Commentary on Pang et al. (2023) Nature

[Pang et al. (2023)][1] present novel analyses demonstrating that brain dynamics can be understood as resulting from the excitation of geometric modes, derived from the shape of the brain. Notably, they demonstrate that linear combinations of geometric modes can reconstruct patterns of fMRI data more accurately, and with fewer dimensions, than comparable connectivity-derived modes. Equipped with these results, and underpinned by neural field theory, the authors contend that the geometry of the cortical surface provides a more parsimonious explanation of brain activity than structural brain connectivity. This claim runs counter to prevailing theories of information flow in the brain, which emphasize the role of long-distance axonal projections and fasciculated white matter in relaying signals between cortical regions ([Honey et al. 2009][2]; [Deco et al. 2011][3]; [Seguin et al., 2023][4]). While we acknowledge that cortical geometry plays an important role in shaping human brain function, we feel that the presented work falls short of establishing that the brain’s geometry is “a more fundamental constraint on dynamics than complex interregional connectivity” ([Pang et al. 2023][1]). Here, we provide 1) a brief critique of the paper’s framing and 2) evidence showing that their methodology lacks specificity to the brain’s orientation and shape. Ultimately, we recognize that the geometric mode approach is a powerful representational framework for brain dynamics analysis, but we also believe that there are key caveats to consider alongside the claims made in the manuscript. ### Competing Interest Statement The authors have declared no competing interest. [1]: #ref-14 [2]: #ref-9 [3]: #ref-5 [4]: #ref-18