The Six Layers of a Neural Interface

Building a neural input wristband isn't one discipline — it's six interdependent ones, stacked. Understanding the stack explains why so few companies ship a working product, and why a decade of vertical integration is hard to replicate.

At the base is band design and materials: a bracelet that's bendable, durable, biocompatible, and comfortable for all-day wear, while housing flexible electronics that survive bending and elongation. Above that sit the sensors and electrodes — the electrode geometry, materials, and arrangement that determine how much signal the device can actually capture from the skin. Then the electronics: a miniaturized flex-rigid PCB that fits a watch-strap form factor.

On the software side comes the algorithm and signal-processing engine — the models that turn a noisy bio-signal into a reliable command, running on low-power hardware with limited CPU and memory. Above that, the software and applications layer: the SDK, the tools, the integrations that let other people build. And at the very top is humanware — the human facet of the design: which gestures map to which functions, and how natural and intuitive they feel to perform.

The reason this matters is interdependence. A decision in the electrode layer changes what the algorithm can do; a constraint in the algorithm layer changes the form factor; a form-factor choice changes comfort, which changes whether anyone actually wears it. You can't optimize one layer in isolation. Getting a neural interface to work means getting all six right together — which is why the discipline is as much systems engineering as it is signal processing.

This is the unglamorous foundation beneath the product. It's also the moat: the expertise compounds across layers and across years, and it doesn't transfer easily to anyone starting from a single layer.