:

• How do the contextual representations in the DMN interact with task-specific processes in other domain-specialized networks (e.g., language, perception)?
• What are the precise mechanisms by which the DMN and MD network coordinate at cognitive transitions?
• Can this framework help explain executive deficits in clinical populations, and guide the development of targeted interventions?

As the field continues to explore the neural underpinnings of executive control, this study’s integrative perspective on domain-general and domain-specific processes will be a valuable guide. By bridging the unity and diversity of executive functions, it points the way toward a more comprehensive understanding of how the brain supports flexible, intelligent behavior.

Conclusion

Using advanced neuroimaging methods, this study has revealed a novel framework for understanding the brain’s implementation of executive functions. It demonstrates that executive activations reflect an interplay between a domain-general MD network and adjacent domain-specific systems, such as the DMN.

The DMN’s contextual representations appear to play a key role in guiding and constraining behavior, complementing the task-specific control mechanisms of the MD network. This interplay may be a crucial neural substrate for flexible, goal-directed cognition.

These findings challenge simplistic views of executive functions and the brain networks that support them. They point the way toward a more nuanced, integrative understanding of how the brain orchestrates complex, adaptive behavior.