Schedule of Events | Symposia

Symposium Session 12 - Advancing global and local theories of DMN function across cognitive domains

Symposium Session 12: Tuesday, April 1, 2025, 1:30 – 3:30 pm EDT, Constitution B

Chair: Ajay Satpute¹; ¹Northeastern University
Presenters: Maureen Ritchey, Ajay Satpute, Lauren DiNicola, Lucina Uddin

The brain is composed of networks of interacting brain regions that support higher-order cognition. Among these, the default mode network (DMN) appears to cut across cognitive domains with contributions to memory, social cognition, language, and emotion. In this symposium, we will discuss recent advances in characterizing the functional architecture of the DMN, highlighting evidence for domain integration and/or specialization. Ritchey will discuss the role of the DMN in flexibly representing different types of information to build and update internal models of the world, here discussed through the lens of episodic memory. Satpute will provide a mechanistic account of the DMN function in affect and emotion based on predictive processing architectures. DiNicola will present recent advances from individualized functional neuroimaging that reveal that the DMN comprises distinct functional networks differentially associated with remembering, social inference and language. Uddin proposes a global model of DMN function, proposes that the DMN functions as a staging ground for goal-directed behavior, interacting with other networks to support complex cognition. These talks will address debates between global and local theories of DMN function, highlighting efforts to pursue a unified account as well as the insights we can gain from its subnetwork organization.

Presentations

Characterizing the representational structure of the default mode network through the lens of episodic memory

Maureen Ritchey¹; ¹Boston College

What can episodic memory tell us about the functions of the default mode network? In the cognitive neuroscience of episodic memory, there has been a long history of characterizing the functions of individual brain regions, for instance, by studying patterns of memory deficits in patients with focal brain damage. Neuroimaging research, in turn, has revealed the reliable engagement of the default mode network (DMN) during memory tasks, leading to a tension between region- and network-focused accounts of memory function. In this talk, I will discuss efforts to understand the roles of individual brain regions in the context of their associated networks, using the domain of episodic memory as a proving ground for integrating these levels of analysis. Within the DMN, there are distinct cortico-hippocampal subnetworks that represent different kinds of information in memory. Recent fMRI work from my lab has suggested that this subnetwork organization largely explains region-level participation in memory tasks, with the functions of individual regions relating to their ability to bridge across networks. Furthermore, DMN subnetworks are differentially modulated by the perceptual specificity and emotional tone of recall, indicating that parallel DMN pathways can be flexibly engaged to represent a range of experiences. This line of research can inform our understanding of the representational structure of the DMN, as it builds and updates internal representations of the world in the service of memory and other forms of cognition.

Integrating emotion and social cognition in a predictive processing framework of default mode network function.

Ajay Satpute¹; ¹Northeastern University

In this talk, I embed DMN function in a broader, predictive processing framework, which suggests that the DMN contributes to driving experience-dependent predictions of sensory inputs to the brain. Functional activity in the DMN is a product of prediction and prediction errors wherein the functional role ascribed to each node is dependent on the content of these predictions in relation to its neuroanatomical position. This view suggests three key hypotheses. First, whether and which DMN areas are associated with affective or social cognitive processing will depend on the experimental context. Consistent with this notion, I will summarize research showing that functional activation patterns associated with both domains vary across study contexts. Second, whether and which DMN areas are engaged will depend on the source of predictions and prediction errors in a given study context. Findings from my lab suggest a tradeoff in functional activity between the precuneus and dorsomedial prefrontal cortex in constructing predictive mental models using existing prior experiences or more distant associations. Third, DMN activity will depend on external validity, such that paradigms that study phenomena “in the world” are more likely to engage DMN activity than decontextualized tasks. The summary of this work suggests that DMN nodes perform domain general computations in contributing to diverse phenomena in a context dependent manner.

Revisiting the default network: Growing evidence for functionally dissociable, parallel association networks

Lauren DiNicola¹; ¹Harvard University

Nearly two decades ago, a striking observation was made: when groups of people performed social inference, remembering, and self-reflection tasks, similar patterns of brain activity appeared. These patterns overlapped regions commonly called the default (or default mode) network (DMN), opening debate about this network’s role in advanced cognition. Initial DMN studies relied on group-averaging, which revealed broad patterns but blurred spatial details. Recently, individualized functional neuroimaging has provided new insights into network organization and functions. In this talk, I will present growing evidence from individualized work that association zones attributed to the canonical DMN comprise at least three distinct, parallel and domain-specialized networks. These networks are tightly interwoven and easy to blur, but reliably identified within individuals. The networks show dissociable responses to tasks targeting remembering, theory of mind and sentence processing. And over the past five years, the networks have been estimated – and functional distinctions observed – in multiple studies and prospective replications, across distributed cortical regions. Canonical DMN estimates thus appear to overlap multiple networks that differentially subserve some of the cognitive abilities that are hallmarks of our species – remembering, social inference and language. These findings further raise the question: how might such parallel networks form? I will propose one possibility: that interwoven networks might originate from a less-differentiated archetype that fractionates and specializes early in development.

What does the DMN do? Insights from network neuroscience

Lucina Uddin¹; ¹University of California, Los Angeles

The default mode network (DMN) is one of the most widely investigated large-scale functional brain networks in neuroscience. Anchored in the medial prefrontal and posterior cingulate cortex, it exhibits unique properties including high metabolic activity at rest and antagonistic relationships with frontoparietal networks underlying executive control. Many clinical conditions are associated with alterations in DMN activation and functional connectivity including autism spectrum disorder and depression. Typical development of the DMN is associated with enhanced performance on tasks of attention and executive function. So, what does the DMN do, exactly? This talk will explore the hypothesis that the main function of the DMN is to act as a staging ground for preparation of goal-congruent behavior, participating in large-scale functional brain network dynamics in ways that impact nearly every domain of cognition.