Caltech scientists have estimated that the speed of human thinking is only 10 bits per second. This data rate is so slow that it calls into question claims made about brain-computer interactions and artificial intelligence. In a paper [PDF] entitled, “The Unbearable Slowness of Being: Why do we live at 10 bits/s?,” published in the journal Neuron, Markus Meister, professor and graduate researcher at Caltech, explores the cognitive conundrum that is the human brain.
They ask why the inner brain processes thought at 10 bits per seconds, while the outer brain, which processes sensory information, operates 100 million times more quickly, at 109 bits per second.
They ask, “The stark contrast between these numbers remains unexplained and touches on fundamental aspects of brain function: What neural substrate sets this speed limit on the pace of our existence? Why does the brain need billions of neurons to process 10 bits/s? Why can we only think about one thing at a time?”
Zheng & Meister arrived at their estimate of 10 bits/s for the speed of thinking by analyzing previous studies conducted in the last century on human behavioral throughput. These include: binary-digit memorization (4,9 bits/s); speaking in 17 different languages (39); listening comprehension in English (13); object recognition (30-50 bit/s); StarCraft (10) and typing (10 bits/s).
The results are consistent with previous research that suggests humans communicate in a rate of 40 bits/s.
The researchers also say that this has implications for estimating the storage capacity of the brain, which they say could be accommodated by a 5GB USB thumb drive if someone absorbed data at 10 bits/s 24 hours a days, for 100 years.
Caltech authors note that people tend to believe their inner lives are too complex to be expressed through real-time speech. This is what a brain pipeline operating at 10 bits/s allows. They say that this is an illusion. In the paper, they say
“Because we could engage in any one of the 2^10 possible actions or thoughts in the next second, it feels as though we could execute them all at the same time,” . They argue that this illusion about mental throughput is harmful because technologists believe in it.
In a 2018 interview, Elon Musk, the co-founder of Neuralink’s brain implant company, described it as follows: “The purpose of neuro-link is to create a high bandwidth interface to the brain such that we can be symbiotic with AI, because we have a bandwidth problem. You just can’t communicate through your fingers. It’s too slow.”
Zheng & Meister do not believe that a high bandwidth interface is needed to communicate with brain.
“Based on the research reviewed here regarding the rate of human cognition, we predict that Musk’s brain will communicate with the computer at about 10 bits/s,” They write. “Instead of the bundle of Neuralink electrodes, Musk could just use a telephone, whose data rate has been designed to match human language, which in turn is matched to the speed of perception and cognition.”
Neuralink didn’t respond to a comment request.
- Neuralink chips head to the Great White North.
- A second patient receives a Neuralink implant.
- China is working on a standard for brain-computer interactions.
- Open source software should be used in your pacemaker.
Meister told The Register by email that he believed the tech community needed to rethink its requirements for direct communication between the brain and computer in Meister said. “That’s almost certainly not going to work. We argued there’s plenty of evidence to say that no matter what, you can’t think faster than 10 bits per second. So yes, massive reevaluation would be useful here.”
Last year, the MIT Technology Review reported that the speed at which we think sets a limit to brain implant bandwidth. However, it also acknowledged that there were potential applications that could benefit from higher throughput, such as restoring bodily function through techniques like motor-neuron monitoring. Zheng and Meister, however, express skepticism regarding the need for high bandwidth brain-computer interfaces for patients with impaired motor control or senses.
The researchers cite efforts to restore sight using electrode arrays implanted into the eye to stimulate the ganglion cells by signals from a camera, which required data speeds on the order gigabits per sec. They note
“While driven by good intentions, this approach has been altogether unsuccessful: After decades of efforts, all the implanted patients remain legally blind,” . “The major companies behind this approach have now gone out of business, and their patients are left carrying abandoned hardware in their eyeballs.”
They say that a more practical (and one successfully deployed in 2018) approach is to have a computer translate a visual scene into speech in real-time and then describe it using words at a cognitively understandable data rate. The authors claim that
“The important principle for both sensory and motor BCIs is that one really needs to convey only a few bits per second to and from the brain, and those can generally be carried by interfaces that don’t require drilling holes in the user’s head,” . The authors argue that, while there is still much to learn about the brain’s actual functioning, there are many opportunities for new discoveries.
“The key is to observe the human brain under conditions of complex behavior where for example the specific task you’re engaged in switches several times a second,” said Meister. “We’re currently starting such experiments with human subjects under realistic conditions like driving a car in simulation and electrode recordings from neurons in different brain areas. But there is room for a great deal more inventiveness, and creativity in designing new experiments here.” (r)