Neural network without neurotransmitters

[Secan]

Could evolution have produced neural network without chemical neurotransmitter (such as serotonin), like our computers? Are their organisms without neurotansmitters?

If neurotransmitter are indispensable. Why don't our computers have components or function similar to neurotransmitters?

 

[BillTre]

Could evolution have produced neural network without chemical neurotransmitter (such as serotonin), like our computers? Are their organisms without neurotansmitters?

There are electrical synapses between neurons.
They can either activate the post-synaptic neuron or inhibit it.
Activating electrical synapses can be retifying (current goes only in one direction).
The negative type electrical synapses are quite rare however.
Here are some pictures from the above article to illustrate how this works:

(See article for text).

Chemical synapses (that use neurotransmitters are much more common in known nervous systems).

If neurotransmitter are indispensable. Why don't our computers have components or function similar to neurotransmitters?

In theory they are not indispensable, but there are no known animals with just electrical synapses.

Computers don't use neurotransmitters because they work on completely different principles and are made of different components.

Electrical circuits work on currents moving around in wires.
Neurons work mostly by propagating transmembrane voltage changes along their membranes. This is not driving currents from one end of a neuron to the other. The membrane change may propagate the length of the neuron, but the currents only flow short distances across the membrane.
Circuits have nothing like neurotransmitters, the vesicles that contain them, and the vesicle release mechanisms of neurons.

 

[jim mcnamara]

I do not know of one. Not knowing fails to prove it cannot happen.

Jellyfish use neurotransmitters.

A very "early" neural system is what jellyfish have - a neural net, with no central brain. It is considered to have have evolved very early on. You can think of it as "simplest possible neural network" for the purposes here.
Even simpler and earlier are the action potentials in single celled beasties. The action potential is the basis for transmission in modern nerves

https://www.cell.com/current-biology/pdf/S0960-9822(13)00359-X.pdf

And I think your assumption: "computers could be the model for neurons" may not be very useful. It really is the other way around. Concepts like parallel programming - or multitasking- are easily seen in modern mammalian brains. For which the primary design evolved 500 million years ago, long before John Von Neumann came up the the cpu concept.

Please read:
https://en.wikipedia.org/wiki/Evolution_of_nervous_systems

Before you start speculating again. Thanks.

 

[atyy]

If neurotransmitter are indispensable. Why don't our computers have components or function similar to neurotransmitters?

It is hard to get a minus sign using gap junctions or ephaptic connections. Computers can do subtraction and also multiple by negative numbers.

As @BillTre writes above, it is in principle possible to get an inhibitory synapse without using neurotransmitters. One possible example of inhibitory ephaptic synapse (not confirmed) is the synapse from retinal horizontal cells to cones: https://en.wikipedia.org/wiki/Retina_horizontal_cell.

Another function that neurotransmitters allow is for the presynaptic neuron to affect each of its postsynaptic neurons differentkly (even though it releases the same neurotransmitters to all its postsynaptic neurons); this is because the receptor may be different in each postsynaptic neuron.

 

[Secan]

There are electrical synapses between neurons.
They can either activate the post-synaptic neuron or inhibit it.
Activating electrical synapses can be retifying (current goes only in one direction.
The negative type electrical synapses are quite rare however.
Here are some pictures from the above article to illustrate how this works:
View attachment 268691
(See article for text).

Chemical synapses (that use neurotransmitters are much more common in known nervous systems).In theory they are not indispensable, but there are no known animals with just electrical synapses.

Computers don't use neurotransmitters because they work on completely different principles and are made of different components.

Electrical circuits work on currents moving around in wires.
Neurons work mostly by propagating changes in membrane along their membranes. This is not driving currents from one end of a neuron to the other. The membrane change may propagate the length of the neuron, but the currents only flow short distances across the membrane.
Circuits have nothing like neurotransmitters, the vesicles that contain them, and the vesicle release mechanisms of neurons.

I mean. Why does nature have to evolve such cumbersome method of impulse propagation using vesicles or neurotransmitters? Should other lifeforms in the universe have such too? Can organisms as complex like humans be made without these vesicles or transmitters too? If designing artificial sentient beings, should they have vesicles or neurotransmitters too and any such implementations or plans in neuroscience now?

 

[BillTre]

Why does nature have to evolve such cumbersome method of impulse propagation using vesicles or neurotransmitters?

Because it works and in the evolutionary lineage of animals on earth, this is what resulted to fulfill this function.

Should other lifeforms in the universe have such too?

There is no reason why not. However, it does not rule out something else.

Can organisms as complex like humans be made without these vesicles or transmitters too?

Maybe, maybe not.

If designing artificial sentient beings, should they have vesicles or neurotransmitters too and any such implementations or plans in neuroscience now?

Different and independent evolutionary paths could possibly take other courses. There are not clear conclusions.

 

[Pythagorean]

A conceptual discussion of neural processing often comes down to "gas" and "breaks". For example, Caffeine doesn't give you energy, it just diminishes a "break" in your brain. Chemical synapses allow this gassing and breaking to happen in a unidirectional way, which makes for a more stable control system.

From a general information processing perspective, this kind of unidirectional adding and subtracting allows you to make a lot of versatile feedback/feedforward loops in an ensemble of such processing elements. It may not be necessary (plenty of organisms live without brains at all) but it led to the neocortex, which has given us an undeniable competitive advantage against other animals - to the point of us endangering other populations and causing extinctions. However, depending on how you measure success, you may consider bacteria the dominant species - and they don't even have neurons.

edit: I should address gap junctions, which BillTre noted can be rectifying. The problem with gap junctions is that they're passive. They follow laws of charge diffusion to match the internal environment of the cell to the outside. Thus, their efficacy can be sensitive to global brain changes, such as peptide release or modulation. Chemical synapses also have refractory mechanisms. As a result of these two properties, chemical synapses tend to be more reliable in a wider range of "environments" (where now the environment is the extracellular regions of the brain outside the neuron)

 

[madness]

A conceptual discussion of neural processing often comes down to "gas" and "breaks". For example, Caffeine doesn't give you energy, it just diminishes a "break" in your brain. Chemical synapses allow this gassing and breaking to happen in a unidirectional way, which makes for a more stable control system.

I'm curious about what you mean here. Networks with symmetric coupling are usually more stable in the sense that they have an energy function (Lyapunov function), as in the Hopfield network. As soon as you make the couplings asymmetric you are no longer guaranteed that an energy function exists, which tends to make things more unstable (in the sense of no fixed points).

Stability in neural networks is more commonly attributed to feedback inhibition which counters feedback excitation, as in inhibitory-stabilised networks. Is that what you are referring to? Even in that case, many classical models ignore the separation into excitatory and inhibitory populations and they are still perfectly stable, even when they have symmetric connectivity (e.g., the famous Ben-Yishai and Sompolinsky model https://www.pnas.org/content/92/9/3844).

 

[Pythagorean]

Stability in neural networks is more commonly attributed to feedback inhibition which counters feedback excitation, as in inhibitory-stabilised networks. Is that what you are referring to? Even in that case, many classical models ignore the separation into excitatory and inhibitory populations and they are still perfectly stable, even when they have symmetric connectivity (e.g., the famous Ben-Yishai and Sompolinsky model https://www.pnas.org/content/92/9/3844).

This seems like a higher level discussion about processing, whereas my answer is framed more in morphological evolution. Chemical synapses (whether inhibitory or excitatory) fire in one direction and have a refractory period, which makes it an ideal processing unit in a "noisy" world.

 

[madness]

This seems like a higher level discussion about processing, whereas my answer is framed more in morphological evolution. Chemical synapses (whether inhibitory or excitatory) fire in one direction and have a refractory period, which makes it an ideal processing unit in a "noisy" world.

I'm not sure I would agree (or maybe I simply don't follow). Refractory periods are normally related to ion channels at the soma AFAIK. What I would agree with (if I understand you) is that there is a huge amount of biochemical machinery in a chemical synapse which can enhance computation, for example short term depression and facilitation, pre and post-synaptic plasticity, synaptic tagging etc. If I had to harazd a guess as to why nervous systems have synapses, it would be that they add extra computational power to the system.

 

[Pythagorean]

I'm not sure I would agree (or maybe I simply don't follow). Refractory periods are normally related to ion channels at the soma AFAIK. What I would agree with (if I understand you) is that there is a huge amount of biochemical machinery in a chemical synapse which can enhance computation, for example short term depression and facilitation, pre and post-synaptic plasticity, synaptic tagging etc. If I had to harazd a guess as to why nervous systems have synapses, it would be that they add extra computational power to the system.

Sure, I'm trying not to get in the woods and confuse OP - seems kind of derailing, but the point, in the frame of a discussion of "why chemical synapses?" is that the refractory period suppresses signal from chemical synapses. As opposed to, say, gap junctions which can passively change cell charge directly on their own.

 

[madness]

Sure, I'm trying not to get in the woods and confuse OP - seems kind of derailing, but the point, in the frame of a discussion of "why chemical synapses?" is that the refractory period suppresses signal from chemical synapses. As opposed to, say, gap junctions which can passively change cell charge directly on their own.

Refractory periods would be indifferent to whether inputs were chemical or electrical. They suppress the neuron from spiking, they don't suppress any particular input signals. Anyway, I agree best not to derail the thread any further.

 

[Pythagorean]

Recractory period suppress the neuron from spiking via some channels; electrical gap junctions are their own channels in the axon though.

 

[atyy]

@Pythagorean is probably using the word "refractory" for a synapse in the plain English sense. I guess he is referring to the depletion of synaptic vesicles or neurotransmitter, which may be a source of short-term synaptic dynamics?
https://cshperspectives.cshlp.org/content/4/8/a013680.full
http://www.scholarpedia.org/article/Short-term_synaptic_plasticity

 

[docnet]

@Pythagorean is probably using the word "refractory" for a synapse in the plain English sense. I guess he is referring to the depletion of synaptic vesicles or neurotransmitter, which may be a source of short-term synaptic dynamics?
https://cshperspectives.cshlp.org/content/4/8/a013680.full
http://www.scholarpedia.org/article/Short-term_synaptic_plasticity

More simply, refractory period refers to the time it takes for the membrane potential to return to its resting state.

 

[Secan]

Refractory periods would be indifferent to whether inputs were chemical or electrical. They suppress the neuron from spiking, they don't suppress any particular input signals. Anyway, I agree best not to derail the thread any further.

No worries. I am trying to understand the language of the brain and all the complexities of the neural network so need all the other facts too.

 

[madness]

I would sum up the ideas like this:

1) Synapses can have all kinds of complicated short and long term changes in their properties. We roughly call these short and long term plasticity. If a neuron fires a lot in a short period of time, it can undergo depression due to depletion of neurotransmitter (synaptic vesicles). But it can also do the opposite and facilitate, become temporarily stronger. As @atyy opinted out, you could call short term synaptic depression a refractory period, but it isn't commonly called that.

2) Neurons often have refractoriness in their spiking. Which means they basically can't be fired for a brief period after they last fired a spike. This refractory period is due to the kinetics of voltage gated ion channels in the membrane. From Wikipedia (https://en.wikipedia.org/wiki/Refractory_period_(physiology)#Neuronal_refractory_period) " The refractory periods are due to the inactivation property of voltage-gated sodium channels and the lag of potassium channels in closing. Voltage-gated sodium channels have two gating mechanisms, the activation mechanism that opens the channel with depolarization and the inactivation mechanism that closes the channel with repolarization. While the channel is in the inactive state, it will not open in response to depolarization. The period when the majority of sodium channels remain in the inactive state is the absolute refractory period. "

Because a neuron needs sodium channels to open in order to fire an action potential, I was/am confused by @Pythagorean's statement that refractory periods only apply to chemical and not electrical inputs.

*Edit: Neurons can also have all kinds of interesting short and long term changes in their behaviour other than refractoriness. Long term changes are often called intrinsic plasticity, and involve (for example) changes in the ion channels in the cell membrane. If you like dynamical systems and want to know about single-neuron models then this is a good (free) book https://www.izhikevich.org/publications/dsn.pdf

 

[Pythagorean]

Yeah, synaptic fatigue was probably kicking around in my subconscious, but with axonic gap junctions, the flood of current can be too much for the neuron to return to rest state, leading to "spikelets" and/or "ripples" [1], because gap junctions are promiscuous and let anything in in a passive way (unlike ion channels which are selective to particular ions). They also can let calcium and other ligands in, triggering second messenger processes.

So it's not the the refractory period isn't happening in the Na and K ions, it's just being washed out by the flood of current coming in axonic gap junctions - typically from coupling with neighboring neurons (a synchronization process). I vaguely remember that this can also lead to extra current adding up in the soma, but from the axon (through gap junction impulses) instead of the dendrites - I can't remember the name of these backward traveling waves. Theoretically, this could lead to interesting processing situations for memory (and such axoaxonic gap junctions are found in hippocampus [1].)

Anyway, my point being that chemical synapses, afaik, are always fixed behind the dendrite therefore always constrained by the soma's threshold and the refractory period of the Na-K channels and thus unidirectional. There's synaptic fatigue that allows for short term plasticity. Gap junctions tend to form between neighboring cells (even non neural cells) and tend to be much more passive to their environment.

[1]

Granule cells in the hippocampal dentate gyrus provide a major source of synaptic excitation to CA3 pyramidal neurons via morphologically complex mossy fiber (MF) terminals that wrap around large spines (thorny excrescences) on the proximal segment of apical dendrites of the postsynaptic neurons (1). Individual granule cells in vivo have low spontaneous firing rates (2), yet they exert powerful effects when they fire a burst of action potentials, causing the discharge of postsynaptic CA3 pyramidal neurons (1, 3). Gap junctions between axons of cortical excitatory (principal) neurons were predicted to exist, based on the rapidly rising upstrokes of putative intracellular coupling potentials [fast prepotentials or spikelets, (4, 5)] during ≈200-Hz ripples in vitro in low-calcium media that blocked chemical synapses (6). Schmitz and colleagues (7) provided electrophysiological and dye coupling evidence for axoaxonic gap junctions in CA1 and CA3 pyramidal cells as well as in dentate granule cells. Subsequently, both modeling and in vitro experimental data suggested that axonal gap junctions could account for very fast oscillations (>70 Hz), including ≈200-Hz ripples (8, 9) as well as play a critical role in the generation of persistent γ (30–70 Hz) (10) and neocortical β2 (20–30 Hz) oscillations (11).

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1924795/

 

[Pythagorean]

A pathological place this could occur, is of course epilepsy, where a neuron that is supposed to be "excitatory" (eg, gets excited then goes back to rest with a nice refractory state) becomes oscillatory (somehow gets flooded by too much current and just starts spiking continuously). That fast spiking and synchronization indicate gap junctions may play a role

https://www.frontiersin.org/articles/10.3389/fphys.2014.00172/fullBut, this can also exist in physiological case where the neuron intended to function with prolonged "afterdischarge" - (as in contractions at birth) and they typically work via peptide bath triggered by neuropeptide release upstream (calcium is no stranger). Such neurons can spike as a normal excitatory neuron, but if it receives a certain kind of input, will reach a tipping point will go into this afterdischarge mode, after which they have a refractory period of their own - in their form of refractory period, they can be excited still, but cannot go back into high frequency spiking.

https://www.researchgate.net/figure/Unlike-adult-cells-juvenile-bag-cell-neurons-do-not-afterdischarge-in-response-to_fig1_14285555
https://pubmed.ncbi.nlm.nih.gov/25185820/

 

[madness]

Yeah, synaptic fatigue was probably kicking around in my subconscious, but with axonic gap junctions, the flood of current can be too much for the neuron to return to rest state, leading to "spikelets" and/or "ripples" [1], because gap junctions are promiscuous and let anything in in a passive way (unlike ion channels which are selective to particular ions). They also can let calcium and other ligands in, triggering second messenger processes.

So it's not the the refractory period isn't happening in the Na and K ions, it's just being washed out by the flood of current coming in axonic gap junctions - typically from coupling with neighboring neurons (a synchronization process). I vaguely remember that this can also lead to extra current adding up in the soma, but from the axon (through gap junction impulses) instead of the dendrites - I can't remember the name of these backward traveling waves. Theoretically, this could lead to interesting processing situations for memory (and such axoaxonic gap junctions are found in hippocampus [1].)

Sure, a gap junction can depolarise the cell during a refractory period but this wont't cause it to spike. Same with chemical synapses.

I agree that a major difference between gap junctions and chemical synapses is that gap junctions are promiscuous. Chemical synapses use neurotransmitters to selectively allow certain ions through. They are more discrete in the sense that they depend on a presynaptic spike/vesicle release whereas a gap junction is more graded.

Anyway, my point being that chemical synapses, afaik, are always fixed behind the dendrite therefore always constrained by the soma's threshold and the refractory period of the Na-K channels and thus unidirectional. There's synaptic fatigue that allows for short term plasticity. Gap junctions tend to form between neighboring cells (even non neural cells) and tend to be much more passive to their environment.

Chemical synapses can also target axons (https://en.wikipedia.org/wiki/Axo-axonic_synapse), and gap junctions can target dendrites (in fact I think they mostly do).

 

[BillTre]

I can't remember the name of these backward traveling waves.

Antidromic.

I suspect that if a nervous system were to evolve with only electrical synapses (no chemical synapses), that it could make adaptations such that the resulting nervous system would be able to achieve useful functions.

 

[madness]

Antidromic.

I suspect that if a nervous system were to evolve with only electrical synapses (no chemical synapses), that it could make adaptations such that the resulting nervous system would be able to achieve useful functions.

Naively, it doesn't seem there is as much you can "do" with an electrical synapse (gap junction). In a chemical synapse you can alter the pre or post-synaptic side in terms of ion channels, number of synaptic vesicles, etc. With a gap junction I can imagine making it bigger or smaller, or putting it in different places on the cell, or perhaps modulating the local chemical environment around the gap junction.

It's interesting though, that there was a historical debate about whether neurons are individual cells or form a "synthesium", which would be a bit like a huge network of cells all joined by gap junctions to form what is effectively one big brain-sized cell (https://en.wikipedia.org/wiki/Reticular_theory). This debate was won over with the development of the so-called "neuron doctrine"(https://en.wikipedia.org/wiki/Neuron_doctrine).

Going beyond the neuron doctrine (supposedly, although I find that to be a of bit sensational way to present the research), are people investigating compartmentalisation of computations in cells, such as in their dendritic trees. Recently it was claimed that a single neuron can act like a small neural network (https://www.quantamagazine.org/neural-dendrites-reveal-their-computational-power-20200114/).

You could put it this way: spiking and chemical synapses allow for digital computation whereas gap junctions and single-neuron spatially extended phenomena allow for analog computation.

If single cells can do this much, then perhaps single cells connected with gap junctions could be evolved to do plenty. But chemical synapses seem to confer a lot of advantages.

 

[Pythagorean]

There's still a lot of difference in dynamical response between an equilibrium seeking leak (gap junctions) and a "prescribed" quantal release. Equilibrium seeking nature of gap junctions is what makes them unstable. For chemical synapses, there's relatively much fewer context in which both the current and amplitude dramatically change direction (e.g. yes, GABAergic neurons are excitatory during development, but gap junctions can change direction and amplitude in a matter of milliseconds).

The way our brain has evolved to use each of their features probably couldn't be beat by one or the other alone. Gap junctions and chemical synapses complement each other wonderfully.

Sure, a gap junction can depolarise the cell during a refractory period but this wont't cause it to spike. Same with chemical synapses. ... Chemical synapses can also target axons

A single gap junction won't of course, but a single channel "causes a spike" is a bit loaded. The question is how much of a contributor it is to a spike. The story with electrical synapses is not straightforward:

These ectopic action potentials are able to spread between neighboring axons via axo-axonic gap junctions. Irregularly occurring bursts of collective activity in principal cells are observed spontaneously (8) and are believed to be mediated by ectopic activity propagating between electrically and dye-coupled axons (6).

https://www.pnas.org/content/100/19/11047

Ectopic action potentials (EAPs) can be initiated in distal parts of the axon and travel antidromically along the fibre to invade the soma. Such ectopic spikes have been recorded during fast hippocampal oscillations in vitro (Dugladze et al. 2012) or during spatial exploration in vivo (Epsztein et al. 2010), but they also occur in several neuronal pathologies such as epilepsies or demyelinating diseases

The precise way EAPs can be initiated is not yet properly defined but it is thought that local depolarization mediated by spiking activity in adjacent axons through gap junctions (Schmitz et al. 2001) or stochastic activation of sodium channels in demyelinated segments of the axon (Hamada & Kole, 2015) might initiate EAPs. Compared to AIS‐evoked action potentials that are classically initiated on the top of a slow subthreshold depolarization such as an excitatory postsynaptic potential, EAPs recorded in the cell body display a much lower threshold (∼ −20 mV) (Hamada & Kole, 2015). This lower threshold is thought to result from the axial current flow initiated distally in the axon that propagates back to the soma. In this case, the spike is initiated without subthreshold depolarization generated in the soma, thus avoiding inactivation of sodium channels in the axon.

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6209742/

 

[Secan]

I once read the book Rhythm of the Brain described thus:

"Studies of mechanisms in the brain that allow complicated things to happen in a coordinated fashion have produced some of the most spectacular discoveries in neuroscience. This book provides eloquent support for the idea that spontaneous neuron activity, far from being mere noise, is actually the source of our cognitive abilities. It takes a fresh look at the coevolution of structure and function in the mammalian brain, illustrating how self-emerged oscillatory timing is the brain's fundamental organizer of neuronal information. The small-world-like connectivity of the cerebral cortex allows for global computation on multiple spatial and temporal scales. The perpetual interactions among the multiple network oscillators keep cortical systems in a highly sensitive "metastable" state and provide energy-efficient synchronizing mechanisms via weak links."

I just want to know where exactly awareness of being aware is encoded in the brain? Can it occur without neurotransmitters? Is it encoded in the rhythms? What you think?

 

[madness]

I once read the book Rhythm of the Brain described thus:

"Studies of mechanisms in the brain that allow complicated things to happen in a coordinated fashion have produced some of the most spectacular discoveries in neuroscience. This book provides eloquent support for the idea that spontaneous neuron activity, far from being mere noise, is actually the source of our cognitive abilities. It takes a fresh look at the coevolution of structure and function in the mammalian brain, illustrating how self-emerged oscillatory timing is the brain's fundamental organizer of neuronal information. The small-world-like connectivity of the cerebral cortex allows for global computation on multiple spatial and temporal scales. The perpetual interactions among the multiple network oscillators keep cortical systems in a highly sensitive "metastable" state and provide energy-efficient synchronizing mechanisms via weak links."

I just want to know where exactly awareness of being aware is encoded in the brain? Can it occur without neurotransmitters? Is it encoded in the rhythms? What you think?

The author of that book wasn't talking about awareness. The book is forwarding a particular view that oscillatory coordination of groups of neurons (cell assemblies as he calls them) provides a mechanism for cognitive processes.

Do you really mean "awareness of being aware" or is that a typo? If you really meant it, I would recommend reading up more on this: https://en.wikipedia.org/wiki/Attention_schema_theory. Graziano basically argues that "consciousness" (or awareness) is to do with the brain having a model of its own attentional system. You are more likely to get an answer to your question by thinking at this more abstract level of information processing than by getting lost in all the messy details of cells and synapses and neurotransmitters.

I would recommend watching this video on David Marr's "Three Levels of Analysis" to get a better idea of why you're barking up the wrong tree .

 

[Physics4Funn]

...
And I think your assumption: "computers could be the model for neurons" may not be very useful. It really is the other way around. Concepts like parallel programming - or multitasking- are easily seen in modern mammalian brains. For which the primary design evolved 500 million years ago, long before John Von Neumann came up the the cpu concept.
...

I agree.
I thing of circuits as components connected by the wires being simplified neurons/synapses. Since synapses have hundreds of neurotransmitters, that can be used as a form of multichannel information flow. For passive components, resistors, capacitors, there can be only one channel of information in the circuit. But interconnected computers approach multichannel processing. Animals have had a 500 million year head start on our electrical technology.

 

[Laroxe]

I think the idea that the brain can be understood as if it was simply a complicated electric circuit with neurones acting as wires is hugely misleading. A neurone doesn't simply function as a switch, its been recognised for some time that neurons can compute a huge range of functions. Each branch of a dendrite acts like a little non-linear output device, summing up and outputting a local spike if that branch gets enough inputs at roughly the same time. The dendrites receive inputs from a very large number of other neurons from far and wide and it seems the distance and direction of the inputs is significant. We also know that neurons continuously send spikes to each other, that appear to be independent of particular inputs a lot of repeated movements are controlled by groups of neurons independently.

Any network is a mixture of excitatory and inhibitory neurons, with the latter making their target neurons less likely to fire a spike. These will tend to use different chemical transmitters, but the brain doesn't just have two, it uses a variety of different transmitters with different qualities, in addition in a nerve synapse there are a range of different binding sites again with different functions. A single neuron can have synaptic connections with up to 10,000 others and not just at the dendrites, connections can be made at the cell body or even along the axon.

We then have to consider that the brain sits in a neurochemical bath and if you alter the chemicals in the bath of fluid, you can change how communication works in specific areas with the right receptor sites or in the whole network at once.

There is an attempt to model some of these processes with Liquid State Machines which try to introduce continuous dynamics in randomly connected networks.

This is important is that the resulting network is almost guaranteed to have chaotic dynamics and chaotic dynamics give us a very rich playground in which to do stuff. Of course evolution could have equipped us with directly connected neurons, but it would be a massive downgrade.

Sorry I couldn't find anything which summarised all of this though there are plenty of individual papers around.