Heading in the Wrong Direction

This is a very readable, informative book by a recognized expert and leader in neuroscience / cognitive science related to the problems of consciousness. It also reflects the esteem and affection held by Christof for his collaborator and friend Francis Crick who, sadly, died recently. Many of their mutual ideas are presented in the book which includes thoughtful, extremely useful recapitulations at the end of each chapter.

Admirably, Christof embraces the “hard problem”, acknowledging as real the issue of how conscious experience in the form of qualia arises from neural activities in the brain. But he avoids an assault, choosing instead to seek the neural correlate of consciousness—the NCC, a term and concept introduced by Crick. “A man’s got to know his limitations” Christof once said about this strategy, quoting Clint Eastwood.

Fair enough. Identifying the precise neuronal activity and location which correspond to consciousness would be a gigantic accomplishment.

Christof points to a variety of nonconscious activities “Zombie modes” which, he argues, are performed in the guise of consciousness. We merely (falsely) believe we are thinking, analyzing, planning and executing many actions consciously, but we are only actually aware of the final big picture. He relates this concept to Ray Jackendoff’s intermediate theory, in which consciousness is sandwiched between raw sensory percepts and executive planning.

I was intrigued by the chronicle of the 40 Hz story. As many people know, Crick and Koch jumpstarted the 40 Hz bandwagon in 1990 with their endorsement of gamma synchrony oscillations (discovered by Lord Adrian, I learned from Christof) popularized by Gray and Singer’s 1985 paper. Borrowing from von der Marlsburg’s notion of correlational binding, Crick and Koch proposed 40 Hz as the NCC, but backed away from it a few years later. In “The quest..” Christof explains that 40 Hz may be useful in promoting a particular coalition of neurons into the lead in competition among coalitions to win the prize of consciousness. But 40 Hz, he says, is not necessary for sustained consciousness. OK, but what is? And what does it mean to be sustained if consciousness is a sequence of discrete events (a property deduced elsewhere in the book)?

Likewise the famous “V1 isn’t conscious” story was enlightening. The notion of the back of the brain “looking” at the front of the brain is fascinating.

Descriptions of coalitions of neurons forming and dissolving like political alliances is nicely written, although I don’t see much difference between these coalitions and Donald Hebb’s neural assemblies, described a half century ago. Coalitions/assemblies are groups of neurons connected by axonal-dendritic (or axonal-somatic) chemical synapses. Axonal action potentials “spikes” within the coalitions/assemblies are the currency and sine qua non of consciousness, according to Christof. But are they?

Cursorily, Christof considers other candidates for the neural-level substrate of the NCC. He mentions that local field potentials (LFPs) are detectable everywhere in the brain, centimeters from their origin. As a global means of communication, however, he concludes that the LFP electromagnetic field is a “crude and inefficient way for neurons to share information”. Perhaps so, but Christof neglects to consider the source of LFPs (i.e. rather than LFPs themselves) as the NCC. This would be prudent if one believes fMRI to be a good indicator of the NCC. Nikos Logethetis and colleagues showed a few years ago that the BOLD signal in fMRI corresponds with LFPs more so than with axonal spikes.

According to Walter Freeman, potentials recorded from scalp or brain surface are not truly LFPs but “open” potentials stemming from dendritic activities of pyramidal cells with axial symmetry. LFPs derive from cortical interneurons with radial symmetry.

Which brings us to Christof’s next NCC candidate: “groups of inhibitory cortical neurons linked by electrical synapses known as gap junctions. Under some conditions all these interneurons trigger action potentials at the same time, acting as a single unit.” But Christof dismisses the possibility: “Not enough is known about this phenomenon to implicate it in conscious perception.” A footnote to the last sentence cites 3 papers showing that cortical interneurons linked by gap junctions generate synchronized activity at around 8 Hz.

But actually, in the past 5 years dozens of papers have shown that cortical interneurons linked by gap junctions are responsible for gamma synchrony/40 Hz. The interneurons are linked by dendro-dendritic gap junctions so that membranes are continuous, and depolarize synchronously, “behaving like one giant neuron” as Eric Kandel says in his textbook. The gap junctions are windows between cell interiors so that cell cytoplasm is also continuous. Gap junction networks are, almost literally, one giant neuron (or a “hyper-neuron” as Roy John christened them), but yet as changeable as a Hebbian assembly. Gap junctions may also link dendrites to glia, glia to glia, axons to axons, and axons to dendrites, over-riding chemical synapses. Some thalamo-cortical neurons are also linked by gap junctions in thalamus. The evidence is unclear regarding gap junctions between cortical interneurons and primary (e.g. pyramidal) neurons, but “hyper-neurons” could include primary cortical neurons via direct gap junctions and/or glial bridges including those bypassing chemical synapses. Though far sparser than chemical synapses, electrotonic gap junctions are prevalent significantly in the mature brain. Gap junction hyper-neurons mediate gamma synchrony/40 Hz via cortical interneurons, and may extend widely through cortical and sub-cortical regions. Could they mediate consciousness?

Christof continues: “How else but with spikes can the highly peculiar character of any one subjective experience–a subtle shade of pink or a rhapsodic waltz tune–be communicated across multiple cortical and subcortical regions?”

This question drips with irony. First, all of a sudden we seem to be tackling the hard problem. Or are we? Does character of qualia imply qualia themselves, or some pertinent information about them?

If “the highly peculiar character” of qualia means qualia themselves, then the question should not be “How else but with spikes?” but rather “How in the world can spikes possibly produce qualia?”

But if Christof means the latter, that the character of qualia involves non-conscious information about them, then, yes, qualia (“The ball has top-spin”) could generate nonconscious instructions by axonal spikes to pre-motor and motor cortex: (“Bend your knees!”).

But whither qualia? Whither consciousness? To me the answer is obvious. Consciousness occurs in dendrites, and the results are conveyed elsewhere by axonal spikes. Eccles thought so, and so does Karl Pribram (“Dendrites, dendrites, dendrites!!”)

Dendrites are missing in action in “The Quest for Consciousness”. They are barely mentioned. Yet a review in Science characterized dendrites (whose potential information processing capacity per neuron can only be guessed at, so vast is it) as the “brain of the neuron”. Christof’s suggestion for producing qualia is neurons whose axons reverberate via cell bodies, bypassing dendrites altogether.

Two more quibbles: First, Christof describes how he once thought that understanding anesthesia would be useful in understanding consciousness, but became disillusioned. The dismay stems from two major points: a) a variety of types of molecules/drugs are used to produce anesthesia, and b) anesthetics act on a variety of brain proteins. Christof concludes anesthetics are “too blunt a tool” to discern the site of consciousness.

I disagree. While various types of drug molecules are used in anesthesia, the volatile, inhaled gas anesthetics are by and large the only true, complete anesthetics. As opposed to a smorgasbord of drugs, anesthetic gases by themselves erase consciousness (though typically we do add others). Anesthetic gases include a variety of types of molecules: halogenated hydrocarbons, ethers, the inert elemental gas xenon, nitrous oxide etc. And they act on a host of proteins. How to make sense of this morass?

It’s been known for a hundred years that anesthetic gases dissolve in a particular non-polar solvent medium, characteristic of olive oil (in fact anesthetic potency correlates with this solubility). This oily solubility medium is found in discrete “hydrophobic pockets” of certain brain proteins. The tiny pockets (~ 1/50 the volume of the protein) act as “the brain of the protein” because subtle (quantum mechanical) forces among intra-pocket electrons (e.g. from benzene-like rings in certain amino acids) control the protein’s conformational state (e.g. channel open or closed). Anesthetics get into the pockets and form their own quantum forces, thereby preventing the conformational activities.

Here’s the rub. At just the right anesthetic concentration, by and large ONLY consciousness is inhibited. Other brain activities (and bodily functions) are unaffected. Thus only proteins required for consciousness have critical hydrophobic pockets (or depend on subtle electron movements within them for function). So the site of action of anesthesia, and the site of consciousness, is not a specific brain location or type of neuron or particular protein. The site/action of anesthesia and consciousness is a distributed phase of non-polar, hydrophobic solubility medium composed of discrete pockets in a group of proteins throughout the brain. In this hydrophobic phase, quantum mechanical forces rule.

Which leads to my next quibble. Christof dispenses with the Penrose-Hameroff quantum consciousness microtubule model early on (page 7 to be exact). According to Christof, Penrose-Hameroff is dead-on-arrival because of “the lack of any biophysical mechanism that would permit neurons, and not just any cells in the body, to rapidly form highly specific coalitions across large regions of the brain on the basis of quantum-coherency effects”.

Well, gap junction networks with a hydrophobic phase describe just such a mechanism (and only in the brain), assuming quantum coherence can be sustained in one neuron. That brings us to Christof’s second point: “All this is supposed to take place at body temperature, a rather hostile environment for sustaining quantum coherency over macroscopic scales.” See Grush and Churchland (1995) for a telling criticism.

It turns out that quantum coherence in some circumstances flourishes at high temperature (in which the coherence is pumped, rather than condensed by cooling). A laser is a longstanding example, and recent evidence shows that organic benzene molecules (identical to those in hydrophobic pockets of brain proteins) mediate quantum states more efficiently at high temperatures, and that noise can enhance quantum processes. Physicists Scott Hagan, Jack Tuszynski and I have shown how biophysical mechanisms (e.g. actin gelation, Debye screening) can sustain quantum coherent superposition in microtubules for hundreds of milliseconds. As for Grush and Churchland, their objections were answered tit-for-tat in an article by Roger Penrose and me in the subsequent issue of the journal in which their attack appeared (many quote the attack; few read the rebuttal). Quantum computation in microtubules is not proven, but it hasn�t been refuted. And unlike other theories, it is falsifiable and actually addresses the hard problem.

Let’s get back to Christof’s book. He’s bet the farm on spikes, and there’s no evidence to support the notion that spikes alone generate consciousness. Every instance of correlation between spikes and consciousness also includes dendritic activities. On the other hand there are reasons to believe (40 Hz, computational processing, dendro-dendritic interactions, post-synaptic site of anesthetic effect) that dendritic activities do generate consciousness. Perhaps quantum computations in microtubules in dendrites linked by gap junctions generate consciousness. But even if not, dendrites are in some way involved. The Quest for Consciousness is heading in the wrong direction.

Stuart Hameroff
Professor, Departments of Anesthesiology and Psychology
Director, Center for Consciousness Studies
The University of Arizona, Tucson, Arizona
www.consciousness.arizona.edu/hameroff
Copyright © 2004 Stuart Hameroff M.D.

References

1. Interview with Christof Koch