A Thousand Brains: A New Theory of Intelligence

by Jeff Hawkins

comrades comes before even your own life.” The conflict between the old reptilian and the new mammalian brain furnishes the answer to such riddles as “Why does pain have to be so damn painful?” What, after all, is pain for? Pain is a proxy for death. It is a warning to the brain, “Don’t do that again: don’t tease a snake, pick up a hot ember, jump from a great height. This time it only hurt; next time it might kill you.” But now a designing engineer might say what we need here is the equivalent of a painless flag in the brain. When the flag shoots up, don’t repeat whatever you just did. But instead of the engineer’s easy and painless flag, what we actually get is pain—often excruciating, unbearable pain. Why? What’s wrong with the sensible flag? The answer probably lies in the disputatious nature of the brain’s decision-making processes: the tussle between old brain and new brain. It being too easy for the new brain to overrule the vote of the old brain, the painless flag system wouldn’t work. Neither would torture.

An animal doesn’t need a neocortex to live a complex life. A crocodile’s brain is roughly equivalent to our brain, but without a proper neocortex. A crocodile has sophisticated behaviors, cares for its young, and knows how to navigate its environment. Most people would say a crocodile has some level of intelligence, but nothing close to human intelligence. The neocortex and the older parts of the brain are connected via nerve fibers; therefore, we cannot think of them as completely separate organs. They are more like roommates, with separate agendas and personalities, but who need to cooperate to get anything done. The neocortex is in a decidedly unfair position, as it doesn’t control behavior directly. Unlike other parts of the brain, none of the cells in the neocortex connect directly to muscles, so it can’t, on its own, make any muscles move. When the neocortex wants to do something, it sends a signal to the old brain, in a sense asking the old brain to do its bidding. For example, breathing is a function of the brain stem, requiring no thought or input from the neocortex. The neocortex can temporarily control breathing, as when you consciously decide to hold your breath. But if the brain stem detects that your body needs more oxygen, it will ignore the neocortex and take back control. Similarly, the neocortex might think, “Don’t eat this piece of cake. It isn’t healthy.” But if older and more primitive parts of the brain say, “Looks good, smells good, eat it,” the cake...

The neocortex is surprisingly different. Although it occupies almost three-quarters of the brain’s volume and is responsible for a myriad of cognitive functions, it has no visually obvious divisions. The folds and creases are needed to fit the neocortex into the skull, similar to what you would see if you forced a napkin into a large wine glass. If you ignore the folds and creases, then the neocortex looks like one large sheet of cells, with no obvious divisions.

What Mountcastle says in these first three sentences is that the brain grew large over evolutionary time by adding new brain parts on top of old brain parts. The older parts control more primitive behaviors while the newer parts create more sophisticated ones. Hopefully this sounds familiar, as I discussed this idea in the previous chapter. However, Mountcastle goes on to say that while much of the brain got bigger by adding new parts on top of old parts, that is not how the neocortex grew to occupy 70 percent of our brain. The neocortex got big by making many copies of the same thing: a basic circuit. Imagine watching a video of our brain evolving. The brain starts small. A new piece appears at one end, then another piece appears on top of that, and then another piece is appended on top of the previous pieces. At some point, millions of years ago, a new piece appears that we now call the neocortex. The neocortex starts small, but then grows larger, not by creating anything new, but by copying a basic circuit over and over. As the neocortex grows, it gets larger in area but not in thickness. Mountcastle argued that, although a human neocortex is much larger than a rat or dog neocortex, they are all made of the same element—we just have more copies of that element.

So, what was Mountcastle’s proposal for the location of the cortical algorithm? He said that the fundamental unit of the neocortex, the unit of intelligence, was a “cortical column.” Looking at the surface of the neocortex, a cortical column occupies about one square millimeter. It extends through the entire 2.5 mm thickness, giving it a volume of 2.5 cubic millimeters. By this definition, there are roughly 150,000 cortical columns stacked side by side in a human neocortex. You can imagine a cortical column is like a little piece of thin spaghetti. A human neocortex is like 150,000 short pieces of spaghetti stacked vertically next to each other.

Let’s review. The neocortex is a sheet of tissue about the size of a large napkin. It is divided into dozens of regions that do different things. Each region is divided into thousands of columns. Each column is composed of several hundred hairlike minicolumns, which consist of a little over one hundred cells each. Mountcastle proposed that throughout the neocortex columns and minicolumns performed the same function: implementing a fundamental algorithm that is responsible for every aspect of perception and intelligence.

In 1971, scientist John O’Keefe and his student Jonathan Dostrovsky placed a wire into a rat’s brain. The wire recorded the spiking activity of a single neuron in the hippocampus. The wire went up toward the ceiling so they could record the activity of the cell as the rat moved and explored its environment, which was typically a big box on a table. They discovered what are now called place cells: neurons that fire every time the rat is in a particular location in a particular environment. A place cell is like a “you are here” marker on a map. As the rat moves, different place cells become active in each new location. If the rat returns to a location where it was before, the same place cell becomes active again. In 2005, scientists in the lab of May-Britt Moser and Edvard Moser used a similar experimental setup, again with rats. In their experiments, they recorded signals from neurons in the entorhinal cortex, adjacent to the hippocampus. They discovered what are now called grid cells, which fire at multiple locations in an environment. The locations where a grid cell becomes active form a grid pattern. If the rat moves in a straight line, the same grid cell becomes active again and again, at equally spaced intervals.

There are two modest-size regions of the neocortex that are said to be responsible for language. Wernicke’s area is thought to be responsible for language comprehension, and Broca’s area is thought to be responsible for language production. This is a bit of a simplification. First, there is disagreement over the exact location and extent of these regions. Second, the functions of Wernicke’s and Broca’s areas are not neatly differentiated into comprehension and production; they overlap a bit. Finally, and this should be obvious, language can’t be isolated to two small regions of the neocortex. We use spoken language, written language, and sign language. Wernicke’s and Broca’s areas don’t get input directly from the sensors, so the comprehension of language must rely on auditory and visual regions, and the production of language must rely on different motor abilities. Large areas of the neocortex are needed to create and understand language. Wernicke’s and Broca’s areas play a key role, but it is wrong to think of them as creating language in isolation. A surprising thing about language, and one that suggests that language may be different than other cognitive functions, is that Broca’s and Wernicke’s areas are only on the left side of the brain. The equivalent areas on the right side are only marginally implicated in language. Almost everything else the neocortex does occurs on both sides of the brain. The unique asymmetry of language suggests there is something different a...

According to linguists, one of the defining attributes of language is its nested structure. For example, sentences are composed of phrases, phrases are composed of words, and words are composed of letters. Recursion, the ability to repeatedly apply a rule, is another defining attribute. Recursion allows sentences to be constructed with almost unlimited complexity. For example, the simple sentence “Tom asked for more tea” can be extended to “Tom, who works at the auto shop, asked for more tea,” which can be extended to “Tom, who works at the auto shop, the one by the thrift store, asked for more tea.” The exact definition of recursion as it relates to language is debated, but the general idea is not hard to understand. Sentences can be composed of phrases, which can be composed of other phrases, and so on. It has long been argued that nested structure and recursion are key attributes of language. However, nested and recursive structure is not unique to language. In fact, everything in the world is composed this way. Take my coffee cup with the Numenta logo printed on the side. The cup has a nested structure: it consists of a cylinder, a handle, and a logo. The logo consists of a graphic and a word. The graphic is made up of circles and lines, while the word “Numenta” is composed of syllables, and the syllables are themselves made of letters. Objects can also have recursive structure. For example, imagine that the Numenta logo included a picture of a coffee cup, on which was...

The long-term goal of AI research is to create machines that exhibit human-like intelligence—machines that can rapidly learn new tasks, see analogies between different tasks, and flexibly solve new problems. This goal is called “artificial general intelligence,” or AGI, to distinguish it from today’s limited AI.

I recently attended a panel discussion titled “Being Human in the Age of Intelligent Machines.” At one point during the evening, a philosophy professor from Yale said that if a machine ever became conscious, then we would probably be morally obligated to not turn it off. The implication was that if something is conscious, even a machine, then it has moral rights, so turning it off is equivalent to murder. Wow! Imagine being sent to prison for unplugging a computer. Should we be concerned about this?

Wiki Earth Letting a distant civilization know that we once existed is an important first goal. But to me, the most important thing about humans is our knowledge. We are the only species on Earth that possesses knowledge of the universe and how it works. Knowledge is rare, and we should attempt to preserve it.