A 2024 study from Markus Meister and Jieyu Zheng proposed that the mind is an oddly slow thing. Compared to other systems of the brain which operate fast and tackle multiple problems all at once, the mind is frustratingly limited to a heavily filtered down experience and a sequential stream of thoughts that trudge along at a relative snails pace.

Others before them have also looked into the speed of certain ways of thinking, notably Daniel Kahneman in his book ‘Thinking, Fast and Slow.’ But Meister and Zheng take a new approach—using information theory to measure the bottleneck between sensory systems and conscious awareness.

They examine performance on a range of different tasks, like solving a Rubik’s cube, typing, visual memory, and language comprehension, and compared them to different features of the brain, like how much information an individual neuron, eye, or sensory system can transmit.

The difference is startling. While the peripheral nervous system (what the authors call the ‘outer brain’) is capable of feeding the brain information at rates of gigabytes per second, the mind itself (the ‘inner brain’) seems to operate around a leisurely 10 bits per second.

It’s a rate slower than those old dial up internet connections that took what felt like hours to download a song, or load an image. Isn’t that a little insulting?

“This defines a paradox”, they write, “the vast gulf between the tiny information throughput of human behavior and the huge information inputs on which the behavior is based. This enormous ratio—about 100,000,000—remains largely unexplained.”

The Inner and Outer Brain

The limitation is so familiar that you wouldn’t be blamed for having never really questioned it—of course we have one thought at a time, passing by one-after-another in a stream of consciousness. How could be be any other way?

Meanwhile the outer brain is taking in swaths of information from the environment, processing it all at once, and just sending the inner brain a little notification if there’s anything important happening.

The ‘cocktail party effect’ highlights this—we can only follow one conversation at a time, yet if we’re having that conversation in a busy bar and someone behind us mentions our name, our attention is diverted. This suggests that while we’re concentrating on one conversation and isolating it from all the background noise, some less conscious elements of the brain are still keeping an ear out for newsworthy items on our behalf.

It’s a good thing those unconscious operations tend to work reasonably well (though Kahneman’s work highlights the flaws and limitations). But is this really the best way to run a mind? Doesn’t it seem like there might be an advantage in a fast and parallel inner brain? If the brain is processing all this information, why is the mind only assessing a tiny fraction?

I wonder if an alternative can be seen in the octopus. Octopuses are “an independent experiment in the evolution of large and complex nervous systems – in the biological machinery of the mind,” says Peter Godfrey-smith.

Their arms have a significant degree of independence, their neural wiring is distributed and decentralised where ours is largely centrally located in the brain. With few exceptions, our arms do what we tell them, but for an octopus, the arm can do it’s own thing.

Who knows what it’s like to be an octopus but an octopus? Is there one expansive experience operating over all the limbs? Does each arm have it’s own mind? Maybe the arms have no inner mental components and only operate in an unconscious fashion.

While octopuses may be an exception, humans, for whatever reason, can’t mentally multitask.

The Rich vs Sparse Debate

The limitations don’t only apply to the number of thoughts but also to the richness of perception. The paper builds on an interesting debate over the fidelity of our inner experiences, concluding that the world we perceive at any one moment is surprisingly limited, despite our sense of a rich and detailed scene.

In the visual realm, when we look out at around us, we only clearly see the very central focal point of our vision—our peripheral vision is poor and blurry. We don’t really notice this because it’s quick and easy to dart our eyes back and forth in movements called saccades, something we do instinctively around 3 to 5 times a second, depending on the task.

“This shocking deficiency in our peripheral vision (all vision except two or three degrees around dead center) is normally concealed from us by the fact that our eyes, unlike television cameras, are not steadily trained on the world but dart about in an incessant and largely unnoticed game of visual tag with the items of potential interest happening in out field of view.” —Daniel Dennett, Consciousness Explained

In fact, many people aren’t aware that each eye has a blind spot—if you close one eye, the visual scene you’re left with has a missing piece due to the optic nerve making it’s way through the retina, but instead of a black hole, your brain fills it in with a blurry guess.

When you have both eyes open they make up for each others’ blind spots, yet even when you close an eye it’s not an easy find. How can we be so unaware of a hole in our visual world? The clarity of our visual world comes more from darting the eyes around—even one eye—while the blurry filling in of the blindspot is good enough for the peripheral, which is bad anyway.

One more point about the lowly levels of information passing form the eyes to the mind—even when we can move our eyes about to watch and interpret a scene, we can be rather poor at picking up significant details, highlighted in cases of both change and inattentional blindness:

• we can fail to notice rather significant changes in a picture.
• when we’re trying to focus on people catching a ball we miss a gorilla walking through the scene.
• sometimes the people we’re talking to can be swapped for someone else, without us picking up on the switch.

It is not uncommon for people to think they’ve seen more than they did, or can see more than they do. But the richness of our experience may be more of an internally generated feeling than an accurate reflection on the world, what some call ‘subjective inflation’.

From Movement to Mind

As the researchers write, all creatures with a nervous system move. The nervous system likely evolved to facilitate coordinated movement, towards food or away from predators. Movement would likely require focusing on a singular goal at a time, and sequential planning of movements to achieve them.

“There is no need or even opportunity to simultaneously process multiple paths to food because only one such path corresponds to the current reality. Accordingly, the cognitive architecture is designed to handle that one problem, which is local in space and in time.”

And this, they reason, evolved into our modern cognitive architecture that navigates through a space of abstract concepts.

“In this argument, human thought has co-opted the brain architecture that was originally designed for a jellyfish to follow the smell of decaying flesh in the ocean.”

By this reasoning, we have the hyper-aggressive filter reducing the vast range of sensory data into a 10-bit trickle because these are the necessary bits for planning and executing the behavioural sequences we’re engaged in, whether moving basic limbs towards rotting fish or navigating our mind through the ideas contained in a recent scientific paper.

The unconscious outer brain mostly stays out of the way, so as not to distract us from the task, unless the distraction is particularly important. This keeps cognition flexible but focused, capable of weaving through complex actions but also changing direction at a moments notice.

Generating Data

I believe there is a good case for the sparse and limited amount of information making it’s way from the senses to the mind, ‘information throughput’ as it’s termed, and I’m fascinated by the theory that our sequential mind evolved through the development of movement, but I’m curious how this paper might relate to the notion of the brain as a prediction engine.

We aren’t passive receivers of outside information, even a drastically filtered bit stream—we generate much of our experience internally. Tasks like reading, solving a Rubiks cube, or studying and memorising pictures, might highlight a low rate of transfer between outside and inside, but there is often more going on than you’d get from watching people do tasks and take tests.

Someone who takes a hallucinogenic will have an experience that doesn’t much correspond to the environment, but of course there is quite an experience taking place—the information throughput is lower, yet the experience vivid. Even when we see a visual illusion, there is data being generated that hasn’t come in through the eyes.

A lot of our experience is prediction, we have models of the world that generate hypotheses which we check against reality as we move our eyes and limbs about. Sometimes these models get totally unhinged from reality, but most the time they’re quite accurate.

To think of the mind and perception as a filter might be a poor metaphor, it’s more like a controlled hallucination, where sensory data fine-tunes an internally generated reality. Maybe the 10-bit limit isn’t a restriction on thought or perception, but the bandwidth of reality correction, the information used to modify our ongoing hallucination.

This would help align our internal sense of richness with the limited data throughput that the study highlights. While the sensory data coming into awareness is sparse, our experience is a rich, multimodal, hierarchical set of predictions.

“By making prediction the common root of both perception and action, predictive processing (active inference) reveals a hidden unity in the workings of the mind. Action and perception form a single whole, jointly orchestrated by the drive to eliminate errors in prediction.” —Andy Clark, The Experience Machine

Share the word