Why doesn’t your brain get overwhelmed when you’re at a noisy party? How is it that so many competing signals resolve into one coherent world rather than a cacophony of competing interactions and interpretations?

Perhaps it’s because your brain isn’t one singular processor at all, but thousands of tiny “mini-brains” working in parallel, weaving their independent models into the unified experience of perception.

A Thousand Brains by Jeff Hawkins proposes a new theory of intelligence centered around the Thousand Brains Theory which posits that our neocortex contains thousands (approximately 150,000) cortical columns in humans. Each cortical column is independently building a complete model of the world using sensory-motor input and map-like reference frames. These columns then engage in a voting process that ultimately creates our unified perception of reality and sense of self.

I read the book A Thousand Brains several months ago, and it remains one of the most interesting and impactful books about the brain I’ve read in years.

In this book review, I want to focus on a short explanation of the book’s core thesis and explore two or three interconnected concepts:

1. Cortical Columns are the Fundamental “Microcircuit” of Higher Order Intelligence
2. Reference Frames as the Foundation of Knowledge (and mental movement as key to learning and understanding)

Below are my explorations of these key concepts, some of my takeaways as well as additional quotes and notes.

Thousand Brains Theory’s Central Thesis: Independent Mini-Brains Operating in Parallel

Our neocortex is the outermost layer of the brain, responsible for higher-level functions like sensory perception, cognition, learning and language.

The Thousand Brains Theory contrasts with the traditional hierarchical view of the neocortex and argues that intelligence isn’t a single entity but an emergent property of many parallel processes. Instead of one giant computer knowing and deciding, our brain’s intelligence depends on “mini computers” or “mini-brains” working together.

Knowledge is not stored in one specific location or part of the brain, but distributed across thousands of these cortical columns. Think of a cortical column like a tiny processor chip that runs its own version of reality. Each is capable of learning and simulating a predictive model through its inputs and within a specific reference frame.

Strikingly, even though we don’t perceive it, the knowledge of any particular item, event, or idea is therefore represented by multiple, complementary models at the same time and within a small fraction of the total number of columns.

Hawkins writes, “Cortical columns create reference frames for every object they know. Reference frames are then populated with links to other reference frames.” (p. 88). Put another way, each mini brain is perceiving, learning and deciding on its own while also within a grouping of other mini brains doing the same.

So, if we have so many mini-brains, why don’t we perceive multiple brains, sensors, etc? Why isn’t perception noisy and chaotic dissonance? Hawkins posits that these independent sensory-motor modeling systems work together through a voting mechanism to create our perception and understanding of the world. This suggests a more decentralized form of intelligence wherein truly intelligent machines must learn models of the world to adapt and solve problems.

My Quick Take Reviews

• I rated this book a 5 out of 5 on Goodreads.
• Would I recommend it? I would definitely recommend this book. In fact, I’d rank this one of the top books I’ve read about the brain and mind in several years.
• Would I read it again? Potentially. There is too much to digest and understand in this book to fully digest in a single reading. I wouldn’t necessarily read it again anytime soon, but I will keep an eye on this researcher and writer and his research group for future work in academic papers and book format.

Cortical Columns: Fundamental “Microcircuit” of Higher Order Intelligence

First identified in 1957, Vernon Benjamin Mountcastle proposed that the fundamental unit of the neocortex is the cortical column, a roughly one square millimeter area that extends through the entire thickness of the neocortex. Structurally, a cortical column consists of a group of neurons shaped like a column and positioned perpendicular to the cortical surface.

SOURCE: p48 of Ebook: Insights from the brain, the road towards Machine Intelligence

Cortical columns are proposed to be the fundamental microcircuits for cognition in neocortex. The fundamental concept is that the neocortex is not a single, unified processor, but rather a collection of approximately 150,000 independent sensory-motor modeling systems, each located within a cortical column. Each operates in parallell and coordinate high-order intelligence including learning, prediction and decision-making.

These columns work semi-autonomously and their collective activity, potentially through a “democratic consensus” or voting mechanism, leads to our overall perception and understanding of the world.

Rather than evolving something new, the evolution of our human intelligence involved duplication and scaling up a network of mini-brains, which allowed human neocortex to “increase in size by almost 3-fold over just the last 3 million years.” As Hawkins puts it, “The neocortex got big by making many copies of the same thing: a basic circuit.” (p. 29).

Reference Frames as the Foundation of Knowledge

Beyond just positing the neurological mechanisms, Hawkins attempts to explain how cortical columns in our neocortex model perception, represent knowledge and think. He posits that all knowledge in the neocortex is stored within map-like structures called reference frames.

Reference frames are used by each cortical column to learn the structure of objects, represent locations, make predictions, plan actions, and even understand abstract concepts. Reference frames help the brain organize knowledge. and, according to the Thousand Brains Theory, are the universal organizing principle for all knowledge.

Based on the sources provided, here is an explanation of how reference frames help the brain organize knowledge, using the specific example of a coffee cup.

The brain organizes all knowledge using reference frames. A reference frame is like an invisible, three-dimensional grid that the brain attaches to objects and concepts. This grid provides structure, allowing the brain to learn the locations of features relative to each other, which is essential for understanding an object’s shape, planning movements, and making predictions. Thinking itself is described as a form of movement, where your brain activates successive locations within these reference frames.

To illustrate this concept, the book uses the example of a coffee cup. When you first encounter a coffee cup, your brain learns a model of it. This isn’t just a list of facts about the coffee cup (like color, funny quote on side, shape, etc.); it’s a structural representation built upon a reference frame that is attached to the cup itself.

Your neocortex through cortical columns processing sensory information establishes a reference frame that is fixed to the cup. You can think of this like an invisible 3D grid surrounding the cup that allows you to mentally rotate the cup. As you interact with the cup in the real world (or mentally through imagination), your brain learns what features exist at specific locations and the overall reference frame of that cup updates and changes. You often can already imagine the backside of that cup even before turning and seeing it. As you manipulate it, your brain creates a three-dimensional shape and structure.

Mental and physical movement might even be a core component of learning and thinking. According to the Thousand Brains Theory, learning and perception are active, sensory-motor processes. Cortical columns learn models of the world by observing how sensory inputs change as we move and interact with our environment. Predictive modeling depends on mental movement, and Hawkins argues that thinking itself is a form of moving through these reference frames.

Beyond physical, real world objects, reference frames can also apply to abstract concepts like government and democracy, because the brain arranges all knowledge using reference frames, not just knowledge about things we can directly touch and see. Obviously the reference frame for democracy cannot correspond to the same one as three-dimensional physical objects, like a coffee cup. You can’t make a simple image of democracy in the same way you can for a cup. That said, democracy and other abstract concepts are more than just a random collection of facts. They also entail relationships, organization, structures and modeling like all knowledge.

How you model, organize and represent concepts matters. Learning what is democracy, software engineering or geology of mountains is more than simply knowing certain facts; it requires the right framework or reference frame for organizing facts, ideas and concepts coherently together. Different people might organize and think about democracy in different ways. For example, one might think of it in terms of a historical timeline or another might interconnect with other concepts like fairness, rights or voting. Interestingly, becoming an expert in a field requires discovering a good framework to represent data and facts and to formulate actions, decisions and solutions.

Additional Book Notes & Quotes

• Neocortex: The Engine of Intelligence: “The neocortex is the organ of intelligence. Almost all the capabilities we think of as intelligence—such as vision, language, music, math, science, and engineering—are created by the neocortex.” (p. 21)
• “All mammals, and only mammals, have a neocortex. The human neocortex is particularly large, occupying about 70 percent of the volume of our brain.” (p. 20)
• Prediction: The Core Function of Intelligence: Our brain constantly makes predictions based on past experience.
• “My brain, specifically my neocortex, was making multiple simultaneous predictions of what it was about to see, hear, and feel.” (p. 36)
• “Prediction was a ubiquitous function of the neocortex.” (p. 37)
• “A dendrite spike occurs when a set of synapses close to each other on a distal dendrite get input at the same time, and it means that the neuron has recognized a pattern of activity in some other neurons.” (p. 51)
• The Role of Movement in Learning: Intelligence and understanding arise through movement and interaction with the world.
• “Thinking occurs when the neurons invoke location after location in a reference frame, bringing to mind what was stored in each location.” (p. 75)
• “Vision is an interactive process, dependent on movement.” (p. 92)
• Knowledge is Distributed Across the Brain: “Our knowledge of something is distributed among thousands of cortical columns.”* (p. 98)
• “Complex systems work best when knowledge and actions are distributed among many, but not too many, elements.” (p. 97)
• Collaborative Process of the Neocortex: Even though neocortex has thousands of models distributed across cortical columns, they constantly communicate and vote on perceptions, leading to stability, unity and consistency of our perceptual experience of the world (and of our self).
• “The input to the neocortex is not like a photograph. It is a highly distorted and incomplete quilt of image patches. Yet we are unaware of the distortions and missing pieces; our perception of the world is uniform and complete.” (p. 94)
• “The voting mechanism of the Thousand Brains Theory explains why we have a singular, non-distorted perception.” (p. 107)
• Human Uniqueness in Language and Abstract Thought: “Our ability to see and hear is similar to a monkey’s, but only humans use complex language, make complex tools such as computers, and are able to reason about concepts such as evolution, genetics, and democracy.” (p. 71)
• Learning and Remember a Model of the World = Conscious Systems “At some point in the future, we will accept that any system that learns a model of the world, continuously remembers the states of that model, and recalls the remembered states will be conscious.” (p. 138)
• Brain Uploading: Brain uploading entails recording all the details of your brain and then using them to simulate your brain on a computer. “The fact is, if we want to upload you, and we want the uploaded brain to be normal, then we have to upload the entire brain, everything.” (p. 192) The process involves making a map of every neuron and every synapse and then recreating all of this structure in software.
• Uploaded brain will be a separate person or self: “‘Uploading your brain’ is a misleading phrase. What you have really done is split yourself into two people.” (p. 194) A mental and intellectual “you” will become separate from your biological body such “you” could then live in the computer.
• Genes vs Knowledge: Which is our future?
• Superiority of knowledge over genes: Genes are simply molecules that replicate. There is no overarching goal beyond replication. By contrast, knowledge has both a direction and an end goal.
• “We have the opportunity to choose between a future where the primary driver is the creation and dissemination of knowledge and a future where the primary driver is the copying and dissemination of genes.” (p. 219)
• “Our superior intelligence is unique, and as far as we know, the human brain is the only thing in the universe that knows the broader universe exists.” (p 232)
• Wiki Earth, an “estate plan for humanity” Goal is preserve human knowledge for potential future intelligent beings, because knowledge is rare and valuable. The aim would be to “preserve our knowledge in a more permanent form, one that could last tens of millions of years,” like Wikipedia floating in space.
• “Reverse engineering the brain and understanding intelligence is, in my opinion, the most important scientific quest humans will ever undertake.” (p. 229)

What I got out of this book?

While the science of the human brain remains deeply mysterious, this book made me rethink many of my commonly held understandings about how our brains work and what intelligence entails. Specifically, it emphasized that the brain is not a single centralized processor but a network of thousands of models working together. If true (and admittedly it is a bit difficult to experientially validate), this is a pretty revolutionary model of intelligence.

I personally find the idea that we are a collection of mini-brains fascinating. It means that our perceptions of the world get sensed, experienced and processed via multiple cortical columns at once and our first-person subjective experiences are built up from multiple angles and sources and distributed across thousands of complementary models in different columns. Our singular, unified experience is actually a consensus reached by these thousands of columns “voting” on what they are sensing.

This decentralized approach corresponds with my long-held belief in multiple selves, our evolving core identity and personality, and possiblity of personal transformation and change too. Underlying our brain’s flexibility and robustness is ambiguity. Our inputs and experiences are ambiguous and sensed through several mini-brains. In turn our knowledge and understanding depends on learning, creating and manipulating reference changes.

Ultimately, this perspective reminds us that we never perceive or experience the real world directly but instead we experience a simulation of the world created by the predictive models in our head. A concept repeated in The Mind Is Flat by Nick Chater. Our brain is a simulation and prediction machine. It is constantly making thousands of predictions about what it will see, hear, and feel, and we only become consciously aware of this process when a prediction is wrong. This continuous, predictive modeling is what allows the brain to learn, adapt, and create the flexible intelligence and selves that makes us human.

A Concluding Thought: How do we as individuals and societies overcome our “old brain”?

Beyond the core ideas and fascinating science, Hawkins also explores several ethical, philosophical, and even futurist ideas. I was particularly struck by his framing of the conflict between our two brains, namely our older, so-called “reptilian” brain and our more recent neocortex.

As he puts it, “We human mammals are the victims of a recurrent dispute: a tussle between the old reptilian brain, which unconsciously runs the survival machine, and the mammalian neocortex sitting in a kind of driver’s seat atop it”. These two brains serve different purposes with the old brain serving reproduction and selfish genes and the new brain serving knowledge. He believes that the battle between genes and knowledge defines our era.

I’d argue this internal conflict is responsible for many of our personal and societal struggles. For example, our neocortex, informed by research, doctors and books, might understand that eating a sugary cake is unhealthy. But our old brain, which evolved over millions of years when sugar was scarce and valuable for survival, simply screams, “Cake. Want cake. Mmmm cake. Gimme”. In this battle, the old brain, which evolved to serve the primal drives, often wins.

Sadly, this same dynamic seems to be playing out on a global scale with issues like climate change, epidemics and even racism. Our neocortex can learn a model of the world and predict the long-term consequences of our actions, but it is often overpowered by the old brain’s fears, pleasure-seeking and drive to replicate. Hawkins also highlights our old brain’s susceptibility to false beliefs including racism and xenophobia.

So, how do we overcome this? Hawkins suggests that humanity faces a profound choice: do we want our future to be driven by the processes that got us here—natural selection and the drive of selfish genes—or by intelligence and its desire to understand the world?. He argues that our newly emerged intelligence, residing in the neocortex, gives us the unique ability to defy the dictates of our selfish genes and our old brain. The key seems to be for the “new brain” to find clever ways to outsmart the “old brain.” He cites examples like birth control and advocates for education, including learning about how our brains work.

While I’m a bit skeptical that we will ever fully overcome the limitations of our older brain without radical and profound transformation on societal and even neurological levels, it is a noble ideal. What if every human understood what was going on inside their own head, might we have fewer conflicts, more knowledge and a better chance at a sustainable future?

Got a comment? Send me an email.

References:

Hawkins, J. (2021). A Thousand Brains: A New Theory of Intelligence. United States: Basic Books.