What's missing from the brain-computer analogy

It's a pretty common idea nowadays - or at least it used to be - that the brain is, essentially, a computer, which makes sense, in a way.

After all, brains and computers do the same job - processing information. They are assembled from discrete components - in the case of the computer, transistors; in the case of the brain, neurons - and these assemblies have amazing properties due to something difficult to see on a mechanical level - in the computer, this is the software program; in the brain this is the mind. And, as far as we can tell, they are both digital, after a fashion - their elements of information (bits or action potentials) are either there are they aren't: 1 or 0. There, however, major differences between a computer and a brain.

For one thing, we know how a computer works - it operates on boolean logic principles, which means all of its operations involve the logical rules AND, OR, and NOT. In contrast, as far as I can tell we have essentially no idea how information is encoded and processed in the brain. There is a correlation between stimulus intensity and rate of firing for certain neurons, and evidence gathered from the physiology of neurons suggests that they are not boolean in nature (due to synaptic weighting) - and that is in a nutshell the extent of our knowledge regarding the framework of neural information processing. So one way in which brains differ from computers is in the fundamental principles of information processing and encoding.

Another way in which they differ is in their architecture. A computer has 3 main parts - a CPU, memory, and a bus to communicate between the two. A brain has at least dozens of parts, each with a specialized function - and information processing and memory are distributed all over the brain, with no central organization. This makes sense from an evolutionary point of view - you can't run a program with half a CPU, but you can think (somewhat) with half a frontal lobe (http://en.wikipedia.org/wiki/Phineas_Gage). Because of the distributed nature of the brain, it is unlikely that information is processed serially, as in a computer. Rather, the nature of information processing in the brain seems to be parallel.

Despite these differences, I don't think the brain-computer analogy is a misleading one. What is fundamentally different between a computer and a brain is the form, not the function. Both are designed to process information; both have emergent properties based on the interaction of their parts. The aesthetic differences between a brain and a computer are minor - the biggest difference lies in what we know about how a computer works, versus what we don't know about how a brain works.

Certainly the brain is radically different from a computer, but that's because it's built out of different components. We have a decent understanding of the components - now we need to learn how to build something useful out of them. The biggest discovery in neuroscience won't be the anatomy of the brain, or the physiology of the neuron - it will be the information theory of the mind.

That theory of information is precisely what is missing from the brain-computer analogy.

What this blog is all about

I'm a blogging neophyte, so I'm not sure how you're supposed to start these things - but the first order of business, it seems, is to explain the name of the blog.

"Encephalogue" is a word I made up with my limited knowledge of language. It's a blend of the Greek root, enkephalos, meaning "brain" or "inside the head", and logos, meaning "word" or "discourse" (Language experts, feel free to correct me if I'm wrong.) The word itself, as I imagined it, means "discourse about the brain".

That's what this blog is for - talking about the brain. I'll try to post a new rant about neuroscience bi-weekly, and will probably post things like reviews of neuroscience books every once in a while too. Depending on how often I can think of interesting brain-related things to rant about, I may toss some material about robotics and AI on here as well. They're not brain related per se,but still relevant enough that I don't want to make a new blog just to talk about them.

I usually try to post interesting neuro- and robotics-related links once a day on my Twitter page, twitter.com/SeanKPatrick, so feel free to check that out.