Molecules in the brain's extracellular space diffuse 2-5 times slower than in free solution. This retardation is traditionally attributed to tortuosity — the winding paths molecules must take around cells. The geometry is complex, so diffusion is slow. It seems obvious.
Gresil et al. (arXiv:2603.18936) use near-infrared carbon nanotubes to track three-dimensional diffusion at nanometer resolution and find that the explanation is wrong, or rather, incomplete. At short length scales — below a characteristic structural scale of the tissue — diffusion is normal. Molecules move freely, as if they were in open solution. The hindrance only appears when molecules try to travel beyond this scale, crossing from one local space into the next.
Tortuosity is not a property of the tissue. It is a property of the measurement scale.
The transition is geometry-controlled, not stochastic. There is no need for anomalous dynamics, fractional diffusion, or scale-free processes — mechanisms frequently invoked to explain non-standard transport in biological tissue. The motion is locally Brownian everywhere. What changes is the geometry the molecule encounters as it tries to move farther. At short scales, the local space is open. At longer scales, the narrow passages between cells create bottlenecks that retard effective transport.
This connects brain tissue diffusion to the general physics of transport in porous media, where the same phenomenon occurs: local diffusion is normal, and macroscopic retardation emerges from the pore network geometry. The brain's extracellular space is, physically, a porous medium with cells as the solid phase.
The slowdown isn't in the medium. It's in the transition between scales. Measure at one resolution and you see free diffusion. Measure at another and you see hindrance. Both are real; neither is the whole story. The tissue teaches a different lesson depending on the ruler you use.