Global workspace theory: consciousness as brain wide information sharing

Lately I’ve been reading up on global workspace theory (GWT).  In a survey published last year, among general consciousness enthusiasts, integrated information theory (IIT) was the most popular theory, followed closely by GWT.  However, among active consciousness researchers, GWT was seen as the most promising by far (although no theory garnered a majority).  Since seeing those results, I’ve been curious about why.

One reason might be that GWT has been around a long time, having first been proposed by Bernard Baars in 1988, with periodic updates all recently republished in his new book.  It’s received a lot of development and has spawned numerous variants.  Daniel Dennett’s multiple drafts model is one.  But perhaps the one with the most current support is Stanislas Dehaene’s global neuronal workspace, which I read and wrote about earlier this year.

All of the variants posit that for an item to make it into consciousness, it has to enter a global workspace in the brain.  This is most commonly described using a theater metaphor.

Imagine a play in progress in a theater.  A light shines down on the stage on the currently most relevant actor or events, the light of conscious.  The backstage personnel enabling the play, along with the director and other controlling personnel, are not in the light.  They’re in the unconscious dark.  The audience, likewise, is in the dark.  That is, the audience members are unconscious information processing modules.

This last point is crucial, because this is not the infamous Cartesian theater, with an audience of one conscious homunculus, a little person observing events.  Such a notion merely defers the explanation.  If the homunculus provides consciousness, then does it too have its own homunculus?  And that one yet its own?  With infinite regression?  By stipulating that the audience is not conscious, we avoid this circular trap.

That said, one issue I have with this metaphor is the passivity of the audience.  Consider instead a large meeting room with a lot of rowdy people.  There is someone chairing the meeting, but their control is tenuous, with lots of people attempting to talk.  Every so often, someone manages to gain the floor and make a speech, conveying their message throughout the room.  At least until the next person, or coalition of people, either adds to their message, or shouts them down and takes over the floor.

Most of the talking in the room is taking place in low level side conversations.  But the general room “consciousness”, that is, the common things everyone is aware of, are only of what’s conveyed in the speeches, even though all the side conversations are constantly changing the tenor and state of people’s opinions throughout the room, and could effect future speeches.

I think this alternate metaphor makes it more clear what it means to enter the workspace.  In all of the theories, the workspace is not a particular location in the brain.  To “enter” it is to be broadcast throughout the brain, or at least the cortical-thalamic system.

Diagram showing the regions of the brain
Lobes of the brain
Image credit: BruceBlaus via Wikipedia

How does a piece of information, or a coalition of information, accomplish this?  There is a competition.  Various modules in the brain attempt to propagate their signals.  In many cases, actually in most cases, they are able to connect up to one or a few other modules and accomplish a task (the side conversations).  If they do, the processing involved is unconscious.

But in some cases, the signal from a particular module resonates with information from other modules, and a coalition is formed, which results in the information dominating one of the major integration hubs in the brain and brings the competition to the next level.

At some point, a signal succeeds in dominating the frontoparietal network, all competing signals are massively inhibited, and the winning signal is broadcast throughout the cortical-thalamic system, with binding recurrent connections forming circuits between the originating and receiving regions .  The signal achieves what Daniel Dennett calls “fame in the brain”.  It is made available to all the unconscious specialty modules.

Many of these modules will respond with their own information, which again might be used by one or more other modules unconsciously.  Or the new information might excite enough other modules to win the competition and be the next broadcast throughout the workspace.  The stream of consciousness is the series of images, concepts, feelings, or impulses that win the competition.

One question that has long concerned me about GWT: why does simply being in the workspace cause something to be conscious?  I think the answer is it’s the audience that collectively makes it so.

Consider Dennett’s “fame in the brain” metaphor.  If you were to meet a famous person, would you find anything about the person, in an of themselves, that indicated fame?  They might be attractive, an athlete, funny, or extraordinary in some other fashion, but in all cases you could meet non-famous people with those same traits.  What then gives them the quality of fame?  The fact that large numbers of other people know who they are.  Fame isn’t something they exude.  It’s a quality they are granted by large numbers of people, which often give the famous person causal influence in society.

Similarly, there’s nothing about a piece of information in the brain, in and of itself, that makes it either conscious or unconscious.  It becomes a piece of conscious content when it is accessible by several systems throughout the brain, memory systems that might flag it for long term retention, affect systems that might provide valenced reactions, action systems that might use it in planning, or introspective and language systems that might use it for self report.  All of these systems end up giving the information far more causal influence than it would have had if it remained isolated and unconscious.

Admittedly, this is a description of access consciousness.  Someone might ask how this implies phenomenal consciousness.  GWT proponents tend to dismiss the philosophical idea that phenomenal consciousness is something separate and apart from access.  I agree with them.  To me, phenomenal consciousness is what access consciousness is like from the inside.

But I realize many people don’t see it that way.  I suspect many might accept GWT but feel the need to supplement it with additional philosophy to address the phenomenal issue.  Peter Carruthers, in his latest book, attempts to philosophically demonstrate how GWT explains phenomenal experience, but since he’s a “qualia irrealist”, I’m not sure many people seeking that kind of explanation will find his persuasive.

There are a lot of nuanced differences between the various global workspace theories.  For example, Baars most often speaks of the workspace as being the entire cortical-thalamic core.  Dehaene tends to emphasize the role of the prefrontal cortex, although he admits that parietal, temporal, and other regions in the frontoparietal network are major players.

Subcortical structures of the brain
Image credit: OpenStax College via Wikipedia

Baars emphasizes that processing in any one region of the cortical-thalamic core can be conscious or unconscious.  Any region can potentially win the competition and get its contents into the workspace.

Dehaene is more reserved, noting that some regions, particularly executive ones, have more connectivity than others, and that very early sensory regions don’t necessarily seem capable of generating workspace content, except indirectly through later sensory layers.

Both agree that subcortical regions generally can’t contribute directly to the workspace.  Although Baars sees the hippocampus as a possible exception.

Both Dehaene and Baars think it’s likely that many other animal species have global workspaces and are therefore conscious.  Baars seems confident that any animal with a cortex or a pallium has a workspace, which I think would include all vertebrates.  Dehaene is again a bit more cautious, but he sees all mammals as likely having  a workspace, and possibly birds.  Peter Carruthers, who converted from his own particular higher order theory to GWT, doesn’t think there’s a fact of the matter on animal consciousness.

A common criticism of GWTs is that they are theories of cognition rather than consciousness.  Since to me, any scientific theory of consciousness is going to be a cognitive one, I don’t see that as a drawback.  And I realized while reading about them that they also function as theories of general intelligence, the holy grail of AI research.  Which fits since GWT actually has origins in AI research.

GWTs also seem able to account for situations where large parts of the cortex are injured or destroyed.  Unlike higher order theories (HOT), most of which seem dependent on the prefrontal cortex, if large parts of the frontal regions were lost, the workspace would be dramatically reduced but not eliminated.  Capabilities would be lost, but consciousness would still exist in a reduced form.

I also now understand why the overview paper earlier this year on HOT classified GWTs as first order theories, since first order representations can win the workspace competition as well as higher order or executive ones.  This allows GWTs to avoid many of the computational redundancies implicit in HOT, redundancies that might seem unlikely from an evolutionary perspective.

And I’ve recently realized that GWT resonates with my own intuition from reading cognitive neuroscience, which I described in a post a while back, that subjective experience is communication between the sensory, affective, and planning regions of the brain.  The broadcasting workspace seems like the medium of that communication.

GWTs are scientific theories, so they’ll either succeed or fall on empirical research.  I was impressed with the wealth of empirical data discussed in Dehaene’s and Baars’ books.  Only time will tell, but I now understand why so many consciousness experts are in this camp.

What do you think?  Does this theory sound promising?  Or do you see problems with it?  What stands out to you as either its strengths or weaknesses?

61 thoughts on “Global workspace theory: consciousness as brain wide information sharing

  1. “That said, one issue I have with this metaphor is the passivity of the audience.”

    Is there really an audience or are there just actors? This theory seems to picture consciousness as observer but misses how it also actively engaged in forming representations – it is writing the play on the fly.

    Liked by 1 person

    1. Maybe the way to think about it is that any audience member has the potential to be an actor. Or at least many of them do; depending on which specific GWT you favor, some may only ever be audience members.

      And definitely the whole thing is a dynamic production. This is why I favor the rowdy meeting metaphor, although the theater one is a good stepping stone. Of course, all the metaphors are only stepping stones to the actual architecture.


  2. A fascinating overview of the extant consciousness theories. GWT does seem to have more empirical meat on its bones that HOT and the other theories. I remember back in my introductory psychology course when the lecturer discussed the chasm between Clinical Psychology (the talking therapies) and Behaviorism, how I was initially struck by the realization that both camps had elements of truth that were unaddressed by the opposing sides and that the truth probably lies somewhere in between. I would suggest that the same might be true of GWT and HOT. Regarding the point you raised that HOT might posit redundant operations among different order layers of cognition which wouldn’t be very economical on evolution’s part, I thought that perhaps Evolution would ensure the unlikelihood of those redundancies among layers, unless some cognitive operations are so trivial (require so little energy to perform) that it wouldn’t matter if they were redundant.

    Liked by 1 person

    1. I think you’re absolutely right in that all the theories seem to me to be mostly right and a little bit wrong. They all seem largely compatible.

      One thing I missed, though, is the redundancy of HOT’s. You or Mike wanna explain that?


      Liked by 1 person

    2. Thanks Mike!

      We have to be careful in discussing HOTs since there are a wide variety of them as well. (Indeed, HOT is actually a little older than GWT, and has older philosophical precedents.) My redundancy remark might not apply to all of them.

      But the main difference between HOT and GWT, it seems to me, is in what it takes for us to be conscious of information. GWT only requires that the information be widely spread among processing modules. If there’s some pathology limiting that spread, as long as there remains a substantial population of receiving modules, consciousness in some form continues. So someone with a frontal lobe pathology can still be conscious. Their consciousness would be missing major elements, but it might be functional enough to still model their environment, respond reflexively and habitually to stimuli, and remember what happened.

      HOT, on the other hand, requires that the information reach higher order processing modules, modules that model the lower level processing or representations. Many HOT proponents take a stance that if that if the higher order representations misrepresent the lower order ones, they’d only be conscious of the higher order versions. This is where the redundancy might come in; it seems to imply a duplicated representation. (This is explicit in HOROR theory.)

      Under GWT, it’s possible that their higher order judgement and feelings about the lower order representations might be wrong, but the lower order representations and processing remain a crucial part of the picture.

      Liked by 1 person

      1. Still having trouble. Tell me if you think the following is correct. Let’s say both a yellow rose and a red rose have some special meaning for me, and I get a particular good feeling when I see a yellow one and a bad feeling when I see a red one. Let’s say I look at a red rose. A module for “red rose” gets activated and represented in the global workspace. For some strange reason (random lesion), the “yellow rose” response is activated from the global broadcast of “red rose” and I get a good feeling. At least some HOTs would say I will be conscious of a yellow rose, whereas GWT would say I am conscious of a red rose.



        Liked by 1 person

        1. “At least some HOTs would say I will be conscious of a yellow rose, whereas GWT would say I am conscious of a red rose.


          With HOT, not in the exact way you describe it, but under some HOTs (such as HOROR – higher order representations of representations), it is conceivable that the signals could get messed up so that the higher order representation is of a yellow flower while the first order one is red. Those HOTs say you’ll experience a yellow flower.

          Under GWT, there’s no reason to posit an additional representation where the visual experience itself could get screwed up. But you could have it with the “wrong” feeling.


  3. [wife and kids are out hiking, so I have a few hours to get into this …. woot!]

    Firstly, I like your metaphor of the meeting where members of the audience can occasionally take the stage, but I would caution that it is more probable that those who take the stage do not so much give a speech as repeat one word or phrase over and over.

    Second, I would like some clarification of your model. As best I can figure, the following is your suggested model [considering only sensations getting into consciousness for now]:
    1. Physical stuff happens at the peripheral nervous system.
    2. Modules in the cortex get activated, causing further modules in the cortex to be activated.
    3. Some subset of activated modules “resonate” (synchronize?) which “results in the information [consisting of the combination of modules?] dominating one of the major integration hubs”. [Are these hubs isolated locations in the brain?Where and how many?]
    4. “At some point, a signal succeeds in dominating the frontoparietal network”
    5. “all competing signals are massively inhibited”
    6. “and the winning signal is broadcast throughout the cortical-thalamic system, with binding recurrent connections forming circuits between the originating and receiving regions .

    Here is the clarification I need. What and where are the Integration hubs? It sounds like modules compete to get into a coalition that gets into a hub, and then hubs compete to get into the frontoparietal network. And what could it mean to “dominate the frontal parietal network”. Would one hub dominate the whole network? And why are competing signals inhibited if one signal has already dominated? [I expect it is the inhibition which determines which hub dominates]. And how could the winning signal be broadcast throughout the cortex from the frontoparietal? Is each signal being sent to each module? Consider the signal “yellow rose” being sent to a few thousand places. Now consider “Elizabeth Warren” being sent to those same places. Whatever system is sending, it has to be able to distinguish a vast number of possible signals. One axon is not going to be able to do it.


    Liked by 2 people

    1. On the metaphor, thanks. That’s mostly the only thing in this post that’s truly original to me. The rest is a consolidation of Baars’ and Dehaene’s models. So I can answer some of your questions with my understanding, but if you really want authoritative answers, their books are worth checking out. Dehaene’s is much easier reading. (Go straight to chapter 5 if you want to quickly get to the theory.)

      Or if you really want to get down and dirty, here’s one of Dehaene’s papers:

      “Are these hubs isolated locations in the brain?Where and how many?”

      No, they are all interconnected. In fact, all the major regions in the cortical-thalamic system are interconnected, but the hubs much more so. The hubs are in the prefrontal cortex, the anterior cingulate cortex, the precuneus, the parietal regions, the middle temporal lobe, and a whole bunch of other regions I’m either forgetting about or haven’t learned about yet. The frontoparietal network could be thought of as the hub of hubs.

      “And what could it mean to “dominate the frontal parietal network”. Would one hub dominate the whole network? And why are competing signals inhibited if one signal has already dominated?”

      Yeah, my language wasn’t precise. It’s aimed more to give the general idea that to precisely convey it. In reality, the neural firing patterns attempt to excite their downstream circuitry while inhibiting the other circuitry around them. (It’s called lateral inhibition.) “Dominating” means winning this battle to the extent that a threshold is reached.

      Once that threshold is reach, it sets off what Dehaene calls the “conscious ignition” where the signal is massively amplified and everything else inhibited. Interestingly enough, the massive P3 wave that shows up on EEG scans, which indicates this ignition, is mostly of positive charge, that is, it mostly comes from the inhibited circuits rather than the winning excited one.

      “Is each signal being sent to each module?”

      Yes, although remember that each module is unconscious. So there isn’t a homunculus in each one saying, “Oh, a flower.” No, each module gets a pattern of signals that pertain to its particular functionality. So, for instance, the hippocampus may receive spikes on axons coming from the occipital lobe (where the representation is), from the parietal lobe (associations), the ventral medial prefrontal cortex (feelings related to the flower), all of which may or may not cause it to flag the experience for later retention. But the hippocampus itself never experiences the flower. But it’s getting signals related to the flower while signals related to, say, a friend talking are being inhibited.

      Liked by 1 person

      1. Are these hubs isolated locations in the brain?Where and how many?”

        No, they are all interconnected.

        Of course they’re interconnected, but are the neurons of one hub distinct in not belonging to any other hub? Also, are there hub neurons?

        Once that threshold is reach, it sets off what Dehaene calls the “conscious ignition” where the signal is massively amplified and everything else inhibited. Interestingly enough, the massive P3 wave that shows up on EEG scans, which indicates this ignition, is mostly of positive charge, that is, it mostly comes from the inhibited circuits rather than the winning excited one.

        This seems kinda self-contradictory. You’re saying the signal is massively amplified (ignition) and yet most of the amplification is suppressive. So the signal is not massively amplified so much as the other signals are suppressed.

        “Is each signal being sent to each module?”

        Yes, although remember that each module is unconscious

        How many “modules” do you think there are? And how many hubs? Presumably you are saying the signal from each hub can be sent to each module. How many axons might it take to send one distinct signal from one hub to one module?



  4. “Of course they’re interconnected, but are the neurons of one hub distinct in not belonging to any other hub? Also, are there hub neurons?”

    You did ask if they were isolated. But pyramidal neurons in each region have long axons that connect either to relay nuclei in the thalamus, or directly to other cortical regions. Each region has local neurons that only communicate locally.

    “So the signal is not massively amplified so much as the other signals are suppressed.”

    Based on what I’ve read, it’s both.

    “How many “modules” do you think there are? And how many hubs?”

    Baars shows a diagram in his book listing 74 interconnecting regions for each cortical hemisphere, although I don’t know how comprehensive that was meant to be. There may be hundreds or thousands for all I know. It does fit with something Damasio wrote saying he thought there the number of CDR (convergence-divergence regions) may be in the many dozens. The hubs are fewer, and as I noted above, I’m not sure how many there are.

    “Presumably you are saying the signal from each hub can be sent to each module.”

    There are hub to module connections, hub to hub ones, and module to module ones. But a lot more around the hubs. Incidentally, almost all of these connections are two way. So signals from one region to another are strengthened if they’re reciprocated, leading to a binding recurrent circuit.

    “How many axons might it take to send one distinct signal from one hub to one module?”
    Don’t know. I suspect the numbers vary widely.


      1. On modules and hubs, not sure. I think you’re attaching too much meaning to these labels. There’s probably not a clean distinction between the two. Some regions are just far more massively connected than others, with no clean break.

        The thalamus is obviously heaving involved in the broadcasts and recurrent circuitry. (Which means the thalamic-loop theories are a part of the picture.) Given how ancient the diancephalon / pallium arrangement is, it may be a crucial piece of the architecture, particularly under Baars’ idea of the workspace. But most of the actual information is going to be in the various regions throughout the system.


        1. I’m trying to work out the mechanics of your model. So I originally took “module” to mean what Ruth Millikan calls a unitracker, which is a mechanism (isolatable set of neurons) whose purpose is to track one concept. But one unitracker for each concept would mean hundreds of thousands to millions of unitrackers, and that’s what I think cortical columns are, but this does not seem to correspond to what you mean by “module”.

          So this brings me too: What is a “module” and what is its relation to a concept?



          1. Again, just so we’re clear, this is not my model. I’m usually pretty clear whenever I put something out that’s mine. I do think it has a lot going for it though.

            I suspect a module is broader than just a single concept. It probably more relates to a type of concepts. Of course, you could model it down to the individual convergence zones, in which case we would be talking about thousands. But in the way Baars or others use it, I think it would relate to a functional collection of concepts.


  5. “Every so often, someone manages to gain the floor and make a speech, conveying their message throughout the room.”

    I like Dennett’s “fame in the brain” better, because it allows for the fact that multiple people can be famous at once. Similarly multiple items and features can be conscious at once. There isn’t just a red rose, there’s my wife giving me a red rose in the rain. It’s interesting how some items compete for consciousness, while others can share the spotlight without any problem.


    1. Definitely multiple items that are compatible with each other can share the spotlight. Which raises the question, how does the brain figure out what is and isn’t compatible? In binocular rivalry, if the pictures are similar, the brain will actually combine them. They have to be different enough to generate the alternating perceptions.

      That said, I think we also have to be cognizant that a conscious moment is very brief. What can appear to be one long conscious event may actually be a series of events that we lump together. So you take in the scene of you and your wife in the rain, focus on your wife, focus on the rose, focus back on your wife, then the rain, then a gist of the situation, etc.


      1. But even within that stretched-out view of a “present” moment, I think that my view of the rose actually depends on my view of my wife, and so on around the round-robin of items within the gestalt. So experience isn’t a salad, where items are fully distinct. It’s at least a stew, where they flavor each other, and sometimes it’s even a melting pot.


        1. I wouldn’t argue otherwise. But the capacity for different simultaneous things is limited. Baars sees it as 1-4 depending on the items. But that’s over a “conscious moment” ranging from 50-250 ms (100 ms being the most common). So within the span of a second or two, it can include a lot.


          1. Maybe not “simultaneous” so much as combined? (Like how waves add together. There’s only one wave present, but it’s a combination of various contributing simpler waves.)


          2. Something like the sight of red is often combined with the feeling of seeing red, with the overall experience being introspectively irreducible. But it often seems like there are related things that can be introspectively distinguished. Although maybe the act of “distinguishing” involves rapidly switching between them.


          3. FWIW, I did mean something akin to semantic pointers, although I don’t necessarily see those as more than a way to model whatever the heck the brain is actually doing.

            Whatever the underlying mechanism, I suspect concepts can combine into new monolithic (albeit compound) concepts. Our attention can tease out the simpler concepts that comprise it if we focus on them.


          4. And if we do focus on them, it’s that thing we’re focusing on that becomes the new dominant workspace item. The result is that information from different layers of the analysis can get their time in the spotlight.


          5. James S,
            I haven’t had a chance to go through this paper yet, but thought you might find it interesting:
            (SPC = semantic pointer competition)

            Although SPC challenges IIT, the idea that consciousness results from semantic pointer competition is compatible with
            theories according to which non-conscious processes operating in parallel compete for access to a global workspace that
            is distributed across many brain areas (Baars, 2005; Dehaene, 2009, 2014; Dehaene & Changeux, 2011). SPC is also compatible with the ideas of Damasio (2012) about how mental maps are formed in convergence-divergence zones in the brain. SPC
            goes beyond these neural theories in being more specific about the representations and mechanisms that constitute the global workspace.

            Click to access thagard.two-theories.consc&cog.2014.pdf


          6. I have seen that paper. My impression was that Thagard didn’t quite “get it”. Which is to say he doesn’t see semantic pointers quite the way I do. It’s not so much that semantic pointers compete as it is that unitrackers (you really should add Ruth Millikan’s most recent book to your reading list) compete to generate associated semantic pointers in the global workspace. Also, unless I missed it, Thagard completely missed the chance to explain how semantic pointers are compatible with IIT.

            Also, Thagard, who is essentially the local philosopher for Eliasmith, both working at U. Waterloo, wants to identify Consciousness with semantic pointers and therefor deny it to anything without the correct neural physiology. I’m not sure he is considering correctly the relation between semantic pointers and representation. Consciousness is about the representation [pretty sure].



          7. I’ll defer to your knowledge on semantic pointers, but if Thagard doesn’t get them, then I fear I don’t either. His and Stewart’s concise descriptions of them seemed clearer than just about anything else I’ve read.

            Thagard and Stewart seem pretty skeptical of IIT. I get the impression they regard it as gobbledygook, quoting Tononi directly to avoid misstating the theory. (It reminds me of the care I had to use when relaying the axioms and postulates in my post on Koch’s book, for exactly the same reasons; the language seemed vague and obfuscated.) Did you see their appendix on IIT’s mathematics?


          8. First, never defer to my knowledge. I am definitely in over my head. I need you to figure out where and why I’m wrong.

            Second, it’s not that Thagard doesn’t get them at all. It’s just that his understanding about some parts seems to be subtly different from mine. So for example, here’s the fundamental statement of his hypothesis:

            The crucial mechanisms for consciousness are: representation by patterns of firing in neural populations, binding of these representations into semantic pointers, and competition among semantic pointers.

            My understanding is that (for the purpose of this discussion) there are two basic kinds of “neural populations” that we care about: 1. a population that always fires in the same pattern when activated (say, when a red ball is in the visual field), and 2. a population which is thoroughly self-integrated and can (and does) fire in a vast array of patterns (think vectors). Figure 1 from the paper depicts type 1 populations at the bottom feeding into (i.e., being bound into) a type 2 population at the top. The representation produced by the type 2 population is the semantic pointer. If a different group of type 1 inputs were activated, they would produce a different semantic pointer in the type 2 population. If exactly one of the type 1 populations was the only input, it would produce its specific semantic pointer in the type 2 population.

            So my question is: in what sense do semantic pointers compete with each other? I can see type 1 populations competing to establish their respective semantic pointers in the type 2 population, but I’m not sure that’s what Thagard thinks. Also, I don’t know how he differentiates between binding and competing. Say there are two type 1 populations, A and B, providing input. I expect sometimes they are bound, giving A+B, and sometimes they compete, giving A or B. Thagard would have to suggest that there is some control mechanism within the type 2 population that determines this. Personally, I think it would be easier to simply suppress one or the other of the type 1 populations.

            And before I forget, I’m postulating:
            type 1 population = unitracker (and maybe cortical column)
            type 2 population = global workspace



          9. Always remember that I’m just a fellow student in all this. Always push back if my logic doesn’t make sense.

            I’m a little uneasy about the population 1 vs population 2 distinction. I don’t know that neural populations ever fall entirely into either one of those types you described. That said, the ones in the hippocampal regions, due to their unique roles in navigation and memory formation, might be the closest to population 2.

            My understanding of Thagard is that he’s using the term “semantic pointer” to refer to the symbolic representation in each region which form from binding signals from many other regions. It’s these patterns of bindings and distributed symbols together which compete for dominance. Each circuit, in typical neural fashion, has lateral inhibition projections. In other words, each circuit attempts to inhibit the alternate ones. The circuits that win are the ones that attract a large enough coalition of sending and receiving regions, which results in the P3 wave and suppression of all the others (or at least all the ones not coherent with the winning one).

            In other words, I don’t think it’s right to envision all the semantic pointers being in one location. They’re probably spread throughout the thalamocortical system. At least that’s what Thagard’s assertion of compatibility with mainstream GWT lead me to believe. His description seems to make SPC a GWT variant, or an extension of it.


          10. Regarding IIT, if you look at the table in the paper which compares the two theories you can make exactly one change to IIT, and then SPC becomes a strict subset of it.

            If instead of saying “Consciousness is Integrated Information”, IIT said “Consciousness is about certain processes, and every conscious process involves integrated information as described in the theory, then everything Thagard proposes is compatible with IIT.



          11. That might be, but you may be describing a version of the theory that Tononi and Koch wouldn’t agree with. Most of the stuff I’ve read about IIT equates consciousness with integration. If they backed off to saying it involves multiple integrating processes, then it might start blurring into GWT territory. T&K knew about GWT, and they always had the option of making their theory an enhancement of it, but that’s not the path they chose. They seem to be convinced that the reality is something very different.

            Everyone agrees integration is a major factor, it’s just the IIT variant that a lot of people are questioning.


          12. I know Tononi and Koch would not agree with the change, and that’s entirely (I will wager) because of the mistaken postulate that “Consciousness is a unified field”. They make this postulate because, you know, it “feels” like a unified thing. It’s this postulate that motivates the goofy idea that as soon as the integration of a whole becomes greater than the integration within any of the parts, Consciousness in the parts winks out. It’s this postulate that motivates saying Consciousness just is Integration.

            But if you just jettison that one postulate, all the rest is still valid, and useful. When you calculate something like Phi from brain activity, that’s real and useful. If you jettison that postulate, IIT becomes compatible with all of those other theories, and augments them. Discarding IIT because of that postulate is like throwing out the baby because it’s contaminated with a little bath water.



          13. That gets back to the point of that paper I shared last summer, which pointed out that many theories of consciousness make useful contributions, if we just let go their claim to be the final theory.

            Although I’m less sure of that assertion for IIT than I used to be. A lot of people seem to be saying its mathematics are ambiguous and difficult to work with. And the evidence Koch claims to corroborate IIT (of wide scale synchronized activity correlating with conscious report) also corroborate GWT.


  6. With some of these theories I feel it is like trying to understand football by reading the rule book. Even if everything that happens in the brain goes according to the rules, that doesn’t mean the rules capture what the game is about.


      1. It seems to me there should be some way of drilling down to a lower level or a logical direction to pursue. But that is in part why I find the EM theories interesting. The EM theories are down at to the physical level. Ultimately whichever of theories – GWT, HOT, etc – you like you would expect something physical to underlie consciousness but as best I can tell they really all have the same idea. They all think firing of neurons causes consciousness and they only differ in how they think the neurons fire and maybe which parts of the brain they think are more important.

        Liked by 1 person

        1. GWT and HOT do take mainstream cognitive neuroscience as their baseline, which itself takes the physics of mainstream neuroscience as its. But GWT and HOT are information processing theories and would also be compatible with an underlying EM field mechanism, if the evidence were to lead in that direction. Of course, they’d also be compatible with silicon or other technological substrates.

          On the other hand, if an EM theory insists that it’s the final answer, then they of course wouldn’t be compatible with that. But I’m pretty skeptical that any one theory is going to be the final answer.

          The question that keeps occurring to me though, is why EM fields are thought to be an improvement over neural processing. I think we can safely say that those fields, along with other environmental factors, perturb neural processing. But what do they bring as a complete alternate substrate? What extra explanatory work do they do?


          1. Do the other theories have a hypothesis about the binding problem or how neurons are synced to fire together in different parts of the brain?

            My hunch to be explained in more detail eventually is they are major facilitators of learning and memory.


          2. Baars, in his book (the last part of which is pretty technical), spends a lot of time on the binding recurrent (two-way) circuits that form between the originating and receiving regions. The connections form and lead to synchronized firing between the regions. They hold until the concept fades or is superseded by another one. I would imagine Dehaene covers similar ground in his more technical stuff.

            I checked them somewhat by reading on attention processing in Marie T. Banich’s Cognitive Neuroscience textbook, which describes similar dynamics, with regions working together syncing their oscillations. (Although Banich labels it for attention and seems to favor IIT for consciousness itself.)

            Remember, the thalamo-cortical system is profoundly interconnected, and the long range connections are mostly reciprocal, allowing recurrent (“reentrant”) processing.


          3. That link goes to “The interplay of attention and consciousness in visual search, attentional blink and working memory consolidation”.

            Is that what is intended? There is a Rube Goldberg like diagram in there showing a bunch of nodes connecting to each other but why consciousness emerges in some of the nodes but not others isn’t clear to me. Maybe it would become more clear if I spent more time with it .


          4. That link does look more relevant.

            What strikes me on initial inspection is that much of it would align perfectly with EM theories. Actually he seems to be presenting ample evidence, including one explicit reference to electromagnetic fields, for concluding that EM fields are involved significantly.

            “A variety of high-level wave-like phenomena emerge in the C-T system, including standing and traveling waves, spiral vortices, centrifugal propagation, phase coupling and decoupling, microstates, cross-frequency coupling, and hemisphere-wide phase transitions at theta-alpha rates (Freeman et al., 2003; Izhikevich and Edelman, 2008). Complex waveforms in the core range from 0.1 to 200 Hz, with momentary spikes up to 600 Hz.

            The basic unit of the C-T is therefore not the single neuron, nor the traditional one-directional sensory pathway. It is rather “a unit of adaptive resonance,” which can be thought of as an artificial neural network with at least two layers. ”

            “This is not to minimize the role of neuronal spike timing, which is known to evoke synaptic plasticity. The C-T core may use multiple neuronal codes, including non-classical propagation via glial cells, electromagnetic field induction, electrical synapses, and even membrane ion currents. In conventional terms, brain rhythms interact with single cell activity. Population rhythms require interacting single cells that are both excitatory and inhibitory. Regional rhythms like theta can also recruit single cells that phase-adapt to the peak of the regional wave ”

            “Thus the waking state appears to enable brain-wide, adaptive resonance in the cortex and thalamus. By comparison, slow-wave sleep – the least conscious state of the daily cycle – may interrupt processing due to massive firing pauses in the trough of the delta wave ”

            “Phase coupling in the gamma range is tightly linked with the waking state. Slow-wave sleep and REM dreaming show a drastic drop in array-to-array gamma coherence. Observed gamma synchrony may reflect underlying theta-gamma cross-frequency coupling, which may in turn be carried by Slow Oscillation-theta-alpha coupling, as suggested by Steriade (2006). Slow Oscillations (SOs) continue throughout the daily cycle at <1 Hz. Widespread SO’s may change the firing threshold among billions of neurons, just as regional theta activity might do at a faster rate. A slow-to-fast wave hierarchy is one way to coordinate large populations of neuronal oscillators in the brain."

            "The most interesting possibility is that reported microstates reflect both binding and broadcasting in the brain, much like double-wing Lorenz attractors. Using a Hilbert analysis of the cortical EEG, Freeman et al. (2003) report that electrophysiological equilibria reflect a hemisphere-wide spatial phase change, occurring at theta rates. "


          5. Baars strikes me as the type to be more open to exotic processing. But his background is psychology and his grasp of neuroscience sometimes has issues. In multiple places in his book he makes statements that outright contradict what’s found from most neuroscience sources. For example, he says multiple times that the cortex as 50-55 billion neurons (it has around 16 billion), and makes some remarks about the thalamus that don’t agree with the vast majority of what’s out there.

            I found Baars writing more interesting for its conceptual content. Overall, I find Dehaene a lot more credible when it comes to the neuroscience itself. In his book, he discussed how various environmental factors, such as temperature, chemicals, and background fields, do perturb neural processing.

            But neither of them discuss EM fields as a communication mechanism. (At least not that I saw.) All of Baars synchronized firing comes from more mundane neural processing.


          6. “The C-T core may use multiple neuronal codes, including non-classical propagation via glial cells, electromagnetic field induction, electrical synapses, and even membrane ion currents.”

            Note electromagnetic field induction as an option among many other forms of non-classical propagation.


          7. Just wanted to throw in that the claustrum might have a role in this. It’s connected to everything. My guess is that it coordinates the synchronization leading to attention.



          8. Could be. The claustrum is heavily interconnected with the major players, prefrontal, parietal, temporal, etc, but it’s very small and its function isn’t well understood. When I do a search for it in my neuroscience books, the only mention is typically a quote from Christof Koch on his and Crick’s interest in it.


    1. Just keep that thought James. Rules are derived from the notion of law. Law is an ancient, primordial intellectual construct which postulates either one of two discrete outcomes, obedience or disobedience. Furthermore, law is an intellectual construction which reduces to absurdity: Law, that mystical, ethereal some “thing” which commands unwavering, unquestioning obedience from its unknowing, unsuspecting subjects. Talk about magic sauce, the notion of law is absurd.

      Human relationships, as well as the meaningfully complex relationship of particles in motion are more sophisticated than a primordial construction called law with its two discrete outcomes. From my short career as a blogger, the notion of law appears to be the most profound, insurmountable impasse for homo sapiens to overcome.



Your thoughts?

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.