ConFusion 2015, Panel 2: Science vs. Fiction

John Scalzi interviewed Dr. Cynthia Chestek about the state of human/machine neural interfaces.

This was a fun panel. I had heard of Scalzi’s legendary moderating style, and here he did not disappoint. He laid out the rules right up front and whenever someone from the audience tried to speak out of turn, he pounced!

The content of the panel was mostly Dr. Chestek talking about her work in the field of neural interfaces. In particular, implanting extremely fine wires in human brains to allow them to control machines or prostheses. Some random points from the talk

* The wires are very fine, on the order of 8 microns in diameter. For comparison, a human hair is between 40 and 50 microns in diameter. The wires need to be so thin so that they can match up with individual neurons, and thus receive discrete signals. Dr. Chestek said they slide into brain tissue “like butter”.

* The bit rate across one of these wires is very slow – in the neighborhood of four bits a second. This rate is per wire, though, so the more wires attached to a brain, the more information can be gathered. The down side is that for and significant amount of resolution and control, you need to have, at minimum, hundreds of wires.

* The advantage of using individual wires attached to neurons as opposed to, say, a mesh laid across the surface of the brain to pick up electrical signals is this: Given the complexity and density of the brain, it is extremely difficult to target the exact neurons producing the electrical signals. Direct contact is much more precise than close proximity.

* Also, the contents of the brain are not rigid. The brain itself is quite soft. It has “plasticity”. Neurons move around; not a lot, granted, but when your work is measured in microns any movement at all is huge. This increases the difficulty of mass-producing an implant by several orders of magnitude.

* Dr. Chestek was very firm on the point of “prosthetics, NOT augmentation”. For her this is an ethical point. Her work is meant to allow people to regain abilities they have lost through illness or injury. It is not meant to be used for powered battle exoskeletons or the like. Not that it ultimately couldn’t be used for such.

* One of the first uses might well be simple telepresence. Given the (relative) ubiquity of wireless broadband and cellular signals, a person confined to a bed could control a communication system of some kind in another room, or another building, or indeed almost anywhere on the planet. The more fidelity and the more senses engaged, the more bandwidth becomes an issue – not just in getting the signal from point A to point B, but in the amount of control a (for instance) completely paralyzed person has over the interface itself.

* John Scalzi’s latest book Lock In addresses some of these issues. Also, it’s a great read!

* I asked what the advantages were of inserting electrodes directly in the brain, instead of at the ends of the nerves in amputees. Dr. Chestek said she would be addresing these issues at another panel. Unfortunately I did not attend that panel.

That’s all I have for this panel. Click here for the introductory article of this series, as well as links to the other articles.

Leave a Reply

Your email address will not be published. Required fields are marked *