Immortality

[Izawwlgood]

I'm not claiming that we know everything about the brain. I'm pointing out that reading a couple papers about how telomeres may protect dividing cells from being unable to replicate the ends of chromosomes in some species doesn't mean that elongating telomeres is going to mean everyone can live forever. Especially considering neurodegeneration having zero links to telomere length. So, while yes, rampant telomerase activity has been implicated in some types of cancer, we cannot safely assume the opposite, that is, cell or organism death is ONLY a result of shortening of telomeres.

Point being that being confident of a lot of conclusions at the moment feels a tiny bit reminiscent of Max Plank's physics professor telling him that physics was a "dead field" and that "pretty much everything has been figured out".

Yes, rather like you asserted:

However, there is no fundamental physical reason humans can't have perfect dna replication until the physical resources are no longer left due to entropy in billions of years (assuming this is, in fact, the inevitable fate of the universe). Point is, is there a fundamental physics reason for ageing? Nope, none whatsoever. So now we have to deal with this just in terms of engineering practicality.

Which is frankly, one of the most patently absurd claims I've ever seen in biology, next to the whole 'microtubule and actin filaments convey quantum data which is what leads to consciousness'.

As an aaside FP, this is the second thread you've launched into a bit of a tirade about telomeres in. Why?

[idobox]

I'm fairly new to neurobiology, so can you linky me to an example of a neuron that potentiates and fires solely with the use of ions? It is my understanding that channels (which allow the flow of ions for potentiation) are entirely predicated on neurotransmitter binding (or in some cases, other stimuli like pressure, heat/cold and various other compounds), which can raise the action potential of the neuron such that it fires. Firing produces a wave of depolarization, and at the synapse, causes the release of neurotransmitters.

AFAIK, there is no neuron that releases ions which somehow directly potentiate the recieving neuron.

There is at least one type of synapse, called electric synapse, that uses ion channels from one cell to the other. But that's not the question

In most exitatory synapses, the presynaptic neuron will indeed release neurotransmitters in the cleft, that will make ion channels open. The depolarisation will then make voltage dependant ion channels open, and create an action potential. So yes, neurotransmitters are implied, but the heart of the mechanism is ionic currents, and everything that can cause a current through the membrane can replace them.
Of course, it's more difficult with neurotransmitters that act by changing the behaviour of other neurotransmitters, or by causing chemical changes in the cell, that's why microfluidics will probably be required for certain task. Neuromodulators and neuro hormones, in particular, will need this kind of technology.

It's not as difficult as it seemed to me, since people are actually pretty close (still one or two decades from implantation in humans,a t least) to make working implants that replace central parts of the brain.

There are implants that deliver metronomic current to different parts of the brain, kind of like a pacemaker, but they're pretty radical and only for things like non-drug responsive epilipsy. Sticking a few electrodes in a brain and pulsing a small current is vastly different from claiming we can rewire or replace parts of the brain.

Again, I think you're understanding the level of complexity involved here on every level; from the organization of the network to the molecular diversity each component of the network responds to and outputs.

That's exactly what people are trying to do with hippocampal prostheses. A circuit that's supposed to replace the whole hippocampus, with cell-to-cell inputs and outputs. And they're experimenting it on animals right now.

And direct stimulation of the visual cortex has been used to restore a limited sight in humans, it uses a matrix of electrodes/pixels. It's nothing like replacing a part of the brain, but neurostimulation is not limited to mono-electrode deep brain stimulation.
You can also buy retinal and cochlear implants that stimulate the retina or accoustic nerve to transmit visual or auditive information to the peripheric nervous system, or brainstem implants that directly transmit sound signals.

This technology is far from being able to perfectly replace pieces of the brain, but we just need to improve it, and to better understand the parts we want to replace. The only part that will be significantly different in far-future neural prostheses will be the electrodes, the rest will be pretty similar, the way the first transistor-based computers are similar to the latest generation of smartphones: the same technology with decades of incremental improvement.

[WarDaft]

Yes, rather like you asserted:

However, there is no fundamental physical reason humans can't have perfect dna replication until the physical resources are no longer left due to entropy in billions of years (assuming this is, in fact, the inevitable fate of the universe). Point is, is there a fundamental physics reason for ageing? Nope, none whatsoever. So now we have to deal with this just in terms of engineering practicality.

Which is frankly, one of the most patently absurd claims I've ever seen in biology, next to the whole 'microtubule and actin filaments convey quantum data which is what leads to consciousness'.

As an aaside FP, this is the second thread you've launched into a bit of a tirade about telomeres in. Why?

One obvious argument against aging... if cells were totally unable to reset the process, children would be born old. So there is definitely some biological process by which things can be reset. Maybe it's not less complicated than re-buillding the human body, but still no particular reason for it to be literally impossible. In fact, there's probably evolutionary precedent for groups which (all other things the same) had older members die off faster adapting to new circumstances faster. Imagine if people didn't die of old age. That many of the people who were around in BC were still alive, and possibly even in power (perhaps due to say thousands of years of advantage in arranging political allegiances.) How quickly do you think society would advance in such a world? A species where the un-adapted stick around to pollute the gene pool is one that will adapt much slower and so succumb quickly to the forces of extinction.

[Izawwlgood]

There is at least one type of synapse, called electric synapse, that uses ion channels from one cell to the other. But that's not the question

Right, I think there is a channel that binds to ions to mediate the activity of the channel itself. But most channels rely on a compound binding to them to open or close their activity. You were proposing (I think anyway, correct me if I'm mistaken) a replacement neuron that has at one end a receiver of signal (which would probably be chemical), a radio emitter, and at the other end, a radio receiver, and then... something? What I'm trying to say is that replacing a neuron with an RF device is forgetting that both the reception and production of signal is mediated on chemical events, not electrical events. Your device is useless unless it can both recieve and release chemical signals. The axon replacement notion of this is fine; signal can be propagated by radio signal, laser beams, whatever.

This technology is far from being able to perfectly replace pieces of the brain, but we just need to improve it, and to better understand the parts we want to replace.

Hippocampus replacement devices aside (that's really cool), I'm inclined to believe you're still over simplifying the matter. Personally, I think we're a ways away from replacing parts of the brain; I think stimulating regions of the brain with electrodes is about as advanced as we'll see for a while.

@WarDaft: There are a couple of organisms that display individual cellular immortality. Some Cniderians (jellyfish) for example, not only exhibit cellular immortality, but lifelong cellular pluripotence. That is, you can take a jellyfish, that normally lives, say, 1 year before reproducing and then falling apart into guck, and you can shove it through a sonicator, dismantling it into individual cells. You can then put those cells into a goop, and they will resort into a new jellyfish. But (!!!), not all the cells go back to where they were originally; so, a tentacle may now be mantle, mantle may now be nerve ring, gonad may now be nematocyst (stinger). So that's pretty rad.

But humans don't work like that. One of the things we gave up when we developed specific immunity was the ability to wantonly regenerate pluripotent tissues; we're so bad at cellular division (compared to other organisms) that cancer was far more threatening than old age. Understand that by 'we', I mean 'vertebrates'. So our tissues fall apart over time. Mechanically, tissue replacement will probably greatly improve our lifespan, especially with the advent of induced pluripotent stem cells (as we already discussed). Truthfully, cancers and organ failure are problems with simpler solutions. Early detection and replacement of cancerous tissues is largely a mechanical issue, an issue we still have lots of work to do on, but one that I feel is reasonable to imagine solved in our lifetime. But that doesn't stop the incredibly looming issue of neurodegeneration, a problem that we don't fully understand, and don't have any real solutions for, nor any early detection methods.

Of course, I'm probably biased towards seeing increased importance in the thing I work on.

[gmalivuk]

Imagine if people didn't die of old age.

Until relatively recently, very few people lived long enough to die of "old age" in the first place.

[Izawwlgood]

@idobox: I'm reading more, and just came across voltage gated ion channels. So, yes, you're right that there is a class of channels that specifically responds to nothing more than charge.

[idobox]

@idobox: I'm reading more, and just came across voltage gated ion channels. So, yes, you're right that there is a class of channels that specifically responds to nothing more than charge.

Outside the synapse, they are the most frequent type, and they're the ones who allow action potentials to exist and propagate.

You were proposing (I think anyway, correct me if I'm mistaken) a replacement neuron that has at one end a receiver of signal (which would probably be chemical), a radio emitter, and at the other end, a radio receiver, and then... something? What I'm trying to say is that replacing a neuron with an RF device is forgetting that both the reception and production of signal is mediated on chemical events, not electrical events. Your device is useless unless it can both recieve and release chemical signals. The axon replacement notion of this is fine; signal can be propagated by radio signal, laser beams, whatever.

Using one RF emitter per pseudo neuron would be absurdly innefficient.
My idea is to use electrodes or chemosensors (like chemFET) in one site to capture local activity, transmit that info to a processing unit, and then transmit stimuli to one or multiple excitatory electrodes arrays. The connections between each block can be made by wire, or could use different forms of wireless, including RF, to avoid having to put cables in sensitive areas.
Today, electrode arrays are usually 1 or 2D, but nanotechnology will probably allow us to build 3D ones someday, with spikes of different length, which would allow us to record activity of the 6 layers of the neocortex separately, for example.
Many neurons release only simple excitatory or inhibitory neurotransmitters, and those we can emulate by simple electrical currents, and its done with a lot of succes in many cases. Emulating neuromodulators, neurohormones, and other substances that do things other than change the immediate post-synaptic polarization will be much more difficult, and will probably require micro(nano?)fluidics to release said substances.
An alternative solution would be to grow the type of neurons you need on an electrode array, and rather than try to grow axons through the brain, implant that device directly where axons are supposed to project to, and use the electrodes to excite the grown cells and release the neurotransmitters you need where you need it.

[Izawwlgood]

Yeah, I was understanding your proposed device correctly. My point was that all you've really done is come up with a way to replace the axon, not the cell body or the synapse.

Outside the synapse, they are the most frequent type, and they're the ones who allow action potentials to exist and propagate.

I've been talking about the synapse though this whole time; propogation of signal down the axon is simple, and can be replaced by wire. It's receiving signal at the cell body, or continuing signal at the synapse that cannot be handled by your device.

Microfluidics only partially addresses this matter; you still need to synthesize the NT that is released, unless you want a very short lived cell. For what it's worth, I've heard neurons described as professional secretory cells - being only interested in the axon is missing the bulk of the cool activity that they actually get up to.

[idobox]

For some reason, you seem to think I obsess over the axon.
When you model a neuron, the value you cannot ignore is the transmembrane potential. Since the very beginning (Hodge et Huxley), people have been studying the electrical properties of neurons. Intracellular electrodes can replicate the behaviour of Ach and GABA with great accuracy, and in most models of neurons, synaptic influence is modelized as transmembrane currents.

Depending on the NT you need, supply might not be a problem. Stuff like NO and CO that you can easily produce, and very potent ones, where delivered doses are tiny.
For the other ones, a good solution could be to grow secretory cells and harvest them within the body, and use microfluidics to deliver them. It seems extremly difficult to do, but compared to making new dopaminergic neurons from the dorsal striatum to grow axons to 7 or 8 different places without connecting to other unrelated parts, or other neurons growing new axons, it seems easier.

[Angua]

Hodgekin and Huxley were mainly looking at propagation though. Neurotransmitters are what create the action potential within the neuron - things like ACh open channels which change the membrane potential enough for the voltage gated channels to open. However, there is a lot of spatial and temporal summation (where different signals are required for a nerve to fire or not) which will be hard to replicated by solely giving the target cells voltages - especially the inhibitory neurotransmitters (eg GABA) which act by lowering the resting membrane potential so it is a lot harder for the activating signals to open enough ion channels to get past the threshold potential required to kick start an AP. You also have to factor in feedback signals from the target cell which act to modulate what is going on as well.

The AP is an all or nothing signal, so a lot of stuff goes into determining when exactly an AP will fire.

[idobox]

Neurotransmitters are what create the action potential within the neuron

Ionic pumps create the polarization, voltage gated channels create the AP, and a number of process can provoke it, including, but not restricted to, NT and transmembrane currents.
Simple NT, like Ach and Gaba open ionic channels, creating transmembrane currents. You don't really need the NT, just the current.

However, there is a lot of spatial and temporal summation (where different signals are required for a nerve to fire or not) which will be hard to replicated by solely giving the target cells voltages - especially the inhibitory neurotransmitters (eg GABA) which act by lowering the resting membrane potential so it is a lot harder for the activating signals to open enough ion channels to get past the threshold potential required to kick start an AP

That's why you use currents, and not voltages. The voltage is the result of intrinsec processes, like ionic pumps and neurohormones, and synaptic activity, all of which are merely currents. Since it's ionic currents, you have to consider all ionic species separately because of osmtic pressure, but it's still just a current, and can be emulated by a current.

You also have to factor in feedback signals from the target cell which act to modulate what is going on as well.

If the feedback happens directly in the synapse, you have to modelize it in the shape of the current pulses you send. If the feedback is more complex, through another synapse for example, then you obviously need to measure that feedback and use that as an input of your processing unit.

The AP is an all or nothing signal, so a lot of stuff goes into determining when exactly an AP will fire.

Not really, it's only true for certain neurons, but it makes a good first level approximation. There are neurons with subthreshold oscillations, variable thresholds, AP of variable lengths, strengths and shapes.
I suggest reading Izhikevich if you really want to learn how to modelize all neurons. To put it simply, neurons are electro-chemical dynamic system that can be modelized with purely electrical models.

NT are the usual way for neurons to communicate, it doesn't mean it is the only possible one, or that it is the best one for something engineered by humans.

On the separate note, this discussion is drifting quite far away from the OP. Should we split threads?

[Angua]

How are you going to get the current to the cells your chip is going to be interacting with at a fine-tuned enough level not to damage the nerve (you're definitely going to need a way to make sure that it can detect what the target nerves are telling it so it knows when they've had enough). The ion currents are also not the only things the NT receptors affect - you're ignoring all the modulation of NTs that happens within the synapse (eg the cells will up/down regulate NT receptors depending on how often they are activated, other signalling pathways that can be activated (or inactivated) by NT receptor activation, the effects of the glial cells in modulating the NT transmission within the synapse, etc).

[Izawwlgood]

Simple NT, like Ach and Gaba open ionic channels, creating transmembrane currents. You don't really need the NT, just the current.

This is why I kept disagreeing with you; you don't need NT for signal propagation along an axon, but you do need NT to transfer signal to the next neuron. Post synaptic activity is entirely predicated on NT release and uptake by the next neuron in the circuit. You keep talking about ways to model/replace individual neurons, forgetting that signal propagation down an axon is pointless if you can't induce a synaptic bulb to release NT onto the next neuron; saying 'microfluidics will deliver small amounts of NT' forgets that neurons have a lifespan of... a lifespan. How do you propose replacing a lifetimes worth of neurotransmitter?

To put it simply, neurons are electro-chemical dynamic system that can be modelized with purely electrical models.

"To put it simply, just assume a spherical cow"

[Angua]

Well, you can start it firing by injecting current directly into the cell (though I don't think cells take that kindly to having things poked into them).

[idobox]

This is why I kept disagreeing with you; you don't need NT for signal propagation along an axon, but you do need NT to transfer signal to the next neuron. Post synaptic activity is entirely predicated on NT release and uptake by the next neuron in the circuit. You keep talking about ways to model/replace individual neurons, forgetting that signal propagation down an axon is pointless if you can't induce a synaptic bulb to release NT onto the next neuron; saying 'microfluidics will deliver small amounts of NT' forgets that neurons have a lifespan of... a lifespan. How do you propose replacing a lifetimes worth of neurotransmitter?

If you stick one electrode in the presynaptic neuron, one in the postsynaptic neuron, assuming you understand the normal dynamics of that kind of synapse (true for some NT, not for others), you can cut the axon, and use a microcontroller to shape pulses sent to the post-synaptic cell depending on the potential of the pre-synaptic one.
How much serotonin does your brain produce every year? I have no idea, but I suppose it is not on the order of liters. Remember it's a very potent neurohormone, you don't need a lot of it. Let's say a container of 10ml lasts for 5 years (no idea what the real figure is, even the order of magnitude), you would need an operation to "refuel" from time to time. It's not ideal, I agree, but early pacemakers had the same kind of issue, and it's much easier than growing an axon from the midbrain to both the cerebellum and the frontal cortex.

To put it simply, just assume a spherical cow

You can't predict anything useful with spherical cows in a perfect vacuum, except maybe orbits. Electrical models of the neurons work, make predictions, and are used by everyone that is not studying a single molecule.

How are you going to get the current to the cells your chip is going to be interacting with at a fine-tuned enough level not to damage the nerve (you're definitely going to need a way to make sure that it can detect what the target nerves are telling it so it knows when they've had enough). The ion currents are also not the only things the NT receptors affect - you're ignoring all the modulation of NTs that happens within the synapse (eg the cells will up/down regulate NT receptors depending on how often they are activated, other signalling pathways that can be activated (or inactivated) by NT receptor activation, the effects of the glial cells in modulating the NT transmission within the synapse, etc).

Electrical stimulation of single neurons, or small groups of neurons, is currently used. In humans. For years.
If you know the dynamics of NT receptor self regulation, you can emulate it by modulating your pulses.

And I already said currents can't emulate all that is not directly affecting the potential of the cell. Obviously, we can't replicate the effect of dopamine or serotonine with electrodes, and that's why I suggest microfluidics.

I'm not saying my solution is easy or perfect, just that it is much simpler than growing new neurons. Neurostimulation has been applied since what, the 60's? with very crude effects that have been refined to the point people are actually considering replacing the hippocampus.
Growing neurons, on the other hand, is something we have no idea how to do. And when we figure that out, we will have to figure out how to direct the growth of the axon, make it split, direct the growth of branches, same thing with the dendrites, all that through a fully grown brain, tightly packed with axons and other stuff, with similar neurons that will have NOT to be disturbed too much by the process.
Repairing the brain will consist largerly in rewiring stuff. A biological process is stochastic, and it is difficult to direct it without direct mechanical intervention.

The brain is not the skin, it's not renewing itself on a continuous basis, and there is no reason to believe it has mechanism to repair eveything magically if we drop stem cells in the right place. We aren't able to repair organs with a much simpler structure, and far less different types of cells. The day we are able to grow a liver or a pancreas from stem cells, with arteries and all, without using organ scafolds and such; this day only, biological solutions will have a chance to compete with engineered systems.

[Angua]

All of the implants people use at the moment are pretty crude (also, are people actually considering replacing the hippocampus with something that will do more than just stimulate things without knowing much about what you're stimulating? - the mechanisms behind how DBS works isn't know, and with the cochlea we were lucky that frequency maps so readily along it). I'm not sure how long the longest one has been inside a person for though (and these don't work by cutting the axon which could be dangerous given the necessity of retrograde transport of neurotrophic factors for the neurons to stay alive!). If you're going to have these things big enough to have different NTs inside them, as well as being able to detect and respond to other NTs given to them, then it might be a problem. How is replacing the hippocampus supposed to word - will it be as plastic as the normal one is?

Also, growing new neurons is something that is actually being done - I personally would put my money on that before these implants.