Can the human mind be simplified to a binary model of ones and zeroes?

[KattyGenovese]

First of all, I apologize if this issue has been covered before.

Would it be accurate to say that, given that a neuron is either in a resting potential or action potential, that a binary model would be representative of the neural structure human mind, or would this be too simplistic of an analogy?

This is of course not to say that such a mechanism is solely responsible for cognition and behaviour--chemical (and non electrical) exchanges do occur between the synapses, for example. However, these actions are triggered by the current flowing to the synapses, and assuming that such cognition operates on a purely biological level, would it be accurate to say that at the lowest level of function, our minds fundamentally operate on ones and zeroes?

 

[apeiron]

Would it be accurate to say that, given that a neuron is either in a resting potential or action potential, that a binary model would be representative of the neural structure human mind, or would this be too simplistic of an analogy?

That's way too simple - but still a good question.

Remember that any living neuron is in fact always firing off action potentials, so what changes is the rate. This can be sped up or slowed down by feedback from other neurons.

Then timing is also crucial, as synchronising with other neurons is more important than just firing.

And the decision of neurons to fire is finely tuned to global attentional states, so that some of the time, a neuron may fire to anything that looks close to its preferred input, at other times (when we are attentively seeking something) it will only fire (speed up, get synchronised) to a far more narrowly specific stimulus. It depends what the other neurons it is connected to are telling it to do.

So the way to look at it is that there are some very basic digital-like features of neurons. Such as spike trains. And also the fact that they have "wiring patterns" which make discrete point-to-point connections. So yes, this makes them seem to be information processing components.

But then there is just a ton of complexity after that. And the language of dynamics may account for it better. Certainly, it is no longer clear that neurons are just doing "information processing" in any strict computational sense.

 

[Pythagorean]

Way too simplistic. There are many intrinsic currents underlying the action potential, and of course the action potential itself is not a step function, the activation function is closer to a hyperbolic tangent.

For a single neuron, this doesn't matter too much, but for large networks of coupled neurons (especially where gap junction coupling is highly expressed) it can make a big difference in the behavior that emerges.

 

[KattyGenovese]

Thanks for the replies.

Also, I'm a bit confused as to the exact meaning of the "all or none law". My textbook states that "either an action potential is fired, or it does not...and the action potential is always the same". Then again, it's a first-year course textbook. What exactly is "meaningful output" (aside from rate) given such a constant sized action potential?

Pythagorean's post seems to imply that the "slope" of the spike matters by stressing that it's akin to a hyperbolic function--as opposed to a stepwise function. So would that mean that at each potential level along the spike (e.g., -70, -40, +40, etc.) these differences generate different outputs?

@apeiron: What do you mean by "information processing"? Do you mean to say that neurons aren't input-processing-output units?

 

[apeiron]

@apeiron: What do you mean by "information processing"? Do you mean to say that neurons aren't input-processing-output units?

They are not signalling 1s by spiking and 0s when not spiking. They are not simple logic gates, despite their "integrate and fire" design. They are not even transforming input to output in a simple sense (a cortical neuron has far more feedback connections than "input" ones).

It is just like the workings of the eye. First thing you learn is how the eye is like a camera. Then you can spend the rest of a career why it is much more complicated than that.

 

[bobze]

Remember that any living neuron is in fact always firing off action potentials, so what changes is the rate. This can be sped up or slowed down by feedback from other neurons.

Sorry I just woke up and my brain isn't running up to full speed yet, so I may not be understanding you here correctly, but are you saying that neurons are only inhibitory here by "off action potentials"?

If so that isn't right, neurons can be inhibitory, excitation or even both at the same time depending on the populations of receptors they are innervating.

 

[Pythagorean]

bobze,
think apeiron meant "firing off" action potentials, not firing "off action potentials".

Pythagorean's post seems to imply that the "slope" of the spike matters by stressing that it's akin to a hyperbolic function--as opposed to a stepwise function. So would that mean that at each potential level along the spike (e.g., -70, -40, +40, etc.) these differences generate different outputs?

It depends on where you measure the "output". In a single neuron, downstream the axon, it shouldn't make a difference for a true all-or-nothing (integrate and fire) neuron (note that not all neurons are integrators, some are resonators; who knows what else is in there. Resonators can have a half-assed action potential).

But prior to the axon hillock, the initial current that was larger or smaller will add to the underlying intrinsic passive currents of the neuron, particularly with respect to reflection of potential waves upstream through the dendrites. (So signals coming from the dendrites, which is what you simulate with injected current, can 'reflect' off of the soma back upstream through the dendrites, the energy doesn't just disappear).

This is important for networks of neurons because you have now modified the interaction between the neuron and neurons upstream of it.

 

[apeiron]

Sorry I just woke up and my brain isn't running up to full speed yet, so I may not be understanding you here correctly, but are you saying that neurons are only inhibitory here by "off action potentials"?

If so that isn't right, neurons can be inhibitory, excitation or even both at the same time depending on the populations of receptors they are innervating.

Yes, I did say both sped up or slowed down.

 

[That Neuron]

The basic concept of an action potential is similar to that of a logic gate persay, but not exactly, because the charge of the interior of the upper axon determines whether the ion channel that allow sodium into the cell activates or "opens", this influx of sodium (which forms a positive ion) increases the positivity of the interior of the axon, which stimulates more channels allowing positively charged ions into the cell, once the total polarity of the interior becomes positive (flips) it deactivates the channels and activates potassium channels to allow positive potassium into the interstitial fluid surrounding the axon, and then the original sodium channels must go a period of deactivation (complexity is increased here) before firing again, of course the rate of firing is key in many neurological functions, however the part where I see complexity arising is during the synaptic-dendritic reactions that occur when increased positivity from the fired action potential causes the production of neurotransmitters that cause the lysogenic excretion of neurotransmitters into the synapse. The amount and type of neurotransmitter depends upon the state of the cell (an oxygen starved cell, or a cell that had undergone repeated firing would likely produce less neurotransmitter than one that had not). Since the amount of neurotransmitter and structure of surrounding neurons affects the consequent firing of connected neurons, this is a seriously complex system no longer described by the logic gate like scenario.