Browsed by
Tag: neurons

Choosing the Right Scale for Brain Emulation – Article by Franco Cortese

Choosing the Right Scale for Brain Emulation – Article by Franco Cortese

The New Renaissance Hat
Franco Cortese
June 2, 2013
******************************
This essay is the ninth chapter in Franco Cortese’s forthcoming e-book, I Shall Not Go Quietly Into That Good Night!: My Quest to Cure Death, published by the Center for Transhumanity. The first eight chapters were previously published on The Rational Argumentator under the following titles:
***

The two approaches falling within this class considered thus far are (a) computational models that model the biophysical (e.g., electromagnetic, chemical, and kinetic) operation of the neurons—i.e., the physical processes instantiating their emergent functionality, whether at the scale of tissues, molecules and/or atoms, and anything in between—and (b) abstracted models, a term which designates anything that computationally models the neuron using the (sub-neuron but super-protein-complex) components themselves as the chosen model-scale (whereas the latter uses for its chosen model-scale the scale at which physical processes emergently instantiating those higher-level neuronal components exist, such as the membrane and individual proteins forming the transmembrane protein-complexes), regardless of whether each component is abstracted as a normative-electrical-component analogue (i.e., using circuit diagrams in place of biological schematics, like equating the lipid bilayer membrane with a capacitor connected to a variable battery) or mathematical models in which a relevant component or aspect of the neuron becomes a term (e.g., a variable or constant) in an equation.

It was during the process of trying to formulate different ways of mathematically (and otherwise computationally) modeling neurons or sub-neuron regions that I laid the conceptual embryo of the first new possible basis for subjective-continuity: the notion of operational isomorphism.

A New Approach to Subjective-Continuity Through Substrate Replacement

There are two other approaches to increasing the likelihood of subjective-continuity, each based on the presumption of two possible physical bases for discontinuity, that I explored during this period. Note that these approaches are unrelated to graduality, which has been the main determining factor impacting the likelihood of subjective-continuity considered thus far. The new approaches consist of designing the NRUs so as to retain the respective postulated physical bases for subjective-continuity that exist in the biological brain. Thus they are unrelated to increasing the efficacy of the gradual-replacement procedure itself, instead being related to the design requirements of functional-equivalents used to gradually replace the neurons that maintain immediate subjective-continuity.

Operational Isomorphism

Whereas functionality deals only with the emergent effects or end-product of a given entity or process, operationality deals with the procedural operations performed so as to give rise to those emergent effects. A mathematical model of a neuron might be highly functionally equivalent while failing to be operationally equivalent in most respects. Isomorphism can be considered a measure of “sameness”, but technically means a 1-to-1 correspondence between the elements of two sets (which would correspond with operational isomorphism) or between the sums or products of the elements of two sets (which would correspond with functional isomorphism, using the definition of functionality employed above). Thus, operational isomorphism is the degree with which the sub-components (be they material as in entities or procedural as in processes) of the two larger-scale components, or the operational modalities possessed by each respective collection of sub-components, are equivalent.

To what extent does the brain possess operational isomorphism? It seems to depend on the scale being considered. At the highest scale, different areas of the nervous system are classed as systems (as in functional taxonomies) or regions (as in anatomical taxonomies). At this level the separate regions (i.e., components of a shared scale) differ widely from one another in terms of operational-modality; they process information very differently from the way other components on the same scale process information. If this scale was chosen as the model-scale of our replication-approach and the preceding premise (that the physical basis for subjective-continuity is the degree of operational isomorphism between components at a given scale) is accepted, then we would in such a case have a high probability of replicating functionality, but a low probability of retaining subjective-continuity through gradual replacement. This would be true even if we used the degree of operational isomorphism between separate components as the only determining factor for subjective-continuity, and ignored concerns of graduality (e.g., the scale or rate—or scale-to-rate ratio—at which gradual substrate replacement occurs).

Contrast this to the molecular scale, where the operational modality of each component (being a given molecule) and the procedural rules determining the state-changes of components at this scale are highly isomorphic. The state-changes of a given molecule are determined by molecular and atomic forces. Thus if we use an informational-functionalist approach, choose a molecular scale for our model, and accept the same premises as the first example, we would have a high probability of both replicating functionality and retaining subjective-continuity through gradual replacement because the components (molecules) have a high degree of operational isomorphism.

Note that this is only a requirement for the sub-components instantiating the high-level neural regions/systems that embody our personalities and higher cognitive faculties such as the neocortex — i.e., we wouldn’t have to choose a molecular scale as our model scale (if it proved necessary for the reasons described above) for the whole brain, which would be very computationally intensive.

So at the atomic and molecular scale the brain possesses a high degree of operational isomorphism. On the scale of the individual protein complexes, which collectively form a given sub-neuronal component (e.g., ion channel), components still appear to possess a high degree of operational isomorphism because all state-changes are determined by the rules governing macroscale proteins and protein-complexes (i.e., biochemistry and particularly protein-protein interactions); by virtue of being of the same general constituents (amino acids), the factors determining state-changes at this level are shared by all components at this scale. The scale of individual neuronal components, however, seems to possess a comparatively lesser degree of operational isomorphism. Some ion channels are ligand-gated while others are voltage-gated. Thus, different aspects of physicality (i.e., molecular shape and voltage respectively) form the procedural-rules determining state-changes at this scale. Since there are now two different determining factors at this scale, its degree of operational isomorphism is comparatively less than the protein and protein-complex scale and the molecular scale, both of which appear to have only one governing procedural-rule set. The scale of individual neurons by contrast appears to possess a greater degree of operational isomorphism; every neuron fires according to its threshold value, and sums analog action-potential values into a binary output (i.e., neuron either fires or does not). All individual neurons operate in a highly isomorphic manner. Even though individual neurons of a given type are more operationally isomorphic in relation to each other than with a neuron of another type, all neurons regardless of type still act in a highly isomorphic manner. However, the scale of neuron-clusters and neural-networks, which operate and communicate according to spatiotemporal sequences of firing patterns (action-potential patterns), appears to possess a lesser degree of operational isomorphism compared to individual neurons, because different sequences of firing patterns will mean a different thing to two respective neural clusters or networks. Also note that at this scale the degree of functional isomorphism between components appears to be less than their degree of operational isomorphism—that is, the way each cluster or network operates is more similar in relation to each other than is their actual function (i.e., what they effectively do). And lastly, at the scale of high-level neural regions/systems, components (i.e., neural regions) differ significantly in morphology, in operationality, and in functionality; thus they appear to constitute the scale that possesses the least operational isomorphism.

I will now illustrate the concept of operational isomorphism using the physical-functionalist and the informational-functionalist NRU approaches, respectively, as examples. In terms of the physical-functionalist (i.e., prosthetic neuron) approach, both the passive (i.e., “direct”) and CPU-controlled sub-classes, respectively, are operationally isomorphic. An example of a physical-functionalist NRU that would not possess operational isomorphism is one that uses a passive-physicalist approach for the one type of component (e.g., voltage-gated ion channel) and a CPU-controlled/cyber-physicalist approach [see Part 4 of this series] for another type of component (e.g., ligand-gated ion channel)—on that scale the components act according to different technological and methodological infrastructures, exhibit different operational modalities, and thus appear to possess a low degree of operational isomorphism. Note that the concern is not the degree of operational isomorphism between the functional-replication units and their biological counterparts, but rather with the degree of operational isomorphism between the functional-replication units and other units on the same scale.

Another possibly relevant type of operational isomorphism is the degree of isomorphism between the individual sub-components or procedural-operations (i.e., “steps”) composing a given component, designated here as intra-operational isomorphism. While very similar to the degree of isomorphism for the scale immediately below, this differs from (i.e., is not equivalent to) such a designation in that the sub-components of a given larger component could be functionally isomorphic in relation to each other without being operationally isomorphic in relation to all other components on that scale. The passive sub-approach of the physical-functionalist approach would possess a greater degree of intra-operational isomorphism than would the CPU-controlled/cyber-physicalist sub-approach, because presumably each component would interact with the others (via physically embodied feedback) according to the same technological and methodological infrastructure—be it mechanical, electrical, chemical, or otherwise. The CPU-controlled sub-approach by contrast would possess a lesser degree of intra-operational-isomorphism, because the sensors, CPU, and the electric or electromechanical systems, respectively (the three main sub-components for each singular neuronal component—e.g., an artificial ion channel), operate according to different technological and methodological infrastructures and thus exhibit alternate operational modalities in relation to eachother.

In regard to the informational-functionalist approach, an NRU model that would be operationally isomorphic is one wherein, regardless of the scale used, the type of approach used to model a given component on that scale is as isomorphic with the ones used to model other components on the same scale as is possible. For example, if one uses a mathematical model to simulate spiking regions of the dendritic spine, then one shouldn’t use a non-mathematical (e.g., strict computational-logic) approach to model non-spiking regions of the dendritic spine. Since the number of variations to the informational-functionalist approach is greater than could exist for the physical-functionalist approach, there are more gradations to the degree of operational isomorphism. Using the exact same branches of mathematics to mathematically model the two respective components would incur a greater degree of operational isomorphism than if we used alternate mathematical techniques from different disciplines to model them. Likewise, if we used different computational approaches to model the respective components, then we would have a lesser degree of operational isomorphism. If we emulated some components while merely simulating others, we would have a lesser degree of operational isomorphism than if both were either strictly simulatory or strictly emulatory.

If this premise proves true, it suggests that when picking the scale of our replication-approach (be it physical-functionalist or informational-functionalist), we choose a scale that exhibits operational isomorphism—for example, the molecular scale rather than the scale of high-level neural-regions, and that we don’t use widely dissimilar types of modeling techniques to model one component (e.g., a molecular system) than we do for another component on the same scale.

Note that unlike operational-continuity, the degree of operational isomorphism was not an explicit concept or potential physical basis for subjective-continuity at the time of my working on immortality (i.e., this concept wasn’t yet fully fleshed out in 2010), but rather was formulated in response to going over my notes from this period so as to distill the broad developmental gestalt of my project; though it appears to be somewhat inherent (i.e., appears to be hinted at), it hasn’t been explicitized until relatively recently.

The next chapter describes the rest of my work on technological approaches to techno-immortality in 2010, focusing on a second new approach to subjective-continuity through a gradual-substrate-replacement procedure, and concluding with an overview of the ways my project differs from the other techno-immortalist projects.

Franco Cortese is an editor for Transhumanity.net, as well as one of its most frequent contributors.  He has also published articles and essays on Immortal Life and The Rational Argumentator. He contributed 4 essays and 7 debate responses to the digital anthology Human Destiny is to Eliminate Death: Essays, Rants and Arguments About Immortality.

Franco is an Advisor for Lifeboat Foundation (on its Futurists Board and its Life Extension Board) and contributes regularly to its blog.

Wireless Synapses, Artificial Plasticity, and Neuromodulation – Article by Franco Cortese

Wireless Synapses, Artificial Plasticity, and Neuromodulation – Article by Franco Cortese

The New Renaissance Hat
Franco Cortese
May 21, 2013
******************************
This essay is the fifth chapter in Franco Cortese’s forthcoming e-book, I Shall Not Go Quietly Into That Good Night!: My Quest to Cure Death, published by the Center for Transhumanity. The first four chapters were previously published on The Rational Argumentator as “The Moral Imperative and Technical Feasibility of Defeating Death”, “Immortality: Material or Ethereal? Nanotech Does Both!, “Concepts for Functional Replication of Biological Neurons“, and “Gradual Neuron Replacement for the Preservation of Subjective-Continuity“.
***

Morphological Changes for Neural Plasticity

The finished physical-functionalist units would need the ability to change their emergent morphology not only for active modification of single-neuron functionality but even for basic functional replication of normative neuron behavior, by virtue of needing to take into account neural plasticity and the way that morphological changes facilitate learning and memory. My original approach involved the use of retractable, telescopic dendrites and axons (with corresponding internal retractable and telescopic dendritic spines and axonal spines, respectively) activated electromechanically by the unit-CPU. For morphological changes, by providing the edges of each membrane section with an electromechanical hinged connection (i.e., a means of changing the angle of inclination between immediately adjacent sections), the emergent morphology can be controllably varied. This eventually developed to consist of an internal compartment designed so as to detach a given membrane section, move it down into the internal compartment of the neuronal soma or terminal, transport it along a track that stores alternative membrane sections stacked face-to-face (to compensate for limited space), and subsequently replaces it with a membrane section containing an alternate functional component (e.g., ion pump, ion channel, [voltage-gated or ligand-gated], etc.) embedded therein. Note that this approach was also conceived of as an alternative to retractable axons/dendrites and axonal/dendritic spines, by attaching additional membrane sections with a very steep angle of inclination (or a lesser inclination with a greater quantity of segments) and thereby creating an emergent section of artificial membrane that extends out from the biological membrane in the same way as axons and dendrites.

However, this approach was eventually supplemented by one that necessitates less technological infrastructure (i.e., that was simpler and thus more economical and realizable). If the size of the integral-membrane components is small enough (preferably smaller than their biological analogues), then differential activation of components or membrane sections would achieve the same effect as changing the organization or type of integral-membrane components, effectively eliminating the need at actually interchange membrane sections at all.

Active Neuronal Modulation and Modification

The technological and methodological infrastructure used to facilitate neural plasticity can also be used for active modification and modulation of neural behavior (and the emergent functionality determined by local neuronal behavior) towards the aim of mental augmentation and modification. Potential uses already discussed include mental amplification (increasing or augmenting existing functional modalities—i.e., intelligence, emotion, morality), or mental augmentation (the creation of categorically new functional and experiential modalities). While the distinction between modification and modulation isn’t definitive, a useful way of differentiating them is to consider modification as morphological changes creating new functional modalities, and to consider modulation as actively varying the operation of existing structures/processes through not morphological change but rather changes to the operation of integral-membrane components or the properties of the local environment (e.g., increasing local ionic concentrations).

Modulation: A Less Discontinuous Alternative to Morphological Modification

The use of modulation to achieve the effective results of morphological changes seemed like a hypothetically less discontinuous alternative to morphological changes (and thus as having a hypothetically greater probability of achieving subjective-continuity). I’m more dubious in regards to the validity of this approach now, because the emergent functionality (normatively determined by morphological features) is still changed in an effectively equivalent manner.

The Eventual Replacement of Neural Ionic Solutions with Direct Electric Fields

Upon full gradual replacement of the CNS with physical-functionalist equivalents, the preferred embodiment consisted of replacing the ionic solutions with electric fields that preserve the electric potential instantiated by the difference in ionic concentrations on the respective sides of the membrane. Such electric fields can be generated directly, without recourse to electrochemicals for manifesting them. In such a case the integral-membrane components would be replaced by a means of generating and maintaining a static and/or dynamic electric field on either side of the membrane, or even merely of generating an electrical potential (i.e., voltage—a broader category encompassing electric fields) with solid-state electronics.

This procedure would allow a fraction of the speedups (that is, increased rate of subjective perception of time, which extends to speed of thought) resulting from emulatory (i.e., strictly computational) replication-methods by no longer being limited to the rate of passive ionic diffusion—now instead being limited to the propagation velocity of electric or electromagnetic fields.

Wireless Synapses

If we replace the physical synaptic connections the NRU uses to communicate (with both existing biological neurons and with other NRUs) with a wireless means of synaptic-transmission, we can preserve the same functionality (insofar as it is determined by synaptic connectivity) while allowing any NRU to communicate with any other NRU or biological neuron in the brain at potentially equal speed. First we need a way of converting the output of an NRU or biological neuron into information that can be transmitted wirelessly. For cyber-physicalist-functionalist NRUs, regardless of their sub-class, this requires no new technological infrastructure because they already deal with 2nd-order (i.e., not structurally or directly embodied) information; informational-functional NRU deals solely in terms of this type of information, and the cyber-physical-systems sub-class of the physicalist-functionalist NRUs deal with this kind of information in the intermediary stage between sensors and actuators—and consequently, converting what would have been a sequence of electromechanical actuations into information isn’t a problem. Only the passive-physicalist-functionalist NRU class requires additional technological infrastructure to accomplish this, because they don’t already use computational operational-modalities for their normative operation, whereas the other NRU classes do.

We dispose receivers within the range of every neuron (or alternatively NRU) in the brain, connected to actuators – the precise composition of which depends on the operational modality of the receiving biological neuron or NRU. The receiver translates incoming information into physical actuations (e.g., the release of chemical stores), thereby instantiating that informational output in physical terms. For biological neurons, the receiver’s actuators would consist of a means of electrically stimulating the neuron and releasable chemical stores of neurotransmitters (or ionic concentrations as an alternate means of electrical stimulation via the manipulation of local ionic concentrations). For informational-functionalist NRUs, the information is already in a form it can accept; it can simply integrate that information into its extant model. For cyber-physicalist-NRUs, the unit’s CPU merely needs to be able to translate that information into the sequence in which it must electromechanically actuate its artificial ion-channels. For the passive-physicalist (i.e., having no computational hardware devoted to operating individual components at all, operating according to physical feedback between components alone) NRUs, our only option appears to be translating received information into the manipulation of the local environment to vicariously affect the operation of the NRU (e.g., increasing electric potential through manipulations of local ionic concentrations, or increasing the rate of diffusion via applied electric fields to attract ions and thus achieve the same effect as a steeper electrochemical gradient or potential-difference).

The technological and methodological infrastructure for this is very similar to that used for the “integrational NRUs”, which allows a given NRU-class to communicate with either existing biological neurons or NRUs of an alternate class.

Integrating New Neural Nets Without Functional Distortion of Existing Regions

The use of artificial neural networks (which here will designate NRU-networks that do not replicate any existing biological neurons, rather than the normative Artificial Neuron Networks mentioned in the first and second parts of this essay), rather than normative neural prosthetics and BCI, was the preferred method of cognitive augmentation (creation of categorically new functional/experiential modalities) and cognitive amplification (the extension of existing functional/experiential modalities). Due to functioning according to the same operational modality as existing neurons (whether biological or artificial-replacements), they can become a continuous part of our “selves”, whereas normative neural prosthetics and BCI are comparatively less likely to be capable of becoming an integral part of our experiential continuum (or subjective sense of self) due to their significant operational dissimilarity in relation to biological neural networks.

A given artificial neural network can be integrated with existing biological networks in a few ways. One is interior integration, wherein the new neural network is integrated so as to be “inter-threaded”, in which a given artificial-neuron is placed among one or multiple existing networks. The networks are integrated and connected on a very local level. In “anterior” integration, the new network would be integrated in a way comparable to the connection between separate cortical columns, with the majority of integration happening at the peripherals of each respective network or cluster.

If the interior integration approach is used then the functionality of the region may be distorted or negated by virtue of the fact that neurons that once took a certain amount of time to communicate now take comparatively longer due to the distance between them having been increased to compensate for the extra space necessitated by the integration of the new artificial neurons. Thus in order to negate these problematizing aspects, a means of increasing the speed of communication (determined by both [a] the rate of diffusion across the synaptic junction and [b] the rate of diffusion across the neuronal membrane, which in most cases is synonymous with the propagation velocity in the membrane – the exception being myelinated axons, wherein a given action potential “jumps” from node of Ranvier to node of Ranvier; in these cases propagation velocity is determined by the thickness and length of the myelinated sections) must be employed.

My original solution was the use of an artificial membrane morphologically modeled on a myelinated axon that possesses very high capacitance (and thus high propagation velocity), combined with increasing the capacitance of the existing axon or dendrite of the biological neuron. The cumulative capacitance of both is increased in proportion to how far apart they are moved. In this way, the propagation velocity of the existing neuron and the connector-terminal are increased to allow the existing biological neurons to communicate as fast as they would have prior to the addition of the artificial neural network. This solution was eventually supplemented by the wireless means of synaptic transmission described above, which allows any neuron to communicate with any other neuron at equal speed.

Gradually Assigning Operational Control of a Physical NRU to a Virtual NRU

This approach allows us to apply the single-neuron gradual replacement facilitated by the physical-functionalist NRU to the informational-functionalist (physically embodied) NRU. A given section of artificial membrane and its integral membrane components are modeled. When this model is functioning in parallel (i.e., synchronization of operative states) with its corresponding membrane section, the normative operational routines of that artificial membrane section (usually controlled by the unit’s CPU and its programming) are subsequently taken over by the computational model—i.e., the physical operation of the artificial membrane section is implemented according to and in correspondence with the operative states of the model. This is done iteratively, with the informationalist-functionalist NRU progressively controlling more and more sections of the membrane until the physical operation of the whole physical-functionalist NRU is controlled by the informational operative states of the informationalist-functionalist NRU. While this concept sprang originally from the approach of using multiple gradual-replacement phases (with a class of model assigned to each phase, wherein each is more dissimilar to the original than the preceding phase, thereby increasing the cumulative degree of graduality), I now see it as a way of facilitating sub-neuron gradual replacement in computational NRUs. Also note that this approach can be used to go from existing biological membrane-sections to a computational NRU, without a physical-functionalist intermediary stage. This, however, is comparatively more complex because the physical-functionalist NRU already has a means of modulating its operative states, whereas the biological neuron does not. In such a case the section of lipid bilayer membrane would presumably have to be operationally isolated from adjacent sections of membrane, using a system of chemical inventories (of either highly concentrated ionic solution or neurotransmitters, depending on the area of membrane) to produce electrochemical output and chemical sensors to accept the electrochemical input from adjacent sections (i.e., a means of detecting depolarization and hyperpolarization). Thus to facilitate an action potential, for example, the chemical sensors would detect depolarization, the computational NRU would then model the influx of ions through the section of membrane it is replacing and subsequently translate the effective results impinging upon the opposite side to that opposite edge via either the release of neurotransmitters or the manipulation of local ionic concentrations so as to generate the required depolarization at the adjacent section of biological membrane.

Integrational NRU

This consisted of a unit facilitating connection between emulatory (i.e., informational-functionalist) units and existing biological neurons. The output of the emulatory units is converted into chemical and electrical output at the locations where the emulatory NRU makes synaptic connection with other biological neurons, facilitated through electric stimulation or the release of chemical inventories for the increase of ionic concentrations and the release of neurotransmitters, respectively. The input of existing biological neurons making synaptic connections with the emulatory NRU is read, likewise, by chemical and electrical sensors and is converted into informational input that corresponds to the operational modality of the informationalist-functionalist NRU classes.

Solutions to Scale

If we needed NEMS or something below the scale of the present state of MEMS for the technological infrastructure of either (a) the electromechanical systems replicating a given section of neuronal membrane, or (b) the systems used to construct and/or integrate the sections, or those used to remove or otherwise operationally isolate the existing section of lipid bilayer membrane being replaced from adjacent sections, a postulated solution consisted of taking the difference in length between the artificial membrane section and the existing bilipid section (which difference is determined by how small we can construct functionally operative artificial ion-channels) and incorporating this as added curvature in the artificial membrane-section such that its edges converge upon or superpose with the edges of the space left by the removal the lipid bilayer membrane-section. We would also need to increase the propagation velocity (typically determined by the rate of ionic influx, which in turn is typically determined by the concentration gradient or difference in the ionic concentrations on the respective sides of the membrane) such that the action potential reaches the opposite end of the replacement section at the same time that it would normally have via the lipid bilayer membrane. This could be accomplished directly by the application of electric fields with a charge opposite that of the ions (which would attract them, thus increasing the rate of diffusion), by increasing the number of open channels or the diameter of existing channels, or simply by increasing the concentration gradient through local manipulation of extracellular and/or intracellular ionic concentration—e.g., through concentrated electrolyte stores of the relevant ion that can be released to increase the local ionic concentration.

If the degree of miniaturization is so low as to make this approach untenable (e.g., increasing curvature still doesn’t allow successful integration) then a hypothesized alternative approach was to increase the overall space between adjacent neurons, integrate the NRU, and replace normative connection with chemical inventories (of either ionic compound or neurotransmitter) released at the site of existing connection, and having the NRU (or NRU sub-section—i.e., artificial membrane section) wirelessly control the release of such chemical inventories according to its operative states.

The next chapter describes (a) possible physical bases for subjective-continuity through a gradual-uploading procedure and (b) possible design requirements for in vivo brain-scanning and for systems to construct and integrate the prosthetic neurons with the existing biological brain.

Franco Cortese is an editor for Transhumanity.net, as well as one of its most frequent contributors.  He has also published articles and essays on Immortal Life and The Rational Argumentator. He contributed 4 essays and 7 debate responses to the digital anthology Human Destiny is to Eliminate Death: Essays, Rants and Arguments About Immortality.

Franco is an Advisor for Lifeboat Foundation (on its Futurists Board and its Life Extension Board) and contributes regularly to its blog.

Bibliography

Project Avatar (2011). Retrieved February 28, 2013 from http://2045.com/tech2/

Gradual Neuron Replacement for the Preservation of Subjective-Continuity – Article by Franco Cortese

Gradual Neuron Replacement for the Preservation of Subjective-Continuity – Article by Franco Cortese

The New Renaissance Hat
Franco Cortese
May 19, 2013
******************************
This essay is the fourth chapter in Franco Cortese’s forthcoming e-book, I Shall Not Go Quietly Into That Good Night!: My Quest to Cure Death, published by the Center for Transhumanity. The first three chapters were previously published on The Rational Argumentator as “The Moral Imperative and Technical Feasibility of Defeating Death”, “Immortality: Material or Ethereal? Nanotech Does Both!, and “Concepts for Functional Replication of Biological Neurons“.
***

Gradual Uploading Applied to Single Neurons (2008)

In early 2008 I was trying to conceptualize a means of applying the logic of gradual replacement to single neurons under the premise that extending the scale of gradual replacement to individual sections of the neuronal membrane and its integral membrane proteins—thus increasing the degree of graduality between replacement sections—would increase the likelihood of subjective-continuity through substrate transfer. I also started moving away from the use of normative nanotechnology as the technological and methodological infrastructure for the NRUs, as it would delay the date at which these systems could be developed and experimentally verified. Instead I started focusing on conceptualizing systems that electromechanically replicate the functional modalities of the small-scale integral-membrane-components of the neuron. I was calling this approach the “active mechanical membrane” to differentiate it from the electro-chemical-mechanical modalities of the nanotech approach. I also started using MEMS rather than NEMS for the underlying technological infrastructure (because MEMS are less restrictive) while identifying NEMS as preferred.

I felt that trying to replicate the metabolic replacement rate in biological neurons should be the ideal to strive for, since we know that subjective-continuity is preserved through the gradual metabolic replacement (a.k.a. molecular-turnover) that occurs in the existing biological brain. My approach was to measure the normal rate of metabolic replacement in existing biological neurons and the scale at which such replacement occurs (i.e., are the sections being replaced metabolically with single molecules, molecular complexes, or whole molecular clusters?). Then, when replacing sections of the membrane with electromechanical functional equivalents, the same ratio of replacement-section size to replacement-time factor would be applied—that is, the time between sectional replacement would be increased in proportion to how much larger the sectional-replacement section/scale is compared to the existing scale of metabolic replacement-sections/scale. Replacement size/scale is defined as the size of the section being replaced—and so would be molecular complexes in the case of normative metabolic replacement. Replacement time is defined as the interval of time between a given section being replaced and a section that it has causal connection with is replaced; in metabolic replacement it is the time interval between a given molecular complex being replaced and an adjacent (or directly-causally-connected) molecular complex being replaced.

I therefore posited the following formula:

 Ta = (Sa/Sb)*Tb,

where Sa is the size of the artificial-membrane-replacement sections, Sb is the size of the metabolic replacement sections, Tb is the time interval between the metabolic replacement of two successive metabolic replacement sections, and Ta is the time interval needing to be applied to the comparatively larger artificial-membrane-replacement sections so as to preserve the same replacement-rate factor (and correspondingly the same degree of graduality) that exists in normative metabolic replacement through the process of gradual replacement on the comparatively larger scale of the artificial-membrane sections.

The use of the time-to-scale factor corresponding with normative molecular turnover or “metabolic replacement” follows from the fact that we know subjective-continuity through substrate replacement is successful at this time-to-scale ratio. However, the lack of a non-arbitrarily quantifiable measure of time and the fact that that time is infinitely divisible (i.e., it can be broken down into smaller intervals to an arbitrarily large degree) logically necessitates that the salient variable is not time, but rather causal interaction between co-affective or “causally coupled” components. Interaction between components and the state transitions each component or procedural step undergo are the only viable quantifiable measures of time. Thus, while time is the relevant variable in the above equation, a better (i.e., more methodologically rigorous) variable would be a measure of either (a) the number of causal interactions occurring between co-affective or “adjacent” components within the interval of replacement time Ta, which is synonymous with the frequency of causal interaction; or (b) the number of state-transitions a given component undergoes within the interval of time Ta. While they should be generally correlative, in that state-transitions are facilitated via causal interaction among components, state-transitions may be a better metric because they allow us to quantitatively compare categorically dissimilar types of causal interaction that otherwise couldn’t be summed into a single variable or measure. For example, if one type of molecular interaction has a greater effect on the state-transitions of either component involved (i.e., facilitates a comparatively greater state-transition) than does another type of molecular interaction, then quantifying a measure of causal interactions may be less accurate than quantifying a measure of the magnitude of state-transitions.

In this way the rate of gradual replacement, despite being on a scale larger than normative metabolic replacement, would hypothetically follow the same degree of graduality with which biological metabolic replacement occurs. This was meant to increase the likelihood of subjective-continuity through a substrate-replacement procedure (both because it is necessarily more gradual than gradual replacement of whole individual neurons at a time, and because it preserves the degree of graduality that exists through the normative metabolic replacement that we already undergo).

Replicating Neuronal Membrane and Integral Membrane Components

Thus far there have been 2 main classes of neuron-replication approach identified: informational-functionalist and physical-functionalist, the former corresponding to computational and simulation/emulation approaches and the latter to physically embodied, “prosthetic” approaches.

The physicalist-functionalist approach, however, can at this point be further sub-divided into two sub-classes. The first can be called “cyber-physicalist-functionalist”, which involves controlling the artificial ion-channels and receptor-channels via normative computation (i.e., an internal CPU or controller-circuit) operatively connected to sensors and to the electromechanical actuators and components of the ion and receptor channels (i.e., sensing the presence of an electrochemical gradient or difference in electrochemical potential [equivalent to relative ionic concentration] between the respective sides of a neuronal membrane, and activating the actuators of the artificial channels to either open or remain closed, based upon programmed rules). This sub-class is an example of a cyber-physical system, which designates any system with a high level of connection or interaction between its physical and computational components, itself a class of technology that grew out of embedded systems, which designates any system using embedded computational technology and includes many electronic devices and appliances.

This is one further functional step removed from the second approach, which I was then simply calling the “direct” method, but which would be more accurately called the passive-physicalist-functionalist approach. Electronic systems are differentiated from electric systems by being active (i.e., performing computation or more generally signal-processing), whereas electric systems are passive and aren’t meant to transform (i.e., process) incoming signals (though any computational system’s individual components must at some level be comprised of electric, passive components). Whereas the cyber-physicalist-functionalist sub-class has computational technology controlling its processes, the passive-physicalist-functionalist approach has components emergently constituting a computational device. This consisted of providing the artificial ion-channels with a means of opening in the presence of a given electric potential difference (i.e., voltage) and the receptor-channels with a means of opening in response to the unique attributes of the neurotransmitter it corresponds to (such as chemical bonding as in ligand-based receptors, or alternatively in response to its electrical properties in the same manner – i.e., according to the same operational-modality – as the artificial ion channels), without a CPU correlating the presence of an attribute measured by sensors with the corresponding electromechanical behavior of the membrane needing to be replicated in response thereto. Such passive systems differ from computation in that they only require feedback between components, wherein a system of mechanical, electrical, or electromechanical components is operatively connected so as to produce specific system-states or processes in response to the presence of specific sensed system-states of its environment or itself. An example of this in regards to the present case would be constructing an ionic channel from piezoelectric materials, such that the presence of a certain electrochemical potential induces internal mechanical strain in the material; the spacing, dimensions and quantity of segments would be designed so as to either close or open, respectively, as a single unit when eliciting internal mechanical strain in response to one electrochemical potential while remaining unresponsive (or insufficiently responsive—i.e., not opening all the way) to another electrochemical potential. Biological neurons work in a similarly passive way, in which systems are organized to exhibit specific responses to specific stimuli in basic stimulus-response causal sequences by virtue of their own properties rather than by external control of individual components via CPU.

However, I found the cyber-physicalist approach preferable if it proved to be sufficient due to the ability to reprogram computational systems, which isn’t possible in passive systems without necessitating a reorganization of the component—which itself necessitates an increase in the required technological infrastructure, thereby increasing cost and design-requirements. This limit on reprogramming also imposes a limit on our ability to modify and modulate the operation of the NRUs (which will be necessary to retain the function of neural plasticity—presumably a prerequisite for experiential subjectivity and memory). The cyber-physicalist approach also seemed preferable due to a larger degree of variability in its operation: it would be easier to operatively connect electromechanical membrane components (e.g., ionic channels, ion pumps) to a CPU, and through the CPU to sensors, programming it to elicit a specific sequence of ionic-channel opening and closing in response to specific sensor-states, than it would be to design artificial ionic channels to respond directly to the presence of an electric potential with sufficient precision and accuracy.

In the cyber-physicalist-functionalist approach the membrane material is constructed so as to be (a) electrically insulative, while (b) remaining thin enough to act as a capacitor via the electric potential differential (which is synonymous with voltage) between the two sides of the membrane.

The ion-channel replacement units consisted of electromechanical pores that open for a fixed amount of time in the presence of an ion gradient (a difference in electric potential between the two sides of the membrane); this was to be accomplished electromechanically via a means of sensing membrane depolarization (such as through the use of reference electrodes) connected to a microcircuit (or nanocircuit, hereafter referred to as a CPU) programmed to open the electromechanical ion-channels for a length of time corresponding to the rate of normative biological repolarization (i.e., the time it takes to restore the membrane polarization to the resting-membrane-potential following an action-potential), thus allowing the influx of ions at a rate equal to the biological ion-channels. Likewise sections of the pre-synaptic membrane were to be replaced by a section of inorganic membrane containing units that sense the presence of the neurotransmitter corresponding to the receptor being replaced, which were to be connected to a microcircuit programmed to elicit specific changes (i.e., increase or decrease in ionic permeability, such as through increasing or decreasing the diameter of ion-channels—e.g., through an increase or decrease in electric stimulation of piezoelectric crystals, as described above—or an increase or decrease in the number of open channels) corresponding to the change in postsynaptic potential in the biological membrane resulting from postsynaptic receptor-binding. This requires a bit more technological infrastructure than I anticipated the ion-channels requiring.

While the accurate and active detection of particular types and relative quantities of neurotransmitters is normally ligand-gated, we have a variety of potential, mutually exclusive approaches. For ligand-based receptors, sensing the presence and steepness of electrochemical gradients may not suffice. However, we don’t necessarily have to use ligand-receptor fitting to replicate the functionality of ligand-based receptors. If there is a difference in the charge (i.e., valence) between the neurotransmitter needing to be detected and other neurotransmitters, and the degree of that difference is detectable given the precision of our sensing technologies, then a means of sensing a specific charge may prove sufficient. I developed an alternate method for ligand-based receptor fitting in the event that sensing-electric charge proved insufficient, however. Different chemicals (e.g., neurotransmitters, but also potentially electrolyte solutions) have different volume-to-weight ratios. We equip the artificial-membrane sections with an empty compartment capable of measuring the weight of its contents. Since the volume of the container is already known, this would allow us to identify specific neurotransmitters (or other relevant molecules and compounds) based on their unique weight-to-volume ratio. By operatively connecting the unit’s CPU to this sensor, we can program specific operations (i.e., receptor opens allowing entry for fixed amount of time, or remains closed) in response to the detection of specific neurotransmitters. Though it is unlikely to be necessitated, this method could also work for the detection of specific ions, and thus could work as the operating mechanism underlying the artificial ion-channels as well—though this would probably require higher-precision volume-to-weight comparison than is required for neurotransmitters.

Sectional Integration with Biological Neurons

Integrating replacement-membrane sections with adjacent sections of the existing lipid bilayer membrane becomes a lot less problematic if the scale at which the membrane sections are handled (determined by the size of the replacement membrane sections) is homogenous, as in the case of biological tissues, rather than molecularly heterogeneous—that is, if we are affixing the edges to a biological tissue, rather than to complexes of individual lipid molecules. Reasons for hypothesizing a higher probability for homogeneity at the replacement scale include (a) the ability of experimenters and medical researchers to puncture the neuronal membrane with a micropipette (so as to measure membrane voltage) without rupturing the membrane beyond functionality, and (b) the fact that sodium and potassium ions do not leak through the gaps between the individual bilipid molecules, which would be present if it were heterogeneous at this scale. If we find homogeneity at the scale of sectional replacement, we can use more normative means of affixing the edges of the replacement membrane section with the existing lipid bilayer membrane, such as micromechanical fasteners, adhesive, or fusing via heating or energizing. However, I also developed an approach applicable if the scale of sectional replacement was found to be molecular and thus heterogeneous. We find an intermediate chemical that stably bonds to both the bilipid molecules constituting the membrane and the molecules or compounds constituting the artificial membrane section. Note that if the molecules or compounds constituting either must be energized so as to put them in an abnormal (i.e., unstable) energy state to make them susceptible to bonding, this is fine so long as the energies don’t reach levels damaging to the biological cell (or if such energies could be absorbed prior to impinging upon or otherwise damaging the biological cell). If such an intermediate molecule or compound cannot be found, a second intermediate chemical that stably bonds with two alternate and secondary intermediate molecules (which themselves bond to either the biological membrane or the non-biological membrane section, respectively) can be used. The chances of finding a sequence of chemicals that stably bond (i.e., a given chemical forms stable bonds with the preceding and succeeding chemicals in the sequence) increases in proportion to the number of intermediate chemicals used. Note that it might be possible to apply constant external energization to certain molecules so as to force them to bond in the case that a stable bond cannot be formed, but this would probably be economically prohibitive and potentially dangerous, depending on the levels of energy and energization-precision.

I also worked on the means of constructing and integrating these components in vivo, using MEMS or NEMS. Most of the developments in this regard are described in the next chapter. However, some specific variations on construction procedure were necessitated by the sectional-integration procedure, which I will comment on here. The integration unit would position itself above the membrane section. Using the data acquired by the neuron data-measurement units, which specify the constituents of a given membrane section and assign it a number corresponding to a type of artificial-membrane section in the integration unit’s section-inventory (essentially a store of stacked artificial-membrane-sections). A means of disconnecting a section of lipid bilayer membrane from the biological neuron is depressed. This could be a hollow rectangular compartment with edges that sever the lipid bilayer membrane via force (e.g., edges terminate in blades), energy (e.g., edges terminate in heat elements), or chemical corrosion (e.g., edges coated with or secrete a corrosive substance). The detached section of lipid bilayer membrane is then lifted out and compacted, to be drawn into a separate compartment for storing waste organic materials. The artificial-membrane section is subsequently transported down through the same compartment. Since it is perpendicular to the face of the container, moving the section down through the compartment should force the intra-cellular fluid (which would have presumably leaked into the constructional container’s internal area when the lipid bilayer membrane-section was removed) back into the cell. Once the artificial-membrane section is in place, the preferred integration method is applied.

Sub-neuronal (i.e., sectional) replacement also necessitates that any dynamic patterns of polarization (e.g., an action potential) are continuated during the interval of time between section removal and artificial-section integration. This was to be achieved by chemical sensors (that detect membrane depolarization) operatively connected to actuators that manipulate ionic concentration on the other side of the membrane gap via the release or uptake of ions from biochemical inventories so as to induce membrane depolarization on the opposite side of the membrane gap at the right time. Such techniques as partially freezing the cell so as to slow the rate of membrane depolarization and/or the propagation velocity of action potentials were also considered.

The next chapter describes my continued work in 2008, focusing on (a) the design requirements for replicating the neural plasticity necessary for memory and subjectivity, (b) the active and conscious modulation and modification of neural operation, (c) wireless synaptic transmission, (d) on ways to integrate new neural networks (i.e., mental amplification and augmentation) without disrupting the operation of existing neural networks and regions, and (e) a gradual transition from or intermediary phase between the physical (i.e., prosthetic) approach and the informational (i.e., computational, or mind-uploading proper) approach.

Franco Cortese is an editor for Transhumanity.net, as well as one of its most frequent contributors.  He has also published articles and essays on Immortal Life and The Rational Argumentator. He contributed 4 essays and 7 debate responses to the digital anthology Human Destiny is to Eliminate Death: Essays, Rants and Arguments About Immortality.

Franco is an Advisor for Lifeboat Foundation (on its Futurists Board and its Life Extension Board) and contributes regularly to its blog.

Bibliography

Churchland, P. S. (1989). Neurophilosophy: Toward a Unified Science of the Mind/Brain.  MIT Press, p. 30.

Pribram, K. H. (1971). Languages of the Brain: Experimental Paradoxes and Principles in Neuropsychology. New York: Prentice Hall/Brandon House.

Immortality: Material or Ethereal? Nanotech Does Both! – Article by Franco Cortese

Immortality: Material or Ethereal? Nanotech Does Both! – Article by Franco Cortese

The New Renaissance Hat
Franco Cortese
May 11, 2013
******************************

This essay is the second chapter in Franco Cortese’s forthcoming e-book, I Shall Not Go Quietly Into That Good Night!: My Quest to Cure Death, published by the Center for Transhumanity. The first chapter was previously published on The Rational Argumentator as “The Moral Imperative and Technical Feasibility of Defeating Death“.

In August 2006 I conceived of the initial cybernetic brain-transplant procedure. It originated from a very simple, even intuitive sentiment: if there were heart and lung machines and prosthetic organs, then why couldn’t these be integrated in combination with modern (and future) robotics to keep the brain alive past the death of its biological body? I saw a possibility, felt its magnitude, and threw myself into realizing it. I couldn’t think of a nobler quest than the final eradication of involuntary death, and felt willing to spend the rest of my life trying to make it happen.

First I collected research on organic brain transplantation, on maintaining the brain’s homeostatic and regulatory mechanisms outside the body (or in this case without the body), on a host of prosthetic and robotic technologies (including sensory prosthesis and substitution), and on the work in Brain-Computer-Interface technologies that would eventually allow a given brain to control its new, non-biological body—essentially collecting the disparate mechanisms and technologies that would collectively converge to facilitate the creation of a fully cybernetic body to house the organic brain and keep it alive past the death of its homeostatic and regulatory organs.

I had by this point come across online literature on Artificial Neurons (ANs) and Artificial Neural Networks (ANNs), which are basically simplified mathematical models of neurons meant to process information in a way coarsely comparable to them. There was no mention in the literature of integrating them with existing neurons or for replacing existing neurons towards the objective of immortality ; their use was merely as an interesting approach to computation particularly optimal to certain situations. While artificial neurons can be run on general-purpose hardware (massively parallel architectures being the most efficient for ANNs, however), I had something more akin to neuromorphic hardware in mind (though I wasn’t aware of that just yet).

At its most fundamental level, Artificial Neurons need not even be physical at all. Their basic definition is a mathematical model roughly based on neuronal operation – and there is nothing precluding that model from existing solely on paper, with no actual computation going on. When I discovered them, I had thought that a given artificial neuron was a physically-embodied entity, rather than a software simulation. – i.e., an electronic device that operates in a way comparable to biological neurons.  Upon learning that they were mathematical models however, and that each AN needn’t be a separate entity from the rest of the ANs in a given AN Network, I saw no problem in designing them so as to be separate physical entities (which they needed to be in order to fit the purposes I had for them – namely, the gradual replacement of biological neurons with prosthetic functional equivalents). Each AN would be a software entity run on a piece of computational substrate, enclosed in a protective casing allowing it to co-exist with the biological neurons already in-place. The mathematical or informational outputs of the simulated neuron would be translated into biophysical, chemical, and electrical output by operatively connecting the simulation to an appropriate series of actuators (which could range from being as simple as producing electric fields or currents, to the release of chemical stores of neurotransmitters) and likewise a series of sensors to translate biophysical, chemical, and electrical properties into the mathematical or informational form they would need to be in to be accepted as input by the simulated AN.

Thus at this point I didn’t make a fundamental distinction between replicating the functions and operations of a neuron via physical embodiment (e.g., via physically embodied electrical, chemical, and/or electromechanical systems) or via virtual embodiment (usefully considered as 2nd-order embodiment, e.g., via a mathematical or computational model, whether simulation or emulation, run on a 1st-order physically embodied computational substrate).

The potential advantages, disadvantages, and categorical differences between these two approaches were still a few months away. When I discovered ANs, still thinking of them as physically embodied electronic devices rather than as mathematical or computational models, I hadn’t yet moved on to ways of preserving the organic brain itself so as to delay its organic death. Their utility in constituting a more permanent, durable, and readily repairable supplement for our biological neurons wasn’t yet apparent.

I initially saw their utility as being intelligence amplification, extension and modification through their integration with the existing biological brain. I realized that they were categorically different than Brain-Computer Interfaces (BCIs) and normative neural prosthesis for being able to become an integral and continuous part of our minds and personalities – or more properly the subjective, experiential parts of our minds. If they communicated with single neurons and interact with them on their own terms—if the two were operationally indistinct—then they could become a continuous part of us in a way that didn’t seem possible for normative BCI due to their fundamental operational dissimilarity with existing biological neural networks. I also collected research on the artificial synthesis and regeneration of biological neurons as an alternative to ANs. This approach would replace an aging or dying neuron with an artificially synthesized but still structurally and operationally biological neuron, so as to maintain the aging or dying neuron’s existing connections and relative location. I saw this procedure (i.e., adding artificial or artificially synthesized but still biological neurons to the existing neurons constituting our brains, not yet for the purposes of gradually replacing the brain but instead for the purpose of mental expansion and amplification) as not only allowing us to extend our existing functional and experiential modalities (e.g., making us smarter through an increase in synaptic density and connectivity, and an increase in the number of neurons in general) but even to create fundamentally new functional and experiential modalities that are categorically unimaginable to us now via the integration of wholly new Artificial Neural Networks embodying such new modalities. Note that I saw this as newly possible with my cybernetic-body approach because additional space could be made for the additional neurons and neural networks, whereas the degree with which we could integrate new, artificial neural networks in a normal biological body would be limited by the available volume of the unmodified skull.

Before I discovered ANs, I speculated in my notes as to whether the “bionic nerves” alluded to in some of the literature I had collected by this point (specifically regarding BCI, neural prosthesis, and the ability to operatively connect a robotic prosthetic extremity – e.g., an arm or a leg – via BCI) could be used to extend the total number of neurons and synaptic connections in the biological brain. This sprang from my knowledge on the operational similarities between neurons and muscle cells, both of the larger class of excitable cells.

Kurzweil’s cyborgification approach (i.e., that we could integrate non-biological systems with our biological brains to such an extent that the biological portions become so small as to be negligible to our subjective-continuity when they succumb to cell-death, thus achieving effective immortality without needing to actually replace any of our existing biological neurons at all) may have been implicit in this concept. I envisioned our brains increasing in size many times over and thus that the majority of our mind would be embodied or instantiated in larger part by the artificial portion than by the biological portions; the fact that the degree with which the loss of a part of our brain will affect our emergent personalities depends on how big (other potential metrics alternative to size include connectivity and the degree with which other systems depend on that potion for their own normative operation) that lost part is in comparison to the total size of the brain, the loss of a lobe being much worse than the loss of a neuron, follows naturally from this initial premise. The lack of any explicit statement of this realization in my notes during this period, however, makes this mere speculation.

It wasn’t until November 11, 2006, that I had the fundamental insight underlying mind-uploading—that the replacement of existing biological neurons with non-biological functional equivalents that maintain the existing relative location and connection of such biological neurons could very well facilitate maintaining the memory and personality embodied therein or instantiated thereby—essentially achieving potential technological immortality, since the approach is based on replacement and iterations of replacement-cycles can be run indefinitely. Moreover, the fact that we would be manufacturing such functional equivalents ourselves means that we could not only diagnose potential eventual dysfunctions easier and with greater speed, but we could manufacture them so as to have readily replaceable parts, thus simplifying the process of physically remediating any such potential dysfunction or operational degradation, even going so far as to include systems for the safe import and export of replacement components or as to make all such components readily detachable, so that we don’t have to cause damage to adjacent structures and systems in the process of removing a given component.

Perhaps it wasn’t so large a conceptual step from knowledge of the existence of computational models of neurons to the realization of using them to replace existing biological neurons towards the aim of immortality. Perhaps I take too much credit for independently conceiving both the underlying conceptual gestalt of mind-uploading, as well as some specific technologies and methodologies for its pragmatic technological implementation. Nonetheless, it was a realization I arrived at on my own, and was one that I felt would allow us to escape the biological death of the brain itself.

While I was aware (after a little more research) that ANNs were mathematical (and thus computational) models of neurons, hereafter referred to as the informationalist-functionalist approach, I felt that a physically embodied (i.e., not computationally emulated or simulated) prosthetic approach, hereafter referred to as the physicalist-functionalist approach, would be a better approach to take. This was because even if the brain were completely reducible to computation, a prosthetic approach would necessarily facilitate the computation underlying the functioning of the neuron (as the physical operations of biological neurons do presently), and if the brain proved to be computationally irreducible, then the prosthetic approach would in such a case presumably preserve whatever salient physical processes were necessary. So the prosthetic approach didn’t necessitate the computational-reducibility premise – but neither did it preclude such a view, thereby allowing me to hedge my bets and increase the cumulative likelihood of maintaining subjective-continuity of consciousness through substrate-replacement in general.

This marks a telling proclivity recurrent throughout my project: the development of mutually exclusive and methodologically and/or technologically alternate systems for a given objective, each based upon alternate premises and contingencies – a sort of possibilizational web unfurling fore and outward. After all, if one approach failed, then we had alternate approaches to try. This seemed like the work-ethic and conceptualizational methodology that would best ensure the eventual success of the project.

I also had less assurance in the sufficiency of the informational-functionalist approach at the time, stemming mainly from a misconception with the premises of normative Whole-Brain Emulation (WBE). When I first discovered ANs, I was more dubious at that point about the computational reducibility of the mind because I thought that it relied on the premise that neurons act in a computational fashion (i.e., like normative computational paradigms) to begin with—thus a conflation of classical computation with neural operation—rather than on the conclusion, drawn from the Church-Turing thesis, that mind is computable because the universe is. It is not that the brain is a computer to begin with, but that we can model any physical process via mathematical/computational emulation and simulation. The latter would be the correct view, and I didn’t really realize that this was the case until after I had discovered the WBE roadmap in 2010. This fundamental misconception allowed me, however, to also independently arrive at the insight underlying the real premise of WBE:  that combining premise A – that we had various mathematical computational models of neuron behavior – with premise B – that we can perform mathematical models on computers – ultimately yields the conclusion C – that we can simply perform the relevant mathematical models on computational substrate, thereby effectively instantiating the mind “embodied” in those neural operations while simultaneously eliminating many logistical and technological challenges to the prosthetic approach. This seemed both likelier than the original assumption—conflating neuronal activity with normative computation, as a special case not applicable to, say, muscle cells or skin cells, which wasn’t the presumption WBE makes at all—because this approach only required the ability to mathematically model anything, rather than relying on a fundamental equivalence between two different types of physical system (neuron and classical computer). The fact that I mistakenly saw it as an approach to emulation that was categorically dissimilar to normative WBE also helped urge me on to continue conceptual development of the various sub-aims of the project after having found that the idea of brain emulation already existed, because I thought that my approach was sufficiently different to warrant my continued effort.

There are other reasons for suspecting that mind may not be computationally reducible using current computational paradigms – reasons that rely on neither vitalism (i.e., the claim that mind is at least partially immaterial and irreducible to physical processes) nor on the invalidity of the Church-Turing thesis. This line of reasoning has nothing to do with functionality and everything to do with possible physical bases for subjective-continuity, both a) immediate subjective-continuity (i.e., how can we be a unified, continuous subjectivity if all our component parts are discrete and separate in space?), which can be considered as the capacity to have subjective experience, also called sentience (as opposed to sapience, which designated the higher cognitive capacities like abstract thinking) and b) temporal subjective-continuity (i.e., how do we survive as continuous subjectivities through a process of gradual substrate replacement?). Thus this argument impacts the possibility of computationally reproducing mind only insofar as the definition of mind is not strictly functional but is made to include a subjective sense of self—or immediate subjective-continuity. Note that subjective-continuity through gradual replacement is not speculative (just the scale and rate required to sufficiently implement it are), but rather has proof of concept in the normal metabolic replacement of the neuron’s constituent molecules. Each of us is a different person materially than we were 7 years ago, and we still claim to retain subjective-continuity. Thus, gradual replacement works; it is just the scale and rate required that are under question.

This is another way in which my approach and project differs from WBE. WBE equates functional equivalence (i.e., the same output via different processes) with subjective equivalence, whereas my approach involved developing variant approaches to neuron-replication-unit design that were each based on a different hypothetical basis for instantive subjective continuity.

 Are Current Computational Paradigms Sufficient?

Biological neurons are both analog and binary. It is useful to consider a 1st tier of analog processes, manifest in the action potentials occurring all over the neuronal soma and terminals, with a 2nd tier of binary processing, in that either the APs’ sum crosses the threshold value needed for the neuron to fire, or it falls short of it and the neuron fails to fire. Thus the analog processes form the basis of the digital ones. Moreover, the neuron is in an analog state even in the absence of membrane depolarization through the generation of the resting-membrane potential (maintained via active ion-transport proteins), which is analog rather than binary for always undergoing minor fluctuations due to it being an active process (ion-pumps) that instantiates it. Thus the neuron at any given time is always in the process of a state-transition (and minor state-transitions still within the variation-range allowed by a given higher-level static state; e.g., resting membrane potential is a single state, yet still undergoes minor fluctuations because the ions and components manifesting it still undergo state-transitions without the resting-membrane potential itself undergoing a state-transition), and thus is never definitively on or off. This brings us to the first potential physical basis for both immediate and temporal subjective-continuity. Analog states are continuous, and the fact that there is never a definitive break in the processes occurring at the lower levels of the neuron represents a potential basis for our subjective sense of immediate and temporal continuity.

Paradigms of digital computation, on the other hand, are at the lowest scale either definitively on or definitively off. While any voltage within a certain range will cause the generation of an output, it is still at base binary because in the absence of input the logic elements are not producing any sort of fluctuating voltage—they are definitively off. In binary computation, the substrates undergo a break (i.e., region of discontinuity) in their processing in the absence of inputs, and are in this way fundamentally dissimilar to the low-level operational modality of biological neurons by virtue of being procedurally discrete rather than procedurally continuous.

If the premise that the analog and procedurally continuous nature of neuron-functioning (including action potentials, resting-membrane potential, and metabolic processes that form a potential basis for immediate and temporal subjective-continuity) holds true, then current digital paradigms of computation may prove insufficient at maintaining subjective-continuity if used as the substrate in a gradual-replacement procedure, while still being sufficient to functionally replicate the mind in all empirically verifiable metrics and measures. This is due to both the operational modality of binary processing (i.e., lack of analog output) and the procedural modality of binary processing (the lack of temporal continuity or lack of producing minor fluctuations in reference to a baseline state when in a resting or inoperative state). A logic element could have a fluctuating resting voltage rather than the absence of any voltage and could thus be procedurally continuous while still being operationally discrete by producing solely binary outputs.

So there are two possibilities here. One is that any physical substrate used to replicate a neuron (whether via 1st-order embodiment a.k.a prosthesis/physical-systems, or via 2nd-order embodiment a.k.a computational emulation or simulation) must not undergo a break in its operation in the absence of input, because biological neurons do not, and this may be a potential basis for instantive subjective-continuity, but rather must produce a continuous or uninterrupted signal when in a “steady-state” (i.e., in the absence of inputs). The second possibility includes all the premises of the first, but adds that such an inoperative-state signal (or “no-inputs”-state signal) undergo minor fluctuations (because then a steady stream of causal interaction is occurring – e.g., producing a steady signal could be as discontinuous as no signal, like “being on pause”.

Thus one reason for developing the physicalist-functionalist (i.e., physically embodied prosthetic) approach to NRU design was a hedging of bets, in the case that a.) current computational substrates fail to replicate a personally continuous mind for the reasons described above, or b.) we fail to discover the principles underlying a given physical process—thus being unable to predictively model it—but still succeed in integrating them with the artificial systems comprising the prosthetic approach until such a time as to be able to discover their underlying principles, or c.) in the event that we find some other, heretofore unanticipated conceptual obstacle to computational reducibility of mind.

Franco Cortese is an editor for Transhumanity.net, as well as one of its most frequent contributors.  He has also published articles and essays on Immortal Life and The Rational Argumentator. He contributed 4 essays and 7 debate responses to the digital anthology Human Destiny is to Eliminate Death: Essays, Rants and Arguments About Immortality.

Franco is an Advisor for Lifeboat Foundation (on its Futurists Board and its Life Extension Board) and contributes regularly to its blog.

Bibliography

Copeland, J. B. (2008). Neural Network. In The Stanford Encyclopedia of Philosophy (Fall 2008 Edition). Retrieved February 28, 2013. from http://plato.stanford.edu/archives/fall2008/entries/church-turing

Crick, F. (1984 Nov 8-14). Memory and molecular turnover. In Nature, 312(5990)(101). PMID: 6504122

Criterion of Falsifiability, Encyclopædia Britannica. Encyclopædia Britannica Online Academic Edition. Retrieved February 28, 2013, from http://www.britannica.com/EBchecked/topic/201091/criterion-of-falsifiability

Drexler, K. E. (1986). Engines of Creation: The Coming Era of Nanotechnology. New York: Anchor Books.

Grabianowski (2007). How Brain-computer Interfaces Work. Retrieved February 28, 2013, from http://computer.howstuffworks.com/brain-computer-interface.htm

Koene, R. (2011). The Society of Neural Prosthetics and Whole Brain Emulation Science. Retrieved February 28, 2013, from http://www.minduploading.org/

Martins, N. R., Erlhagen, W. & Freitas Jr., R. A. (2012). Non-destructive whole-brain monitoring using nanorobots: Neural electrical data rate requirements. International Journal of Machine Consciousness, 2011. Retrieved February 28, 2013, from http://www.nanomedicine.com/Papers/NanoroboticBrainMonitoring2012.pdf.

Narayan, A. (2004). Computational Methods for NEMS. Retrieved February 28, 2013, from http://nanohub.org/resources/407.

Sandberg, A. & Bostrom, N. (2008). Whole Brain Emulation: A Roadmap, Technical Report #2008-3. Retrieved February 28, 2013, from Whole Brain Emulation: A Roadmap, Technical Report #2008-3.

Star, E. N., Kwiatkowski, D. J. & Murthy, V. N. (2002). Rapid turnover of actin in dendritic spines and its regulation by activity. Nature Neuroscience, 5 , 239-246.

Tsien, J. Z., Rampon, C., Tang,Y.P. & Shimizu, E. (2000). NMDA receptor dependent synaptic reinforcement as a crucial process for memory consolidation. Science, 290 , 1170-1174.

Vladimir, Z. (2013). Neural Network. In Encyclopædia Britannica Online Academic Edition. Retrieved February 28, 2013, from http://www.britannica.com/EBchecked/topic/410549/neural-network

Wolf, W. & (March 2009). Cyber-physical Systems. In Embedded Computing. Retrieved February 28, 2013, from http://www.jiafuwan.net/download/cyber_physical_systems.pdf