In mice lacking the Eml1 gene, cortical neurons that should migrate to the cerebral cortex during development end up in the wrong place — displaced beneath the cortex in heterotopic masses. The standard prediction is that misplaced neurons would be dysfunctional: the cortex is organized into layers and columns, and this architecture is believed to be essential for proper sensory processing. Neurons in the wrong location should lack the right connections, the right inputs, and the right context to do their job.
Published in Nature Neuroscience, the study showed that the displaced neurons retained their molecular identities, formed appropriate long-range connections, and exhibited coherent electrophysiological properties. They organized into functional sensory-processing centers that mirrored their cortical counterparts, with preserved somatotopic mapping — a spatial map of the body — and responsiveness to sensory stimuli. Most remarkably, when the researchers silenced the normal cortex, sensory discrimination was unimpaired. The heterotopic neurons were the primary drivers of the behavior. The misplaced neurons built a working brain from the wrong address.
The structural insight is about the relative importance of position versus identity. The cortex's layered architecture was thought to be constitutive of its function — the position determines the wiring, the wiring determines the computation, the computation produces behavior. The Eml1 result shows that position is not constitutive. Identity is. The neurons knew what they were supposed to be (molecular identity), found the right connection partners (connectivity), and performed the right computations (functional maps) — all without being in the right place. Position is the normal scaffold, but the neurons can build without it.
This does not mean cortical architecture is unimportant. The normal brain uses position to organize efficiently — layers and columns provide spatial structure that simplifies wiring. But when position is disrupted, the neurons adapt. The organizational principle is robust to positional error because the principle is encoded in the cells, not in their coordinates.
The implication for neuroscience is that the brain's functional unit is the connection, not the location. Two neurons that connect correctly compute correctly, regardless of where they are. The atlas — the detailed map of which neurons are where — is a description of the normal case, not a specification of the necessary case. The map is typical, not required.