Neural dust · Grey Matter

Neural dust is the idea of reading the brain with many tiny, wireless, battery-free implants scattered through the tissue instead of one large wired array, and it is a research route toward recording from far more sites than today's implants reach.

Neural dust. Neural dust is the idea of reading the brain with many tiny, wireless, battery-free implants scattered through the tissue instead of one large wired array, and it is a research route toward recording from far more sites than today's implants reach.

Each mote is a sensor with almost no electronics: it picks up the voltage at its location and reports it by modulating energy sent in from outside. In the ultrasonic version demonstrated in 2016, millimetre-scale motes on a rat's sciatic nerve were powered by ultrasound and answered by changing how they reflected it (backscatter), with no battery and no wire. The radio version, neurograins, shrank each sensor to roughly the size of a grain of salt; in 2021 forty-eight of them were placed on a rat's cortex, powered and read by a thumb-sized patch on the scalp acting as a small base station that gave each grain its own address, and they both recorded and stimulated.

The appeal is scaling by number. Small separate sensors could in principle be spread over many areas without the bulk of one big array, with the network design supporting hundreds of grains and the stated aim of many thousands.

The limits are physical. Each mote must be powered through tissue, which caps the energy it can spend (heat budget); its signal must share a wireless channel with every other mote (recording bandwidth); and it must be placed and later found or removed.

It samples, it does not cover. Thousands of motes, or even millions, would still be far from tens of billions of neurons, so neural dust is a way to sample many more places at once.

Questions: What is hard about consent for implants scattered through the brain in hundreds? Neural dust and neurograins have so far been tested only in animals, so this is a question asked ahead of time. A single array can be removed if it fails or if the person changes their mind, while hundreds of grain-sized motes spread over the cortex would be hard to find, retrieve or switch off individually. Consent for such a system would have to cover removal and long-term support in advance, the point that reviews of neural device trials already make about post-trial responsibilities for ordinary implants.