August 18th, 2010

This term is sometimes used to describe the annoying experience we have when some catchy tune or ditty gets stuck in our brains and then cycles over and over. This is especially infuriating when it’s some commercial jingle:

“Mr. Clean gets dirt and grime and grease in just a minute;
Mr. Clean will clean your whole house, and everything in it!”

Augghhh! I try to distract myself with something else: perhaps something erotic (never works; Mr. Clean always steals the show from Bubbles LeBreast). So I try to fight fire with fire by singing some other tune, something long and complicated like a Beethoven symphony. That works as long as I keep at it. Ten minutes after I’ve stopped, Mr. Clean slips back, whispering in my ear, “Mr. Clean gets dirt and grime and grease...”

Earworms are obviously brain glitches, some flaw in the design of our brains that plagues us, rather like a mental version of hiccups. Whence comes this evil distraction? I can offer a guess: it’s a byproduct of language. There are two areas of the brain that are deeply involved in the handling of language: Broca’s Area and Wernicke’s Area. They both reside on the left side of the brain in most people. Of course, language handling involves a great deal more than what Broca’s Area and Wernicke’s Area handle, but these two areas have been identified as crucial to particular aspects of language. But their locations on the left side of the brain lead to a question: what to the corresponding areas on the right side of the brain do? We know that they don’t process language. Much evidence suggests that perception of music is concentrated on the right side of the brain. Thus, we end up with a reasonable symmetry: both language and music are forms of auditory expression and perception. So it makes some sense that we’d see language processing and music processing in similar locations on opposite sides of the brain.

Cautionary note: brain localization is tricky business. There are always exceptions: people who seem to process information in somewhat different locations. The brain can sometimes compensate for damage to one area by developing the lost function in a new area. And a great many mental functions show some localization while also showing up all over the brain. So don’t get too comfortable with precise localization in the brain; functions are localized in the brain in much the same way that teenage girls are localized at the shopping mall on Saturday afternoons. Sure, you’ll find a lot of them there, but there will always be others in other places.

But now let’s think not so much in terms of localization and instead consider the subtasks involved in language perception. One crucial task is the retention of a spoken sentence in the brain while the language faculties chew it over. In many cases, you can’t really figure out a sentence until you reach its end. A classic example is something like this: “The dog took the bone that the neighbor across the street from Tom gave him.” Your brain has to store away all those secondary phrases until it gets the verb “gave”, which contains the core meaning of the sentence. Your brain must remember all those words. We know that much short-term memory is what I’ll call “circulating memory”. Neurons themselves don’t directly store short-term information; instead, they fire a pulse down their axons. Once the pulse has gone, the neuron has no memory of it. How do you store a pulse? It’s rather like the problem that a computer might have when the information is stored in pulses -- once those pulses have pulsed, they’re gone forever. Fortunately, there is a rather clumsy way around this problem: circulating memory. Here’s a huge oversimplification of the process: a neuron that means “dog” sends a pulse to another neuron that also means “dog”; the second neuron sends a pulse to a third neuron that also means “dog”; then the third neuron sends a pulse to the first neuron that means “dog”. Around and around the word “dog” goes, and so long as the cycle continues, the word is “remembered”. In practice, the real process is immensely more complicated, involving thousands of neurons for each little fragment of meaning. But the basic idea behind all this complicated processing is the notion of circulating memory.

Your brain has to store a LOT of information to remember a sentence, so there’s a big circulating memory system set up on the left side of the brain for that task. But if the brain preserves some degree of symmetry, then we’d expect there to be some sort of circulating memory system over on the right side of the brain, set up to handle musical perception. Musical “sentences” would circulate through this system as part of the “music comprehension” faculty. After all, we know that we are capable of understanding musical sequences in which relationships between notes are widely separated in time (consider that our mind catches the rhyme between Mr. Clean’s ‘minute’ and ‘in it’) But there’s one horrible difference between the two circulating memory systems. The language circulating memory must have foolproof termination systems to insure that a sentence is cleared out of circulating memory the instant it has been comprehended. After all, another sentence is likely to be coming in immediately after that one. But clearing out musical sentences is not so critical to music comprehension. A fragment of music can enter musical circulating memory and just go round and round and round. There’s no sure-fire system for terminating it. This explains why distractions such as Bubbles LeBreast don’t help: they don’t clear out the musical memory. It also explains why singing some other form of music does work: it replaces the earworm with another musical fragment. Of course, echoes of the original earworm are still reverberating through your brain, so you really have to keep going with that other music until all those echoes have died out.

You want evidence? I don’t got none! Believe it or don’t -- I don’t care. I’m just putting this down as a hypothesis for those people who are receptive to the idea.