HEALTHY AGING AS INFORMATION DIVERGENCE IN THE MULTIPLEX BRAIN

Healthy Aging Defined by Progressive 'Untethering' of Brain Function from Structure

As humans age, the way our brain's electrical signals travel becomes increasingly disconnected from the physical wiring that supports them. This "untethering" is most intense in subcortical areas like the thalamus. Meanwhile, memory centers like the hippocampus remain remarkably stable. This mismatch may mirror the decline in certain cognitive skills. These include rapid problem-solving and motor adaptation seen in older age.

The search for a unified architecture of aging

Understanding how the human brain changes across the adult lifespan is a fundamental challenge. We know that aging involves the degradation of grey matter (neuronal cell bodies). It also involves changes in functional activation (patterns of electrical activity). However, a central question remains. How do these two processes interact? Specifically, how does functional "traffic" navigate a physical "scaffold" of white-matter fibers that is progressively eroding?

To answer this, we must look at the brain as an integrated system. Scientists seek to understand the joint trajectory of the structural connectome (the physical wiring diagram) and the functional connectome (the map of active communication). Until now, the field has lacked an integrative framework. Such a framework must track how these two layers co-evolve as a single, interdependent entity.

The limitations of the isolationist paradigm

Historically, our understanding of the aging brain has been fragmented. Researchers have typically studied structural changes and functional changes as distinct, isolated phenomena. This "isolationist" approach has produced valuable data. It has documented white-matter degradation and network reorganization. However, it fails to explain the tension at the heart of cognitive aging. It cannot fully explain how functional integration is maintained while physical infrastructure breaks down.

Previous studies have attempted to relate these two layers using simple correlation coefficients. These essentially check if a strong physical connection always corresponds to a strong functional one. But these methods are limited. They assume a linear relationship. They are also often blind to higher-order topological reorganizations (complex changes in network shape). They can tell you if two regions are talking. They cannot easily quantify the statistical "distance" between the physical map and the functional traffic.

Modeling the brain as a multiplex network

To move beyond these gaps, the authors of this study modeled the brain as a multiplex network. In this framework, the brain is a two-layer system. The same anatomical regions exist in both the structural and functional layers simultaneously. This allows for a mathematical interrogation of how much information is shared between the two.

Using a cohort of 589 healthy individuals aged 18 to 88 from the Cam-CAN dataset, the researchers applied information-theoretic metrics. These metrics quantify the "distance" between these layers. They utilized Jensen-Shannon Divergence ($d_{JS}$), which measures the statistical dissimilarity between the two connectivity distributions. They also used relative information loss ($q(C)$), which assesses information lost when collapsing the two layers into a single, simplified network .

This approach allowed the researchers to observe how the entire system-wide architecture diverges over time.

The emergence of a global information divergence

The investigation revealed a striking organizing principle of healthy aging. It is a progressive, linear "untethering" of functional dynamics from their structural constraints. The authors report that as age increases, functional traffic operates with increasing independence from its physical scaffolding .

Specifically, the relative information loss ($q(C)$) rises from approximately 1% in late adolescence to nearly 2% by age 89. Similarly, the Jensen-Shannon divergence shifts from roughly 5% in early adulthood to 8% in later life . This is a robust, systemic trend. The data suggest that in youth, the functional brain is tightly "tethered" to its physical wiring. This ensures efficient, direct communication. As we age, this alignment erodes. Functional interactions increasingly traverse pathways not strictly governed by the underlying fiber bundles.

Crucially, this decoupling is not uniform across the brain. The researchers found that subcortical hubs—the brain's central "switchboards"—are the primary epicenters of this divergence .

These include the putamen, pallidum, caudate, and thalamus. In these regions, a geometric paradox emerges. While the physical structural scaffold undergoes severe "constriction" (shrinking and isolating the nodes), the functional communities actually expand and "smear" their boundaries .

This creates a mismatch. Functional communication becomes increasingly unconstrained compared to the narrowing physical pathways.

Implications for cognitive resilience and disease

This "untethering" provides a biological marker that correlates with cognitive changes. The authors found that this divergence in subcortical hubs operates collinearly with the decline of fluid intelligence and motor adaptation. However, the relationship is complex. When controlling for chronological age, the global associations with intelligence and reaction time largely attenuated .

This indicates that at a whole-brain level, this untethering acts as a proxy for aging. It is not an independent driver of cognitive decline.

Interestingly, while the subcortical switchboards lose their structural-functional fidelity, the limbic core remains remarkably stable . This includes the hippocampus and entorhinal cortex. This suggests a hierarchy of resilience. The brain appears to prioritize "representational fidelity" in the limbic system. It protects core memory circuits from corruption. This happens even as subcortical "routing" systems undergo radical reorganization.

If this pattern holds, the degree of subcortical untethering could serve as a powerful baseline. While the current study focuses on healthy aging, this divergence provides a normative standard. In the future, clinicians might use these metrics to distinguish "normal" topological drift from the specific, pathological signals of neurodegenerative diseases. A logical next step would be to apply these multiplex metrics to task-based fMRI data. This could reveal if decoupling becomes even more pronounced during active cognitive processing.