ok, so I am total newby, never got further than high school biology. Question:

what exactly is the mechanism behind a neuron's decision to fire or not to fire due to a particular event involving signals from some subset of incoming synapses? So suppose there are 100 incoming synapses, and maybe 20 of them fire in some specified way. How does this neuron decide whether to become activated or not?

innervating axons (inputs) release neurotransmitters when an action potential invades their terminals, resulting in presynaptic release of neurotransmitter into the synaptic cleft, which binds to postsynaptic receptors, causing changes in ion conductances in the postsynaptic neuron, which usually causes depolarizations or hyperpolarizations in said neuron. If the level of depolarization at the soma or axon hillock exceeds threshold for spike initiation, then that neuron fires an action potential (which itself, is just non-decrementallly propagating ion conductance change). Ultimately, it all comes down to ion conductances.

ok, so is each synapse equivalent? E.g. when we have one neuron firing. Suppose it has 10 outgoing synapses. Does it then release neurotransmitters to all 10 clefts?

And is the releasing of neurotransmitters essentially an on/off operation, or can it release more/less neurotransmitters, or maybe several different neurotransmitters depending on circumstances?

no. Release is a probabilistic event, in part, because the action potential does not propagate down every axon brach due to propagation failures (which commonly occur at branch points in the axon).


It's not a binary on/off operation. There are presynaptic autoreceptors that detect quanta release from their terminal and produce feedback effects. There is also vesicle depletion when quanta are released at high frequency. Also, it's no uncommon to have peptides released along with smaller neurotransmitters. If this sounds uncomfortably complicated, it's even more so when the diversity of postsynaptic targets are considered.

is there some organism or a subsystem of an organism (like maybe a single human finger or something) for which the behavior of the neural system is understood so well that it can be accurately computer simulated? Or is the research so far limited to neuronal topology and physiology of separate neurons, not on the system's level interactions?

On a related note, I understand that neuron polarization/depolarization can be tracked at least in some cases, but can neurotransmitter releases into particular clefts be tracked?

the complete nervous system of the worm, C. elegans, which contains about 302 neurons, has been ultrastructurally mapped (i.e., using electron microscopy) at synapse-resolution (published by White, et al in 1986), and so this simple nervous system presents an ideal 'toy' case for modeling. Interestingly, I have seen few complete simulations of C. elegans, even though the complete ultrastructural mapping of its nervous system has been around for over 20 years. There is definitely interesting simulation work that could be done with C. elegans, and the fact of the matter is that if you can't accurately simulate the super-simple nervous system of C. elegans, then you have no hope of accurately simulating more complicated mammalian nervous systems.

If a computer model of the C. elegans nervous system were to accurately replicate the worm behavior, that would prove that there is no complex hidden functioning of neurons. That is, it would disprove the Stuart Hameroff hypothesis of orchestrated objective reduction in microtubules (an internal complex function of neurons).

I don't think there are tools for resolution into individual clefts yet?, but there are techniques for measuring neurotransmitter release with either good time or substance resolution in vivo- chronoamperometry and microdialysis.

a synaptic cleft is 10 nm wide, so you need electron microscopy to visualize it, which means you can't view events, like transmitter release into the synaptic cleft, in real time.

how about tracking of neuron polarization/depolarization? Suppose we have a bunch of neurons intertwined in a motor nerve, let's say in a reptile's leg. Are there ways to track the firing of these neurons, e.g. when the reptile is doing something?

2-photon and optical microscopy using voltage-sensitive dyes lets you track neuronal activity (over a relatively small volume of tissue).

are these studies with "2-photon and optical microscopy" very expensive? Or straightforward and cheap? In terms of what is actually happening in the discipline, is studying neural activity with resolution to a single neuron a significant area of research now or is it essentially impractical? What is the typical resolution level in the study of neural system outside the brain?

it's very expensive and technically demanding, and few universities have the facilities. It's at the cutting edge of neuroscience. To my knowledge, these techniques have not been applied outside the brain as it is not really of interest unless you were looking at the spinal cord.

how do you go about applying this technique in the brain? Is it done with brain tissue extracted onto a dish and kept alive for awhile? Or is there a way to observe a part of an intact functioning brain?

both. Pubmed: "2-photon microscopy"

Why need very accurate simulate neurons in C. elegans? And that would be if not very accurate? And what scientists was hoping from this simulation? Still need connect at least virtual "legs" to C elegans and if they don't do legs to C. elegan then how they suspect to understand that simulation was successful or not?
And I can't understand why for each sinapse need each individual weight? If many synapses going from one neuron then this neuron must fireing in all his synapses connections with equal frenquency and weights, etc. Or transmision from fireing neuron depending on condition of neuron, which receiving signal?