The global neuronal workspace

Global Workspace Theory - tutorial I first mentioned the global neuronal workspace in the post Toward a Culture of Neurons.   It is also discussed in Consciousness, Confessions of a Romantic Reductionist. Stanislas Dehaene and his colleagues have done much to enhance and improve the GNW Model, but original credit goes to B.J. Baars who was the author of the 1997 book, In the Theater of Consciousness:  The Workspace of the Mind. The GNW model relies upon a few simple assumptions. Its main premise is that conscious access is global information availability: what we subjectively experience as conscious access is the selection, amplification and global broadcasting, to many distant areas, of a single piece of information selected for its salience or relevance to current goals. Although today the relevance to judgment and decision making may be indirect, it certainly shares the limited capacity of consciousness and vast capacity of the unconscious with parallel constraint satisfaction theory.

According to Dehaene et. al., from a neuronal architecture standpoint, two main computational spaces are distinguished within the brain, each characterized by a distinct pattern of connectivity (see Fig. 1): (1) a processing network, composed of a set of parallel, distributed and functionally specialized processors or modular subsystems that “encapsulate” information relevant to its function. Processors typically operate non-consciously and in a bottom-up manner, although local top-down projections may also contribute to their operation by providing local predictions and prediction errors; and (2) a global neuronal workspace (GNW), consisting of a distributed set of cortical neurons characterized by their ability to receive from and send back to homologous neurons in other cortical areas horizontal projections through long-range excitatory axons. Such long-range cortico-cortical connections include callosal connections and mostly originate from the pyramidal cells of layers 2 and 3 that are particularly elevated in prefrontal, parieto-temporal and cingulate associative cortices, together with their thalamo-cortical relationships . GNW neurons typically accumulate information through recurrent top–down/bottom–up loops, in a competitive manner such that a single representation eventually achieves a global conscious status. Because GNW neurons are broadly distributed, there is no single brain center where conscious information is gathered and dispatched but rather a brain-scale process of conscious synthesis achieved when multiple processors converge to a coherent metastable state. According to the GNW hypothesis, conscious access proceeds in two successive phases. In a first phase, lasting from ~100 to ~300 ms, the stimulus climbs up the cortical hierarchy of processors in a primarily bottom–up and non-conscious manner. In a second phase, if the stimulus is selected for its adequacy to current goals and attention state, it is amplified in a top–down manner and becomes maintained by sustained activity of a fraction of GNW neurons, the rest being inhibited. The entire workspace is globally interconnected in such a way that only one such conscious representation can be active at any given time. This all-or-none invasive property distinguishes it from peripheral processors in which, due to local patterns of  connections, several representations with different formats may coexist.

Dehaene explains that an important statement of the GNW model is that the GNW network is the seat of a particular kind of brain-scale activity state characterized by spontaneous “ignitions” similar to those that can be elicited by external stimuli, but occurring even in the absence of external inputs. A representation that has invaded the workspace may remain active in an autonomous manner and resist changes in peripheral activity. If it is negatively evaluated, or if attention fails, it may, however, be spontaneously and randomly replaced by another discrete combination of workspace neurons, thus implementing an active “generator of diversity” that constantly projects and tests hypotheses on the outside world.  When the ascending vigilance signal is large, several such spontaneous ignitions follow each other in a never-ending stream and partially prevent access to external stimuli, possibly capturing empirical observations of inattentional blindness and mind wandering. The dynamics of workspace neuron activity is thus characterized by a constant flow of individual coherent episodes of variable duration, selected by specialized reward processors.

With these ideas, the idea of consciousness as emergent from simpler phenomena seems appropriate.  Looking at the complex patterns that can emerge from simple rules of cellular automata or “the game of life” especially when there is intermediate interdependence, I can see parallels with the upward and downward electrical signals.

Explicit formulations and computer simulations of the GNW architecture and physiology have been developed, leading to specific experimental predictions that have been confirmed with various types of imaging. Simulations have explored the sequence of activity leading to conscious access. When sensory stimulation was simulated, using a brief  input at the lowest thalamic level, activation propagated according to two successive phases: (1) initially, a brief wave of excitation progressed, with an amplitude and duration directly related to the initial input; and (2) in a second stage, mediated by the slower feedback connections, the advancing feed-forward wave amplified its own inputs in a cascading manner, quickly leading the whole network into a global self-sustained reverberating or “ignited” state. Thus, this second phase of the simulation reproduces the signatures of conscious access – late, all-or-none, cortically distributed potentials involving prefrontal cortex (PFC) and other high-level associative cortices, with simultaneous increases in high frequency power and synchrony. Even at higher stimulus amplitudes, the second global phase could also be disrupted if another incoming stimulus had been simultaneously accessed. Such a disruption occurs because, during ignition, the GNW is mobilized as a whole, some GNW neurons being active while the rest are actively inhibited, thus preventing multiple simultaneous ignitions. A strict seriality of conscious access and processing is therefore predicted and has been simulated.

In summary, simulations of the GNW architecture explain the close similarity of the brain activations seen during (1) conscious access to a single external stimulus, (2) modulation of conscious access by changes in vigilance and (3) effortful serial processing, even in the absence of an external stimulus. In all of these cases, the GNW provides a working memory space that can be temporarily detached from incoming stimuli and operates in an autonomous manner.

Dehaene asks if the emerging understanding of conscious processing eventually propose novel therapeutic and pharmacological tools for patients suffering from impaired consciousness? He answers that deep brain stimulation (DBS) in the thalamus, upper brain stem, and associated targets is advanced as a method to restore consciousness after loss of consciousness (LOC) due to severe brain injury.

Dehaene, S., Changeux, J-P., & Naccache, L.(2011). “The Global Neuronal Workspace Model
of Conscious Access: From Neuronal Architectures to Clinical Applications.” S. Dehaene and Y. Christen (eds.), Characterizing Consciousness: From Cognition to the Clinic? Research and Perspectives in Neurosciences, Springer-Verlag: Berlin Heidelberg.