Thanks to Christof Koch for his response to my critique of his book “The Quest for Consciousness”. I believe this type of debate is quite useful, and take the opportunity here to reply to the three major points he raised.

Gap junction hyper-neurons

Regarding my suggestion that gap junction-connected neurons (“hyper-neurons”) may be the neural correlate of consciousness, Christof raises the issue of connexin-36 (a brain gap junction protein) knockout mice who appear relatively normal from a cognitive standpoint, and are presumably conscious. (This exact point was debated on PSYCHE-B a year or so ago, raised by Johnjoe MacFadden). Christof notes that gamma synchrony continued in the knockout mice, though reduced.

As Christof notes, there at least ten types of connexins. Additional There at least ten types of connexins connexins are being discovered all the time. Further, another family of gap junction proteins – the pannexins – has been uncovered. So when Christof says: “The most important connexin of the adult brain is Cx36”, this is not necessarily the case. And to say that connexin-36 knockout mice lack functional gap junctions in their brains is an extremely weak contention (e.g. shown by the occurrence of even weak gamma synchrony).

Christof: “no positive evidence links neuron-to-neuron coupling in cortex via gap junctions to perception, let alone consciousness.”

Stuart: a) Extremely strong evidence links gap junction hyper-neurons to gamma synchrony. b) Extremely strong evidence links gamma synchrony to perception and consciousness. Ergo, gap junction hyper-neurons are linked (through gamma synchrony) to perception and consciousness.

Dendrites and consciousness

Christof: “Stuart criticizes my neglect of dendrites, the extended appendages formed by all nerve cells and that carry the majority of synapses. He states his opinion flat out as “Consciousness occurs in dendrites, and the results are conveyed elsewhere by axonal spikes.

I do not understand this statement. Does he mean to imply that the dendritic tree of any one neuron is sufficient for consciousness? Or that dendrites, collectively, are essential to understanding the neuronal correlates of consciousness?”

Stuart: My statement implies that the collective dendritic tree of a hyper-neuron – a set of neurons linked by dendro-dendritic gap junctions – is essential to understanding the neuronal correlate of consciousness (NCC). This may include the dendrites of tens of thousands or more individual neurons. Because gap junctions change dynamically, a different particular (Hebbian) hyper-neuron set of dendrites can be the NCC at any one time, each lasting, say, hundreds of milliseconds..

Christof: Furthermore, I have never stated “that axons reverberate via cell bodies, bypassing dendrites altogether”.

Stuart: I said: “Christof’s suggestion for producing qualia is neurons whose axons reverberate via cell bodies, bypassing dendrites altogether.”

On page 103 in Chapter 5 (“What are the neuronal correlates of consciousness”) of The Quest for Consciousness, Christof states:

“I therefore look for particular mechanisms that confer onto coalitions of neurons properties that correspond to attributes of conscious percepts. One possibility might be small sets of cortical pyramidal cells that receive strong excitatory synaptic input from another set of pyramidal cells directly onto their cell bodies in a reciprocal manner. Such an arrangement might instantiate a loop, a set of neurons that, once triggered, would keep on firing until actively inhibited by another coalition of neurons. The firing dynamics of such a group might be close to that of consciousness.

The action of anesthesia

Christof: Stuart disagrees with my statement that “So far, they (i.e. anesthetics) have proven to be too blunt of a tool to help in our quest, though that may change in the future. “ (p. 96), in particular as applied to volatile, inhaled gas anesthetics (e.g., ether, nitrous oxide).

All of the recent evidence implicates specific proteins, in particular voltage- and ligand-gated ionic protein channels as the site action of such volatile agents (e.g. Sonner et al., 2003). This does not rule out that some anesthetics may exert some of their diverse effects in a global, unspecific manner, compatible with the Meyer-Overton lipid solubility relationship (see, for example, Tang and Xu, 2002). Whether or not quantum mechanical effects play any role at all in their action, as asserted by Stuart, remains speculative.

StuartAnesthetics act on proteins precisely according to the Meyer-Overton lipid solubility relationship : I agree completely that sets of specific proteins mediate anesthetic effects. (But in addition to membrane proteins, don’t forget the cytoskeleton. Both actin and microtubules are affected.) Sets of specific proteins means that numerous proteins mediate effects of any one gas anesthetic. At first glance this seems like “too blunt a tool”. But anesthetics act on proteins precisely according to the Meyer-Overton lipid solubility relationship.

All proteins affected by gas anesthetics (and thus proteins which mediate consciousness) have lipid-like (”hydrophobic”) pockets within them, where anesthetics act. And they bind in these pockets by Van der Waals London forces which are quantum effects. This is not speculation. It is also not speculation that (in the absence of anesthetics, i.e. during consciousness) quantum London forces in hydrophobic pockets can regulate conformational dynamics of certain proteins.

What is speculation is that, by forming their own London forces in the pockets, the anesthetics are disturbing normally occurring London forces necessary for protein function and consciousness. It is also speculation that quantum states in hydrophobic pockets form (near) brain-wide macroscopic quantum states through coherence/entanglement/condensation. But these speculations are falsifiable, and generate testable predictions.

Christof: Stuart claims that at just the right concentration, gas anesthetics only inhibit consciousness. This is news to me. More than one hundred years of experience has shown that the majority of anesthetic agents act in a dose-dependent manner, first causing analgesia, then amnesia, followed by loss of purposeful response, immobility and finally autonomic stability.

Stuart: These speculations are falsifiable, and generate testable predictions(I’ve only been practicing anesthesiology for 30 years.) But what Christof is correctly describing is the well known four stages of ether anesthesia discovered in the last century (except Christof left out an excitatory phase between loss of purposeful response and immobility). Autonomic (in)stability is below the level of anesthesia. By and large these stages still hold true for modern gas anesthetics.

Christof: Somewhere along this concentration gradient consciousness begins to be progressively reduced until it is finally abolished.

Stuart: I agree that “somewhere along this concentration gradient” consciousness is lost. But how do you know consciousness is progressively reduced? How do you know it doesn’t cease abruptly underneath the unconscious behaviors you describe?

Christof: At any given therapeutic concentration, a host of bodily processes is affected (e.g. cardiac or pulmonary function).

Stuart: (I said “by and large” only consciousness is affected.) For any single patient there is precisely one therapeutic concentration. But finding the precise therapeutic concentration – just enough to erase consciousness with minimal side effects – is sometimes a bit tricky (for one thing, it changes during an anesthetic for various reasons, e.g. temperature). So we tend to err on the side of slightly too much anesthetic to avoid awareness. But at just the right anesthetic amount, amazingly few bodily processes are affected (and the extent depends on a host of factors like hydration, and intrinsic organ function etc.). Plus, consciousness includes putting out various hormones like catecholamines (e.g. adrenaline) which are not put out during anesthesia.

Christof: I challenge Stuart to provide evidence of loss of consciousness without concomitant loss of pain perception, memory impairment, pulmonary function and so on.

Stuart: OKAY.

Pain perception: Patients under anesthesia may respond to pain with autonomic changes in blood pressure, heart rate, pupillary size, lacrimation, sweating, mucous secretion etc. Pain from tourniquets placed on arms or legs (to reduce blood flow in the surgical field) cause increases in blood pressure and heart rate. As anesthesiologists, we use these autonomic signs as early warnings to deepen the anesthetic level. They are nonconscious responses.

Patients under anesthesia also have relatively intact sensory evoked potentials, and visual evoked potentials (auditory evoked potentials are the most sensitive to anesthetic effect).

Memory: Christof is operating from the assumption that consciousness involves a specific set of neurons, a specific type of receptor, a specific type of circuitNumerous studies have shown that implicit learning/memory occurs under anesthesia. Phil Merikle, for example, did such studies years ago. (Phil, however, put the nonconscious pain responses noted above together with the implicit learning and occasional reports of awareness under anesthesia to suggest – incorrectly in my view – that patients do not lose consciousness under anesthesia, they just suffer and don’t remember. There are many arguments against this idea, but because consciousness is unmeasurable it is impossible to completely refute. But it is also impossible to prove that anyone other than one’s self is conscious.)

Pulmonary function: Anesthetized patients routinely breathe on their own during general anesthesia. The gas anesthetics shift the slope of the carbon dioxide response curve somewhat (we breathe in response to carbon dioxide), and slightly impair ciliary function (cilia are made of microtubules, and act to expel mucous etc. from the lungs). But breathing is relatively unaffected at therapeutic concentration (unless we add muscle paralyzing agents of course, in which case breathing stops entirely).

Christof: The discovery of an agent that would turn consciousness off and back on again, without affecting other brain and body functions, would revolutionize the practice of anesthesiology and the study of the neuronal roots of consciousness.

Stuart: We basically have that now: the gas anesthetics. They are not perfect, but close.

The problem is that Christof (like most people in the anesthetic mechanism field) is operating from the assumption that consciousness involves a specific set of neurons, a specific type of receptor, a specific type of circuit. That’s what they are looking for and its not there. The answer is right in front of us. As I said in my review:

”…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.”

In my opinion, consciousness happens in quantum pockets in hyper-neuron dendrite proteins. That is where the evidence points. That is where the quest for consciousness should be looking.

Stuart Hameroff

Homepage

Copyright © S. Hameroff

Related Articles

1. November 2004: Vol 1: The quest for consciousness: An interview with Christof Koch by Thomas Zoega Ramsoy.
2. November 2004: Vol 2: Heading in the wrong direction: Review of Christof Koch’s “The quest for consciousness” by Stuart Hameroff.
3. December 2004: Vol 2: The quest for consciousness: Response to Hameroff ’s ”Heading in the wrong direction” by Christof Koch.
4. Hameroff S (1998) Anesthesia, consciousness and hydrophobic pockets-A unitary quantum hypothesis of anesthetic action Toxicology Letters100/101:31-39 (1998)