I think it might be important to include some additional components to the GNW theory. For one thing, it is important to recognize that while the GNW is based upon the global workspace (GW) theory of Baars, it is not identical and the differences are crucial for two reasons: 1) GW is, like so many theories of consciousness, very abstract compared to your typical research in computational neuroscience, neurobiology, etc., and 2) GNW concerns a single brain region that dominates the entire theory.

So while we might speak about the global workspace and its "contents" or "access to the global workspace" in abstract terms (i.e., we don't need to talk about specific brain regions or types of neurons), we can't do this if we are using or referring to the GNW. We must refer to workspace access and signals in terms of the prefronto-parietal network. This is not to say we must become bogged down in the details concerning pyramidal neurons and other evidence put forward as supporting the theory, merely that the two conditions which must be satisfied in order for stimuli to access the global workspace in GNW refers to access to the prefronto-parietal network.

Also, as one might guess from the last statement above, we should also be aware of what the conditions are that allow access to the GNW. The strength of the signal is certainly one of the conditions. However, no matter how strong a signal some stimulus creates/causes, there is no guarantee that it will be capable of "entering" or accessing the GNW. That's because the second condition requires an "opening" or "free space" in the GNW. Put differently, the GNW is already too loaded/busy for some other stimulus to access it. Put differently yet again, the signal created by a stimulus may be strong enough such that it could activate neurons, neural populations, or neuronal networks that are connected to the prefronto-parietal cortex such that the resulting signals could change the activity in this area of the cortex, but it so happens that the current activity in the prefronto-parietal cortex makes this impossible (somewhat akin to the way neurons usually require a small amount of time after generating an action potential to send any signals, no matter what the nature the neurons dendritic activity/state).

A final issue is whether or not there is any basis for the "mind-as-computer" metaphor. For example, are there any neurophysiological processes that are comparable to "threads"? Clearly, we can't concentrate or be conscious of everything, and just as clearly much brain activity is unrelated anything we are aware/conscious of. But this is so trivially true that we do not need to invoke the notion of "threads" simply because of it. It may be, though, that the ways in which the brain processes information, and in particular how information processing activity with the potential to correspond to something we are conscious of actually does or does not become something we are conscious of (in GNW terms, whether the mechanisms whereby a stimulus accesses the GNW is sufficiently conceptually similar to threads to warrant the comparison).

"The asymmetry between the brain and the computer is complete, all comparisons are flawed and the idea of computer-generated consciousness is nonsense."

Torey, Z. (2009). The crucible of consciousness: An integrated theory of mind and brain. Cambridge: MIT press.

“The brain is not a computer, nor is the world an unambiguous piece of tape defining an effective procedure and constituting “symbolic information.” Such a selectional brain system is endlessly more responsive and plastic than a coded system.”

Edelman, G. M. (1999). Building a Picture of the Brain. Annals of the New York Academy of Sciences, 882(1), 68-89.

"Why would the mind work like a computer? This book is aimed—like some other recent books (e.g., Kelso, 1995; Port & van Gelder, 1995; see also Fodor, 2000)—at responding to that question with the following answer: 'It doesn’t.'”

Spivey, M. (2007). The continuity of mind. Oxford University Press.

“Today’s programs—at best—solve specific problems. Where humans have broad and flexible capabilities, computers do not.

Perhaps we’ve been going about it in the wrong way. For 50 years, computer scientists have been trying to make computers intelligent while mostly ignoring the one thing that is intelligent: the human brain. Even so-called neural network programming techniques take as their starting point a highly simplistic view of how the brain operates.”

Hawkins, J. (2007). Why Can't a Computer be more Like a Brain?. Spectrum, IEEE, 44(4), 21-26.

Or for some meta-citing:

“Referring to the ‘widespread belief ... in many scientific circles ... that the brain is a computer,’ neurobiologist Gerald Edelman (2006) insists that ‘this belief is mistaken,’ for a number of reasons, principal among which are that ‘the brain does not operate by logical rules’ (p. 21). Jerome Bruner (1996), a founder of cognitive science itself, yet, coincidentally, a key figure in the emergence of narrative psychology, challenges the ability of ‘information processing’ to account for ‘the messy, ambiguous, and context-sensitive processes of meaning-making’ (p. 5). Psychologist Daniel Goleman (1995), author of the popular book Emotional Intelligence, asserts that cognitive scientists have been so ‘seduced by the computer as the operative model of mind’ (pp. 40f.) that they have forgotten that, ‘in reality, the brain’s wetware is awash in a messy, pulsating puddle of neurochemicals’ (p. 40f.) which is ‘nothing like the sanitized, orderly silicon that has spawned the guiding metaphor for mind’ (pp. 40–41).”

Randall, W. L. (2007). From Computer to Compost Rethinking Our Metaphors for Memory. Theory & psychology, 17(5), 611-633.

I have a running bibliography on the Brain-Computer/Mind-as-Machine metaphor, but dozens of quotes isn't just largely useless, it's also off topic. Suffice it to say that there are a large number of experts in the cognitive sciences, from neuroscientists to philosophers, who believe comparisons between the brain and computers are fundamentally and fatally flawed. There was a time when this would be rather unthinkable due to the origins of cognitive science "proper"; namely, computer science, the algorithmic nature of then-newly formed generative linguistics and its status as essentially "modern linguistics" insofar as by "linguistics" we mean an account of language that seeks to understand its cognitive bases, and the general dominance of formal (logical/mathematical) systems accounts of everything from vision to memory to consciousness.

This makes one final notion important for any discussion of GW theory or GNW theory: what are the nature of other models of consciousness and the evidence for them? I hope nobody will mind if I forego consideration of the soul or similar ideas about the nature of consciousness, as I consider even quantum theories of consciousness exotic and only a "theory" at all because were one to require theories of consciousness to have the kind of evidence we'd expect, that would pretty much leave only those theories that defined consciousness in ways that allow just about anything and everything to be conscious (see Tononi's Integrated Information Theory (IIT) or Conway & Kochen's "free will theorem").

It's far easier to address the evidence for individual theories because we are pretty clueless in general, so evidence for any theory is usually some combination of a metaphysical & philosophical notions (such as the theory's definition of consciousness) along with an abstract model and usually some discussion of neurophysiological processes that may have absolutely no relevance to consciousness.

For example, Penrose and Hameroff's ORCH-OR model combines metaphysical ideas about free will and consciousness in relation to classical physics in order to give a general motivation for looking into quantum processes, followed by a description of sub-cellular neurobiology that has no known relation to consciousness or really any cognitive function. Stapp's theory is similar, but lacks the pointless inclusion of neurobiology that requires quantum mechanics regardless of whether it has any relevance to cognition at all.

Sir John Eccles (Nobel prize-winner in neurobiology) relies on things he refers to as psychons. These are "elemental or unitary mental events" (How the SELF Controls Its Brain), and thus we arrive at a dualistic picture.

Libet (the psychologist most responsible for the empirical evidence against the possibility of free will, despite the fact that he describes a possible mechanism for free will in his book on the subject) proposed that consciousness arises out of what he calls "the conscious mental field" or CMF, "which is in a phenomenological category not reducible to (although intimately related with) neuronal processes. There are no doubt rules for (at least much of ) the relationship between the CMF and the physically (externally) observable neural processes."

Libet, B. (2004). Mind Time: The Temporal Factor in Consciousness (Perspectives in Cognitive Neuroscience). Harvard University Press.

Hofstadter's "strange loops" have several sibling theories (e.g., those of Jackendoff and Pinker), in which somehow a hierarchically ordered collection of neural processes in particular brain regions or involving particular neuronal networks somehow generate algorithms that generate algorithms or nested functions/subroutines that may not be strictly linearly ordered.

There are many others I could add, but I my attempt is not an exhaustive survey but to provide some bases for comparison by instances of theories that fall into different categories (e.g., quantum, computational/algorithmic, information-theoretic, etc.).

Now for an actual response: it is trivially true that something about the processes that result in conscious experiences require both pre-conscious (or pre-emptive) signals and that the strength of these matter. This is regardless of whether or not a global workspace or global neuronal workspace exists. The pre-emptive part comes from the fact that we cannot be conscious of e.g., visual stimuli without neural firing in the PNS component(s) of the visual system. That the strength of signal matters is somehow trivial and yet complicated. One issue is that there are various theories in neuroscience concerning what a "signal" is. Ignoring action potentials that are nothing other than noise, there exist different theories concerning the nature of how spike trains or more simply action potentials "work". The brain is constantly active. For us to become aware of some visual stimulus, something has to change and unless virtually all of brain research is wrong, that something is mostly or entirely based upon action potentials. However, single action potentials don't really matter. What matters may be the rate at which sequences of action potentials generated by various neurons connected to another neuron's dendrite or dendrites (in terms of the cortex, the plural "dendrites' doesn't quite cut it; we are talking about tens of thousands of dendritic connections or more). Or it may be that the rate doesn't matter but the timing does. Or both. Or neither. Also, the most widely used neural model (the one upon which basically all others are somehow based on) doesn't work via all-or-nothing spikes. And there are other complications.

That said, we can demonstrate the triviality of signal strength by going to the PNS again. Specifically, nociceptors or "pain neurons". Most of the time that we experience pain it is mechanical (this covers everything from bumping your elbow to being having limbs chopped off). There are, however, other types (e.g., chemical and thermal), but there is in particular a rather unique kind of mechanical (presssure) induced pain- the mechanical/pressure pain that we experience due to mechano-insensitive nociceptors. These are interesting and ideal here because unlike most pain signals they do not begin as noxious. In order for mechano-insensitive nociceptors to generate a signal powerful enough to be perceived as painful, there must be an initially non-noxious pressure applied repeatedly and over time (tonic pressure) such that eventually the repetition causes mechano-insensitive nociceptors to encode a signal perceived as pain. Normally, things work the other way around. Stimuli initially perceived as noxious, whether from walking out into the freezing cold from a warm house or being pinched, nociceptors using A-fibers or C-fibers initially send a strong signal. Also, the noxious stimulus generally causes the same or mostly the same nociceptors to fire, and unless the stimulus is significant the repeated activation of these nociceptors will be modulated by the CNS (i..e, the CNS says "stop paying attention to these signals, we're doing fine"). Finally, most of the time signals from these nociceptors aren't consciously experienced.  So we require a sufficiently strong initial signal to experience thermal, chemical, & most mechanical stimuli as noxious, but for mechano-insensitive  nociception we the signals strength is locally modulated by repetition and time (i.e., no signal will be generated locally until the stimulus has been applied locally for a sufficiently long time interval and with a particular regularity).

The question "Is the Neuronal Workspace theory correct in assuming that there is a

"strength" difference between modules, or is there some other decision making process involved such as Pre-emptive Selection?"

Actually involves several questions. First, nothing about the particulars matter if the theory is simply wrong. Second, there must exist these "modules". Third, there is the question of whether or not pre-emptive selection is non-trivial, incompatible with either GW theory or GNW theory, or actually descriptive of the theory. Fourth, there is the question that, granting a suitably non-trivial definition of pre-emptive, is it or are other processes involved in the mechanisms underlying what determines what we consciously experience.

At the moment, I can't see a way in which "pre-emptive" could be defined non-trivially or interpret the way it is defined ("something must be 'known'", which suggests it is already something one is conscious of) such that it isn't paradoxical. Nor am I quite sure how the question actually relates to and distinguishes between GW theory, GNW theory, and "Neuronal Workspace theory". But I don't think much of cognitive theories involving (massive) modularity (in the evolutionary & cognitive psychology sense, i.e., "massive modularity" or extreme domain specificity). Also, as both GW theory and GNW theory require domain general mechanisms, and in fact are rather fundamentally based upon the idea of highly synchronized activity between nonlocal neural populations, I am not sure that how "modules' may play a role in these theories (thought I can imagine ways in which they might).

Thank you for the involved answer. I gather my confusion as to cognitive science and Neuro-Science was obvious from my question. We need to move beyond the assumption that individual neurons are conscious, and deal with tissues of neurons.

The Fronto-Parietal Network, (Not the Pre-Fronto-Parietal network), works with the Demand Mode Network and the Dorsal Activation Network to create "Trains of thought". It is thought by those doing DSI and DTI that there are two central networks running down the central fissure and these networks are known to reach out and coordinate activity between clumps of neurons called "Modules" by some.

Whether these two "Central Networks" actually correspond to the Fronto-Parietal Network, is one of those things that is poorly defined by the literature. But then work being done on the human connectome is as yet poorly documented. And, even if it weren't I have limited resources to follow it.

I understand that you bridle at anything that smacks of interpreting the brain in terms of computers. Most of the brain does not operate like computers and even the human brain project with its massive Neuro-Morphic computing has been brought up sharply against the difference between early models of invertebrate cells, and the models that would be needed to model mammilian cells adequately.

However my background is computers, and I hope you can accept that when I am dealing with thorny problems in architecture I tend to fall back on my understanding of computers to offer analogs.

GW theory as suggested by BAARS and implemented by Franklin claims to define a conscious machine. Philosophers are talking about Access Consciousness as a result of this work. While I am extremely skeptical that Franklins Robots are conscious, they do implement Baars theory.

DeHeane's work on GNW is an attempt to co-opt GW for neuroscience, and while I remain relatively skeptical, work on "Trains of thought" has taken on new significance with the work by the MRI teams working on Mind Wandering.

One understands the Fronto-Parietal Network is something like Working Memory where there are two locations where neural involvement in Memory are focussed, the prefrontal cortex and the parietal cortex in the area around the Angular Gyrus.

It seems to me that the parietal lobe being close to both the lateral and dorsal connections to the hippocampus, is the output buffers, while the prefrontal cortex, are the executive buffers for working memory, although this model smacks of Millers early work on cognitive science where he described a central processor to do executive functions. I don't ascribe to that model, instead I am looking at a distributed model where some parts of the brain have similar functions. This Fronto-Parietal Network might therefore be equivalent in some sense to a Global Workspace.

However the networks are curtailed and do not stretch fully across the brain merely running down the central fissure, and capturing a small subset of neural processes on the way.

Now on to the question of whether you would be conscious of information that could be used to pre-empt selection.

My answer is of course, Not Necessarily. The information needed to make the judgement for selection could be in a separate network, that is not accessed by consciousness. This simply assumes that the Fronto-Parietal Network is not a GLOBAL network but that processing goes on outside it. In this model, Access Consciousness might only exist at a certain stage of processing where information pre-exists the process in other networks that condition the processing in the Workspace.

In essence only that information the is "In Band" for access consciousness would be conscious, this is why we need the separation between GNW and NW. GNW is obviously flawed, but an NW theory might not be.

""We need to move beyond the assumption that individual neurons are conscious, and deal with tissues of neurons.""

Yes, which Is why I said so.

"The Fronto-Parietal Network, (Not the Pre-Fronto-Parietal network), works with the Demand Mode Network and the Dorsal Activation Network to create "Trains of thought". 

You haven't offered a single source for either Baars' global workspace theory or the GNW theory (or any theory). Yet you refer to these theories, and at least some of your statements such as that quoted above are unsubstantiated. The importance in GNW of the Prefronto-Parietal Network is at the core of GNW theory:

For example:

"Dehaene and Changeux (2005, 2011) take the assumption of a global workspace of consciousness as starting point and determine it in more neuronal detail when suggesting what they call the global neuronal workspace theory (GNW). They postulate that neural activity in the prefrontal-parietal cortical network is central for yielding consciousness. More specifically, the prefrontal-parietal cortical network has to be recruited by the single stimulus in order to link and recruit the different neural networks."

Northoff, G. (2013). Unlocking the Brain Vol. II:: Consciousness. Oxford University Press.

Or see Dehaene, S., & Changeux, J. P. (2011). Experimental and theoretical approaches to conscious processing. Neuron, 70(2), 200-227.

"I understand that you bridle at anything that smacks of interpreting the brain in terms of computers"

Not at all. I just haven't found the evidence to support this. I do bridle (reasonably or no) at those who describe neural dynamics, neurobiology, consciousness research, cognitive science, etc., reaching conclusions about the fundamental nature of the brain with an elementary familiarity of almost every field in the cognitive sciences. I do not know how much literature in the neurosciences, cognitive psychology, cognitive linguistics, etc., you know or are familiar with. However, it seems that you may be approaching this question based upon a rather limited exposure to the research. It's hard to tell, given that you cite nothing and do not seem  to present the frameworks you are using accurately. If you can't even present the work since Baars' accurately, then I feel it important to contribute some additional views and perspectives.

"One understands the Fronto-Parietal Network is something like Working Memory where there are two locations where neural involvement in Memory are focussed, the prefrontal cortex and the parietal cortex in the area around the Angular Gyrus."

Nothing about this is accurate.

"Now on to the question of whether you would be conscious of information that could be used to pre-empt selection"

Actually, before we address this it would be nice if you could demonstrate you have an elementary familiarity with basic neurobiology and consciousness research, including the various confused references to the tiny sample you present.