Human brain connectivity geometry predicts cortical task transformations

A functional magnetic resonance imaging study of human participants performing 16 diverse cognitive tasks found that the dimensionality of intrinsic brain connectivity predicts representational changes across the cortex. Regions reached through lower-dimensional connectivity showed greater compression of task representations and increased similarity across tasks, whereas higher-dimensional connectivity was associated with representational expansion and conjunctive coding of task features.

The finding offers a systems-level account of how the cortex transforms information as activity moves between regions. Earlier work had shown that task representations are compressed in association cortex and expanded in sensory and motor systems, but the mechanisms underlying those patterns were unclear. The researchers asked whether the geometry of intrinsic connectivity constrains how task representations change between cortical areas.

The investigators combined functional MRI measurements from the 16-task experiment with activity flow modeling. Connectivity dimensionality served as the central measure of network organization. Lower-dimensional connectivity was associated with compressed representations and greater similarity among tasks. Higher-dimensional connectivity was associated with expanded representations and combinations of task features. These relationships were observed across the sensory–association–motor hierarchy, linking a property of network organization to differences in cortical representational geometry.

Activity flow modeling reproduced the observed compression-to-expansion pattern along that hierarchy. Importantly, the modeled representations more closely resembled the geometry of target regions than that of source regions. The authors interpret this result as consistent with transformation rather than simple transfer: the structure of a representation changes as it moves through a network, rather than merely being relayed unchanged from one region to another. This interpretation comes from a computational model built around the imaging results, not from a direct experimental manipulation of connectivity.

Within the study’s scope, the results identify network geometry as a candidate systems-level principle shaping cortical representations relevant to flexible cognition. The analysis does not demonstrate that changing connectivity dimensionality would cause a particular change in thinking, nor does it establish behavioral or clinical benefits. Its evidence comes from functional MRI acquired during 16 tasks and from activity flow modeling, so the conclusions are bounded by the experimental tasks, the measured population and the assumptions of the model. Even with those limitations, the study presents a coherent link between the dimensional structure of cortical connectivity and the compression, expansion and feature combination of task representations across the human cortex.

Subject of Research: How intrinsic connectivity geometry predicts representational transformations across the human cortex

Article Title: Network geometry shapes the transformation of task representations across the human cortex

Article References: Chakravarthula, L. N. C., Ito, T., Tzalavras, A., & Cole, M. W. (2026). Network geometry shapes the transformation of task representations across the human cortex. Nature Neuroscience. https://doi.org/10.1038/s41593-026-02486-8

Image Credits: AI Generated

DOI: 10.1038/s41593-026-02486-8

Keywords: functional MRI, human cortex, network geometry, connectivity dimensionality, task representations, representational compression, representational expansion, activity flow modeling, conjunctive coding, flexible cognition, Network, geometry