The acute-to-chronic gap
The field's biggest honesty problem is the distance between a beautiful saline or day-0 measurement and the reality of months in vivo. Chronic signal degradation and limited chronic recording longevity arise from protein adsorption and encapsulation slowly raising impedance, from coatings that delaminate, and from the tissue reaction closing in.
The foreign-body response is the real ceiling
Materials can soften the tissue reaction but cannot abolish it. Reactive gliosis wraps the site in an insulating glial scar, and the peri-electrode kill zone removes the very neurons you want to hear. The mitigations are real — drug-eluting bioactive coatings, ultrasmall and ultrasoft probes, anti-inflammatory surface chemistry, collectively foreign-body-response mitigation — but no material yet makes the brain treat an implant as self.
This is the deepest point of the track: below all the electrochemistry sits a biological ceiling. The best imaginable material still leaves an object in living tissue, and living tissue reacts. Progress here is measured in slowing the decline, not stopping it.
Sealing resistance and the intracellular frontier
How large a signal you record depends on the sealing resistance between the membrane and the electrode — how tightly the cleft is closed. In an idealised point-contact view, the recorded fraction of the transmembrane signal grows with that resistance:
Idealised point-contact model: as sealing resistance grows relative to the electrode impedance, the recorded amplitude approaches the full transmembrane signal — the promise of array-scale intracellular-quality recording.
Patch clamp reaches gigaohm seals and enormous signals but does not scale to thousands of channels. Nanostructures such as nanopillars and mushroom electrodes chase intracellular-quality signals at array scale, yet membrane resealing, gradual loss of intracellular access, and chronic cell viability remain unsolved. Truly chronic, array-scale intracellular recording is still an open frontier.
Recording and stimulation pull apart
A material tuned for low recording noise (capacitive, high area, low impedance) is not automatically the best for stimulation (high reversible CIC, tolerant of large potential swings) — and vice versa. A bidirectional BCI that both records and stimulates must either compromise on one material or use different materials at different sites, and interleaving the two operations forces read–write blanking constraints on the front-end.
Standardisation, manufacturing, and where it's going
Two unglamorous problems gate real progress. First, standardisation: without agreed protocols for CSC, CIC, impedance, and failure-mode reporting, the literature's numbers do not compare, and accelerated aging tests do not yet reliably predict multi-year behaviour. Second, manufacturing: nanostructured and polymer coatings must be made reproducibly at wafer scale with clinical-grade reliability — a very different problem from a hero device in a lab.
An honest reading of where this goes: not a single miracle material, but incremental, compounding gains — soft mixed conductors, bioactive surfaces, and smaller sites — converging with the other Vol III frontiers. Optically transparent graphene serves the optical interfaces; high-CIC coatings enable dense stimulation for visual and auditory prostheses; low-impedance sites make million-channel arrays thinkable. The direction is a slow migration from inert metal to a living-tissue-mimetic interface.