Johanne Marais’s Review of Chirimuuta, M., The Brain Abstracted

BOOK REVIEW

The Brain Abstracted: Simplification in the History and Philosophy of Neuroscience

by M. Chirimuuta

MIT Press
March 5, 2024, 376 pp.
Paperback ISBN: 9780262548045

Reviewed by Johanné Marais (University of the Witwatersrand)

In the exceedingly complex natural world, reduction and simplicity have been dominant scholastic approaches for scientific advancement—particularly among experimental neuroscientists. In The Brain Abstracted: Simplification in the History and Philosophy of Neuroscience, M. Chirimuuta describes the practice of biological reduction as the “separation of a part of the organism from the rest, performance of precise examinations and interventions feasible only under such conditions of isolation” (p. 15). In the introduction, Chirimuuta supposes that for seeking simplicity, one of two justifications surface: either the reasons are psychological, making experiential life more digestible, or universal natural laws are inherently simple. This book, concisely and confidently, argues that a sounder approach to accepting the brain as infinitely complex requires remaining skeptical about the accuracy of various neuroscientific models in communicating the workings of the brain. Chirimuuta does this by exemplifying the history of simplicity, challenging dogmatic theory, embodying complexity, and retaining neutrality.

Simplicity in its etymology means “one-foldedness” (p. 22)—a definition boldly applied by theologians, philosophers, and scientists who map simplicity as an extension of God, with the argument that a Creator would rely on minimal design effort (hence, simplicity) for maximal impact (or, “natural wonder”). This “one-foldedness” evokes the philosophical approach of scientific realism, which believes that scientific inquiry is a means to observe all structured, natural processes, and events such that entire natural knowledge systems are ultimately accessible through science. Realist stances informed historical theories of cognitive neuroscientific discovery; one early example is the reflex theory, which suggests a conditioned reflex—as a functional unit—is the basis of all behavior.

The continuity or circularity of reflexes themselves challenged the simplicity of such a theory, giving rise to the famous Hodgkin-Huxley model: framing a neuron as analogous to a circuit in the firing frequency and action potentials that traverse neuronal units. The computational theory of the brain slowly developed with support from models such as these to become the dominant theoretical framework of neuroscience by the 1970s, allowing the investigation of neuroscience in a modular, compartmentalized, and mechanistic way. Despite the pursuit of experiment-to-model approaches in neuroscientific research that strive to observe and model what is called the natural ideal pattern (derived from a series of observations, reductions, and computations) in neuronal circuitry, Chirimuuta reminds the reader that even computational theories retain human influence in what is looked for, what is modeled, and how it is represented.

Neuroscientific computational models evolved from single-unit neuronal doctrines to “big data, ethological turn” practices, predominantly thanks to instrumental advances making multiple-channel neuronal measurement possible (p. 139).[1] Neuroscientists developed their understanding that neurons may not be the smallest operating unit of a behavioral model, but that behavior relates to connectivity between multiple neurons. Chirimuuta begins with measurements of a single neuron’s voltage patterns in the 1970s; by fifty years later, 65,000-channel neuronal imaging was possible, demonstrating how scalable techniques of neuroscientific experimentation made certain theoretical frameworks more feasible than others by the volume and detail of data that could be captured.[2]

What Chirimuuta conveys in refined detail between mapping the theoretical movements just described is how reductive approaches to cognition inevitably fall short, and that emergent, ethological practice strengthens the representation of neural networks. In emergentism, one learns more about the brain through observing neuronal connectivity in a “wild-type” environment, which accommodates the natural behaviors of a subject such that observations from multiple constituent parts emerge, rather than being deduced from one component’s contributions as reductionism.

The next substantive landmark in neuroscientific research was the conception of neural representations. Again, the polarities of causality-based mechanistic philosophy contrasted with the “proximity principle” tell the reader that although neural representations may exist as a multidimensional construction of neuronal activity, brains are heterogenous; we either become too wrapped up in mechanics when operating distally, or too reductive of meaning when we operate proximally (p. 153). Here, Chirimuuta emphasizes again the power of embracing Docta Ignorantia—knowledge ignorance: “What I have not made, I do not understand” (p. 147)—in acknowledging that although there is observational value in the data of neural representations, those representations remain simply representative and inconclusive.[3]

This neutrality-preferring approach to neuroscientific progress leans into a near-intersecting philosophy of two established philosophies: what Chirimuuta calls “transcendental idealism” (an adaption of Kantian formal idealism) and “perspectival pluralism” (pp. 41, 42). In transcendental idealism, a scientist invariably contributes their viewpoints, bias, and experimental reasonings for inquiry to scientific “fact” that emerges, while in perspectival pluralism, scientists can hold observations from many viewpoints. To me, these philosophies of science are the ground for Chirimuuta’s assertiveness in the book, where she writes, “I wish to prompt scrutiny into our concept of science,” entreating philosophers and neuroscientists to maintain metaphysical neutrality in understanding the ever-changing brain (p. 209).

Chirimuuta does not advocate for the total eradication of abstraction (not in the Kantian nor Platonist stances); indeed, benefits of simplification include making the natural (and abstract) more comprehensible, manipulatable, and communicable. Computational- and bioengineering tools can facilitate understanding of complicated neural processes, but the brain cannot and should not be reduced to a machine. Chirimuuta’s comparison between artificial neural networks (ANNs) and biological neural networks (BNNs) reminds the reader that computation is capable of prediction, but not understanding; BNN learning does not distinguish training from performance, whereas ANNs have a finite training-capacity derived from their programming.

Functionalism within the metaphysics of mind forms the backbone of computationalism-centered theories of consciousness, such as those endorsed by David Chalmers and Susan Schneider.[4] Functionalism itself discounts the importance of the composition—the cellular constituents—of neurobiological systems in giving rise to mental states. The looming query of the phenomenology of consciousness is frequently posed as the greatest enigma of modern science. Chirimuuta’s response to unraveling mysteries of consciousness is to question defining artificial models as intelligence at all; she encourages a biological naturalism that acknowledges that the perfect replicas and recreations of neuronal cells are biological duplicates, and not computational models.

Chirimuuta’s refining of biological naturalism becomes what she describes as an “embodied, embedded, non-computational neuroscience” that sees consciousness intrinsically connected to the matter and makeup of the brain as part of the body and the body as part of the environment (p. 278). Chirimuuta also indirectly prompts the reader to reflect on whether knowing consciousness—how it arises and what it is—is even possible at all. She challenges the pattern of accepting science as true and correct; urges curiosity among philosophers, neuroscientists, and society; and stresses that substantial progress can come from remaining teachable and humble in the quest for neuroscientific understanding.

Notes

[1]. For single-unit neuronal doctrines, see Horace Barlow, “Single Units and Sensation: A Neuron Doctrine for Perceptual Psychology?” Perception 1 (1972): 371-94.

[2]. On single neuron’s voltage patterns in the 1970s, see for example David H. Hubel and Torsten N. Wiesel, “Receptive Fields and Functional Architecture of Monkey Striate Cortex,” Journal of Physiology 195 (1968): 215-44; for 65,000-channel neuronal imaging, see for example Kunal Sahasrabuddhe, Aamir A. Khan, Aditya P. Singh et al., “The Argo: A High Channel Count Recording System for Neural Recording in Vivo,” Journal of Neural Engineering 18 (2021):015002, doi: https://doi.org/10.1088/1741-2552/abd0ce.

[3]. Nicolas Cusanus, Of Learned Ignorance, trans. Germain Heron (London: Routledge & Kegan Paul, 1954).

[4]. For computational theories of consciousness, see, for example, David J. Chalmers, “A Computational Foundation for the Study of Cognition,” Journal of Cognitive Science 12 (2012): 323-57; David J. Chalmers, “Uploading: A Philosophical Analysis,” in Intelligence Unbound: The Future of Uploaded Machine Minds, ed. Russell Blackford and Damien Broderick (Chichester, UK: John Wiley & Sons, 2014), 102-18 ; and Susan Schneider, Artificial You (Princeton, NJ: Princeton University Press, 2019).

[This work from H-Net is reproduced here under a Creative Commons License]