Hello, Monty and all other members of Tonmo. First of all, I want to thank Monty for the great review--and excellent summary--of the talk I gave at Caltech a couple of weeks ago. I think I enjoyed giving the talk as much as Monty enjoyed attending it!
Monty's synopsis of the major points I made was dead-on, but I'll reiterate the most central ones to keep potential extensions of this thread on point:
1) Owing to the influence of the American Behaviorist school (initially more theoretical; later, more methodological), as founded by John Watson and B.F. Skinner, the scientific study of consciousness was largely off the radar for much of the twentieth century. It's only been within the last 15 or 20 years that researchers have begun to revisit the field.
2) Based on findings from the study of conscious humans (both healthy and brain-damaged), it seems clear that certain anatomical regions (e.g., cortex and thalamus) and electrophysiological signatures (widespread, low-amplitude electrical activity in cortex) are important in the generation/maintenance of mammalian conscious states.
3) Contrary to traditional--and largely obsolete--nomenclature, bird brains contain "higher" structures that are homologous to those found in mammals (i.e., the avian hyperpallium corresponds to mammalian neocortex). Moreover, the genetics of avian brain development resembles that of mammalian brain development. Finally, many birds seem to be capable of quite complex behaviors (tool manufacture and use, logic-based problem solving, etc.) involving planning and even formulating predictive schema based on a "theory of mind"--that is, what the other guy could/might do. Given all of this, it wouldn't be a stretch to argue that birds are probably conscious; at the very least, the prospect and/or nature of avian consciousness is well worth investigating.
4) The principles underlying the generation of consciousness may not be confined to vertebrate groups. Clearly, specific evolutionarily conserved vertebrate structures and their electrophysiological interactions are at the core of conscious states. But it is possible that certain key vertebrate electrical properties and functional neural circuitry have analogs within the nervous systems of some invertebrates. Although mammalian and avian brains share deep genetic, anatomical, and physiological homologies, such homologies--at least at the level of gross observable structure--may not be required for consciousness. Rather, it may be that functional analogy is important here. For example, while one wouldn't find cortex or thalamus in jumping spiders, octopus, cuttlefish, or squid, it is possible that functional equivalents of such structures exist in these vertebrates. And, following from this, if such functional analogs exist, one might find evidence of interactions between them, in the form of signature electrical properties, that resemble those recorded from mammalian brains during conscious experience.
5) Given the overwhelming evidence for the behavioral complexity of many cephalopods (which, of course, I need not go into in detail for this group!), as well as recent evidence for the presence of something in the octopus vertical lobe that looks very much like long term potentiation (thought by many to be a fundamental physiological basis for learning in vertebrates), I believe that this group of animals bears further scrutiny by consciousness researchers.
6) Perhaps the most critical point I tried to make in my talk is the idea that consciousness could have arisen convergently within at least a few--quite disparate--animal radiations. Certain principles of gross inter-areal wiring (thalamocortical analogs), activity across circuits (e.g., widespread, low amplitude electrical activity, the functional mappings discussed by Gerald Edelman (yeah, my dad) in the books Neural Darwinism, Topobiology, and The Remembered Present, etc.), and finally, the manner in which variable and rapidly changing input from convergently evolved complex sensory structures such as eyes (particularly eyes!) gets processed and integrated may in fact be universal among [sufficiently] complex nervous systems. In sum, given enough time and a sufficient elaboration of complex ecologies in which sensory and motor adaptations (i.e., speedy detection, speedy movement) among predators and prey are under intense selection pressure, I think it's a fair bet that consciousness emerged in diverse animal forms on more than a few occasions.
Now to a couple of important bits that I didn't really cover:
First, some important things to consider when thinking about the evolution of nervous systems:
First, the world is an unlabeled place.
Second, no two signals are exactly the same across time or space, whether perceived by different nervous systems or by the same nervous system at different times. That is, the qualities of the physical world are constantly shifting, subtly or not so subtly. By their nature, nervous systems parse signals into categories of "same" or "different"; in effect, nervous systems impose meaning on the world, not vice-versa.
Third, in complex animals (think perhaps beyond the level of, say, C. elegans), no two nervous systems are exactly the same, even those of identical twins (clones). Myriad epigenetic interactions during development assure such individuality. So, for example, neurons never undergo arborization--they never connect up--the same way twice.
So, now we have a bit of a matching problem. How do highly individual, idiosyncratic, nervous systems interact with a constantly shifting world--a category-free continuum--and successfully do the things they need to do to maintain survival?
Here, I think, is part of the answer:
Starting about thirty years ago, my father, Gerald Edelman, suggested that, in brains comprising vast and degenerate repertoires of neurons that are hyper-densely wired together, higher functioning (learning, memory, and even consciousness) emerge via principles akin to Darwinian evolution. Over the years, the Theory of Neuronal Group Selection (TNGS), or "Neural Darwinism" has been elaborated upon, to some degree; more often than not, though, it has been met with vocal, and often caustic, criticism. In any case, here's a "nutshell" precis:
By birth or shortly thereafter, the nervous system comprises a vast primary repertoire of wiring combinations that has been shaped by epigenetic interactions during development. As an animal interacts with its world--maybe quite randomly at first (think of a flailing baby)--that world selects for particular functional sets of neurons; clusters of neurons that perhaps responded more quickly or efficaciously than others to salient stimuli (To a degree, one might think of "salience" as the one aspect of the system that is hardwired, in the sense of some very basic properties, such as extreme temperatures and eating vs. not eating, having evolutionarily programmed negative or positive salience; so, complex brains might have emerged with specialized salience detectors that regulate synaptic strengths "up" or "down" among/between neurons in "higher" centers like cortex). The synaptic connections between neurons in the selected clusters become strengthened. As a result, the next time a similar stimulus is presented to the animal, the likelihood of the same cluster of neurons firing together is much greater (think, "fire together, wire together"--a phrase most associated with Donald Hebb). At this point, a secondary repertoire emerges consisting of the functional groups that have been selected over the course of the animal's experience.
The nice aspect of such a system is that its strength is really in its degeneracy. That is, if the wiring of brains was point-to-point and quite specific (like a digital computer), it's hard to see how such brains could deal with the vast amounts of novelty they would encounter continually in the [unlabeled] world. A selection-based nervous system, in which no two neurons are exactly alike and no two circuits are identical, would seem to be able to handle novelty much better than a system in which constituent elements are identical and there is a hyperspecificity of wiring patterns. As is the case among evolving populations of animals, too much specificity--or specialization--is probably a bad thing over time; less specificity--or degeneracy--is ultimately a good thing.
In all of the foregoing, the notion of "degeneracy" is really critical. This is the idea, again, that there is no exact duplication, either in the phenotypes of individual neurons or in the functional circuitry of whole nervous systems. Think of this as a pool of variation that allows for a rich repertoire of neuronal interactions from which experience can select the "right" combinations of neuronal connectivity. It's very much like Darwin's notion of "species" (e.g., groups of [interbreeding] organisms that, while resembling one another, show a fair degree of variance).
I won't make the jump from selection-based brains to consciousness today--I've babbled on far too long as it is. But consider this idea as the basis for my optimism that cephalopod brains--though not nearly as complex as mammalian brains--are sufficiently complex to yield the sort of interactions among vast repertoires of functional neuronal groups that is necessary--and sufficient--for the emergence of conscious states.
Oh, one more bit: Daremo's question about comparing reentrant pathways (i.e., thalamocortical connectivity) in mammals to the anatomy of the cephalopod CNS is a really good one. Honestly, given the difference in CNS organization, it is hard to know where to look for such functional circuitry in cephalopods. But, as they say, "I'm on it like a cheap suit." Hopefully, we'll be able to say something about this in the near future.
For now, 'nuff said. Thanks for reading my screed!