Convergent cortical temporal axis: common cortical oscillatory modes

Abstract

Brain dynamics provide the most direct substrate for neural function and representation. Extensive prior work has demonstrated that the human brain is organized into hierarchical structures across molecular, cellular, microstructural, and macroscale network levels. However, whether these hierarchies can be directly linked through shared patterns of neural dynamics remains unresolved, and clear evidence from data-driven analyses is still lacking. A central open question is whether brain dynamics vary along a single continuous dimension, or whether they are instead composed of a limited number of reproducible coordination modes that could provide a unifying basis for cross-scale integration. Here, using source-resolved resting-state magnetoencephalography (MEG), we adopt a fully data-driven approach that imposes no a priori assumptions about cortical hierarchy or gradient structure. By characterizing full-spectrum power relationships across cortical regions, we identify a set of stable and reproducible spectral coordination modes. These modes capture how multiple frequency components jointly co-vary across space, defining distinct dynamical configurations rather than conventional band-limited oscillations or continuous gradients. The identified spectral coordination modes exhibit pronounced spatial differentiation across the cortex, distinguishing regions dominated by narrowband rhythmic activity from those characterized by broadband or multi-frequency coordination. Importantly, these modes do not correspond to predefined hierarchical axes, nor can they be reduced to existing MEG gradients or functional connectivity organizations. Instead, they emerge directly from the data, revealing a finite set of shared organizational patterns underlying cortical dynamics. Multimodal analyses further demonstrate that individual spectral coordination modes show selective associations with neurotransmitter system distributions, laminar microarchitecture, and cell-type-specific gene expression, suggesting that these dynamical patterns may serve as linking mechanisms between micro- and macroscale brain organization. Computational modeling supports this interpretation, showing that differences in local circuit parameters are sufficient to generate distinct spectral coordination states without invoking a single global dynamical hierarchy. Across the adult lifespan, these coordination modes exhibit frequency- and region-specific reorganization, indicating multiple parallel trajectories of dynamical aging. In Parkinsons disease, alterations are observed in specific spectral coordination modes, particularly within higher-order cortical regions, suggesting that the disorder preferentially disrupts distinct dynamical configurations rather than inducing a global breakdown of cortical dynamics. Together, these findings indicate that large-scale brain dynamics are not organized along a single pre-existing hierarchical axis, but instead are structured by a limited number of data-driven spectral coordination modes. This framework provides a new dynamical perspective for linking brain organization across modalities and scales, and offers a principled avenue for understanding systematic changes associated with aging and neurological disease.