The question in the title now seems foolish or pointless. Of course what has historically been called “consciousness” is a techno-scientific object of analysis and research! For decades now. Even a quick search online would yield such a definition of cognitive science:
Cognitive science is the interdisciplinary scientific study of the mind and its processes. It examines what cognition is, what it does, and how it works. It includes research on intelligence and behavior… and consists of multiple research fields, including psychology, artificial intelligence, philosophy, neuroscience, linguistics, and anthropology…
Many translate the term cognition as “knowledge.” Certainly this term has been chosen carefully and such is a typical interpretation of it; whereas in English the term for consciousness is different: conscience. However, the “knowledge” implied in cognition is not that “knowledge” which has been objectified / neutralized as information. It is, if we may put it this way, the subjective side of “knowing,” together with the active dimensions of “doing,” “thinking,” etc. Because, moreover, the Greek word “συνείδηση” (consciousness) pertains to this subjective side of knowing, we are entitled to claim that cognitive science indeed has as its goal the “analysis and action” of consciousness, while avoiding the weight of philosophical, ethical and political issues that have historically been associated with consciences. After all, this is not denied by many specialties of cognitive scientists either, as we shall see in this report and in others, in future issues of cyborg.
This long-term, multi-level techno-scientific engagement with consciousness may not have yielded the spectacular wealth of applications seen in information technology or biotechnologies, but that does not reflect its social and political significance. On the contrary, many fields of the cognitive sciences—neuroscience, for example—have managed to avoid any scrutiny. The common notion of what “new technologies” are is usually confined to gadgets and how they are used, which amounts to profound ignorance of the real technological stack of the new capitalist paradigm—let alone informed critique.
The general idea of what “consciousness” itself is, from the perspective of the cognitive sciences, can be considered old: it is a neurobiological process that takes place in the brain (of humans as well as other living beings) on the basis of stimuli and of the “history” of each organism—whether individual or social/species history. However, in the theoretical starting points of recent decades, the old “electro-chemical” notion of the linear transmission of “messages” from the sensory organs to “specific centers” of the brain, and the point-to-point mapping of these centers onto corresponding neurotransmitters, has been replaced by a far more complex representation. The king of this modern representation is the neuron (a particular kind of brain cell), though not as an “individual” but as the basic element of groups, networks, “cliques.” The term “neural networks,” used just as readily for your/our heads as for organizational principles of electronic computers and/or electronic networks, comes from the foundational base of most cognitive sciences. And it is this term, with its dual usage (in living beings and in modern machines), that is one of the many expressions of the modern ideo-politico-technical convergences between what is still called “life” (including human life) and the new techno-mechanical substrate of capitalism.
Despite the gloom—or even the anxiety—that might be triggered by realizing that (also) the cognitive sciences have strategic importance in the evolution of the new capitalist paradigm, it is worth the effort, as “unskilled” people, to approach this field too, one that evolves somewhat invisibly inside techno-scientific institutes around the planet, inside research projects and experiments. It concerns us, and it concerns us far more directly than we think (or would wish). Rather than “cursing” the techno-scientists in the name of an imaginary primitivism, we are obliged to know—to the greatest extent allowed by our non-specialized study amid the ranks of the hyper-specialized idiots (to recall the Situationists).
Here, for example, is a small triumphal advocacy piece in one of the branches of the cognitive sciences, the neurosciences, published in a specialized American journal seven years ago:
…
In recent years, neuroscience has concluded that every mental state corresponds to a precise pattern of neuronal action that can be studied with accuracy. We now know enough—and are expected to learn more in the near future—about the intracerebral route of sensory, motor and cognitive information, as well as about the correlation between experiences and neuronal events. Precise and sophisticated scanners now provide striking images (processed in false color) of the brain regions activated during the performance of any task, mental or otherwise.
The information from brain images taken while a person sees, hears, moves, ejaculates, gets angry or prays has obvious scientific value.
…
The “obvious scientific value” is of course not obvious at all, unless one is talking about techno-scientific careers, prestige and money. However, as we shall see both in what follows and in the future, there is a range of pursued techno-political goals that indeed “have value,” and indeed great value—provided one finds out for whom.
you are your neurons
Below follow long excerpts from a techno-scientific report written from within the cognitive sciences, concerning memory1 – the emphases and comments are ours. Investigating the “mechanisms” through which human memory is constituted and operates is a core area of the techno-scientification of consciousness. The reason is easy to understand: without memory, no enduring relation with any environment (with the World) can be woven.
The spirit of this report (as well as the spirit of the neurosciences) is not (yet) to explain what exactly (animal, biotic) memory is. But to represent it as accurately as possible, so that in a next phase (which is much closer than one might think) techno-scientists can intervene in it “correctively,” either neurosurgically, chemically, or magnetically. Consequently, the ideological model is simple and old: correlating causes with specific outcomes. What strongly differentiates previous practices of surveillance and “correction” of brain functions, such as lobotomy or electroshock, from those already outlined on the horizon, is the detail of the recordings and the analogies with contemporary information machines.
By way of introduction to this report, the editor of the following article points out the main ideas below. We republish them because they are useful.
KEY CONCEPTS
To represent and form memories of an individual’s experiences, the brain relies on the coordinated activity of large populations of neurons.
It has been shown that in the hippocampus of the mouse [an area of great importance for memory formation], subsets of such populations – which we call “brain cliques” – respond to different aspects of the same event. Some cliques encode abstract, general information about the incident; others indicate its more specific characteristics.
When the brain converts sets of electrical impulses into perception, knowledge, and behavior, it may be applying the same hierarchical organization used during memory formation. If this is true, then work in the field of memory brings researchers closer to uncovering the universal neural code—the rules the brain uses to recognize and give meaning to bodily experiences.
The author and his collaborators have converted recordings of neural-click activity into binary code. Such digitization techniques of brain signals could form the foundation for compiling a mind codebook—a tool for cataloguing thoughts and experiences and for comparing them across different individuals or even across different species.
It may be unpleasant, but that’s (according to the tech pundits) how it works. While you are reading these lines, electrical impulses in your head create “material” that can be recorded electromagnetically. If these “imprints” are archived and classified, a “code dictionary” of your mind can be built, in case you lose it (your mind…).
On the other hand, since it is not easy to grasp the great benefit such a list will bring to one’s life, let us place ourselves in the hands of those who know what they are doing:
The code of memory
Anyone who has ever experienced an earthquake retains vivid memories of the event: the ground shakes, trembles, quivers, ripples; cupboards open and close, books, utensils and small objects tumble off shelves; the air fills with noise and the sound of breaking glass. We remember such episodes with remarkable clarity even years later, because our brain evolved to do exactly that: to extract information from significant events so that this stored knowledge can be used in the future to guide our reactions to similar situations. This ability to learn from past experiences gives all animals the capacity to adapt to a complex and constantly changing world.
For several decades, neuroscientists have been trying to unravel how the brain forms memories. Recently, my collaborators and I combined the design of original experiments with new techniques for simultaneously recording the activity of hundreds of neurons in awake mice, as well as the application of powerful mathematical analysis methods.
What we ultimately discovered, we believe, constitutes the basic mechanism the brain uses to extract vital information from the organism’s experiences and convert it into memories. Our research findings add to a growing body of evidence showing that the linear flow of signals from one neuron to the next is not a sufficient explanatory model for how the brain represents perceptual data and memories.
Moreover, our studies show that the neuronal populations involved in memory encoding extract from our daily experiences those abstract concepts that allow us to transform these experiences into knowledge and ideas. Our findings bring biologists closer to deciphering the universal neuronal code: that is, the rules the brain follows to convert sets of electrical impulses into perception, memory, knowledge, and, ultimately, behavior. Understanding it could give researchers the ability to build fully compatible brain-machine interfaces, design an entirely new generation of intelligent computers and robots, and perhaps even compile a lexicon of the mind that would enable us to decode what someone remembers or thinks by monitoring their neuronal activity.
It should come as no surprise that what this particular researcher and his team (and many others, for that matter) are ultimately after is the improvement of “brain–machine” interfaces, the design of even more intelligent computers and robots, and the espionage of other people’s thoughts. Within the brutality of this admission lies that peculiar kind of “honesty” of experts who know they face no serious (obviously social) opposition that would “cut off” funding for their research—or, at the very least, raise serious doubts about their aims.
We must, then, note here this relentless inclination toward the absolute osmosis of human and machine, an inclination that cannot be explained “ontologically” but only institutionally and organizationally. As Kuhn2, Feyerabend3 and several others have shown convincingly for years, no scientific research program (and, even more categorically: no research program that becomes known, even within the “communities” of the respective specialists) is carried out in a vacuum, on the basis of (supposed) “human curiosity,” on the basis of the (well-meaning) “restlessness of the human spirit.” Like everything else within the ideo-politico-technical ecosystem of universities and research institutes, the quest for the “truth of the brain” is funded by… (whatever you like). And the funders have sometimes specific, sometimes unsettled interests, yet always within definite orientations.
Consequently, when some people search for and claim to find the rules the brain follows to turn sets of electrical impulses into perception, memory, knowledge and, ultimately, behaviour, one must add, beyond the research goal itself (so that we understand the kind of results it is supposed to deliver…), this permanently invisible datum: the possibility that “independent philosophers” might wonder about one issue or another has long vanished. Instead, there are more or less daring, more or less conscious contractors of segments of the quest for a specific holy grail: the universal mechanisation and/or the universal mechanical “reading” and “correction” of human behaviours. Those who “deviate”, but also the “normal” ones.
the mouse that is afraid
Our team’s research on the brain’s code grew out of earlier work that focused on the molecular basis of learning and memory. In the fall of 1999 we created a genetically engineered strain of mice with enhanced memory ability. This “smart” mouse—nicknamed “Doogie”—learns faster and retains memories longer than wild-type mice. Our work generated great interest and much discussion. Yet the findings left me with the question, “What exactly is memory?”
Scientists already knew that the hippocampus—a brain region—was essential for turning perceptual experiences into long-term memories. We also knew which molecules play a decisive role in this process, such as the NMDA receptor, which we modified to create Dougie. Yet no one knew exactly how neuronal activation mapped onto memory processes.
…
To learn more about the neural code of memory, we first had to design better tools for monitoring brain activity… So Longnian Lin, then a postdoctoral researcher in my lab, and I developed a recording device that let us track the activity of a far larger number of individual neurons in awake, freely moving mice.Our next step was to design experiments that took advantage of something the brain seems to do very well: retain memories of dramatic events that can deeply affect a person’s life. The experience of the September 11 terrorist attacks, surviving an earthquake, or even a plunge from a height of thirteen stories at the Disney-MGM Studios’ “Tower of Terror” are experiences that are hard to forget. We therefore created tasks that mimicked such emotionally charged episodes. Experiences like these would likely create strong and long-lasting memory traces. According to our reasoning, the encoding of such resilient memories would involve a larger number of hippocampal neurons, which would increase the chances of finding cells activated by the specific experience and collecting sufficient data to reveal any patterns and organizational principles underlying the process.
Among the events we selected were a laboratory version of an earthquake (produced by shaking the small cage in which the mouse was located), a sudden gust of wind across the animal’s back (intended to mimic an aerial attack by an owl), and a brief free-fall inside a tiny “elevator” (which, in the early stages of these experiments, happened to be a cookie container we had in the lab). Each animal underwent seven episodes of every event, separated by rest intervals of several hours. During the episodes—and during the rest periods—we recorded the activity of up to 260 cells in the hippocampal CA1 region, an area critical for memory formation in both animals and humans.
…
We discovered that every particular episode is always represented by a set of neuronal cliques [Eds. note: a “neuronal clique” is a group of neurons that respond in synchrony] that encode different features of the event, from the general to the specific. Specifically, an earthquake episode activates:
– a general-startle clique (a group of neurons that responds to all three startling stimuli);
– a second clique that responds only to events that disturb the animal’s balance (i.e., during the earthquake and the elevator drop);
– a third clique activated exclusively by the vibration;
– and a fourth clique that signals where the event took place (we placed the animal in one of two different cages before each earthquake).Thus, the information about these episodes is represented by assemblies of neuronal spikes that are hierarchically organized in an invariant way (from general to specific). We perceive this hierarchical arrangement as a feature-encoding pyramid, whose base encodes a general feature (such as “sudden event”) while its apex encodes a more specific piece of information (such as “vibration” or “vibration inside the black cage”).
…
Let us describe in other words this “experimental / research” setup, which aims to contribute to answering the question “what is memory” – even if only for the tinkered mice.
In principle, an advanced imaging device is required, which we can imagine as a magnetic brain scanner that will allow (based on some technical specifications that are irrelevant for our purposes here) the “representation” of the activity of neuronal cells, with some kind of “energy radar”. We assume that there are indeed variations in these representations, which will be attributed to variations in neuronal activities that are considered the (stereotyped) outcomes of some very specific artificially induced “stimuli”.
Next, they need intense stimuli of this kind, which in this particular case can be described in two words: fear provocation. The researchers want to “stress” the laboratory animals by scaring them in three different ways, assuming that different fear “stimuli” will activate different brain areas; different “neuronal clicks”.
Up to this point there doesn’t seem to be much distance from Pavlovian “science” in the study of “conditioned reflexes.” There is a basic initial assumption that the laboratory animal (its brain) will react in such a way as to justify a linear correlation: cause – effect.
The differentiation from the Pavlovian “function” lies in the details, which in turn are products of the technical surveillance of the “results,” that is, of the capabilities of the recording machines (the MRI scanner we assumed earlier). While “external observation” would yield the data earthquake: fear, owl: fear, fall: fear (i.e., useless conclusions, since the provocation of fear is presupposed anyway), the new machine permits the “cerebral analysis” (of the representation) of each fear separately. Thus, earthquake: fear “activates” the “cerebral cliques” A, B, and D (the names are arbitrary, of course, but that’s how the techno-scientist works); owl: fear “activates” the “cerebral cliques” A, C, D, and E; and fall: fear “activates” the “cerebral cliques” A, B, D, and E.
What has the neuroscientist achieved so far? Nothing spectacular, apart from having classified the images produced by his machine and calling them encodings of the brain (of the lab animal). In other words, he has meticulously (and in a “scientifically” acceptable way) transformed the technical representations of his “target” (the mouse brain) into properties of that target. This shift, and the conflation of “representation” and “represented,” would have sparked fierce arguments in other “scientific disciplines,” perhaps in other times—but not here and now, when neuroscientists are hunting the most secret of life’s secrets. Today, for example, one is allowed to speak of “mnemonic representations in the brain,” whereas the only thing one would actually be entitled to talk about is “technical representations produced by the machine.”
Consequently, the neuroscientist can go on “analyzing memory,” getting ever closer to where he was destined (historically, genealogically, by the division of labour, knowledge, money and prestige) to arrive:
Our findings seem to lead to certain conclusions about the organizational principles governing memory encoding. First, we believe that neuronal cliques act as functional encoding units for memory formation and are quite consistent in representing information, even if the activity of some individual neurons diverges somewhat from that of the group.4 … The brain relies on memory-encoding cliques to record and retrieve different features of the same event, essentially arranging the relevant information in a pyramid of hierarchically structured levels, from the most general, abstract characteristics to the most specific aspects of the event.
… This combinatorial, hierarchical approach to memory formation allows the brain to generate an almost unlimited number of unique (at the network level) patterns to represent the infinite variety of experiences an organism can have during its lifetime – just as the four “letters”, or nucleotides, that make up the DNA molecule can be arranged in a virtually boundless number of combinations to produce the seemingly infinite diversity of organisms living on our planet. And because the mnemonic code is categorical and hierarchical, representing new experiences may simply entail replacing the specific cliques at the apex of the memory pyramid with others, so as to indicate, for instance, that the dog barking behind the fence this time is a Pomeranian and not a Husky, or that the earthquake occurred in California and not in Indonesia.
…
Consider, for example, the concept “bed.” Any person, entering any room in any hotel in the world, will immediately recognize the bed, even if they have never before seen that particular piece of furniture. What enables us to retain not only the image of a specific bed, but also the general knowledge of “what a bed is,” is the architecture of our neuronal populations for memory encoding.
The neuroscientist seems to have discovered (or, more accurately, invented) the following “truth” about what memory ultimately is. It is the “storage” of life’s stimuli within neuronal cliques (or through them…), arranged in a hierarchical “coding” system, where at the base of this hierarchy are the “general memories” (the general memory bed, the general memory table) and at the top the “specific memories” (my childhood bed, my grandmother’s old table)… What happens (from the perspective of memory, that is, the behavior of neuronal cliques) when someone lies down to sleep on a table or eats in bed, has not yet been investigated.
duality; why not?
The particular contribution of Tsien’s experiments and his team lies, on the one hand, in the idea of this arrangement, and on the other, in the mapping of neural cliques that are responsible for each individual “encoding”—such groups of neurons were termed “encoding units.”
Therefore, the conclusion of this particular report is reasonable:
Our research work with the mice also yielded a way to compare patterns between different brains—as far as passing information from the brain to the computer. Using a mathematical method called matrix inversion, we succeeded in translating the activity of neuronal-clique assemblies into strings of binary code, with 1 representing the active and 0 the inactive state for each encoding unit of a given assembly. For example, the memory of the earthquake could be represented by the string “11001”, where
– the first 1 represents the activation of the clique for the general startle state;
– the second 1 denotes the activation of the clique that responds to balance disturbance;
– the first 0 indicates absence of activation of the clique for wind gusts;
– the second 0 corresponds to absence of activity of the elevator-drop clique;
– and the last 1 shows activation of the earthquake clique.We applied a similar binary encoding to the collective neural activity we had recorded from four different mice, and in this way we succeeded in predicting with 99 % accuracy which event the animals had experienced. In other words, by reading the binary code we were able to read and mathematically compare the animals’ minds.
A binary code of the brain like this could perhaps offer a unifying framework for studying cognitive functions even in different animal species, and it could also facilitate the design of devices more compatible for real-time brain-machine communication. We, for example, have built a system that converts the neural activity of a mouse experiencing an earthquake into binary code, which then commands the opening of an escape door, allowing the animal to exit the vibrating cage.
…
The practical application the neuroscience team arrived at while studying “what memory is” is certainly impressive. What is particularly interesting about it is what is not said. Namely, that the “escape hatch” does not open through any deliberate effort (by the lab animal) but simply and solely because it is afraid. In its generality, this scheme means that the machine does not need thought (to function); emotional / neuronal “flashes” suffice.
As for the binary “translation” of the successive subtractions and “codifications” of the scientist (and not of the animal), of his professional / scientific brain and not of the living brain in general, there are many interesting, political and (class) competitive aspects—dare we say. For the moment, and as a down payment for future references to the cyborg, we must note the following. There is a permanent constant in the evolution of new technologies, which we shall call languagification. Whether it concerns the binary 01 system, the pairwise quaternary system attributed to the bases of DNA, or a decimal system that may prove more useful (as it is more detailed) in quantum informatics, whatever is to be transferred to the new machines must first become language.
“Language,” in this context, means the encoding onto a relatively small number of symbols (they can be numbers, they can be letters, they can be newly-coined mathematical symbols) whose combinations can produce a large variety of different final formations. Linguistification is the governing principle of all (new) techno-scientific truths, the necessary and sufficient complement of measurability, and we have well-founded grounds to believe that this constitutes a specific political innovation of control.
As an indicative preview of a future critical analysis, here is a short excerpt from a text by other neuroscientists5 (emphasis mine):
…
The work of our research team on deciphering the neural code already allows us to apply this rough understanding of the language in practice, reading the firing patterns of neurons from the motor cortex of monkeys or using computer algorithms to translate this information, in real time, into instructions for moving an artificial limb… Our hope is that one day soon we will also possess sufficient knowledge of the syntax of the brain’s language, so that we can speak it, which would allow us, for example, to build a human prosthetic limb equipped with sensors that send tactile information back to the user’s somatosensory cortex.
…
Some form of cyborg, in any case…
Because languagification is the unspoken yet powerful prerequisite of contemporary techno-scientific analysis in almost every field, as happens everywhere, so too in this particular research ploy of Tsien’s neuroscientists the binary “translation” of “brain functions” allegedly appears at the end of the research process, whereas it was there from the very beginning. Whoever looks for binary code will find binary code.
Ziggy Stardust
cyborg #03 – 06/2015
- From the Greek edition of Scientific American magazine, October 2007. Title: the code of memory. Author Joe Z. Tsien, professor of pharmacology and biomedical sciences at the “Center for Systems Neurobiology” of Boston University. He has founded the “Institute of Brain Functional Genomics” at East China Normal University in Shanghai. ↩︎
- Thomas Samuel Kuhn, physicist, historian, and philosopher of science. In his book The Structure of Scientific Revolutions (1962 – also available in Greek), he demonstrated the general reasons why scientific research moves in one direction or another, demystifying the (supposed) epic of the “scientific spirit” in its quest for the discovery of Truth. The concept of Paradigm Shift (in the evolution of scientific doctrines) is his own idea. ↩︎
- Paul Karl Feyerabend, epistemologist. In his book Against Method: An Anarchist Theory of Knowledge (1975 – also published in Greek) he demystified, in a well-documented and entertaining way, the authenticity of various scientific truths. ↩︎
- This “detail”—that ultimately within the “neuronal cliques” there are some differentiations in “reactions”—shows what is common among techno-scientists: whatever does not fit the desired conclusion is set aside… ↩︎
- Looking for the Neural Code, Miguel Nicolelis and Sidarta Ribeiro. ↩︎