Neurocentrism, Hypes and the Illusion of Knowledge

Why neuroscience failed to explain consciousness

Erik Hoel wrote an interesting piece on why neuroscience is ‘pre-paradigmatic.’ I’m glad to see that also someone else brought up this issue that I addressed a couple of years ago in the chapter of my book “Spirit calls Nature.” It was entitled “Neurocentrism, Hypes and the Illusion of Knowledge.” You might like to read first his post and then proceed to this one and make a comparison. I believe there are sufficient commonalities to state with confidence that we are seeing a real issue not only with neuroscience but with science in general. And, if you are still with me once gone through all that, you might understand better what motivated me to publish this article that made a summary and critical review of the mind-brain identity theory.

---

To sum up, we saw how science searched for the seat of consciousness or the mind and found no reliable locus, brain region, or neural correlate that could neither shed light on its whereabouts nor tell us how it comes into being. We saw how surprisingly plastic the brain must be to recover its cognitive and mental functions despite extreme damage, surgical intervention, or deformation and wondered if the relatively intact mind and consciousness of these altered brains could be explained away only by the magic wand of neuronal plasticity.

Moreover, paradoxically, we found that plants — organisms without neurons — show behaviors that display an elementary associative intelligence, a form of ‘proto-awareness.’ Plants have no neurons. Yet, they are capable of actions that we previously thought were possible only in an organism with a central nervous system. We went further and discovered that even single-cells must possess some basal cognition. This is a property that non-neuronal cells, such as bacteria, reinforce by creating collaborative networks. All this looks as if the mind, at least an elementary form of mind, is already present in a ‘pre-neuronal’ form. The bottom line is that the deeper we dig into the microscopic biological realm, the more we find mental signatures still present and active.

At this point, the question I would like to ponder first is: Given the above findings and new insights, is the dominant physicalist science in its present format, which conceives of mind and consciousness only as an epiphenomenon of the brain — that is, a science based on an exclusively neuronal paradigm — still tenable? Is the ‘you are your brain’ dogma still a ‘fact’ we shouldn’t doubt?

I want to maintain a conservative attitude and refrain from replying with a too-self-assured negative answer. After all, the fact remains that any higher-level cognitive function is always associated with neural activity.

However, I believe that it is time to question the exclusively reductionist approach that cognitive neuroscience has taken, especially in the last decades. There is no doubt that neuroscience, especially in the last three or four decades, has made enormous progress. From medicine, neurology, neuropsychology, and cognitive sciences to computational modeling or AI research, these have had a — and may have an even more — dramatic impact on our knowledge in the coming decades. Further technological advances in neuroimaging will undoubtedly continue to shed light on the brain’s activity. The importance and usefulness of these neuroscientific advancements, especially as diagnostic tools, is undeniable. Due to this success, neurosciences experienced a revival that was to be expected and is partially justifiable. I made ample reference to neuroscientific findings.^[1] There is a takeaway from the advances in this important field, which tells us something about how we perceive the world or, more precisely, how we construct the world.

However, rather than resorting to an ever-increasing magnification of the details and failing to notice what the whole is about, we must finally recognize its limit. One point in question is what these empirical data are supposed to explain, suggest, or imply from an ontological standpoint. The uncritical acceptance of the (more or less unaware and metaphysical) neuronal premise, particularly that of the relationship between neural activity and mind or consciousness being an undeniably given datum, should be questioned.

This ‘neurocentric’ attitude is not only ideological; it also has its roots in the political and financial structure of the academic and research system. Nowadays, neuroscientific achievements must be overemphasized because a research project on something ‘neuronal’ has more chances to attain funding than projects that do not. These ample financial concessions are, however, not due to past scientific findings. The willingness of governments, foundations, and society to finance costly brain research projects based on a purely neuro-reductionist paradigm is not triggered by previous results which, in most cases, did not go much further than correlational evidence between what happens in the brain and what one subjectively reports feeling, seeing, thinking, and perceiving. The neurosciences did not transform traditional psychology and therapeutic approaches as was expected and didn’t disclose the mysteries of human consciousness that could close the explanatory gap. In many respects, especially when it comes to potential applications for neurodegenerative diseases, such as Alzheimer’s or Parkinson’s, neuroscience did not meet the expectations, either. Technologies such as MRI, fMRI, MEG, EEG, PET, and other non-invasive scanning techniques are valuable diagnostic tools and have given us important insights into the anatomy, physiology, and function of the brain and other organs, allowing for the study of memory, learning, drug effects, and diseases such as schizophrenia, autism, or clinical depression, or can be vital for preparing surgical interventions. However, despite 40 years of hopes and promises, these did not lead to tangible progress in the treatment of mental disorders or other practical applications. The failure of a European mammoth project such as the Human Brain Project only further testified to this. I’m waiting to see how a similar US-based ‘BRAIN initiative’ that foresees “fundamental new discoveries about the brain” by 2025, or the Chinese ‘China Brain Project’ and the Japanese ‘Brain/MINDS’ mapping program, are going to develop. They may gain some insights into the information processing of the human brain, but any claim that they will allow us to manufacture drugs to treat neurodegenerative diseases, let alone contribute to explaining and reproducing conscious and mental functions in any form, is unfounded.

Nowadays, we hear people talking about artificial neural networks implemented on supercomputers that simulate millions of neurons. What we are rarely told is that an artificial neuron is an extremely simplified version of a real biological neuron. The simulated units are sum-up-inputs and spit-out-spike boxes. To simulate a single neuron, one would need a far more complex functional object. Strictly speaking, a single neuron must be simulated by a neural network itself. All the billions of neurons in our brains are not just neural networks. The brain is a neural network of neural networks (for a more in-depth analysis of this, see link, link).

A pyramidal cell (you have 80 billion of these in your brain) and their electrical neural spikes. The black blob in the middle is the neuron’s body; the rest of the wires are its dendrites. Credit: Alain Dexteshe

Making things even worse is the fact that the connection between neurons, the dendritic arms of some human neurons, can perform logic operations that require whole neural networks. This means that a realistic simulation of a biological neural network should be simulated by tens of billions of neural network of neural networks, each connected to tens of thousands of other neural networks. Even if the most optimistic predictions in the progress of supercomputer technology for the next hundred years came true, we would come nowhere near to such computing power.

The effort and enthusiastic drive that stands behind this ‘neuromania’ is neither scientific nor based on research findings or justified by its potential pragmatic applications. It is based mostly on a hegemonic ideological and cultural belief system anchored in the physicalist, reductionist, and material monist ideology, also not rarely backed by financial interests and policies that are prone to the same collective suggestion. However, that does not prevent society from continuing to support and fund the same rational and scientific reductionist and neuro-centric paradigm. For example, according to the American psychiatrist Allen Frances, former chairman of the Department of Psychiatry at Duke University School of Medicine in North Carolina and member of the task force that produced the fourth edition of the Diagnostic and Statistical Manual of Mental Disorders, or DSM-IV, since 1990, “the National Institute of Mental Health (NIMH) suddenly and radically switched course, embarking on what it tellingly named the ‘Decade of the Brain.’ Ever since, the NIMH has increasingly narrowed its focus almost exclusively to brain biology — leaving out everything else that makes us human, both in sickness and in health. Having largely lost interest in the plight of real people, the NIMH could now more accurately be renamed the ‘National Institute of Brain Research.’” While, as of 2022, a review showed that “…despite three decades of intense neuroimaging research, we still lack a neurobiological account for ANY psychiatric condition”(emphasis mine).

This state of affairs is also reflected in popular media and the collective culture. If something is explained in neuronal terms or brain areas subjected to some bio-physical reaction such as the cause of an ailment or disorder, or supposedly explaining anything about our human nature, it is much more likely to be accepted by the popular audience as final scientific proof. The prefix ‘neuro’ is second only to that of ‘quantum.’ Everything that is labeled with some ‘neuro’-specification is instinctively considered to be a source of ultimate scientific truth that explains everything, obviously from a purely reductionist and physicalist perspective. While in physics, the media outlets’ hyperbolic announcements contribute to a vast ‘quantum woo’ movement, we fail to notice how much ‘neuro woo’ is pervading it as well. A title that slaps the prefix ‘neuro,’ ‘genetic,’ or ‘molecular’ onto a finding or account of things creates a sense of increased explanatory power and ‘truth-value,’ which almost subconsciously lures us into the illusion which transforms a correlation into a causal claim.

Moreover, it has been shown that overestimates of effect size and low reproducibility of results in neuroscience are much more common than assumed. Because of a ‘publish or perish’ policy which is nowadays the almost exclusive directive and evaluation parameter for academic career success, researchers must publish as soon as possible in order to succeed, which leads to a hasty overestimation of their own assumptions and claims with statistically non-significant results reflecting effects that are only apparently true. This is not specific to the neurosciences but generally to medical research. Linguistic spin — that is, reporting strategies aimed at convincing the reader that the beneficial effects and efficacy of the experimental interventions are greater than those shown by the results — was found in more than half of the abstracts of medical peer-reviewed papers published in scientific journals. On top of that, overstatements or even misinformation proliferate on internet-based health news, using causal language and suggesting a strength of inference in academic and media articles where there is none. It turns out that 58% of media articles had inaccurately reported the results of the academic studies.

However impressive and thrilling the techniques of neuroimaging are, when it comes to the question of the mind-body problem and the nature and origin of consciousness, especially to the question of why neural firings become phenomenal experiences, no progress at all has been made. The findings of the last decades didn’t yield any understanding, and the brute force identity between mind and matter is not supported by facts but has always been posited as an a priori metaphysical premise.

The beautiful images of the brain tell us something about the neural activity but nothing about the mind in itself. The fMRI images have now reached a resolution of about a few millimeters, which is indeed an impressive technological achievement, but this pixelated volume (a so-called ‘voxel’ — that is, a cubic volume of a few mm³) still contains millions of neurons and billions of synapses and doesn’t tell us how these neurons and synapses are doing their job. It is like reducing the map of a city with a million inhabitants into a pixel and trying to deduce from its flickering the social behavior of its population. These increasingly precise mappings do not directly measure the neural electric activity. Rather, they are pictures of the fluctuations in the metabolic changes of the neurons involved in a task. MRI or fMRI imaging determined real progress in enabling more accurate neurosurgery, the mapping of internal injuries, or the detection of tumors. These images, however, did not provide any insight into what mental functions do, how they do it, or what they are. For a specific motor-sensory action or cognitive task, it is great to see how a specific brain area ‘lights up.’ However, this, per se, does not tell us much about how these tasks are achieved.

fMRI brain scans are also vitiated by a low signal-to-noise ratio, meaning that they display a signal affected by high noise levels and only 1% change from the baseline. Picking the statistical threshold, which determines what counts as a significant signal, can be a matter of preference. Each scan produces enormous amounts of data, which processing aimed at sorting out signal from noise is left to the discretion of the individual researcher who is already stressed by the pressure to publish as much as possible, and potentially leading to misapplied statistics. In fact, it was shown that several studies misidentified random noise as signals where there isn’t really one, correlating brain activation with emotion, personality, and social cognition measures, portraying implausibly high correlations. Even in a dead salmon’s brain, neuroimaging voxels can exhibit activity if not corrected for multiple comparisons. But repeating MRI experiments is expensive and time-intensive. This is a reason more that forces researchers to play around with analysis parameters until they see a signal where there is none.

Recent findings have shown that things might be even worse than that.

When 70 independent teams were asked to analyze identical brain images, no two teams chose the same approach, and their conclusions were highly variable. Human error and subjective methodological decisions lead to serious issues with scientific reproducibility.

Moreover, the supposed correlation between brain activity and specific tasks identified by fMRI brain scan imaging turns out to be much less reliable than previously thought. A US research group from the Department of Psychology and Neuroscience at Duke University has demonstrated that common task-fMRI measures are unlikely to indicate the real locus of brain activity associated with some specific cognitive task. It turns out that the regional brain activity for a given task, such as reading a text, observing an object, or performing some specific motoric task, not only changes from subject to subject but changes over time in the same subject performing it. If one performs an fMRI of the brain of someone who is, say, reading aloud a meaningful series of sentences, the fMRI images taken only a couple of hours apart while the person is performing exactly the same task will look completely different from each other.

Moreover, it turns out that neuroimaging studies are massively affected by replication failures and are hyped by inflating small effects with dubious statistical significance. Reproducibility requires samples from thousands of individuals. Something that is usually much too expensive to do. In other words, what you read about brain-scan research (beyond its diagnosing purposes) is much too often nonsense. When all things are considered, the frustrating verdict is that “We can count on less than a hand the number of these studies that have held up under scrutiny and are really driving treatment,” as postdoctoral fellow at Harvard Medical B. Tervo-Clemmens said.

The bottom line is that brain scanning images perform very poorly in terms of identifying the centers of brain activity responsible for a particular cognitive process. Despite all the extraordinary progress of neuroscience, there remains an enormous gap dividing us from connecting a clump of neurons acting like a flaming marvel of electrical discharges to the comprehension of how and why they are supposed to forge a mental and emotional life inside a subjective phenomenal experience.

At any rate, even if we could find a way out from these technical oversights, cognitive neuroscience is mainly about the relationship between our higher cognitive functions (such as attention, memory, response inhibition, etc.) and the structural characterization, localization, activation, and temporal changes of specific brain areas seemingly responsible for it. It tells us how some regions of the circuitry underlying the performance of some executive functions can be described by neuronal short- or long-range connectivity and changes, local patterns of activation, or changes in anatomical structures in the brain, eventually including other genetic, physical, physiological, and developmental factors. Nonetheless, all these fantastically complex phenomena still don’t tell us much about how and what the brain is really doing and, ultimately, what or who we are.

In hindsight, that shouldn’t come as a surprise. It is like knowing everything about the electronics of a computer and being able to map all the internal electrical current flows in the CPU, the memory, and inside all the circuitry in its chips down to the microscopic level — but that will not tell you much about what the computer is doing. Trying to reconstruct a complicated AI compiled software running on a computer only by observing where the current flows in its integrated circuits and seeing which group of transistors and logic gates flicker up is an almost impossible task. While, recreating the exact physical structure of a computer would not lead us to a machine making anything meaningful or useful. The software–that is, a running code written by an intelligent external agent– is needed. Here also a computer is only an instrument, a means of expression for a cognitive entity, not its origin or source. In fact, studying a microprocessor with the same criteria employed by modern neuroscience, trying to reverse-engineer its functions, would fail: We would not be able to explain how it works, let alone reveal anything about the running code, and which is the real ‘agent’ that causes the behavior of the machine.

The lesson is that knowing everything about the hardware won’t tell you much about the software running on it. Whatever functionalists might tell us, even if we were able to trace back our ‘mental software’ from the brain’s extremely complex activity, that would furnish us with only a functional description of what is going on. Why should a sequence of internal physical states, however complicated, complex, intricate, convoluted, or multilayer-networked, give rise to a subjective experience? If you love the technical jargon so in fashion nowadays, which has only a seemingly explanatory power and adds nothing to a real understanding, choose labels such as ‘recursive feedback loops,’ ‘self-modeling,’ ‘resonances,’ ‘information integration,’ ‘synchrony,’ ‘phase-locking,’ ‘re-entrant circuits,’ ‘self-referential processes’, or whatever. That didn’t lead to any new insight regarding the question of how phenomenal consciousness and its qualia come into being, either. Anything that loops could be a vessel, but not a producer. The heart ‘loops’ blood throughout the body, but doesn’t produce it. The reason for the blood to go back to the heart has nothing to do with the ‘production’ of blood. There is no rationale in believing things to be different with the ‘production’ of consciousness.

With the gigantic progress of microscopy, we could build microscopes that can almost see molecules reacting inside a living cell. We could look into the new world of molecular biology and make astonishing discoveries. And yet, by going deeper into the elementary structures of our biological existence, we found how things became increasingly complex. It is a complexity that grew exponentially with the degree of magnification and of which we frequently have no understanding, let alone control. The reductionist assumption that everything can be explained by a bottom-up approach that dissects and maps everything in the minutest details has delivered a false sense of ‘knowledge’ which is not, in fact, knowledge but, rather, a superficial representation at the level of raw data–only the tip of an iceberg that we misinterpret for the whole. Knowing everything down to the neuronal, cellular, or even molecular level of a brain function or the molecular description of the physiology of a cell no doubt tells us something. But it is only a partial understanding which becomes an illusionary knowledge when it is mistaken for the whole of knowledge. It delivers a false sense of assuredness based on the assumption that once a territory is mapped, it is also known. The idea that knowing the map is knowledge is an illusion hard to eradicate. It is like analyzing every molecule and elementary particle of the printer’s ink on a piece of paper without realizing it to be Shakespeare’s sonnets.

That’s why the assumption that dominated especially the 1980-90s, according to which an increasing knowledge of the neural correlates of phenomenal consciousness would have explained it, elicited a sort of ‘gold rush’ that revealed itself to be delusional. In this sense, neuroscience did not progress much further than where Descartes was, other than telling us that the mind is not in the pineal gland. The rebuttal of dualism, which replaces it with a mind-brain monism, has no evidence whatsoever supporting it. The fact that world-known high-ranking hardcore physicalists claim otherwise, stating this to be something self-evident that no longer needs any explanation, doesn’t make it true. The mind-body problem and the hard problem of consciousness remain a controversial issue more than ever. On the contrary, instead of leading us nearer to the solution, modern neuroscience only deepened the mystery.

This is also the reason why some scientists resort to extreme attempts to save, in my opinion, an untenable position proclaiming the inexistence of phenomenal experience in the first place as a meaningless concept and wrong construct and going so far as to deny their own experience and phenomenal consciousness as a ‘category mistake.’ There is a modern movement in the philosophy of mind that clings, at any cost, to the functionalist or physicalist paradigm, the so-called ‘eliminative materialism’ or ‘illusionism’ or just ‘eliminativism.’^[2] According to this school of thought, consciousness is ‘eliminated’ from the outset as an illusionary false representation of the brain.

With all due respect, the author’s feeling is that these are only pseudo-philosophical reasonings that resort to obscure and nonsensical babble. That wouldn’t be worrisome if these theories were not taken very seriously by many academics of high-ranking universities and if they were not considered the state-of-the-art of the modern philosophy of mind. It is the easiest shortcut: If you can’t explain something that doesn’t fit into your ideological belief system, instead of reconsidering your metaphysical premises, you simply deny the existence of the thing that must be explained in the first place. This is what we see with climate change denialism and conspiracy theorists. Eliminative materialism is intellectually no better than a flat Earth theory of consciousness. It is the last desperate attempt of the new Ptolemaic who don’t know themselves and try to save appearances and a crumbling anthropomorphic paradigm at any cost. It is to expect that the attempt of physicalism to explain consciousness will go into the history of science like logical positivism in philosophy, or behaviorism in psychology, or like alchemy tried to transmute base metals into gold. To date, these are still the mainstream approaches that attract grant money in ways that the more serious theories cannot. One can only wonder how deep the contemporary philosophy of mind has fallen and it even prides itself for that.

That might, however, lead to an unexpected positive outcome. An increasing number of scientists are realizing how the attempt to answer the questions about mind and consciousness from an exclusive standpoint is leading to untenable, sometimes even ridiculous, lines of thought. Many are now becoming aware of how the vast amount of data and functional insights on brain activity, interpreted only through the purely reductionist and physicalist neuroscientific glasses, delivered a false feeling of certainty and confidence: an illusion of knowledge. Becoming aware of an illusion is the first step to getting rid of it.

---

[1] The author’s animal rights bent struggled with this. I would love to see less animal experimentation and more alternative forms of investigation.

[2] For reasons of personal mental hygiene, I won’t furnish a precise account and literature of eliminativism. The interested reader can search online names like Daniel Dennett, Paul and Patricia Churchland, Susan Blackmore. The following is just a personal summary inspired by an online debate over an article by Keith Frankish, and an article by Michael S. A. Graziano.