Stuart Hameroff MD

What is Consciousness?

Table of Contents

The problem of consciousness1.Microtubules2.Pan-experiential philosophymeets modern physics

3.

Quantum computing andconsciousness

4.

Roger Penrose's 'objectivereduction'

5.

Are proteins qubits?6.Microtubule quantum automata- the 'Orch OR' model

7.

Orch OR, cognition and free will8.Consciousness and evolution9.Conclusions10.Acknowledgements/References11.

1. The Problem of Consciousness

Conventional explanations portray consciousness as an emergent property of classical computer-like activitiesin the brain's neural networks. The prevailing views among scientists in this camp are that 1) patterns ofneural network activities correlate with mental states, 2) synchronous network oscillations in thalamus andcerebral cortex temporally bind information, and 3) consciousness emerges as a novel property ofcomputational complexity among neurons.

However, these approaches appear to fall short in fully explaining certain enigmatic features of consciousness,such as:

The nature of subjective experience, or 'qualia'- our 'inner life' (Chalmers' "hard problem");Binding of spatially distributed brain activities into unitary objects in vision, and a coherent sense of self,or 'oneness';Transition from pre-conscious processes to consciousness itself;Non-computability, or the notion that consciousness involves a factor which is neither random, noralgorithmic, and that consciousness cannot be simulated (Penrose, 1989, 1994, 1997);Free will; and,Subjective time flow.

Brain imaging technologies demonstrate anatomical location of activities which appear to correlate withconsciousness, but which may not be directly responsible for consciousness.

Figure 1. PET scan image of brain showing visual and auditory recognition (from S Petersen, NeuroimagingLaboratory, Washington University, St. Louis. Also see J.A. Hobson "Consciousness," Scientific AmericanLibrary, 1999, p. 65).

Figure 2. Electrophysiological correlates of consciousness.

How do neural firings lead to thoughts and feelings? The conventional (a.k.a. functionalist, reductionist,materialist, physicalist, computationalist) approach argues that neurons and their chemical synapses are thefundamental units of information in the brain, and that conscious experience emerges when a critical level ofcomplexity is reached in the brain's neural networks.

The basic idea is that the mind is a computer functioning in the brain (brain = mind = computer). However in

fitting the brain to a computational view, such explanations omit incompatible neurophysiological details:

Widespread apparent randomness at all levels of neural processes (is it really noise, or underlying levelsof complexity?);Glial cells (which account for some 80% of brain);Dendritic-dendritic processing;Electrotonic gap junctions;Cytoplasmic/cytoskeletal activities; and,Living state (the brain is alive!)

A further difficulty is the absence of testable hypotheses in emergence theory. No threshold or rationale isspecified; rather, consciousness "just happens".

Finally, the complexity of individual neurons and synapses is not accounted for in such arguments. Since manyforms of motile single-celled organisms lacking neurons or synapses are able to swim, find food, learn, andmultiply through the use of their internal cytoskeleton, can they be considered more advanced than neurons?

Figure 3. Single cell paramecium can swim and avoid obstacles using its cytoskeleton.

Are neurons merely simple switches?

2. Microtubules

Activities within cells ranging from single-celled organisms to the brain's neurons are organized by a dynamicscaffolding called the cytoskeleton, whose major components are microtubules. Hollow, crystalline cylinders 25nanometers in diameter, microtubules are comprised of hexagonal lattices of proteins, known as tubulin.Microtubules are essential to cell shape, function, movement, and division. In neurons microtubulesself-assemble to extend axons and dendrites and form synaptic connections, then help to maintain andregulate synaptic activity responsible for learning and cognitive functions. Microtubules interact withmembrane structures mechanically by linking proteins, chemically by ions and "second-messenger" signals,and electrically by voltage fields.

Figure 4. Schematic view of two neurons connected by chemical synapse. Axon terminal (above) releasesneurotransmitter vesicles which bind receptors on post-synaptic dendritic spine. Within neurons are visiblecytoskeletal structures microtubules ("MTs" - thicker tubes) as well as actin, synapsin and others whichconnect MTs to membranes. Also, MT-associated proteins ("MAPs") interconnect MTs.

Figure 5. Immunoelectron micrograph of dendritic microtubules interconnected by MAPs. Some MTs have

been sheared, revealing internal hollow core. The granular "corn-cob" surface of MTs is barely evident to closeinspection. Scale bar, lower left: 100 nanometers. With permission from Hirokawa, 1991.

Figure 6. Crystallographic structure of microtubules.

While microtubules have traditionally been considered as purely structural elements, recent evidence hasrevealed that mechanical signaling and communication functions also exist:

MT "kinks" travel at 15 microns (2000 tubulin subunits) per second. Vernon and Woolley (1995)Experimental Cell Research 220(2)482-494MTs vibrate (100-650 Hz) with nanometer displacement. Yagi, Kamimura, Kaniya (1994) Cell motilityand the cytoskeleton 29:177-185MTs optically "shimmer" when metabolically active. Hunt and Stebbings (1994), Cell motility and thecytoskeleton 17:69-78Mechanical signals propogate through microtubules to cell nucleus; mechanism for MT regulation ofgene expression. Maniotis, Chen and Ingber (1996) Proc. Natl. Acad. Sci. USA 94:849-854Measured tubulin dipoles and MT conductivity suggest MTs are ferroelectric at physiological temperature(Tuszynski; Unger 1998)

Current models propose that tubulins within microtubules undergo coherent excitation, switching between twoor more conformational states in nanoseconds. Dipole couplings among neighboring tubulins in the microtubulelattice form dynamical patterns, or "automata," which evolve, interact and lead to the emergence of newpatterns. Research indicates that microtubule automata computation could support classical informationprocessing, transmission and learning within neurons.

Figure 7. Left: Microtubule (MT) structure: a hollow tube of 25 nanometers diameter, consisting of 13 columnsof tubulin dimers arranged in a skewed hexagonal lattice (Penrose, 1994). Right (top): Each tubulin moleculemay switch between two (or more) conformations, coupled to London forces in a hydrophobic pocket. Right(bottom): Each tubulin can also exist (it is proposed) in quantum superposition of both conformational states.

Figure 8. Microtubule automaton simulation (from Rasmussen et al., 1990). Eight nanosecond time steps of asegment of one microtubule are shown in "classical computing" mode in which conformational states oftubulins are determined by dipole-dipole coupling between each tubulin and its six (asymmetrical) latticeneighbors. Conformational states form patterns which move, evolve, interact and lead to emergence of newpatterns.

Microtubule automata switching offers a potentially vast increase in the computational capacity of the brain.

Conventional approaches focus on synaptic switching at the neural level which optimally yields about 1018

operations per second in human brains (~1011 neurons/brain with ~104 synapses/neuron, switching at ~103

sec-1). Microtubule automata switching can explain some 1027 operations per second (~1011 neurons with

~107 tubulins/neuron, switching at ~109 sec-1). Indeed, the fact that all biological cells typically contain

approximately 107 tubulins could account for the adaptive behaviors of single-celled organisms which have no

nervous system or synapses. Rather than simple switches, neurons are complex computers.

3. Pan-experiential philosophy meets modern physics

Still, greater computational complexity and ultra-reductionism to the level of microtubule automata cannotaddress the enigmatic features of consciousness, in particular the nature of conscious experience. Somethingmore is required. If functional approaches and emergence are incomplete, perhaps the raw components ofmental processes (qualia) are fundamental properties of nature (like mass, spin or charge). This view has longbeen held by pan-psychists throughout the ages, for example Buddhists and Eastern philosophers claim a"universal mind." Following the ancient Greeks, Spinoza argued in the 17th century that some form ofconsciousness existed in everything physical. The 19th century mathematician Leibniz proposed that theuniverse was composed of an infinite number of fundamental units, or "monads," with each possessing a formof primitive psychological being. In the 20th century, Russell claimed that there was a common entityunderlying both mental and physical processes, while Wheeler and Chalmers have maintained that there existsan experiential aspect to fundamental information.

Of particular interest is the work of the 20th century philosopher Alfred North Whitehead, whosepan-experiential view remains most consistent with modern physics. Whitehead argued that consciousness is aprocess of events occurring in a wide, basic field of proto-conscious experience. These events, or "occasions ofexperience," may be comparable to quantum state reductions, or actual events in physical reality (Shimony,1993). This suggests that consciousness may involve quantum state reductions (e.g. a form of quantumcomputation).

But what of Whitehead's basic field of proto-conscious experience? In what medium are the "occasions ofexperience" (?quantum state reductions) occurring? Could proto-conscious qualia simply exist in the emptyspace of the universe?

What is empty space? Historically, empty space has been described as either an absolute void or a pattern offundamental geometry. Democritus and the Michaelson-Morley results argued for "nothingness" while Aristotle("plenum") and Maxwell ("ether") rejected the notion of emptiness in favor of "something" - a backgroundpattern. Einstein weighed in on both sides of this debate, initially supporting the concept of a void with histheory of special relativity but then reversing himself in his theory of general relativity and its curved spaceand geometric distortions-the space-time metric. Could proto-conscious qualia be properties of the metric,fundamental space-time geometry?

What is fundamental space-time geometry? We know that at extremely small scales, space-time is notsmooth, but quantized. Quantum electrodynamics and quantum field theory predict virtual particle/waves (orphotons) that pop into and out of existence, creating quantum "foam" in their wake. The presence of virtualphotons in space-time has been verified (Lamoreaux, 1997).

Figure 9. Quantum electodynamics (QED) predicts a foam of erupting and collapsing virtual particles whichmay be visualized as topographic distortions of the fabric of spacetime. Adapted from Thorne (1994) by DaveCantrell.

Figure 10. A: The Casimir force of the quantum vacuum zero point fluctuation energy may be measured byplacing two macroscopic surfaces separated by a tiny gap d

1. As some virtual photons are excluded in the gap,

the net "quantum foam" pressure forces the surfaces together. In Lamoreaux's (1997) experiment, d1 was in

the range of 0.6 to 6.0 microns (~1500 nanometers). B: George Hall (1996; 1997) has calculated the Casimirforce on microtubules. As the force is proportional to d

-4, and d

2 for microtubules is 15 nanometers, the

predicted Casimir force is 106 greater on microtubules (per equivalent surface area) than that measured byLamoreaux. Hall calculates a range of Casimir force on microtubules (length dependent) from 0.5 to 20atmospheres.

At the basic level, this granularity has been modeled by Roger Penrose as a dynamic web of quantum spins.

These "spin networks" create an array of geometric volumes and configurations at the Planck scale (10-33 cm,

10-43 secs) which dynamically evolve and define space-time geometry.

Figure 11. A spin network. Introduced by Roger Penrose (1971) as a quantum mechanical description of thegeometry of space, spin networks describe spectra of discrete Planck scale volumes and configurations (withpermission, Rovelli and Smolin, 1995).

If spin networks are the fundamental level of space-time geometry, they could provide the basis for proto-conscious experience. In other words, particular configurations of quantum spin geometry would conveyparticular types of qualia, meaning and aesthetic values. A process at the Planck scale (e.g. quantum statereductions) could access and select configurations of experience.

For illustration, 4 dimensional space-time geometry is often portrayed as a 2 dimensional "space-time sheet."

Figure 12. According to Einstein's general relativity, mass is equivalent to curvature in spacetime geometry.Penrose applies this equivalence to the fundamental Planck scale. The motion of an object between twoconformational states of a protein such as tubulin (top) is equivalent to two curvatures in spacetime geometry

as represented as a two-dimensional spacetime sheet (bottom).

4. Quantum computing and consciousness

If proto-conscious information is embedded at the near-infinitesimal Planck scale, how could it be linked tobiology? To begin, Penrose extends Einstein's theory of general relativity (in which mass equates to curvaturein space-time) down to the Planck scale. As a result, specific arrangements of mass are, in reality, specificconfigurations of space-time geometry. Events at the very small scale, however, are subject to the seeminglybizarre goings-on of quantum theory.

A century of experimental observation of quantum systems have shown that, at least at small scales, particles(mass) can exist in two or more states or locations simultaneously (quantum superposition). Penrose takessuperposition (e.g. a mass in two places simultaneously) to be simultaneous space-time curvature in oppositedirections - a separation, or bubble ("blister") in underlying reality.

Figure 13. Mass superposition, e.g. a protein occupying two different conformational states simultaneously(top) is equivalent, according to Penrose, to simultaneous spacetime curvature in opposite directions - aseparation, or bubble ("blister") in fundamental spacetime geometry.

Figure 14. Spacetime superposition/separation bubble (bottom) will reduce, or collapse to one or the otherspacetime curvatures (top).

Superposition and subsequent reduction, or collapse, to single, classical states may have profoundly importantapplications in technology, as well as toward the understanding of consciousness. In the 1980s Benioff,Feynman, Deutsch and other physicists proposed that states in a quantum system could interact (viaentanglement) and enact computation while in quantum superposition of all possible states ("quantumcomputing"). Classical computing processes bits (or conformational states) as 1 or 0, quantum computationsinvolve the processing of superpositioned "qubits" of both 1 and 0 (and other states) simultaneously.

Figure 15. Qubits useful in quantum computation may exist in two or more ("both") states simultaneously priorto collapse, or reduction (left), and then in single, classical ("either, or") states after reduction (right). Spin,quantum dots and photon polarization qubits have been proposed and/or demonstrated in prototype quantumcomputers, and tubulin proteins and spacetime geometry are proposed in the Orch OR model to perform as

qubits also.

Quantum theory also tells us that two or more particles, if once together, will remain somehow connected("entangled"), even when separated by great distances. Qubits can interact by quantum entanglement, so thatquantum computing is able to achieve a nearly infinite parallel computational ability. Quantum computers, ifthey can be constructed, will be able to solve imprtant problems (e.g. factoring large numbers) with efficiencyunattainable in classical computers (Shor, 1994).

Researchers have developed a "Figure of Merit" M for proposed quantum computing technologies (Modifiedfrom Barenco, 1996 & DiVincenzo, 1995). M is related to the number of elementary operations performed perqubit before the superposition/computation is disrupted by decoherence (or in the case of microtubules in theOrch OR proposal, before objective reduction terminates the superposition).

Technologytelem(sec)

Tdecoherence(sec)

M(operations/qubitpre-decoherence)

Mossbauer nucleus 10-19 10-10 109

Electrons GaAs 10-13 10-10 103

Electrons Au 10-14 10-8 106

Trapped ions 10-14 10-1 1013

Optical cavities 10-14 10-5 109

Electron spin 10-7 10-3 104

Electron quantumdots 10

-6 10-3 103

Nuclear spin 10-3 104 107

Superconductorislands 10

-9 103 106

Microtubule tubulins 10-9 10-1 108

Results, or solutions in quantum computing are obtained when, after a period of quantumsuperposition/computation, the qubits "collapse", or reduce to classical bit states ("collapse of the wavefunction"). As quantum superposition may only occur in isolation from environment, collapse (reduction) maybe induced by breaching isolation (this is what is envisioned in technological quantum computers - making ameasurement). But what about quantum superpositions which remain isolated, for example Schrodinger'smythical cat which is both dead and alive? This is the famous problem of collapse of the wave function, orquantum state reduction.

5. Roger Penrose's 'objective reduction' OR

How or why do quantum superpositioned states which avoid environmental interactions become classical anddefinite in the macro-world? Many physicists now believe that some objective factor disturbs the superpositionand causes it to collapse. Roger Penrose proposes that this factor is an intrinsic feature of space-timegeometry itself - quantum gravity. According to Penrose's interpretation of general relativity, quantumsuperposition (e.g. separation of mass from itself) is equivalent to separation in underlying space-timegeometry-simultaneous space-time curvatures in opposite directions. Penrose argues that these separations infundamental reality, ("bubbles, or blisters") are unstable-even when isolated from the environment-and willreduce spontaneously (and non-computably) to specific states at a critical threshold of space-time separation(thereby avoiding the need for "multiple worlds"). This objective threshold is defined by the indeterminacyprinciple:

E = h/T

where E is the gravitational self-energy of the superposed mass separated from itself, h is Planck's constantdivided by 2pi, and T is the coherence time until collapse occurs. Thus, the size and energy of a system insuperposition, or the degree of space-time separation, is inversely related to the time T until reduction. (E canbe calculated from the superposed mass m and the separation distance a. See e.g. Hameroff and Penrose,1996a.)

Assuming isolation, the following masses in superposition would collapse at the designated times, according toPenrose's objective reduction:

Mass (m) Time (T)

Nucleon 107 years

Beryllium ion 106 years

Water Speck

10-5 cm radius Hours

10-4 cm radius 1/20 second

10-3 cm radius 10-3 seconds

Schrodinger's cat (m=1kg,a=10 cm) 10

-37 seconds

If quantum computation with objective reduction were occurring in the brain, enigmatic features ofconsciousness (see Section above - The Problem of Consciousness) could be explained:

By occurring as a self-organizing process in what is suggested to be a pan-experiential medium offundamental spacetime geometry, objective reductions could account for the nature of subjectiveexperience by accessing and selecting proto-conscious qualia.By virtue of involvement of unitary (entangled) quantum states during pre-conscious quantumcomputation and the unity of quantum information selected in each objective reduction, the issue ofbinding may be resolved.Regarding the transitions from pre-conscious processes to consciousness itself, the pre-consciousprocesses may equate to the quantum superposition/quantum computation phase, and consciousnessitself to the actual (instantaneous) objective reduction events. Consciousness may then be seen as asequence of discrete events (e.g. at 40 Hz).As Penrose objective reductions are proposed to be non-computable (reflecting influences fromspace-time geometry which are neither random, nor algorithmic) conscious choices and understandingmay be similarly non-computable.Free will may be seen as a combination of deterministic pre-conscious processes acted on by anon-computable influence.Subjective time flow derives from a sequence of irreversible quantum state reductions.

Could objective reduction be occurring in the brain? If so (from E = h/T) time T would be expected to coincidewith known neurophysiological processes with time scales from tens to hundreds of milliseconds (e.g. 25 msecfor coherent 40 Hz, 100 msec for alpha EEG, 500 msec for sensory threshold events such as Libet's famous1979 experiments). In what types of brain structures might quantum computation with objective reductionoccur? For T in this range we can calculate (from E = h/T, and with E related to mass m as described inHameroff and Penrose, 1996a) that superpositioned mass m in the nanogram range would be required forconscious events of 40 to 500 msec. What brain components in nanogram quantitites could support quantumcomputation and objective reduction? What is m?

6. Are proteins qubits?

Biological life is organized by proteins. By changing their conformational shape, proteins are able to perform awide variety of functions, including muscle movement, molecular binding, enzyme catalysis, metabolism, andmovement. Dynamical protein structure results from a "delicate balance among powerful countervailing forces"(Voet & Voet, 1995). The types of forces acting on proteins include charged interactions (such as covalent,ionic, electrostatic, and hydrogen bonds), hydrophobic interactions, and dipole interactions. The latter group,also known as van der Waals forces, encompasses three types of interactions:

permanent dipole - permanent dipole,permanent dipole - induced dipole, andinduced dipole - induced dipole (London dispersion forces)

As charged interactions cancel out, hydrophobic and dipole - dipole forces are left to regulate protein structure.While induced dipole - induced dipole interactions, or London dispersion forces, are the weakest of the forcesoutlined above, they are also the most numerous and influential. Indeed, they may be critical to proteinfunction. For example, anesthetics are able to bind in hydrophobic "pockets" of certain neural proteins andablate consciousness by virtue of disrupting these London forces. London force attraction between any twoatoms is usually less than a few kilojoules; however, since thousands occur in each protein, they add up tothousands of kilojoules per mole, and cause changes in conformational structure. As London forces areinstrumental in protein folding (a problem intractable to conventional computational simulation), proteinconformation and folding may be quantum computations.

Figure 16. A type of van der Waals force, the London dispersion force, is quantum mechanical and governsboth protein conformation.

Figure 17. A. An anesthetic gas molecule (A) in a hydrophobic pocket of critical brain protein (receptors,channels, tubulin etc.) prevents normally occurring London forces, preventing protein conformational dynamicsand superposition necessary for consciousness. B. A psychedelic hallucinogen (P) acts in hydrophobic pocket incritical brain protein to promote and sustain superposition, 'expanding' consciousness (see Figure 25).

Figure 18. A. Protein qubit. A protein such as tubulin can exist in two conformations determined by quantumLondon forces in hydrophobic pocket (top), or superposition of both conformations (bottom). B. The proteinqubit corresponds to two alternative spacetime curvatures (top), and superposition/separation (bubble) of bothcurvatures (bottom).

If proteins are qubits, arrays or assemblies of proteins in some type of organelle or biomolecular structurecould be a quantum computer. Ideal structures would be:

Abundant;Capable of information processing and computation;Functionally important (e.g., regulating synapses);Self-organizing;Tunable by input information (e.g., microtubule-associated protein orchestration);Periodic and crystal-like in structure (e.g., dipole lattice);Isolated (transiently) from environmental decoherence;Conformationally coupled to quantum events (e.g., London forces);Cylindrical wave-guide structure; and,Plasma-like charge layer coating.

While various structures/organelles have been suggested (e.g., membrane proteins, clathrins, myelin,pre-synaptic grids, and calcium ions), the most logical candidates are microtubule automata.

Figure 19. The Penrose-Hameroff Orch OR model was hatched on a hike in the Grand Canyon following theTucson I conference in April, 1994. From left: David Chalmers, Rhett Savage, Marie-Francoise Insinna, SeamusO'Morain, Stuart Hameroff, Roger Penrose, Vanessa Penrose, Jeff Tollaksen. Photo by Ezio Insinna.

7. Microtubule quantum automata - The 'Orch OR' model

The Penrose - Hameroff model of "orchestrated objective reduction" (Orch OR) proposes that:

Quantum superposition/computation occur in microtubule automata within brain neurons and glia.;

Tubulin subunits within microtubules act as qubits, switching between states on a nanosecond (10-9 sec)scale governed by quantum London forces in hydrophobic pockets;Tubulin qubits interact computationally by nonlocal quantum entanglement according to the Schrodingerequation;

Figure 20. The basic idea in the Orch OR model is that each tubulin in a microtubule is a qubit.

Figure 21. Microtubule automaton sequence simulation in which classical computing (step 1) leads toemergence of quantum coherent superposition (steps 2-6) in certain (gray) tubulins due to pattern resonance.Step 6 (in coherence with other microtubule tubulins) meets critical threshold related to quantum gravity forself-collapse (Orch OR). Consciousness (Orch OR) occurs in the step 6 to 7 transition. Step 7 represents theeigenstate of mass distribution of the collapse which evolves by classical computing automata to regulateneural function. Quantum coherence begins to re-emerge in step 8.

pre-conscious processing which continues until the threshold for objective reduction (OR) is reached byE = h/T;At that instant collapse, or OR occurs which is an actual event in fundamental space-time geometry.This event selects a particular configuration of Planck-scale experiential geometry, enacting a "momentof awareness," "occasion of experience" or conscious event.

Figure 22. Schematic graph of proposed pre-conscious quantum superposition (number of tubulins) emergingversus time in microtubules. Area under curve connects superposed mass energy E with collapse time T inaccordance with E=(h/T. E may be expressed as n

t, the number of tubulins whose mass separation (and

separation of underlying space time) for time T will self-collapse. For T = 25 msec (e.g. 40 Hz oscillations), nt

= 2 x 1010 tubulins.

Figure 23. Schematic of quantum computation of three tubulins which begin (left) in initial classical states,then enter isolated quantum superposition in which all possible states coexist. After reduction, one particularclassical outcome state is chosen (right).

Figure 24. Schematic quantum computation in spacetime curvature for three mass distributions (e.g. tubulinconformations in Figure 23) which begin (left) in initial classical states, then enter isolated quantumsuperposition in which all possible states coexist. After reduction, one particular classical outcome state ischosen (right).

A sequence of OR events (e.g. at 40 Hz) provides a forward flow of subjective time and "stream" ofconsciousness;

Figure 25. Quantum superposition/entanglement in microtubules for 5 states related to consciousness. Areaunder each curve equivalent in all cases. A. Normal 30 Hz experience: as in Figure 22. B. Anesthesia:anesthetics bind in hydrophobic pockets and prevent quantum delocalizability and coherent superposition. C.Heightened Experience: increased sensory experience input (for example) increases rate of emergence ofquantum superposition. Orch OR threshold is reached faster, and Orch OR frequency increases. D. AlteredState: even greater rate of emergence of quantum superposition due to sensory input and other factorspromoting quantum state (e.g. meditation, psychedelic drug etc.). Predisposition to quantum state results inbaseline shift and collapse so that conscious experience merges with normally sub-conscious quantumcomputing mode. E. Dreaming: prolonged sub-threshold quantum superposition time.

At the nanoscale each event determines new classical states of microtubule automata which regulatesynaptic and other neural functions;During the pre-conscious quantum superposition/computation phase, oscillations are "tuned" and"orchestrated" by microtubule-associated proteins (MAPs), providing a feedback loop between thebiological system and the quantum state (hence Orch OR);Quantum states in microtubules may link to those in microtubules in other neurons and glia by tunnelingthrough gap junctions, permitting extension of the quantum state throughout significant volumes of thebrain.

Figure 26. Schematic of proposed quantum superposition and entanglement in microtubules in threedendrites interconnected by tunneling through gap junctions. Within each neuronal dendrite, microtubule-associated-protein (MAP) attachments breach isolation and prevent quantum coherence; MAP attachment sitesthus act as "nodes" which tune and orchestrate quantum oscillations and set possibilities and probabilities forcollapse outcomes (orchestrated objective reduction: Orch OR). Gap junctions may enable quantum tunnelingamong dendrites resulting in macroscopic quantum states.

From E = h/T we can calculate the size and extension of Orch OR events which correlate with subjective orneurophysiological descriptions of conscious events.

Event T E

Buddhist "moment of awareness" 13 ms4 x 1015 nucleons

(4 x 1010 tubulins/cell ~ 40,000neurons)

"Coherent 40 Hz" oscillations" 25 ms2 x 1015 nucleons

(2 x 1010 tubulins/cell ~ 20,000neurons)

EEG alpha rhythm (8 to 12 Hz) 100 ms5 x 1014 nucleons

(5 x 109 tubulins/cell ~ 5000 neurons)

Libet's sensory threshold (1979) 500ms1014 nucleons

(109 tubulins/cell ~ 1000 neurons)

But how could delicate quantum superposition/computation be isolated from environmental decoherence in thebrain (generally considered to be a noisy thermal bath) while also communicating (input/output) with theenvironment? One possibility is that quantum superposition/computation occurs in an isolation phase whichalternates with a communicative phase, for example at 40 Hz. One of the most primitive biological functions isthe transition of cytoplasm between a liquid, solution ("sol") phase, and a solid, gelatinous ("gel") phase due toassembly and disassembly of the cytoskeletal protein actin. Actin sol-gel transitions can occur at 40 Hz or

faster, and are known to be involved in neuronal synaptic release mechanisms.

Mechanisms for enabling microtubule quantum computation and avoiding decoherence long enough to reachthreshold may include:

Sol-gel transitions;

Figure 27. Immunoelectron micrograph of cytoplasm showing microtubules (arrows), intermediate filaments(arrowheads) and actin microfilaments (mf). Dense gel of actin (lower left) completely obscures (?isolates)microtubules. Actin sol-gel transitions can occur at 40 Hz or faster. Scale bar (upper right): 500 nanometers.With permission from Svitkina et al, 1995.

Plasma phase sleeves (Sackett);

Figure 28. Dan Sackett at NIH recently described a plasma-like sleeve of charged ions surroundingmicrotubules at precisely optimal pH.

Quantum excitations/ordering of surrounding water (Jibu/Yasue/Hagan);Hydrophobic pockets;Hollow microtubule cores;Laser-like pumping, including environment (Frohlich, Conrad).Quantum error correcting codes

Another apparent obstacle to the Orch OR proposal is how the weak energy involved in the gravitationalcollapse can be influential. For a detailed description of this problem and potential solutions, see Hameroff,1998c. One possibility is that the gravitational self-energy is delivered to the involved tubulins via Londonforces virtually instantaneously (e.g. within one Planck time) so that the power (energy/time) is significant -approximately one kilowatt per tubulin per Orch OR event.

8. Orch OR, cognition and free will

Quantum computation with objective reduction (Orch OR) is potentially applicable to cognitive activities. Whileclassical neural-level computation can provide a partial explanation, the Orch OR model allows far greaterinformation capacity, and addresses issues of conscious experience, binding, and non-computability consistentwith free will. Functions like face recognition and volitional choice may require a series of conscious eventsarriving at intermediate solutions. For the purpose of illustration consider single Orch OR events in these twotypes of cognitive activities.

Imagine you briefly see a familiar woman's face. Is she Amy, Betty, or Carol? Possibilities may superpose in aquantum computation. For example during 25 milliseconds of pre-conscious processing, quantum computationoccurs with information (Amy, Betty, Carol) in the form of "qubits"3/4superposed states of microtubule tubulinsubunits within groups of neurons. As threshold for objective reduction is reached, an instantaneous consciousevent occurs. The superposed tubulin qubits reduce to definite states, becoming bits. Now, you recognize that

she is Carol! (an immense number of possibilities could be superposed in a human brain's 1019 tubulins).

Figure 29. Face recognition. A familiar face induces superposition (left) of three possible solutions (Amy,Betty, Carol) which "collapse" to the correct answer Carol (right). Volitional choice. Three possible dinnerselections (shrimp, sushi, pasta) are considered in superposition (left), and collapse via Orch OR to choice ofsushi (right).

In a volitional act possible choices may be superposed. Suppose for example you are selecting dinner from amenu. During pre-conscious processing, shrimp, sushi and pasta are superposed in a quantum computation.As threshold for objective reduction is reached, the quantum state reduces to a single classical state. A choiceis made. You'll have sushi!

How does the choice actually occur? In a conventional neural network scheme, the selection criteria can bedescribed by a deterministic algorithm which precludes the possibility of free will. The non-computableinfluence in Orch OR may be useful in understanding free will.

The problem in understanding free will is that our actions seem neither totally deterministic nor random(probabilistic). What else is there in nature? As previously described, in OR (and Orch OR) the reductionoutcomes are neither deterministic nor probabilistic, but involve a factor which is "non-computable." Themicrotubule quantum superposition evolves linearly (analogous to a quantum computer) but is influenced atthe instant of collapse by hidden non-local variables (quantum-mathematical logic inherent in fundamentalspacetime geometry). The possible outcomes are limited, or probabilities set ("orchestrated"), byneurobiological feedback (in particular microtubule associated proteins, or MAPs). The precise outcome3/4ourfree will actions3/4are chosen by effects of the hidden logic on the quantum system poised at the edge ofobjective reduction.

Figure 30. Free will may be seen as the result of deterministic processes (behavior of trained robotwindsurfer) acted on repeatedly by non-computable influences, here represented as a seemingly capriciouswind.

Consider a sailboard analogy for free will. A sailor sets the sail in a certain way; the direction the board sails isdetermined by the action of the wind on the sail. Let's pretend the sailor is a non-conscious robot zombie runby a quantum computer which is trained and programmed to sail. Setting and adjusting of the sail, sensing thewind and position, jibing and tacking (turning the board) are algorithmic and deterministic, and may beanalogous to the pre-conscious, quantum computing phase of Orch OR. The direction and intensity of the wind(seemingly capricious, or unpredictable) may be considered analogous to Planck scale hidden non-localvariables (e.g. "Platonic" quantum-mathematical logic inherent in space-time geometry). The choice, oroutcome (the direction the boat sails, the point on shore it lands) depends on the deterministic sail settingsacted on repeatedly by the apparently unpredictable wind. Our "free will" actions could be the net result ofdeterministic processes acted on by hidden quantum logic at each Orch OR event. This can explain why wegenerally do things in an orderly, deterministic fashion, but occasionally our actions or thoughts are surprising,even to ourselves.

9. Consciousness and evolution

When in the course of evolution did consciousness first appear? Are all living organisms conscious, or didconsciousness emerge more recently, e.g. with language or toolmaking? Or did consciousness appearsomewhere in between, and if so, when and why? The Orch OR model (unlike other models of consciousness)is able to make a prediction as to the onset of consciousness. Based on E = h/T we can ask, for example, is itfeasible for single cell organisms such as paramecium (which exhibit complex behavior such as gracefulswimming, mating and learning) to be conscious? Single cells including paramecium should contain

approximately 107 tubulins, so T would be 50,000 msec, or nearly one minute. This seems unlikely. Larger

organisms such as the nematode worm (e.g., C. elegans) with 300 neurons (3 x 109 tubulins) would need to

maintain quantum isolation for only 133 msec - not unreasonable. Such organisms (tiny worms and urchins)were prevalent at the beginning of the "Cambrian explosion," a burst of evolution which occurred 540 millionyears ago. Did primitive consciousness (via Orch OR) accelerate evolution and precipitate the Cambrianexplosion?

Figure 31. A time-line of when consciousness could have arisen.

The Cambrianexplosion was aburst of evolution540 million yearsago. Organismspresent at theCambrian onsetincluded smallworms andurchins. Didconsciousness(Orch OR) causethe Cambrianexplosion?

Figure 32. Organisms present at the early Cambrian explosion (e.g. tiny urchins, worms and suctorians) arethe right size for primitive consciousness by Orch OR.

Figure 33. Actinosphaerium is a tiny urchin like those present at the early Cambrian explosion. Each has

about one hundred rigid axonemes about 300 microns long, made up of a total of about 3 x 109 tubulins (withpermission from L.E. Roth).

Figure 34. Cross-section of single axoneme of actinosphaerium - a double spiral array of interconnectedmicrotubules. Scale bar: 500 nm (with permission from L.E. Roth).

Would consciousness be advantageous to survival (above and beyond intelligent, complex behavior)? It seemsthat, yes, consciousness would indeed be advantageous to survival, and hence capable of acceleratingevolution. Non-computable behavior (unpredictability, intuitive actions) would be beneficial in predator-preyrelations. Having conscious experience of taste would promote finding food; the experience of pain wouldpromote avoiding predators. And the pleasurable qualia of sex would promote reproduction.

So "what is it like to be a worm?" Lacking our sensory apparatus, associative memory and complex nervoussystem such primitive consciousness would be a mere glimmer, a disjointed smudge of reality. Butqualitatively, at a basic level, such primitive consciousness would be akin to ours.

What about future evolution? Will consciousness occur in computers? The advent of quantum computers opensthe possibility, however as presently envisioned quantum computers will have insufficient mass insuperposition (e.g. electrons) to reach threshold for objective reduction. Instead, superpositions will bedisrupted by environmental decoherence. Conceivably, future generations of quantum computers could satisfyrequirements for objective reduction and consciousness.

10. Conclusions

Brain processes relevant to consciousness extend downward within neurons to the level of cytoskeletalmicrotubules.An explanation for conscious experience requires (in addition to neuroscience and psychology) a modernform of pan-protopsychism in which proto-conscious qualia are embedded in the basic level of reality, asdescribed by modern physics.Roger Penrose's physics of objective reduction (OR) connects brain structures to fundamental reality,leading to the Penrose-Hameroff model of quantum computation with objective reduction inmicrotubules (orchestrated objective reduction: Orch OR).The Orch OR model is consistent with known neurophysiological processes, generates testable

Stuart Hameroff, M.D.Department of Anesthesiology

Arizona Health Sciences Center

Tucson, AZ 85724(520) 626-5605(520) 626-5596 FAX

predictions, and is the type of fundamental, multi-level, interdisciplinary theory which may account forthe mind's enigmatic features.

11. Acknowledgments & References

Some of the newer ideas expresssed here may not necessarily reflect Roger's view. Thanks to Dave Cantrell forillustrations, and Carol Ebbecke for expert technical assistance. For a list of testable predictions of the Orch ORmodel, see Hameroff, 1998c or e.

References

Penrose, R. (1989) The Emperor's New Mind, Oxford Press, Oxford, U.K

Penrose, R. (1994) Shadows of the Mind, Oxford Press, Oxford, U.K.

Penrose, R., Hameroff, S.R. (1995) What gaps? Reply to Grush and Churchland. Journal of ConsciousnessStudies 2(2):99-112.

Hameroff, S.R., and Penrose, R., (1996a) Orchestrated reduction of quantum coherence in brain microtubules:A model for consciousness. In: Toward a Science of Consciousness - The First Tucson Discussions and Debates,S.R. Hameroff, A. Kaszniak and A.C. Scott (eds.), MIT Press, Cambridge, MA.pp. 507-540. Also published inMathematics and Computers in Simulation 40:453-480.

Hameroff, S.R., and Penrose, R. (1996b) Conscious events as orchestrated spacetime selections. Journal ofConsciousness Studies 3(1):36-53.

Penrose, R. (1996) On gravity's role in quantum state reduction. General relativity and gravitation. 28(5):581-600

Penrose, R. (1997) On understanding understanding. International Studies in the Philosophy of Science11(1):7-20.

Penrose R. (1998) The large, the small, and the human mind

Hameroff, S. (1998a) Did consciousness cause the Cambrian evolutionary explosion? In: Toward a Science ofConsciousness II - The Second Tucson Discussions and Debates. Eds S Hameroff, A Kaszniak, A Scott. MITPress, Cambridge MA pp 421-437

Hameroff, S. (1998b) "More neural than thou": Reply to Churchland's "Brainshy" in: Toward a Science ofConsciousness II - The Second Tucson Discussions and Debates. Eds S Hameroff, A Kaszniak, A Scott. MITPress, Cambridge MA pp 197-213

Hameroff, S. (1998c) Funda-mental geometry: The Penrose-Hameroff Orch OR model of consciousness. In:The geometric universe - Science, geometry and the work of Roger Penrose. Eds. S.A. Huggett, L.J. Mason,K.P. Tod, S.T. Tsou, and N.M.J. Woodhouse. Oxford Press, Oxford, U.K. pp 135-160

Hameroff, S. (1998d) "Funda-Mentality" - Is the conscious mind subtly connected to a basic level of theuniverse? Trends in Cognitive Science 2(4):119-127

Hameroff, S. (1998e) Quantum computation in microtubules? The Penrose-Hameroff 'Orch OR' model ofconsciousness. Philosophical Transactions of the Royal Society A (London)356:1869-1896

Hameroff, S. (1998f) Anesthesia, consciousness and hydrophobic pockets - A unitary quantum hypothesis ofanesthetic action. Toxicology Letters 100/101-31-39.