Hacker News Reader

ok, that is a bit creepy. Wondering when the bandwidth gets to the point where you can send multivalued data.

The bandwidth of TMS is extremely low and limited by physics. To get high bandwidth, you'd need implantable devices for stimulating small groups of neurons, which is possible today, but not ethical with human subjects.

Do you think bandwidth is a limiting factor here in practice? FiOS, for example, has plans ranging from 25 to 75 M bits/sec.

My impression is that having a sensing apparatus to sense and process would be more of a limiting factor. A skull sensor that detects, say 20K neurons, with 1 ms resolution would only use 20M bits/sec.

With 20Mbits you would get only 1bit resolution, and since we're talking about analog data that's not enough. If we went to 12bit ADC we would need 320Mbits/sec.

You make a good point about needing more bits of resolution, but that's only a factor of ~10. I admit that I don't know the limits of the current sensors and rest of the system, but my guess is that they are not at the point where they are maxing out the bandwidth.

I expect that a full-bandwidth continuous surface-level brain scan would exceed traditional bandwidth.

Reply on news.ycombinator.com