(Warning: neuroscience weeds)
This is an interesting study getting attention on social media: Does the Prefrontal Cortex Play an Essential Role in Consciousness? Insights from Intracranial Electrical Stimulation of the Human Brain. Ned Block is one of the authors. (Warning: paywalled, but you might have luck here.)
The study looks at data from epileptic patients who had a net of electrodes implanted across various portions of their cortex. The electrodes are used to stimulate cortical regions to figure out where drug resistant seizures are coming from.
From the abstract. (Note: I took the liberty of breaking this into paragraphs.)
A central debate in philosophy and neuroscience pertains to whether PFC activity plays an essential role in the neural basis of consciousness. Neuroimaging and electrophysiology studies have revealed that the contents of conscious perceptual experience can be successfully decoded from PFC activity, but these findings might be confounded by postperceptual cognitive processes, such as thinking, reasoning, and decision-making, that are not necessary for consciousness.
To clarify the involvement of the PFC in consciousness, we present a synthesis of research that has used intracranial electrical stimulation (iES) for the causal modulation of neural activity in the human PFC. This research provides compelling evidence that iES of only certain prefrontal regions (i.e., orbitofrontal cortex and anterior cingulate cortex) reliably perturbs conscious experience. Conversely, stimulation of anterolateral prefrontal sites, often considered crucial in higher-order and global workspace theories of consciousness, seldom elicits any reportable alterations in consciousness.
Furthermore, the wide variety of iES-elicited effects in the PFC (e.g., emotions, thoughts, and olfactory and visual hallucinations) exhibits no clear relation to the immediate environment. Therefore, there is no evidence for the kinds of alterations in ongoing perceptual experience that would be predicted by higher-order or global workspace theories.
Nevertheless, effects in the orbitofrontal and anterior cingulate cortices suggest a specific role for these PFC subregions in supporting emotional aspects of conscious experience.
Overall, this evidence presents a challenge for higher-order and global workspace theories, which commonly point to the PFC as the basis for conscious perception based on correlative and possibly confounded information.
If you’re not familiar with the anatomical references, stimulating the upper and front portions of the prefrontal cortex had limited or no effect on conscious experience. However, stimulating the lower regions did have effects on emotional (affective) experiences. The paper can make for pretty dry reading, but it does have some very interesting diagrams, particularly on page 8.
The authors may have a point that this puts pressure on at least some HOTTs (higher order thought theories), particularly the ones that put the full conscious perception in the PFC. HOTTs posit that for anything to be conscious, there must be a higher order representation or thought about the first or lower order sensory representation. These higher order structures are thought to reside in the PFC, particularly including the regions which when stimulated lead to few or no reported effects. (Although I expect HOTT advocates to mount a response.)
However, I think the conclusions about GWT (global workspace theory) are hasty, and may involve an incorrect understanding of this class of theories. It begins with this statement.
It is important to distinguish between the neural constitution of consciousness and contingent global enabling factors of consciousness (Boly et al., 2017). For example, blood flow enables conscious processing but does not constitute it (but see Moore and Cao, 2008). Therefore, advocates of first-order recurrent activation can allow that PFC activity may control arousal, thus enabling sensory cortices to be activated, yet nonetheless argue that PFC activity is not constitutive of consciousness. Similarly, proponents of prefrontal involvement can allow that, without sensory input, there would be no perceptual consciousness, yet argue that only PFC activity plays a constitutive role. Throughout this Viewpoints article, we are concerned with the neural constitution of consciousness, not the enabling conditions.
This is a problematic standard for GWT, because GWT sees the entire thalamo-cortical system as being constitutive of consciousness, as well as seeing it all, along with numerous subcortical regions, as enabling. Put another way, under GWT, there is no particular region of the brain which is conscious. Consciousness is a system wide effect. For something to be conscious, it must be widely broadcast in the thalamo-cortical system and reacted to by most, or at least a substantial portion of the various subsystems. Just about any region in this system can be one of the sources of the broadcasted content in any particular instant, but no region is always conscious.
It is true that Dehaene’s GNW (global neuronal workspace) theory has historically seen a special role for the PFC, but I think this is more about its role in broadcasting content than where consciousness takes place. It’s worth noting that GNW also sees the parietal regions as part of that broadcast hub. But as noted in the last couple of studies I’ve shared, this role for the PFC as historically described looks increasingly unlikely.
The results of this study actually seem in accord with the study that identified various subcortical locations for the global workspace broadcast hubs. As the authors of that study noted, their findings don’t contradict Dehaene and Changeux’s overall model so much as fine tune it. The other study which identified possible stages of workspace, proposing the concept of a playground, was seen by Dehaene as confirmative, if again fine tuning.
All of which is to say, the data presented in this paper doesn’t seem, to me, to put much pressure on GWTs, except in terms of particular aspects of global neuronal workspace theory, aspects which are already having to be adjusted. It does give possible insights into which regions may be more connected to the broadcast hubs, and that in and of itself is very interesting.
More broadly, I think the authors’ conclusions show how much they’re operating under the theater of the mind paradigm and so attempting to find that theater, to find the presentation, the movie. (Admittedly, many HOTT advocates are in a similar mode, albeit seeing the movie in a very different place.) I’m not sure whether GWT is the answer. Even if it is true it’s likely only part of the overall answer. But even if it isn’t, I think looking for the presentation will be unproductive.
But maybe someone will find evidence tomorrow and prove me wrong. What do you think? Am I overlooking what the data in this paper means for GWT?