The location of the global workspace

(Warning: neuroscience weeds)

I’ve discussed global workspace theories (GWT) before, the idea that consciousness is content making it into a global workspace available to a vast array of specialty processes. More specifically, through a neural competitive process, the content excites key hub areas, which then broadcast it to the rest of the specialty systems throughout the cerebral cortex.

A key question is where these hub areas might be. Strictly speaking, in Bernard Baars’ original conception, the location of the workspace is the entire thalamo-cortical system. But often discussion of where the workspace might be is in terms of where the core broadcast hubs are located. Baars, at one time, thought it might be in the midbrain and thalamus complex, although the data hasn’t seemed to bear that out. Stanislas Dehaene, in his global neuronal workspace variant, saw the prefrontal cortex and parietal regions as likely candidates.

A new paper in Nature claims to have narrowed down the locations by studying the flow of information across the brain using a new technique they call NDTE (normalized directed transfer entropy). They use this technique to identify a set of regions they refer to as FRIC:

 For this purpose, we propose the concept of a ‘functional rich club’ (FRIC) as the core set of regions, an array of functional hubs that are characterized by a tendency to be more densely functionally connected among themselves than to other brain regions from where they receive integrative information.

So what are the locations of these regions? The results are somewhat surprising:

The GW was found to consist of a core subset of brain regions including the precuneus, posterior and isthmus cingulate, nucleus accumbens, putamen, hippocampus and amygdala.

The first two regions are cortical ones, locations inside the midline area between the cerebral hemispheres, and overall part of the parietal regions. The rest are all subcortical, although part of the forebrain, and heavily interconnected with cortical regions. I find the absence of the thalamus and prefrontal cortex here somewhat striking. The authors go on to mention what the functional roles of these regions are thought to be:

This core functional ‘club’ of integrative brain regions is consistent with the original proposal by Dehaene and Changeux15, which suggests that the global neuronal workspace must integrate past and present through focusing and evaluation. Indeed, Dehaene and Changeux proposed that associative perceptual, motor, attention, memory and value areas interconnect to form a higher-level unified space. For the integration of the past, the hippocampus has been shown to play a key role in many aspects of memory (see, for example, refs. 44,45,46). Similarly, the evaluation of value has been shown to involve the nucleus accumbens (see, for example, refs. 47,48,49), putamen (see, for example, refs. 49,50) and amygdala (see, for example, refs. 49,51,52,53,54). The integration of the past, present and future by processing and attending perceptual information has been strongly associated with the precuneus (see, for example, refs. 55,56,57) and the posterior and isthmus cingulate cortices (see, for example, refs. 49,57,58,59,60). Interestingly, the functions of the precuneus have also been shown to be compromised in coma and vegetative state61.

So, assuming these results hold up, we have an idea of where the central regions of the workspace may reside. (Given that this was published in Nature and likely heavily peer reviewed, I have no particular reason to think they won’t hold up, but this is science and replication will be crucial.) Even if the GWT turns out to be wrong, it seems evident that these regions will need to have crucial roles in whatever theory, or more likely collection of theories, does eventually turn out to be right.

What do you think of these results?

23 thoughts on “The location of the global workspace

  1. First reactions.

    The conclusions seem to be drawn from a highly massaged and statistical analysis of fMRI data.

    How is this not subject to some of same issues found in the Duke study?

    Also, if we do assume the brain activity is correctly identified, it just seems like they are simply arbitrarily assigning it to the GW. But on what basis? How do we do not that most of the activity isn’t completely unconscious processing? What necessarily ties it to GW?

    I do note that the results all correlate with electromagnetic activity:

    “This clearly demonstrates the robustness of the NDTE framework for both haemodynamic and direct electromagnetic measures of brain activity”.

    Liked by 1 person

  2. Actually aren’t most of the areas the same as those frequently identified as part of the default mode network.

    ” some structures that are generally included are the medial prefrontal cortex, posterior cingulate cortex, and the inferior parietal lobule. A few of the other structures that may be considered part of the network are the lateral temporal cortex, hippocampal formation, and the precuneus.

    Not 100% match but enough so that maybe they have just done a more precise mapping of the DFN.

    Liked by 2 people

    1. It looks to me like they just found the average most active brain regions across various tasks and resting and decided to call it the GW. It’s not surprising there is a lot of overlap with the DMN. Whether they discovered something else, I’m doubtful.


    2. On how it escapes the Duke critique, I don’t know. I’d have to parse both the Duke issue and the methodology far more than I have. But I think the criticisms from the Duke study are being considered in all current research. And it’s worth noting that the Duke criticisms aren’t put forth as applying to all fMRI research, only to those using, or perhaps more accurately, over using, certain techniques.

      Not sure on unconscious processing. But given the information moving through these regions, as I noted in the post, even if the GWT ends up being wrong, it seems like these regions will be important.

      It makes sense that there’d be overlap with the DFN. I guess the question is whether there’s also overlap with the DAT (dorsal attention network), since those are supposed to alternate.


      1. Apparently the prefrontal and parietal structures are part of DAT so it doesn’t seem to be an overlap. What that means if the regions in this study are the GW doesn’t seem particularly clear. It seems it would mean that if we are concentrating really hard on a task then the GW isn’t involved, which wouldn’t make a lot of sense. 🙂


        1. I think it’s important to remember how interconnected the nucleus accumbens, putamen, and amygdala are with cortical regions, like the prefrontal cortex. I suspect the activity in cortical regions go through these subcortical ones to produce the synchronous effects across regions that we see.


  3. I find the absence of the thalamus and prefrontal cortex here somewhat striking.

    Thalamus, I get. But while your PFC is deeply engaged in some calculation, it would be nice if some other part of your brain could alert the global workspace that the sexiest woman in the tribe just smiled at you. Or that a lion is sneaking up.

    Liked by 1 person

    1. Yeah, a lot to be said for making sure your perceptual and evaluative systems are working. The nucleus accumbens, putamen, and amygdala are all heavily interconnected with the prefrontal cortex, so I think a lot of the evaluative stuff flows through them. The lion would definitely be an amygdala event.


  4. Imagine if my computer had to broadcast a digital sound signal, convert it to analog, transmit it through the sound speakers, re-up take these sound waves again by a microphone, to continue it’s information processing. It would be a pretty appaling computer. I’m against every interpretation that consideres consciousness as part of information processing.

    Liked by 1 person

      1. First of all I consider it still quite a mysterie. But nowadays, just like Socrates, I try to ask questions. I ask people: “if you would find out that your wife or best friend or everybody you know for that matter is a super highly functional robot, and you are the one or only one with real consciousness. Could you continue just living. After all, you lived your whole life already as it was. Or would there be something really bothering you?” Find out what that “really” is and you will find out the meaning of consiousness.

        Liked by 1 person

        1. That question assumes that your wife or best friend could be everything they are without being conscious. It essentially assumes that philosophical zombies are possible.

          But here’s a question. Suppose that your p-zombie wife or friend computed that they were conscious, and would become very angry if you told them that they weren’t. If you asked them to describe their experience to you, they could do so in a convincing manner.

          Furthermore, suppose they asserted that it was in fact you who was the p-zombie, and they went on to deny every statement you made that you actually are conscious. Indeed, they told you that the consciousness you’ve always experienced is not the real consciousness, just a fake version, that you’ve never actually known the real version.

          By what standard would the two of you adjudicate who was conscious and who was just computing that they were conscious? Would you, at any point, doubt which category you might fall in? If not, why not?

          Liked by 1 person

    1. Interestingly enough, for much of human history, the soul was thought to exist in the heart, although even in ancient times there were people who argued that the brain was where it was at. A consensus on the brain didn’t form until the 1600s during the scientific revolution.

      Democritus, the inventor of atomic theory, thought that the intellect was in the brain but emotion in the heart. The reasons we still talk metaphorically that way may be because of a pre-Socratic philosopher in the 400s BC.

      Liked by 1 person

        1. I lean Epicurean so I’ve definitely perused that work. Pretty cool how it was lost to history until a single copy was discovered in a monastery in the early 1400s. You have to wonder what the story of that manuscript was, how it survived all that time.

          Liked by 1 person

          1. There was a story written about its discovery. Poggio comes to mind. Look up “The Swerve…” Stephen Greenblatt.
            Yeah, I’m 1/3 stoic, 1/3 epicurean and 1/3 nihilist/existentialist (or there abouts). Or, in other words, just bloody mixed up and confused.

            Liked by 1 person

  5. What strikes me about this study is the mundane fact that a human brain is about the size of a small globe. If you were mapping the various structures mentioned in the article to an actual globe, there’d be a bunch of nueroscientists saying things like “the Global Workspace region is definitely in Scotland, Tazmania or the lower mantle.”

    Me thinks we have a lot to learn about the brain.

    Liked by 1 person

    1. I don’t know that that’s fair to the study, or to neuroscience in general. If someone could measure the human information flow around the globe, you don’t think they’d come up with meaningful answers on the location of cultural or economic centers?


      1. I think that’s a good example. If aliens observed us from afar, they might decide that New York and Tokyo were important based on their heat signatures, electricity consumption and the amount of signals they get from other centers. And this would be true.

        It would just be a very rough grain picture, I think.

        Liked by 1 person

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.