Philosopher wins NSF grant to study how our brains are like analog computers


Mon, 04/02/2018

author

George Diepenbrock

 

LAWRENCE — If our brains are computers, they are analog and not digital, according to research by a University of Kansas philosopher.

Some researchers have argued there is a close analogy between the all-or-nothing nature of neural spikes in the brain and the binary nature of digital computer logic. But Corey Maley, assistant professor of philosophy, said other recent neuroscience findings cement the implausibility of that argument, though they don't discredit that brains can still function like computers.

"If you expand what you mean by 'computer' to include analog computers, which existed before digital computers were ever built, then the brain might be a computer after all," said Maley, whose research focuses on the philosophy of science — particularly cognitive science — neuroscience and philosophy.

Maley made the argument in his paper "Toward Analog Neural Computation," published recently in the journal Minds and Machines. He also has received a $155,121 National Science Foundation Scholar Award to work on a book manuscript detailing how some neural processes might be analog computations. The grant will allow Maley to travel to academic conferences featuring research on computational neuroscience, cognitive science and philosophy of science to discuss this type of research with experts from a wide variety of disciplines.

This concept has garnered attention in the philosophy of science recently because of the interest in how to explain how biological organs or parts of organisms function, he said.

"For example, an essential feature of what it is for something to be a heart is that it is a kind of pump," Maley said. "In a case like this, we think of a natural organ as having the same kind of function as, say, a water pump, even though the water pump was explicitly designed to have a function, but the heart was not."

Computation is interesting because it lives partly in the material world of engineered artifacts designed to have particular functions and also in the abstract world of mathematics and logic, he said.

"Alan Turing and others formalized the notion of computation long before any physical devices could implement computations, and the theory of computation is very much a brand of mathematics," Maley said. "It would be useful to have a clear example of a kind of natural computation: a natural system that really performs computations, not just metaphorically, and not because we have designed it in a certain way."

Most of the sciences also use computation methods to study many kinds of phenomena. Every day, viewers see computer simulations that meteorologists use for weather forecasts, often in graphics and maps displayed on the local news or The Weather Channel.

"Although meteorologists use computers to simulate the weather, it would be a mistake to say that a weather pattern itself is a computer, or that it is performing computations," Maley said. "When it comes to the brain, however, neuroscientists and psychologists very often speak of the brain as literally performing computations. Upon a bit of investigation, it becomes clear that, whatever that means, it does not mean that they are performing computations of the kind studied in theoretical computer science."

In his argument that the brain functions not like a digital computer but possibly an analog one, the central point focuses on neural spikes in the brain.

Digital computing logic relies on a binary scheme and has a finite set of possible values, while most analog computers use continuously changeable aspects of physical phenomena to model something.

One way to illustrate the difference would be to think of neural spikes in the brain like a smoke detector that would beep once it detects a certain level of smoke.

"The assumption in neuroscience has been that neural spikes are like a detector. Either they go off, or they don't," Maley said. "All you know if you hear a smoke detector beep is that there is some smoke present above a certain threshold; but you don't know how much smoke, or what kind of smoke, or how hot the smoke is, etc."

However, others' work has suggested neural spikes in the brain sometimes act like a much more sophisticated version of a smoke detector, which would make the brain similar to an analog computer.

"The research I highlight suggests that neural spikes sometimes act like a more sophisticated smoke detector, where the beep gets louder if there is more smoke, or the beeps are longer if the smoke is hotter, for example," Maley said. "In this example, the characteristics of the individual beeps would carry more information than just simply whether smoke is present or not. Similarly, the neuroscience findings I focus on suggest that individual neural spikes may carry more information in a similar manner."

Other than the implications in philosophy of how to better explain biological functions, this idea of the brain as an analog computer could also illuminate other fields as well, he said.

"For neuroscience, it really depends on just how much the results I have highlighted generalize," Maley said. "But if this kind of continuous, analog neural representation is widespread, it may be that the brain is processing much more information — and doing so much more efficiently — than we previously thought."

Top right: Credit Modup.net, via Flickr Middle left: Illustration by Corey Maley on research that the brain might not function like a digital computer. His work in philosophy examines instead that our brains could function like analog computers though. Credit: Corey Maley Bottom right: Corey Maley, University of Kansas assistant professor of philosophy

 

Mon, 04/02/2018

author

George Diepenbrock

Media Contacts

George Diepenbrock

KU News Service

785-864-8853