"r norman" <rsnorman_ at _comcast.net> wrote in message
news:m07lav4v8sn776q16k86qhjsm1k5k9fcn5 at 4ax.com...
| On 26 Apr 2003 07:10:09 -0700, y.k.y at lycos.com (yan king yin)
|| >I have another question which is related to making brain-computer
| >interfaces. Suppose I make some nano-scale tubes of diameters on
| >the order of 10-100nm (assuming we know how to do that). Inside
| >these tubes would be an ionic solution. The tubes contain no
| >ionic channels; basically they are closed. Will these tubes
| >conduct electric signals efficiently? Given that ions in solution
| >are different from electrons in metals, what would be the
| >differences between these tubes and electrodes? In terms of
| >conductance, and frequency-dependent impedence?
| >(Another assumption is that the surface of the tubes are
| >electrical insulators -- but this is probably not true
| >because carbon nanotubes, for one, are known to be conductive.)
| >Thanks =)
|| I am not sure what you are really trying to accomplish here.
| To build a brain-computer interface, it is sufficient to put
| electrodes into the proper locations. The electrode are
| interfaces between two conducting media -- the metallic (or
| semiconductor) materials that connect to the computer and that
| function with mobile electrons (or "holes"), and the saline
| of the brain that conduct with ion movements.
|| There are two types of electrode. In one, there is a redox
| that occurs at the surface of the electrode translating electron
| movement into ionic movement. The tradional system for
| work is to use a silver wire electrode covered with insoluble
| chloride. Then at the interface, Ag(s) gives up an electron to the
| metallic side and binds a Cl- from the solution to form AgCl(s).
| the other electrode, the reverse reaction accepts an electron from
| metal to reduce AgCl to Ag(s) and release Cl- into the solution.
|| In the other kind of electrode, there is no redox reaction at the
| electrode surface. Instead, there is an electrical bilayer,
| essentially a capacitance, so that electrons arriving at the
| in the metal accumulate a negative charge on the electrode surface,
| driving away Cl- ions in the solution and attracting Na+ (assuming
| extracellular NaCl solution -- other ions are involved inside the
| cells). Similarly, electrons moving in the other direction allow a
| positive charge to accumulate on the electrode and drive the ions
| solution the other direction.
|| The first, reversible, electrode is suitable for recording and
| stimulating all frequencies down to DC or steady currents and
| potentials. In order to record a cell resting potential, you need
| system that works this way. Similarly, in order to voltage clamp a
| membrane, you need this kind of electrode to be able to pass steady
| currents for some substantial time period.
|| The second kind of electrode is suitable only for AC stimulation
| recording. It acts like a capacitor in series with the circuit.
| lower frequency limit depends on the details of electrode size and
| construction. However, this is perfectly adequate for measuring
| potential changes associated with action potentials and, in many
| cases, even synaptic potentials. It is also quite suitable for
| stimulating neurons with brief "biphasic" pulses -- that is, pulses
| with no DC component which is the usual method of stimulation.
|| The type of electrode you need depends also on the stability of the
| electrode over time. The electrodes with the reversible redox
| reactions tend to be unstable over long periods of time and also to
| potentially toxic or damaging to nerve tissue. "Inert" materials
| suitable for long-term implantation are inert specifically because
| they dont react! Therefore virtually all long term electrodes work
| the second way. I really am not sure just how semiconductor
| electrodes work or their long-term stability and toxicity (I mean
| growing tissue culture neurons directly onto semiconductor chips or
| implanting naked silicon devices directly into the brain). Still
| mechanism of transduction with these interfaces must be one of the
| methods described above (or, more realistically, some combination
| the two mechanisms).
|| Carbon nanotubes would be no different. No matter what the
| method used inside the nanotube, the interaction between the
| "electrode" and the brain will either be through a redox reactions
| a capacitative process at the interface.
|| To see whether any scheme will work, you replace the biological
| with an equivalent electrical circuit (a voltage source with a
| characteristic source impedance). Your electrode also has an
| equivalent electrical circuit (ordinarily a series resistance with
| stray capacitance to ground). Finally, the electrode-solution
| interface has an equivalent circuit (a series resistance and
| source for the reversible electrode, a series capacitance for the
| non-reversible). Add the electrical circuit for your stimulating
| recording device and you can easily figure out just how well the
| system will work.
As I've discussed in the past, there are enormous problems inherent
in artificial brain interfaces.
They virtually always lesion their implant sites, degrading
But, more importantly, they do not replicate the brain's plasticity,
so their relatively-static functionality tends to 'take over control'
of the brain because the brain cannot extend its global TD
E/I-minimization into them.
It's probable that, in the future, these problems will be overcome
via strictly-Biological approaches - but even these will
=necessarily= be governed by global TD E/I-minimization, or they'll
simply not work.
Because 'the brain' is such an awesomely-integrated organ, it's
functioning can only be 'tweaked' via globally-integrated means which
preserve globally-integrated TD E/I-minimization as it's discussed in
But, why 'engineer' foreign stuff into an organically-intact instance
of 'the brain' when everything that's necessary for its
're-engineering' is already in-there and functioning robustly?
Within an organically-intact brain, 'normal' learning is already
capable of accomplishing everything that any proposed 'implant' can
Folks just haven't discerned the True-Wonder stuff that's innately
in-there, so folks've been allowing it to 'slip-by', unrecognized,
Implants for organically-damaged brains should be explored, but, to
the degree that they do not replicate 'normal' plasticity [to the
degree that they do not replicate globally-integrated TD
E/I-minimization] they'll always be encumbered, and encumbering, as
ken [K. P. Collins]
"Schmitd! Schmitd! Ve vill build a Shapel!"