Functional hierarchy of the human neocortex across the lifespan
Article Date: 25 March 2026
Article URL: https://www.nature.com/articles/s41586-026-10219-x
Article Image: Figure 1
Summary
This paper builds a unified, vertex-wise atlas of three principal functional-connectivity gradients across the human lifespan (16 days to 100 years), using 3,972 gradient sets from 3,556 individuals. The three axes—sensory–association (SA), visual–somatosensory (VS) and modulation–representation (MR)—are aligned to a common template and modelled with generalized additive mixed models (GAMMs). The authors chart when and how these gradients mature, show how global embedding measures (range, dispersion, template similarity) evolve with age, link gradient metrics to cognition and microstructure, and validate functional meaning with meta-analytic and transcriptomic analyses. Key findings: gradient topography changes most rapidly in early life (first ~4 years); SA and MR expand through childhood and peak in early adulthood (~19 years) then contract in later life; VS peaks in early childhood (~5 years) and declines thereafter; structure–function coupling declines with age; and SA fidelity robustly predicts cognitive performance in young adults.
Key Points
- Three reproducible FC gradients (SA, VS, MR) provide a compact coordinate system for cortical organisation across the lifespan.
- Big developmental change happens early: gradient topography shifts most markedly in the first four years of life, with continued childhood refinement and protracted tuning through adolescence.
- Global measures diverge by axis: SA and MR expand into early adulthood (peaks ~18–19 years) then contract, while VS peaks in early childhood (~5 years) and steadily contracts.
- Gradient dispersion (embedding differentiation) rises through childhood, peaks in adolescence (~14 years), then declines with ageing, reflecting dedifferentiation.
- Structure–function coupling is moderate overall and declines with age, strongest for SA (myelin and thickness align best) and weaker for VS and MR.
- SA topographical fidelity (cosine similarity to template) is the most consistent predictor of broad cognitive performance in young adults; MR shows early associations with motor/skill emergence in infants.
- Neurosynth meta-analysis supports functional interpretations: SA maps to association vs unimodal tasks, VS to visual vs somatosensory distinctions, MR to control/modulation vs representation.
- Transcriptomic enrichment links SA to synaptic signalling/vesicle-cycle genes, VS to mRNA/transcription themes, and MR to ion-transport/excitability programmes—gene–gradient coupling is stronger early in life.
- Large harmonised lifespan sample and consistent pipeline (WPCA template + Procrustes alignment + GAMMs) provide a normative reference but are mainly cross-sectional and anchored to an adult-referenced template.
Content summary
The authors computed vertex-wise functional-connectivity gradients for nearly 4,000 acquisitions spanning neonates to centenarians, aligned individual gradients to a lifespan-weighted WPCA template and modelled age trajectories with GAMMs. They focused on three dominant axes: SA (unimodal→transmodal), VS (visual↔somatosensory) and MR (modulation↔representation). Topographic and global metrics (range, dispersion, template similarity, mean FC degree) show nonlinear, axis-specific lifecourse profiles: SA and MR broaden through childhood, peaking in early adulthood, then narrow; VS is strongest in early childhood and diminishes thereafter.
Network-level analyses (Schaefer 7-network parcellation) reveal unimodal systems mature earlier than association, limbic and attention networks, and distinct maturational ages per network and axis. Structure–function coupling computed using morphometric-similarity-derived structural gradients shows only modest alignment with functional gradients and a general decline with age, suggesting increasing functional autonomy from coarse microstructure. Behaviourally, SA fidelity predicts multiple cognitive domains in young adults and broader cognitive relevance consolidates from infancy to early adulthood. Meta-analytic (Neurosynth) and transcriptomic (AHBA + PLS) analyses independently validate gradient interpretations and reveal molecular themes that couple to gradient topography most strongly in early life.
Context and relevance
This study integrates decades of developmental functional-connectivity research into a single normative chart for gradient architecture across the full lifespan. For researchers and clinicians, it supplies concrete timelines for when hierarchical motifs (sensorimotor→association; representation→modulation) consolidate, a reference for interpreting deviations in neurodevelopmental or neurodegenerative conditions, and multiscale links—from genes to cognition—that help explain why and when functional organisation matters.
Why should I read this?
Because if you work with brain development, ageing or brain disorders, this paper gives you a tidy, data-driven map of how the brain’s large-scale functional hierarchies form, peak and unravel across life. It saves you trawling dozens of papers by offering a single harmonised atlas and clear behavioural and molecular anchors. Also — early life is where the action is: lots of reorganisation happens in the first few years, which matters if you study infancy or early intervention.
Author style
Punchy: the authors present a bold, large-scale normative resource and back it up with multilevel validation. If you care about how macroscale hierarchy links to cognition, microstructure or genes, read the figures and methods—there are actionable metrics (range, dispersion, cosine similarity) you can reuse.
Limitations & takeaways
Key caveats: the atlas is mainly cross-sectional (not within-subject longitudinal), alignment to an adult-referenced template may bias early-life interpretation, and AHBA transcriptomics are adult-derived so molecular inferences describe adult-like scaffolds rather than dynamic gene expression. Still, the curated gradients, normative trajectories and cognitive links make this a practical reference for studies of typical and atypical brain development and ageing.