Representation as Networks of Cells:

Representation making are stigmergic phenomena in cell structures and in individual cells.   Stigmergic phenomena of the cells produces asemic changes in cell networks.  

This relationship between cells, an input and an output cell, is direct.  Inputs are outputs.  input = output.  functionally, inputs -> outputs.

this network contains an input, a node, and output.  This network is a representation.   the node is input as output   
(input ; output) = node    functionally:  inputs ->  node    node -> output

If N signals R when it receives a signal from input and sends a signal to output, then R reflects N.  The structure of the network shows that R is a representation of N.  
R ; N   N = (input ; output)   

Networks which engage in this kind of reflective function are networks which are making representations.  The structure of the network itself is the representation of the fact R represents N.   And the signaling of N to R is the instance of that representation.  At this level, R;N doesn't mean anything, or rather, the meaning of R;N is asemic.  It is only if the network builds more and more of these reflecting  or representations that what R;N means begins to take shape.  

When N signals a group of these reflecting nodes, then the meaning of that reflection can only be ascertained by what other network connections those nodes are signaled by (here represented by X1-3).  What is not shown is that the R nodes must themselves signal other nodes, so that those structures become functional representations of the R nodes of N.  

Here (N X1) ; R1   (N X2) ; R2   (N X3) ; R3

If the R nodes themselves do not signal other nodes, then the R nodes are representationally dead ends.  The whole network instantiates the representations of (N X3) ; R3 for instance, but that representation produces no changes in the network at large.  It is the function, particularly the reflective function and the structure of the network, the Connectome at large that instantiates the totality of representations for that organism.  If there is network structure that can represent some feature of fact of the organism experience, then the organism cannot experience it.  Conversely, the structure of the network can fool the organism into action or can be a source of illusion.  

If N fires R1 for instance, but does not excite R2 or R3 to fire, then R1 may produce a signal to a downstream node which indicates some activity.  For instance, seeing a mouse or spider out of the corner of your eye.   There phenomena of representation are constantly going on. That the brain fills in missing information is a well documented fact.  This filling in does not originate in at the edges of the network, at the input nodes, but is further in.  

Dream states are explicitly situations when input and output activity is inhibited but the brain is active, and the representations the brain produces happen.  We encounter the dead, we have adventures, we solve problems in our dreams.   This fact alone demonstrates that what is key to experience is the representational activity of the brain.  The inputs and outputs exist, but the effect they have is to signal what otherwise would be a continuous dream experience.  Like we see in the cell, it is the representational functioning of the brain that is important, the dreaming is the key feature.  The development of inputs and outputs is a side-effect of the basic representational process.  A side-effect which offers homeostatic advantage.  But the key activity is the representation making, the dreaming.  

(input Rf output) ; N
(input Rf) -> N
N -> output

Now let's look at how representational nodes, impact function.  the structure of the Rf node and whatever inputs the Rf node receives determine when the Rf node signals N.  That signal is stigmergic.  It depends on the internal conditions and structure of N if Rf will excite N to signal an output, or if Rf inhibits N from signaling to output, or if Rf is necessary to signal to output upon receiving the input signal (co-occurrent).  

A key feature of Rf like nodes is that some may have no inputs, or they may have no inputs that change their function.  or Rf nodes may signal based on local conditions.  

As a hypothetical example, a cell node that is sensitive to Ghrelin (or Ghrelin mediated hormones) may signal directly into the network when Ghrelin levels drop or rise.  When the Ghrelin steady state changes, the membrane of the cell changes which leads to an alteration of the cells action potential inducing it to automatically signal.   In this way, Rf acts as a kind of local input too the N network. 

 [Influence of membrane ion channel in pituitary somatotrophs by hypothalamic regulators.  Yang SK1, Steyn F, Chen C.] [Upregulation of voltage-gated calcium channel cav1.3 in bovine somatotropes treated with gherkin.   Salinas Zarate VM1, Magdaleno Méndez A1, Domínguez Mancera B1, Rodríguez Andrade A2, Barrientos Morales M1, Cervantes Acosta P1, Hernández Beltrán A1, Romero Salas D1, Flores Hernández JL3, Monjaraz Guzmán E3, Félix Grijalva DR4.] [Mechanism of spontaneous and receptor-controlled electrical activity in pituitary somatotrophs: experiments and theory.   Tsaneva-Atanasova K1, Sherman A, van Goor F, Stojilkovic SS.]

In this network the different Rf nodes will have affect N differently.  Why?  because the signals the Rf nodes send are different.  For instance, Rfe may be a dopaminergic neuron the releases that produce an excitatory response by N.  Whereas Rfi maybe a GABAergic neuron which may inhibit signaling by N.  Rfc may simply be another neuron like N.  Where N is such that without a co-occurrent firing from Rfc, N would never signal the output because N's action potential falls back to a base state too rapidly, or the input signal is simply too weak by itself to induce N to signal the output. 

This triangle network shows how N will signal R reflecting when it signals output.  this reflection will in tern signal Rf which acts as a signal regulator to N.  As above, Rf may inhibit or excite N.  But in this network the Rf signal to N is regulated by N's prior reflected signaling of R. 

As above, this network just shows that N's regulation may be driven by a cloud of representations that are signaled by N.  Any node in these diagrams may in fact be it's own network, so that networks are nested inside each other representationally.  But if we look at nodes in terms of cells and diagrams, what it means when the node functions is going to appear mysterious.  We should expect to be able to discover what nodes do, but what the network structures mean in very complex networks will remain a kind of mystery.  

Why?  Because the node is instantiating a representations and representational functions.  But explicitly, ideas, representations are not features of the physics of nodes and their connections.  the meaning of network is something that can only be understood representationally.  We may be able to correlate physical effects of node manipulation with examination and experimentation of nodes, but meaning, ideas, will be a much more complex problem.  

We can cluster certain kinds of representations into network areas of the brain for instance, but as we peer closer and closer to individual nodes, the meanings will become more and more invisible.   This happens for two reasons:  One, representations are constantly occurring.  Two, the stigmergic/chemical interaction of cells is constantly occurring.  The two go hand in hand.  At some point in the development of cellular networks we shift our attention not to the network development but to the representational development, and only when representational development is impaired do we look deeper into the cellular stigmergic functioning and structure of the organism to try to understand what is producing that impairment.  

Interestingly, we are all impaired, it's just that because we share so many representational experiences in common, that what impairs us all are invisible representations we do not apprehend because we do not have the network structures in place to instantiate those representations.  

Here we see a complete network which shows both representation and representational function.   Any signal that N1 receives signals the R1,2 nodes.  These nodes mean different things based on the inputs from the networks X1,2.   

input ; N1
N1 ; R1  N1 ; R2
(X1 R1) ; Rf1   (X2 R2) ; Rf2
(Rf1 N1 N2) ; output
(Rf2 N1 N2) ; no output (hidden condition)

Lets say this circuit represents the mastication process for a rabbit.  Where the input correlates to grass in the rabbit's mouth, and the output is chewing the grass.  

this circuit can describe both eating when hungry and stopping to eat when fleeing a predator.  If X2 represents a predator and X1 represents hunger, then the action of mastication by the rabbit will be inhibited when (X2 R2) signal Rf2.  

For all of the representational activities and ideas an organism has, there must be corresponding neural networks that instantiate those representations.  At the edges of the network are the same nodes for input and for output (mostly motor actions).  There are also nodes in the nervous system that act as chemical interaction points with the rest of the body.  So that chemical changes in the organism at large become representational structures by the networks structures of neurons.  That is, chemical changes in the body are also input/output nodes in the neural network.

What we see here is a simplified model of the action of the N2 network in more detail.  For representational purposes, nodes in a network diagram may in fact be sub-networks.  

When C1 signals the muscles to contract and close the jaw, C1 also signals node J1.  Because there are network hops from C1 through J1 to C2, there is a delay before C2 is signaled.  It takes some amount of time for signals to traverse between nodes and that means a process like opening and closing a jaw for chewing can be regulated simply by timing the signals.  Of course real nervous system networks and muscle regulation is more complex than this, but it is these simple network structures that demonstrate the simple elements of time and action. 

Note:  Wikipedia lists the innervation of different human mandible control muscles which are signaled in the much more complex network of the human nervous system.  But the basic functions of timing and signaling and network inter-regulation  must exist for proper mandibular control.  thus even though the listed network structures above may be overly simplistic for human morphology, the necessary representational functions exist in more complex neural networks such as the innervation pathways listed below.  And it is the structure of the networks, the placement of nodes and types of nodes and how nodes regulate other nodes that effects morphological possibility.  [see: Biology Seminar Series: Franz Goller, Professor, Department of Biology, University of Utah
"Motor control of birdsong: Peripheral answers to central questions"] 
Innervation of Masseter muscle []
Along with the other three muscles of mastication (temporalismedial pterygoid and lateral pterygoid), the masseter is innervated by the mandibular division (V3) of the trigeminal nerve. The innervation pathway is: gyrus precentralis > genu capsula interna > nucleus motorius nervi trigemini > nervus trigeminus > nervus mandibularis > musculus masseter.
Innervation of Temporal muscle []
As with the other muscles of mastication, control of the temporal muscle comes from the third (mandibular) branch of the trigeminal nerve. Specifically, the muscle is innervated by the deep temporal nerves.
Innervation o f Medial pterygoid muscle []
Unlike the lateral pterygoid and all other muscles of mastication which are innervated by the anterior division of the mandibular branch of the trigeminal nerve, the medial pterygoid is innervated by the main trunk of the mandibular branch of the trigeminal nerve (V), before the division.
Innervation of Laterla pterygoid muscle []
The mandibular branch of the fifth cranial nerve, the trigeminal nerve, specifically the lateral pterygoid nerve, innervates the lateral pterygoid muscle.


These diagrams of networks are illustrative. But we know that neurons construct networks that do demonstrate these representational properties.  The key argument is that though networks are created from the molecular processes of cells, the networks continue to exist because the networks serve some homeostatic purpose.  And that the representation making that is achieved through network constructions serve some homeostatic purpose for the organism.