I want a 1M channel amplifier
From Bayesian Behavior Lab
No seriously, why can we not have a 1M channel amplifier? I am just following up a discussion I had with @stochastician. Any input that people could give to help understand this would be awesome. And obviously my interests are related to the ongoing quest to record all neurons from a mouse brain .
State of the art in Neurophysiology
Intan (and many others) produce amplifiers that take neurophysiological signals and amplify/ digitize them. Here is an example. an intan amplifier They have a couple dozen channels.
State of the art in Imaging
An imaging chip pretty much takes 10 million weak inputs and amplifies each of them while reading it out in a massively multiplexed format so that we get high precision readings from each of the channels. Some such chips even do supercool spike coding (Tobi Delbruck).
Why the difference
Why can we not have digitizing amplifiers that have millions of channels? What if it would be worth a million dollars? There must be a fundamental reason why this isnt possible. Is there an issue with noise? How much more noise would such a high density amplifier have? Or is there an impedance issue? How? Somehow I do not understand that difference. Please email/facebook/tweet if you can help me understand.